phonological effects in intrasentential codeswitching ?

39
Phonological Effects in Intrasentential Codeswitching ? Sonia Colina and Jeff MacSwan Arizona State University/University of Arizona and Arizona State University

Upload: sora

Post on 07-Feb-2016

46 views

Category:

Documents


0 download

DESCRIPTION

Phonological Effects in Intrasentential Codeswitching ?. Sonia Colina and Jeff MacSwan Arizona State University/University of Arizona and Arizona State University. Research Questions. How does phonology affect the syntax of codeswitching? - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Phonological Effects in Intrasentential Codeswitching ?

Phonological Effects in Intrasentential Codeswitching?

Sonia Colina and Jeff MacSwanArizona State University/University of Arizona

and Arizona State University

Page 2: Phonological Effects in Intrasentential Codeswitching ?

Research Questions

How does phonology affect the syntax of codeswitching? What are the restrictions on language mixing

within the phonological component of the grammar?

Can restrictions on codeswitching be accounted for in terms of architectural (design) considerations?

Page 3: Phonological Effects in Intrasentential Codeswitching ?

Codeswitching in Morphophonology

We can’t switch before a bound morpheme (Poplack, 1980, 1981)

*Estoy eat-iendo*Juan love-ó a Maria

(Nonce) borrowing versus word-internal switching Lonchamos con Maria [borrowing]*Lunch-amos con Maria [word-internal codeswitch] ‘Let’s have lunch with Maria’

In fact, we can’t mix phonology within words at all.Spanish: g ---> ƒ between vowelsEnglish: o ---> ow [+round] word finallyEnglish “ago” as [aƒow] in Spanish-English is ill-formed.

Page 4: Phonological Effects in Intrasentential Codeswitching ?

Phrasal Affixation

English genitive -’s is a phrasal (XP) affix[Tom and Mary]’s house[the man from Nebraska]’s hat

Spanish-English (MacSwan, 2004)

Su novia’s coche está nuevo‘His girlfriend’s car is new.’

Mi cuñado’s motorcycle is in the driveway‘My brother-in-law’s motorcycle is in the driveway.’

Croatian-English (Hlavac, 2003)

… imam, moja mamin’s sestra jet u i … sve moj tata’s family je sve u Zagreb …

… I have, my brother’s sister is here and … all my dad’s family is all in Zagreb …

Page 5: Phonological Effects in Intrasentential Codeswitching ?

Codeswitching in Head Movement Contexts Language Switching in Restructuring Contexts

(Italian-French)Si è dato un regalo [Italian]Un regalo si è dato [Italian]Si è donné un cadeau [Italian-French]*Un cadeau si è donné [Italian-French]

‘A gift is given’ Language Switching in Negation Contexts (Spanish-

Nahuatl)*No nitekititoc

‘I’m not working’Amo estoy trabajando

‘I’m not working’

Page 6: Phonological Effects in Intrasentential Codeswitching ?

Descriptive Generalization

Codeswitching cannot occur word-internally in head movement contexts

In other words: Codeswitching cannot occur internally within a head (X0), whether simple or complex.

Counter-examples? Finnish-English case morphology Dutch agreement suffix -e on French adjectives

Viable counter-examples stimulate further study, leading to theoretical refinements

Page 7: Phonological Effects in Intrasentential Codeswitching ?

The PF Disjunction Theorem (MacSwan 1999, 2000)

Goals Rule out head-internal codeswitching without positing

codeswitching-specific mechanisms. PF Disjunction Theorem

i. The PF component consists of rules which must be (partially) ordered with respect to each other, and these orders vary cross-linguistically.

ii. Codeswitching entails the union of at least two (lexically-encoded) grammars.

iii. Ordering relations are not preserved under union.iv. Therefore, codeswitching within a PF component is not

possible. Heads are inputs to phonology. PFDT is an instantiation of Full Interpretation (FI)

Words (X0s) which contain objects associated with distinct phonological systems lack sensorimotor interpretations at the PF Interface.

Page 8: Phonological Effects in Intrasentential Codeswitching ?

Previous Studies in the Phonology/ Phonetics of Codeswitching Recent studies of phonological aspects of codeswitching

have focused on phonetic, gradient effects: Grosjean and Miller (1994)

• Voice Onset Time (VOT) measurements in French/English codeswitches. Lexical switches.

• Findings: “base language” has no impact on the production of codeswitches for VOT; the shift from one language to the next is total and immediate.

Botero et al. (2004) VOT measurements in Spanish/English codeswitches. Sentential

switches. Findings: trend towards convergence in the VOT values of

voiceless stops in Spanish/English codeswitches; perseverative phonetic effects in codeswitching contexts.

Page 9: Phonological Effects in Intrasentential Codeswitching ?

Previous Studies in the Phonology/ Phonetics of Codeswitching Bullock et al. (2004)

also focuses on phonetics (phonetic convergence), despite the introduction of allophonic distribution.

does not present a theoretical account of the phonological facts. sounds under investigation are switch-internal (vs. at the point of

switch). findings:

hypothesis (English allophones may be substituted for the Spanish ones and/or vice versa) is not borne out; no allophonic substitution, no convergence at the phonological level of representation; phonetic convergence found instead.

acoustic values of [l] in codeswitched bilingual speech are closer to those of an English dark /l/ than those of monolingual Spanish, but still below the threshold level necessary to be perceived as dark. Although subjects show coda neutralization of the l/r contrast in Spanish (lambdacism), no such process was found in codeswitching into English.

Page 10: Phonological Effects in Intrasentential Codeswitching ?

Goals of the Present Study

1. Contribute to the literature on the phonology of codeswitching by examining phonological, categorical phenomena.

2. Determine empirically whether in mixed-language contexts there is a “base-language effect” in the postlexical

phonological component or a sudden switch in phonological systems

3. Refine/reformulate the PF Disjunction Theorem using insights from Optimality Theory.

Here we report on initial efforts with 1 and 2 will not address 3 for the moment.

Page 11: Phonological Effects in Intrasentential Codeswitching ?

The Relevance of Optimality Theory to the Present Study Crosslinguistic variation (language typology) is the result of language-

specific rankings of universal constraints. Two languages may have opposite rankings of the same

constraints: e.g., languages with consonant epenthesis have ONSET >> DEP, but those without it have the ranking DEP >> ONSET.

Ranking paradoxes (opposite rankings of the same universal constraints) are not permitted.

Conflating phonological systems in bilingual codeswitching could result in a ranking paradox.

OT predictions: In the case of a ranking paradox, since only one ranking is possible,

there must be a sudden switch in phonology (from the “base” to the “receptor language”) at the point of switch.

When no conflicting rankings are involved, the facts could be different. A feature in the receptor language could serve as a trigger for a process in the base language.

Page 12: Phonological Effects in Intrasentential Codeswitching ?

Two Experiments

Experiment 1 Intervocalic allophones of /b, d, g/

Conflicting rankings in English and Spanish. Switches between a Spanish trigger and an English lexical

item test whether Spanish phonology persists into the English lexical item.

Experiment 2 Voicing of syllable-final /s/ in Spanish.

No conflicting rankings involved. Switches between an English trigger and a Spanish lexical

item test whether English phonology persists into the Spanish lexical item.

Page 13: Phonological Effects in Intrasentential Codeswitching ?

Experiment 1: Spanish Approximants (Fricatives) /b, d, g/ are realized as [-continuants] when following a stop or a

pause (or /l/ in the case of /d/):cuando [ku9an5do] ‘when’tengo [tego] ‘I have’cambio [kambi9o] ‘change’caldo [kal5do] ‘broth’

All Spanish dialects with [+continuant] allophones of the voiced stops have a [+ continuant, -vocalic] segment in intervocalic contexts (Lipski, 1994):haba [aBa] ‘bean’hada [aDa] ‘fairy’haga [aa] ‘do-SUBJ.3SG’

In English, by contrast, intervocalic allophones of /b, d, g/ are always [-continuant].

Page 14: Phonological Effects in Intrasentential Codeswitching ?

Relevant Constraints

AGREE (stricture): Adjacent segments must agree in degree of stricture (Díaz-Campos & Colina, 2004; Steriade, 1993)

IDENT-IO (continuant): Corresponding segments are identical with regard to their [+/- continuant] specification (i.e., [+/- continuant] specification in the output must match that of the input and vice versa).

IDENT-IO (sonorant): Corresponding segments are identical with regard to their [+/- sonorant] specification (i.e., [+/- sonorant] specification in the output must match that of the input and vice versa).

Page 15: Phonological Effects in Intrasentential Codeswitching ?

Other Relevant Constraints

FAITHFULNESS: Any segment present in the input must also be present in the output (MAX-IO); any segment present in the output must have a correspondent in the input (DEP-IO).

Stricture theory (cf. Steriade, 1993) A0: maximal stricture (non-continuants, stops,

nasals and laterals) Af: medium aperture (fricatives) AMAX: minimal stricture (approximants and

vowels)

Page 16: Phonological Effects in Intrasentential Codeswitching ?

Constraint Ranking Differences

SpanishAGREE (stricture)>> IDENT-IO

(continuant), IDENT-IO (sonorant) English

IDENT-IO (continuant), IDENT-IO (sonorant) >> AGREE (stricture)

Note: this process affects only voiced obstruents in Spanish (not voiceless). An undominated constraint against voiceless approximants is responsible for the exclusion of the voiceless counterparts.

Page 17: Phonological Effects in Intrasentential Codeswitching ?

Tableaux

Monolingual Spanish

/lagala/ [laÂala] ‘the event’AGREE

(stricture) IDENT-IO(continuant)

IDENT-IO (sonorant)

F a. [laÂala] *b. [lagala] *!

Page 18: Phonological Effects in Intrasentential Codeswitching ?

Tableaux

Monolingual English

/eigoust/ [ei9gou9st] ‘a ghost’

IDENT-IO(continuant)

IDENT-IO(sonorant)

AGREE

(stricture)F [ei9gou9st] *

b. [ei9Âou9st] * *!

Page 19: Phonological Effects in Intrasentential Codeswitching ?

Methods

Participants: 5 simultaneous Spanish-English bilinguals from Central Arizona (students at Arizona State University).

Task: Participants were asked to pronounce 27 sentences three times each in non-sequential order using Presentation, a stimulus-delivery software package, in a sound booth.

Items: Items involved codeswitches from Spanish into English (voiced stop-initial English noun at the onset of the switch, preceded by a vowel-final Spanish determiner). E.g.,Hablamos de mi ghost yesterday

‘We talked about my ghost yesterday’Hablamos de mi disk yesterday

‘We talked about my disk yesterday’Hablamos de mi book yesterday

‘We talked about my book yesterday’

Page 20: Phonological Effects in Intrasentential Codeswitching ?

Methods

The sentences were recorded, transcribed and subjected to spectrographic analysis (using Praat) to determine the continuancy (stop versus approximant) of /b, d, g/.

The codeswitched samples were compared to monolingual Spanish intervocalic contexts to rule out convergence towards English in participants’ Spanish phonology.

Page 21: Phonological Effects in Intrasentential Codeswitching ?

Results

*Note: S4 excluded based on monolingual Spanish results.

Figure 1: Results for Spanish-English bilingual speech sample

total tokens stops approximant Per centstops

other

s2 52 50 2 96.15%s3 55 49 4 89.09% 2s4(excluded)*

54 52 2 96.29%

s5 53 52 1 98.11%

s6 61 48 1 78.68% 2

Page 22: Phonological Effects in Intrasentential Codeswitching ?

Results

Figure 2: Results for monolingual Spanish speech sample

total tokens stops approximant Per centapproximants

other

s2 50 3 29 91% 18 deletionss3 55 0 47 100% 8 deletionss4* 54 51 0 0% 3 1 deletion; 3

stops inducedby pause

s5 53 1 49 98% 3 unintelligible

s6 61 5 27 84% 13 deletions; 16unintelligible

Page 23: Phonological Effects in Intrasentential Codeswitching ?

Bar Graph Representation of Results for Spanish-English Bilingual Speech Sample

Per cent stops

0.00%

20.00%

40.00%

60.00%

80.00%

100.00%

120.00%

s2 s3 s5 s6 average

Per cent stops

Page 24: Phonological Effects in Intrasentential Codeswitching ?

Experiment 2: /s/-voicing in Spanish In Spanish /s/ is realized as [z] when followed by a voiced

consonantdesde [dezD e] ‘from’

mismo [mizmo] ‘same’

las dos [lazDos] ‘the two’

tres manos [trezmanos] ‘three hands’

tres osos [tresosos] ‘three bears’

Note: This process is postlexical and therefore not obligatory. However, the presence of at least some /s/-voicing in a codeswitching context, in particular when /s/ is followed by an English word that starts with a voiced consonant, would indicate that the feature [+voice] can serve as a trigger for the /s/ voicing, independendently of the fact that [+voice] belongs to an English lexical item.

Page 25: Phonological Effects in Intrasentential Codeswitching ?

Relevant Constraints

AGREE (voice): Adjacent segments must agree voicing

IDENT-IO Coda (voice): Corresponding coda segments are identical with regard to their [+/-voice] specification. (i.e., [+/-voice] specification in the output must match that of the input and viceversa).

IDENT-IO Onset (voice): Corresponding onset segments are identical with regard to their [+/-voice] specification. (i.e., [+/-voice] specification in the output must match that of the input and viceversa).

Page 26: Phonological Effects in Intrasentential Codeswitching ?

Constraint Ranking

IDENT-IO Onset (voice) >> AGREE (voice) >> IDENT-IO Coda (voice)

Page 27: Phonological Effects in Intrasentential Codeswitching ?

Tableaux

Monolingual Spanish

/mismo/ [mizmo] ‘same’

IDENT-IOONSET

(voice)AGREE (voice) IDENT-IOCODA

(voice)F a. [mizmo] *b. [mismo] *!

Page 28: Phonological Effects in Intrasentential Codeswitching ?

Tableaux

Code-switched utterance (Spanish > Eng)/mis || goust/ [mi gou9st] ‘my ghosts’

IDENT-IOONSET

(voice)AGREE

(voice)IDENT-IOCODA

(voice)F a.[miz gou9sts] *b. [mis gou9sts] *!

Page 29: Phonological Effects in Intrasentential Codeswitching ?

Methods

Participants: 5 simultaneous Spanish-English bilinguals from Central Arizona (students at Arizona State University).

Task: Participants were asked to pronounce 9 sentences three times each in non-sequential order using Presentation, a stimulus-delivery software package, in a sound booth.

Items: Items involved codeswitches from Spanish into English (voiced stop-initial English noun at the onset of the switch, preceded by a s-final Spanish determiner). E.g.,Hablamos de mis ghosts yesterday

‘We talked about my ghosts yesterday’Hablamos de mis disks yesterday

‘We talked about my disks yesterday’Hablamos de mis books yesterday

‘We talked about my books yesterday’

Page 30: Phonological Effects in Intrasentential Codeswitching ?

Methods

The sentences were recorded, transcribed and subjected to spectrographic analysis (using Praat) to determine voicing of /s/.

The codeswitched samples were compared to monolingual Spanish contexts to rule out convergence towards English in participants’ Spanish phonology.

Page 31: Phonological Effects in Intrasentential Codeswitching ?

Results

Figure 3: Results for Spanish-English bilingual speech sample

total tokens [s] [z] Per cent [z] others2 28 18 (64.28%) 6 21.42% 4 (14.28%)s3 26 5 (19.23%) 18 69.23% 3 (11.53%)s4* 27 26 (96.29%) 1 3.70%s5 25 13 (52%) 8 32% 4 (partial voicing)

(16%)s6 32 5 (15.62%) 16 50% 11 (34.37%)

Page 32: Phonological Effects in Intrasentential Codeswitching ?

Results

Figure 4: Results for monolingual Spanish speech sample (e.g. hablamos de)

total tokens [s] [z] Per cent [z] others2 28 14 (50%) 13 46.42% 1 unintelligibles3 26 1(3.84%) 15 57.59% 10s4* 27 26 (96.29%) 0 1 [s] before

pauses5 25 2(8%) 21 84% 2

s6 32 0 29 90.62% 3 2 deletion; 1unintelligible

Page 33: Phonological Effects in Intrasentential Codeswitching ?

Bar Graph Representation of Results for Spanish-English Bilingual Speech Sample

[z] percent

0.00%

10.00%

20.00%

30.00%

40.00%

50.00%

60.00%

70.00%

80.00%

s2 s3 s5 average

[z] percent

Page 34: Phonological Effects in Intrasentential Codeswitching ?

Conclusions

The presence of the relevant structural description does not always trigger application of Spanish phonology across word boundaries.

The data reveal that simultaneous Spanish-English bilinguals shift suddenly from Spanish to English phonology between word boundaries when there are conflicting rankings (/b,d,g/).

Page 35: Phonological Effects in Intrasentential Codeswitching ?

Conclusions

An English segment may serve as a trigger for Spanish phonology across word boundaries in a Spanish lexical item when no conflicting rankings are involved.

These results are consistent with the theory that phonological systems may be switched at word boundaries but not within words (heads). The defining characteristics of the switch (sudden or not) will depend on the specific processes involved as well as the theoretical account used to explain them.

Page 36: Phonological Effects in Intrasentential Codeswitching ?

Further Research

Another experiment needs to be done to test the opposite directionality in /s/-voicing. In other words, is the English contrast between /s/ and /z/ affected by contact with a [+voice] C in a Spanish lexical item, E.g, price de ‘price of,’ [praisDe] or [praizDe]?

This study seeks to contribute to a model of bilingual codeswitching in which grammaticality facts are substantially explained by conditions on the syntax-phonology interface.

Page 37: Phonological Effects in Intrasentential Codeswitching ?

Acknowledgments

Funding This research was funded by a grant from the

National Academy of Education with funding from the Spencer Foundation.

Participants We thank the several bilingual students at

ASU who participated in the study. Graduate Assistants

Kara McAlister and Peter Sayer assisted in this research.

Page 38: Phonological Effects in Intrasentential Codeswitching ?

BibliographyBotero, Ch., Barbara Bullock, Kristopher Davis, and Almeida Jacqueline Toribio. 2004. Perseverative

phonetic effects in bilingual code-switching. Paper presented at the 34 th Linguistic Symposium on Romance Languages, University of Utah, March.

Bullock, Barbara, Almeida Jacqueline Toribio, Kristopher Allen Davis and Christopher Botero. 2004. Phonetic convergence in bilingual Puerto Rican Spanish. In Benjamin Schmeiser, Vineeta Chand, Ann Kelleher and Angelo Rodríguez, eds. WCCFL 23 Proceedings, pp.101-103. Somerville, MA: Cascadilla Press.

Chomsky, N. 2000. Minimalist inquiries: The framework. In R. Martin, D. Michaels, & J. Uriagereka, eds., Step by Step: Essays on Minimalist Syntax in Honor of Howard Lasnik, pp. 89-155. Cambridge: MIT Press.

Chomsky, N. 2001. Derivation by phase. In M. Kenstowicz, ed. Ken Hale: A life in language, pp. 1-51. Cambridge: MIT Press.

Colina, Sonia and Díaz-Campos, Manuel. 2004 The interaction between faithfulness constraints and sociolinguistic variation: The acquisition of phonological variation in first language speakers. Paper presented at the 34th LSRL, Salt Lake City, March 2004.

Grosjean, François and Joanne Miller. 1994. Going in and out of languages: An example of bilingual flexibility. Psychological Science 5 (4): 201-206.

Hlavac, J. 2003. Second-generation speech: Lexicon, code-switching, and morpho-syntax of Croatian-English bilinguals. Berlin and New York: Peter Lang.

Lipski, John. 1994. Latin American Spanish. Cambridge and New York: CUP.

Page 39: Phonological Effects in Intrasentential Codeswitching ?

Bibliography

MacSwan, J. 1999. A minimalist approach to intrasentential code switching. New York: Garland Press.

MacSwan, J. 2000. The architecture of the bilingual language faculty: Evidence from codeswitching. Bilingualism: Language and Cognition, 3(1), 37-54

MacSwan, J. (2004). Code switching and linguistic theory. In T. K. Bhatia & W. Ritchie (Eds.), Handbook of bilingualism. Oxford: Blackwell.

Poplack, S. 1980. "Sometimes I'll start a sentence in Spanish y termino en Español": Toward a typology of code-switching. Linguistics 18:581-618.

Poplack, S. 1981. The syntactic structure and social function of code-switching. In Durán, R., ed. 1981. Latino language and communicative behavior. Norwood, NJ: Ablex.

Steriade, Donca. 1993. “Closure, release and nasal contours.” In Phonetics and Phonology, MarieHuffman and Rena Krakow (eds), 401-470. San Diego, New York: Academic Press.