Linguistic term paper

LingC140_class2_Oct1_GS.pdf

Home >English homework help >Article writing homework help >Linguistic term paper

Ling C140. Fall 2019 G. Silvestri, Oct 1

The starting point: first language acquisition (II)

1 Stages of language development (idealized picture): Age 0-1 perception and production of speech sounds Age 1-2 lexical/phonological development Age 2-3 morphosyntax and early parameter setting Age 3> everything else 2 PERCEPTION and PRODUCTION of speech sounds (Year 1)

v during year 1: from being able distinguish a wide range of phonetic contrasts to being

able to discriminate only phonemic distinctions in the target language (=adults) v their productions (babbling) increasingly shaped by the ambient language (Davis &

MacNeilage 1995) v children's perceptual and productive abilities adapt to the demands of the native

language input v shift from a psychophysical basis to phonetic or phonological basis for speech sounds 2.1 PERCEPTION • Infants of 4 days old show a preference for the parental language (Mehler, Jusczyk,

Lambertz, Halsted, Bertoncini & Amiel-Tison 1988) • Infants at 3 days show a preference for mother's voice over other females (deCasper &

Fifer 1980) • Preferences due to greater familiarity with rhythm and intonation • Fetus can hear rhythm and intonation and possibly has a vague perception of vowel

differences 2.2.1 Categorical perception of consonants (1-6 months) v Adult speakers perceive speech categorically: they ignore distinctions that are non-

phonemic (Liberman et al 1957) v Stops distinguish themselves according to Voice Onset Time (VOT; Lisker &

Abramson 1964)

Ling C140. Fall 2019 G. Silvestri, Oct 1

• Voiceless aspirated stops [ph,th,kh]: vocal cord vibration delayed till after release; delay creates the perceptual effect of aspiration)

• Voiced stops [b, d, g]: vocal cords vibrate (almost) simultaneously with release. • Prevoiced: vocal cords vibrate long before release (up to 100 msecs) of the stop

• Some languages contrast only two of these:

- voiceless aspirated and voiced - English, Cantonese. - prevoiced with voiceless unaspirated - Dutch, French, Tamil, Hungarian - other languages (eg. Thai) contrast all 3

v English boundary falls at VOT+30 (1)

[b] [p] ________________________________________________________________________ _ -50 -40 -30 -20 -10 0 10 20 30 40 50 VOT

v Categorial perception: adult speakers perceive contrasts that are phonemic in their

language, i.e., perceive differences that straddle the phonemic boundaries, but not within categories.

v Feature detector hypothesis (P. Eimas et al. 1971) - categorical perception has strong biologically determinants, which are later modified by experience

- Eimas et al. (1971) tested for discrimination of synthetic speech along VOT continuum in 1 and 4 month olds within and across English boundary, using a High Amplitude Sucking Technique (HAS)

Results • English speaking children as young as 1 month discriminate pairs which cross the

English boundary • they do not discriminate other pairs within the short lag /b/ or long lag /p/ region • Spanish boundary is at +/-20:

- Guatemalan infants perceive English contrasts (+20/+60), but not the Spanish ones (+/-20; Lasky, Lasky & Klein 1975; Heart Rate technique)

Ling C140. Fall 2019 G. Silvestri, Oct 1

• children perceived the prevoice/voice (-20/-60) contrast found in Thai, but not in Spanish or English

• English babies respond to the prevoice/voice contrast (Aslin et al. 1981).

Further studies - stops vs. nasals (Eimas & Miller 1980) - stops vs. glides (Hillebrand et al. 1979) - [r]/[l] contrast (many languages don't have it; Eimas 1975) - changes in pitch contour (Morse 1972) - syllable length (Spring and Dale 1977)

2.2.2 The shaping effect of the environment v Infants universally respond to specific categories:

i. VOT: central parameter used by languages to encode contrasts ii. perceptual features: their perceptual apparatus that is fine-tuned to the requirements of human languages iii. babbling: (second half of the first year) infants begin to lose perceptual abilities for non-native sounds, just as their babbling begins to be shaped on target language

v Werker and Tees 1984 - American infants and American adults are given sounds from Hindi and Thompson Salish (a Canadian American Indian language): - Hindi - dental vs. retroflex stops - TS - velar vs. uvular ejectives - American adults do not discriminate these sounds - American infants do, up until about 10 months

Ling C140. Fall 2019 G. Silvestri, Oct 1

(2) 100 ___ ______

___ 70 ___

___ 20 ___

___ 0

H TS H TS H TS H TS ______________________________________________________________________

6-8 mos 8-10 10-12 H/TS 11-12 years Figure 1. Discrimination of Hindi and TS contrasts by English speaking adults and children and H and TS speaking young adults (Werker and Tees 1984:61) • "selective tuning of initial sensitivities in accordance with specific phonology...occurs at

about the age that the child is beginning to understand and possibly produce sounds appropriate to his/her native language." (Werker & Tees 1984: 62) • This same pattern of fine-tuning is shown in production

2.3 PRODUCTION v Sounds constrained by anatomy of the vocal tract:

children’s vocal tract configured for vegetative requirements, esp. sucking and breathing, not very good for talking.

2.3.1 Babbling v Early Babbling (4-7 months)

- Vowel and consonant sounds appear but they lack the mature regular-syllable timing characteristics of canonical babbling (Oller 1980)

- Babbling stimulated by speech of others and social interaction, as well as recordings of speech - associated with physical movement

v Canonical, reduplicated babbling (7-12 months) - emergence of adultlike canonical syllables

Ling C140. Fall 2019 G. Silvestri, Oct 1

- repetition of CVCV or single CV , or VCVC, dadada, gaga, mama, etc., with temporal regularity

§ Variegated babbling (> 10 months) - greater variety of consonants and vowels in same utterance - children acquiring different languages have similar segmental inventories - stops, nasals glides, and [h] are the core (Locke 1983)

§ Language specific differences by age 10 mos.:

- babbling begins to sounds like possible words of the language - intonation very appropriate to the particular language - native adults can identify the babbling of their language (French, Cantonese, Arabic) by 8-10 month olds (cf. Boysson-Bardies et al. 1984) - Language-specific differences in frequency of consonants (in babbles and first words), e.g. English, French, Japanese and Swedish children show significant differences in production of labials (Boysson-Bardies & Vihman 1991) - progression from universal to language particular : labials and dentals are most frequent in all languages studied (French, English, Japanese, Swedish, but labials drop off in English and Japanese reflecting differences in adult languages - deaf children babble but do not produce canonical (verbal) babbling - deaf children babble with their hands in the same repetitive fashion as hearing children at around the same age (10 months. Pettito 1991) - babbling is biologically triggered, and shaped by the environment - Babbling is a property of language, not speech

2.3.2 Prosody: rhythmic constraint in children v English as a stress-timed language:

- rhythm based on an opposition between strong (S) and weak (W) syllables - S syllables contain full vowels; weak syllable contain reduced vowels, usually schwa.

(3) SW paper

WS giraffe S dog W the/a

v Infants highly sensitive to the stress-based rhythm of English:

- characteristic rhythm of speech salient to the newborn - Neonates synchronize their movements with speech (live or recorded; native or non- native; Condon & Sanders 1974) - very young infants discriminate stress contrasts (Spring & Dale 1977; Juscyzk and

Ling C140. Fall 2019 G. Silvestri, Oct 1

Thomson 1978) - rhythmic structure one of the language-specific properties that babies get quite early (second half of 1st year) - 9 month old English kids prefer to listen to lists of bisyllabic words with initial stress (former, cable) than final stress (across, before, delay) (Jusczyk, Cutler & Redanz 1993; “trochaic bias”)

§ “Balanced” bilinguals

use segmentation procedure based on their dominant language: eg. English dominant bilinguals show stress-based constraint (for English and French); French dominant bilinguals show syllable-based (for French and English) (Cutler, Mehler, Norris, Segui 1989)

v Parameter-setting model of rhythm (Cutler & Mehler 1993)

Hypothesis: the choice of one or the other rhythm is made early in development and it happens only once

2.4 Lexical/phonological development (year 2)

v children figure out phonemic contrasts, phonotactics, phonological rules as they acquire

lexicon of target language (4) Acquiring phonemic contrasts

labial non-labial p t

labial coronal doral p t k

labial coronal dorsal p t k (b) (voice?)

labial coronal dorsal p t k b d g voice

Ling C140. Fall 2019 G. Silvestri, Oct 1

labial coronal dorsal p t k b d g voice m n nasal

v step-wise building up of contrasts causes the child's phonology to look different from the

adult's: - some features are not used contrastively - Child doesn't have the full set of contrasts (one segment represents more than one underlying form), e.g. Japanese:

(5) Child Adult

neto neko cat tama tama ball k => t

2.4.1 Common phonological processes v Children simplify adult pronunciation by using common phonological processes

§ Syllable structure processes (towards least marked syllable type CV)

cluster reduction 1;11 play --> [pey] train --> [ten] 1;11

deletion of syllable-final C (CVC --> CV) bib ---> [bi] 1;5

reduplication water --> [wawa] 1;9

§ Assimilation processes

final consonant devoicing pig ---> [bIk] 1;5

voicing before vowel paper --> [be:b] 2;3

Ling C140. Fall 2019 G. Silvestri, Oct 1

consonant harmony

duck ---> [gÃk] 1;7 tub ---> [bub]

§ Substitution processes

stopping of fricatives see ---> [ti] 2;9

fronting goat --> [dut] 2;0

gliding of liquids leg ---> [jek] 2;1

2.4.2 The nature of underlying representations v Hypothesis: rules operate on adult-like surface representations (Smith 1972; Stampe,

1969)

- Children perceives correct adult pronunciation, but unable to produce it for articulatory or linguistic reasons (insufficient contrasts) - they create rules to simplify pronunciation (realization rules)

§ Child perceives the difference but fails to produce it:

(6) Adult: Did you say 'wellow'?

Child: No, I said 'wellow.' (Instrumental analysis reveals more phonological information) § Stampe (1969)

- universal set of hierarchically ordered procedures that children use to simplify speech - to produce more oppositions, they must suppress the processes that do not make up part of the phonology of the native language, or limit them to certain contexts

Ling C140. Fall 2019 G. Silvestri, Oct 1

3. Morphosyntax and early parameter setting v Telegraphic speech (roughly age 2)

(7) a. Me/I like coffee

b. It fall down c. Gia write finger d. Papa change pants e. Choo-choo red f. Doggie big g. No car going there h. Shoe on i. Mommy red shoe

§ Brown (1973): “the 14 grammatical morphemes” - use of different functional morphemes

invariant across different children § "telegraphic" children know a lot about the morphosyntax during this stage 3.1 Root infinitives § "telegraphic" children produce root infinitives (RIs): they do not omit verbal morphology (8) a. Papa schoenen wassen.

Daddy shoes wash-inf. (Dutch, Weverink 1989) b. Michel dormir.

Michel sleep-inf. (French, Pierce 1992) c. Thorstn das haben.

Thorstn that have-inf. (German, Poeppel & Wexler 1993) d. Jag också hoppa där å där.

I also hop-inf. there and there (Swedish, Santelmann 1995) § RIs co-occur with finite verbs (9) [+finite] [-finite]

Elle a pas la bouche. Pas la poupée dormir. she has not a mouth not the doll sleep-inf. Veux pas lolo. Pas manger la poupée. (I) want not water not eat-inf the doll Marche pas. Pas casser.

Ling C140. Fall 2019 G. Silvestri, Oct 1

(she) walks not not break-inf. Ça tourne pas. Pas tomber bébé. that turns not not fall baby (Pierce 1992)

§ Adult form-position correlations are observed: Table 1. Finiteness and position of negation in French (Pierce 1992)

finite verb

non-finite verb

neg V

V neg

185

Nathalie and Daniel corpora: Lightbown 1977

3.2 Subject-verb agreement (10) a. Tu leggi il libro (1;11)

'You [nom.] read [2nd per. sing.] the book' b. Io mangio la pera (1;11)

'I [nom.] eat [1st per. sing.] the pear' c. Cosa fa il tore? (2;1)

'What does [3rd per. sing.] the tape recorder (do)?' d. Chelo micino no é piccino (1;11)

'That kitty is [3rd per. sing.] not little'

Ling C140. Fall 2019 G. Silvestri, Oct 1

Table 3. Percentage of subject-verb agreement errors in early language. child

Language

Age

%error

source

Simone

German

1;7-2;8

1732

Clahsen & Penke 1992

Martina*

Italian

1;8-2;7

478

1.6%

Guasti 1994

Diana*

Italian

1;10-2;6

610

1.5%

Guasti 1994

Guglielmo*

Italian

2;2-2;7

201

3.3%

Guasti 1994

Claudia

Italian

1;4-2;4

1410

Pizzuto & Caselli 1992

Francesco

Italian

1;5-2;10

1264

Pizzuto & Caselli 1992

Marco

Italian

1;5-3;0

415

Pizzuto & Caselli 1992

Marti*

Cat/Span

1;9-2;5

178

.56%

Torrens 1992

Josep*

Cat/Span

1;9-2;6

136

Torrens 1992

Gisela*

Catalan

1;10-2;6

1.2%

Torrens 1992

Guillem*

Catalan

1;9-2;6

129

2.3%

Torrens 1992

(*data available on CHILDES, MacWhinney & Snow 1985: Martin, Guglielmo, Diana corpora ; Cipriani et al 1991; Marti, Josep, Guillem, Gisela corpora: Serra & Solé 1992) 4. UG as a system of parameters v parameters of UG express a limited range of possible variation - Some parameters § Head direction parameter (OV vs. VO) § V to I parameter (cf. 4-6 above) § V2 (V to C) parameter (cf. 7-8 above) § Null subject parameter 4.1 Explanatory potential of the parameter model § Rich deductive structure

- several properties tied to one setting: e.g. null subject parameter:

Ling C140. Fall 2019 G. Silvestri, Oct 1

- null referential, subjects, post-verbal subjects, null expletives, Overt Pronoun Constraint (OPC) - Children acquire a lot based on one p-setting, contributes towards solving logical problem of language acquisition

§ Stages of child language fall within a principled range, tied to typological variation 4.2 How to set parameters § Parameters set on the basis of positive evidence § No negative evidence hypothesis entails that there are learnability principles such as the

subject principle, i.e. given two (or more) parameter values choose the value that generates the smaller language unless you have evidence to the contrary.

§ Parameter mis-setting hypothesis children omit subjects because of a misset parameter, i.e. pro- drop is a default setting (Hyams 1983/6), which in non-pro-drop languages such as English, is reset on the basis of positive evidence in the input.

Table 4. Percentage of null and overt subjects in finite clauses and root infinitives

Finite verb Infinitive_____ Language overt null overt null Flemish 75% 25% 11% 89% (Kramer 1993) German 80% 20% 11% 89% (Behrens 1993) French 74% 26% 7% 93% (Krämer 1993) Dutch 68% 32% 15% 85% (Haegeman 1994) § Table 4 shows that in many languages, null subjects are related to the finiteness of the

verb. v Early parameter setting

Ø The available evidence supports the hypothesis that children set the core parameters of UG at a very early age - prior to their first multi-word utterances.

Ling C140. Fall 2019 G. Silvestri, Oct 1

5. Theories of telegraphic speech v Prefunctional Grammar Hypothesis

(aka Small Clause Hypothesis; Radford 1986; Lebeaux 1987; Guilfoyle & Noonan 1989) - Children (to age 2 roughly) have lexical, but not functional categories - utterances structured as adult small clauses

v How adult structure develops

- Maturation Hypothesis - Lexical Learning hypothesis (11) XP (=SC)

NP XP

X = A, N, P, V (12) a. I heard [VP John sing]

b. I consider [AP John stupid] c. I saw [PP John on the boat] d. We elected [NP John president]

v Full clause hypothesis (Hyams 1992; Wexler 1994; Poeppel & Wexler 1993 a.o.) Children have a full adult-like clause

v Underspecification Hypothesis (Wexler 1994; Hyams 1994) one or more I category, e.g. Tense, Person, Number, can be left unspecified, in the early grammar

Ling C140. Fall 2019 G. Silvestri, Oct 1

v Truncation Hypothesis (Rizzi 1992) (13)

CP adult: root = CP SPEC C'

C IP

SPEC I'

I VP child: root = XP

SPEC V'

V NP

v Predictions § Small Clause Hypothesis

children should show no knowledge of functional elements such s agreement, or verb movement (contrary to fact).

§ Full Clause Hypothesis + underspecification functional elements are optional in early grammar.

§ Full clause hypothesis + truncation Non-finite clauses should not contain higher functional material, e.g. wh phrases.

6. Grammatical dependencies: Reflexive anaphors There are elements in a sentence that depend on other elements for their form or interpretation. (14) Ernie’s brother cut himself § Clausemate condition

A reflexive must take an antecedent in its same clause

§ C-command condition

Ling C140. Fall 2019 G. Silvestri, Oct 1

The antecedent must c-command the reflexive - c-command: - The c-command domain of an XP is the first constituent containing the XP - An XP c-commands everything in its c-command domain.

(15) YP

XP ZP

X Y Z (16) TP

DP2 VP

DP1 NP V DP3

Ernie’s brother cut himself him - DP2 c-commands DP3 - DP1 does not c-command DP3 § Chien and Wexler (1988): tested c-command condition through a picture ID study

- Results: by age 5 children are reliable in choosing the correct antecedent for the anaphor; children under 4 perform at about 50%.

§ Hypothesis: children initially treat ‘himself/herself’ as a special kind of possessive

construction, like ‘his/her body’ (children often say ‘hisself’), a lexical error. (17) Ernie’s brother washed his body

a. = Ernie’s brother washed Ernie’s body. b. = Ernie’s brother washed his own body

§ Chien and Wexler (1991): act-out task tested for clausemate condition.

Ling C140. Fall 2019 G. Silvestri, Oct 1

(18) Mulan says that (Sarah) should point to herself. - Results: By 51/2 children reliably choose the local antecedent - At age 21/2 - 3 children chose the long-distance antecedent about 80% of the time (!!) § Hypothesis: Children have mis-analyzed anaphor as a pronoun of the kind you find in

Icelandic (19) Pétur segir að Jón pvoi bílinn í hverri viku.

Peter says that John washes car-the his in every day. ‘Peter says that John washes his (= Peter’s) car every day’

v Lexical learning (McCune & Vihman 2001)

Children must learn properties of particular words - phonological, semantic, syntactic. Lexical properties may take time to learn.

7. Semantic development v children take time to learn the pragmatic and discourse-related properties of language

- contrast to their syntactic and morphological development (very early) (20) Ernie kissed him

Adults: him = someone other Ernie (Condition B of the Binding Conditions) Children (under 5): him = Ernie him = someone other than Ernie

- optional use of pronouns and determiners affects specificity and reference - also, pronouns and determiners used ‘out of the blue’ without a discourse antecedent. 8. Implications for L2 acquisition/bilingualism Issues to think about with respect to L2 acquisition:

• If phonetic perception and production is fixed very early by ambient language, can

Ling C140. Fall 2019 G. Silvestri, Oct 1

this be "undone" for L2? • Is there some relation between the point of acquisition in L1 and the flexibility

(sensitive period) with respect to L2, e.g. is L2 phonology harder because it is fixed earliest in L1?

• If parameters of UG are fixed very early, can they be reset for L2? • Do L2ers use, i.e. transfer, their L1 parameter settings or do they start from default

setting? • Do bilingual children have two sets of parameters? If so what is the implication for

L2 acquisition? • Does adult/child L2 acquisition show the characteristics of L1 acquisition, e.g.

phonological substitutions, pro-drop, telegraphic speech, similar developmental sequences?

• Do hypotheses about L1 acquisition (e.g. prefunctional grammar hypothesis, truncation, etc.) also apply to L2?

Suggested readings (*), assigned readings (**) ** Hummel, K., 2014. Introducing second language acquisition. Chichester, England: Wiley Blackwell. Chapter 9 (pp. 223-246) ** Werker, J. and R. Tees, 1984. Cross-language speech perception: evidence for perceptual reorganization during the first year of life. Infant behavior and development 7, 49-63. * Stoel-Gammon, C., 2011. Relationships between lexical and phonological development in young children. Journal of Child Language 38 (1), 1-34.