Download as pdf or txt
Download as pdf or txt
You are on page 1of 17

Code Switching and the Complement/Adjunct Distinction

Author(s): Shahrzad Mahootian and Beatrice Santorini


Source: Linguistic Inquiry, Vol. 27, No. 3 (Summer, 1996), pp. 464-479
Published by: The MIT Press
Stable URL: http://www.jstor.org/stable/4178946 .
Accessed: 14/06/2014 11:53

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to Linguistic Inquiry.

http://www.jstor.org

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions
464 REMARKS AND REPLIES

Code Switching and the Complement/Adjunct Distinction

ShahrzadMahootian
Beatrice Santorini

Belazi, Rubin, and Toribio (1994) propose two universal syntactic


constraints on intrasententialcode switching: the Functional Head
Constraint,which prohibits switches between functional heads and
theircomplements,andthe Word-Grammar IntegrityCorollary,which
requiresall words of a language to obey that language's grammarin
code-switchingcontexts. After rejectingboth constraintson empirical
as well as conceptualgrounds,this articleoutlinesan alternativeanaly-
sis thatrelies only on generalprinciplesof phrasestructureandrejects
constraintsspecific to code switching.The analysisproposedprovides
strongsupportfor the projectionof syntacticstructurefromthe lexicon
and the complement/adjunctdistinction.

Keywords:argument/adjunct distinction,code mixing, code switching,


complement/adjunctdistinction,phrase structuretheory

Belazi, Rubin, and Toribio (1994) propose two universal syntactic constraintson intrasentential
code switching: the Functional Head Constraint,which prohibits switches between functional
heads and their complements, and the Word-GrammarIntegrityCorollary,according to which
all words of a language obey that language's grammarin code-switching contexts., 2 Section 1
reviews counterevidenceto these two constraintsand arguesagainstBelazi, Rubin, and Toribio's
analysis on conceptual grounds. Section 2 outlines an alternativeanalysis of code switching,
developed in more detail elsewhere (Mahootian 1993a,c, to appear, Mahootian and Santorini
1994, Santoriniand Mahootian 1995), that relies only on general principlesof phrase structure
and rejects constraintsspecific to code switching. Section 3 summarizesthe main points of the
discussion.

1 Belazi, Rubin, and Toribio's Analysis


1.1 The Functional Head Constraint
Belazi, Rubin, and Toribio (1994) argue that syntactic constituentsbear a language feature and
that code switching is constrainedby f-selection, the relationshipbetween functionalheads and

We would like to thank Young-Suk Lee, Philip Miller, Michael Niv, Owen Rambow, and three anonymous LI
reviewersfor valuablediscussion and commentson previousdraftsof this article.We are also indebtedto AravindJoshi,
Anthony Kroch, and GregoryWardfor good advice. Needless to say, we alone are responsiblefor any shortcomingsof
the article.
1 The materialpresentedhere is based in large part on Mahootian 1993c:39-40, 114-129.
2
Rubin and Toribio (1995) and Toribio and Rubin (to appear)reformulatethe approachin Belazi, Rubin, and
Toribio 1994 to conform to the frameworkof Chomsky 1993. Since the more recent work presentsno new evidence and
the reformulationdoes not affect the conceptualcore of the analysis, we restrictour discussion to the analysis presented
in Belazi, Rubin, and Toribio 1994.

Linguistic Inquiry,Volume 27, Number 3, Summer 1996


464-479
?) 1996 by the MassachusettsInstituteof Technology

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions
REMARKS AND REPLIES 465

their complements(Abney 1987). Specifically, they propose the constrainton code switching in
(1) (= their (16)).
(1) The Functional Head Constraint
The language featureof the complementf-selected by a functionalhead, like all other
relevantfeatures,must match the correspondingfeature of that functionalhead.
The FunctionalHead Constraintrules out switches between Comp and IP, Infl and VP, and Det
and NP, among others. However, as we will show in the remainderof this section, switches
between functionalcategories and their complementsare well attested.
(2) gives examples of switches between Comp and IP.
(2) a. I seen everything 'cause no cogi na'.
not I took nothing
'I saw everythingbecause I didn't take anything.'
(English-Spanish;Sankoff and Poplack 1981:6)
b. Ye juri vanemud mikone ke I'm stupid.
a way indicate does that
'He acts as if I'm stupid.'
(Farsi-English;Mahootian 1993b)
c. lorsquej'ai vu que mabqas'
when I have seen that there was nothing left
'when I saw that there was nothing left'
(French-MoroccanArabic;Bentahilaand Davies 1983:310, (22))
Belazi, Rubin, and Toribio (1994:226) interpretthe repetitionof the complementizerin the
repairsequence in (3) as evidence for the strengthof the FunctionalHead Constraint.3
(3) C'est le fer qui donne ... illi yi-ha:rib l'anemie.
it's the iron that gives that it-fights the anemia
'It's iron that gives, that fights anemia.'
(French-TunisianArabic;Belazi, Rubin, and Toribio 1994:226, (13a))
But sequences structurallyparallelto the allegedly prohibitedqui yi-ha:rib occur naturally,as in
(4), and arejudged to be "perfectly natural"by MoroccanArabic-Frenchbilinguals (Bentahila
and Davies 1983:311).
(4) bzzaf djal 1 hmi:r daba lli ignorent 1 m-ya:rba
many of the fools now who do not know the Moroccans
'many of the fools now who do not know Moroccans'
(MoroccanArabic-French;Bentahila and Davies 1983:311, (26))

3Following the authors we cite, we use the following phonetic symbols: T = voiceless retroflex plosive, y =
voiced velar fricative, T = voiced pharyngealfricative, h = voiceless pharyngealfricative. We also use the following
abbreviationsin word-for-wordglosses: ART = article, COND = conditional,FOC = focus, NOM = nominative,Poss =
possessive, REFL = reflexive, TOP = topic.

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions
466 REMARKS AND REPLIES

The repetitionof the complementizerin (3) thereforedoes not representconclusive evidence in


favor of the Functional Head Constraint,but might simply reflect the speaker's choice of a
particularrepairstrategyfamiliarfrom monolingualperformance-namely, anticipatoryretracing
instead of instantrepairing,to use the terminologyof Levelt (1989:490). In instantrepairing,the
speaker replaces only the problem element (the verb of the relative clause in (3)), whereas in
anticipatoryretracing,the speakerbacks up past the problemelement in repairingthe error.
(5) shows examples of switches between free morphemes in Infl and VP (note also the
switchesbetweenCompandIP in bothexamples).Forfurtherexamples,see the studyof Moroccan
Arabic-French code switching by Bentahilaand Davies (1983:315, (52)-(55)).
(5) a. No, parce que hanno donne des cours.
no because have given of the lectures
'No, because they gave lectures.'
(Italian-French;Di Sciullo, Muysken, and Singh 1986:15, (37a))
b. Oui, alors j'ai dit que si potev aller comme ca.
yes so I have said that REFL could walk like that
'Yes, so I said that we could go like that.'
(French-Italian;Di Sciullo, Muysken, and Singh 1986:15, (37b))
Belazi, Rubin, and Toribio's analysis (1994:231) also prohibitsswitches involving bound mor-
phemes in Infl. Indeed,in light of numerousproposalsto representboundinflectionalmorphemes
as independentfunctionalheads in the syntax, they propose extendingthe FunctionalHead Con-
straintto subsume the Free MorphemeConstraint,which prohibitsswitches between "a bound
morphemeand a lexical formunless the latterhas been phonologicallyintegratedinto the language
of the boundmorpheme" (Sankoff and Poplack 1981:5). However, if the syntacticrepresentation
of inflectional morphology is on the right track,then Belazi, Rubin, and Toribio's extension of
the FunctionalHead Constraintis actuallyunderminedby the extensive body of counterevidence
against the Free MorphemeConstraint(Clyne 1967, 1987, Nartey 1982, Bentahila and Davies
1983, Nishimura 1985, Bokamba 1988, 1989, Eliasson 1989, Stenson 1990, Mahootian 1993c,
Myers-Scotton1993).
Switches are also possible between determiners(including quantifiersand numerals) and
nominal projections,as in (6) and (7).
(6) a. I'll take some naemaek.
salt
(English-Farsi;Mahootian 1993c:121, (67))
b. In kitchen xeyli kaesif- e.
this very dirty is
'This kitchen is very dirty.'
(Farsi-English;Karimi 1990:12, (6))
c. those ahoo- ga iru kara
idiots NOM be from
'since those idiots are around'
(English-Japanese;Nishimura 1985:202)

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions
REMARKS AND REPLIES 467

(7) a. vaeli aegaer tu- ye ye dangerous situation bezarisun


but if inside Poss a you put them
'but if you put them in a dangeroussituation'
(Farsi-English;Mahootian 1993b)
b. ga c'est le highest class fran9ais.
that that is the French
'That's the Frenchupperclass.'
(French-English;Brown 1986:404, cited in Myers-Scotton 1993:152, (40a))
For furtherexamples of switches after determiners,see Bentahila and Davies 1983:316-317,
321, (64)-(71), (91)-(92), Mahootian 1993c:119-120, (58)-(60), (62)-(63), Pfaff 1979:306,
(52)-(53), and Woolford 1983:527, 533, (9)-(10), (53)-(54).
A particularlytelling piece of evidence againstthe FunctionalHead Constraintcomes from
code switching between Moroccan Arabic and French. Noting that certain Arabic determiners
take full noun phrasecomplementsratherthan lower nominal projections,Bentahilaand Davies
(1983:317) observe that switches like those in (8) (= their (69)-(71)) are possible and indeed
common; see also Bentahila and Davies 1992:449-451.
(8) a. dak la chemise
that the shirt
'that shirt'
b. wahed le liquide
one the liquid
'some liquid'
c. wahed une cousine
one a cousin
'a cousin'
MoroccanArabic-Frenchexamples of this sort are also reportedto be common by Nait-M'barek
and Sankoff (1988:148-150). Since analyzing the full noun phrase complement of the special
determineras a borrowing is unmotivated,as Nait-M'barekand Sankoff (1988:149) note, the
FunctionalHead Constraintis unable to accommodatethese switches.

1.2 The Word-GrammarIntegrityCorollary


Not all switches fall under the purview of the FunctionalHead Constraint.In orderto account
for contrasts like the one in (9), Belazi, Rubin, and Toribio (1994:232) thereforepropose the
additionalcode-switching constraintin (10) (= their (28)).4

4 Belazi, Rubin, and Toribio (1994:232) introduce the Word-GrammarIntegrity Corollary as a corollary of the
association of lexical entries with morphologicaland syntactic features;it is conceptuallyindependentof the Functional
Head Constraint.

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions
468 REMARKS AND REPLIES

(9) a. J'ai une voiture mizyaena.


I have a car nice
'I have a beautifulcar.'
(French-TunisianArabic;Belazi, Rubin, and Toribio 1994:232, (27a))
b. *fand-ikarhba belle.
at-I car nice
'I have a beautiful car.'
(TunisianArabic-French;Belazi, Rubin, and Toribio 1994:232, (27b))
(10) The Word-GrammarIntegrityCorollary
A word of language X, with grammarGx, must obey grammarGx.
According to their analysis, (9a) is acceptablebecause adnominaladjectives in both Frenchand
Arabic follow the noun they modify. By contrast,they rule out (9b) on the groundsthat beau,
belle 'beautiful'belongs to an exceptionalset of Frenchadjectivesthatarerestrictedto prenominal
position.5Hence, the syntacticrequirementsof the Arabic noun are met, but those of the French
adjective are not.
Belazi, Rubin,andToribio(1994) proposeextendingthe Word-Grammar IntegrityCorollary
to rule out English-Marathiswitches like *some chairs-war 'on some chairs' (Joshi 1985:195,
(8c)).6 According to them, "an English NP looks to its left for Case" (1994:233), and Joshi's
sequence is ruled out because the MarathiCase assigneris on the wrong side of the noun phrase.
In fact, however, examples like Joshi's, in which an English noun phrasereceives Case from a
postposition,are found in naturallyoccurringdata, as shown in (11) (we addressthe discrepancy
between Joshi's introspection-basedjudgmentand the existence of structurallyparallel,naturally
occurringexamples like (11) in section 2.4).7
(11) sorekara, his wife ni yattara
in addition to give-COND
'in addition,if we give it to his wife'
(Japanese-English;Nishimura 1986:129, (5))

I Wilmet
(1981:25) reports33 postnominalout of 447 total instances of beau (7%) in a sample of contemporary
French literature.The postnominal placement of beau cannot be attributedto the register of Wilmet's sources, since
formal French usage favors prenominaladjective placement (Grevisse 1986:537, ?321). Since beau also fails to satisfy
the remainingdiagnosticsfor the class of exceptionalprenominaladjectivesin Romancediscussedby Bernstein(1993)-it
appearsin elliptical constructions,allows modification,and functions as a predicate-we conclude that it is an ordinary
adjective that favors prenominalposition, but is not absolutely restrictedto it. The analysis we present in section 2.3
thereforeallows (9b); see section 2.4 for discussion of the discrepancybetween Belazi, Rubin, and Toribio's analysis
and ours.
6 According to Belazi, Rubin, and Toribio (1994:233), the unacceptabilityof some chairs-war 'on some chairs'
"led Joshi to posit his Constrainton Closed-Class Items." In fact, Joshi's constraintfails to rule out this sequence, as
he himself notes (1985:196). Joshi's tentative solution-namely, to disallow switches between noun phrases and case
markers-incorrectly fails to accommodateexamples like (11).
7 For furtherswitches involving Case-assigningpostpositions,see Nishimura1985, 1986 and Sankoff, Poplack, and

Vanniarajan1990.

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions
REMARKS AND REPLIES 469

In addition to ruling out switches like the one in (11), Belazi, Rubin, and Toribio's argument
based on Case assignmentincorrectlyprohibitsexamples like (12a-b), where noun phrasesfrom
a VX language occupy the preverbalslot of XV clauses.8
(12) a. Ten dollars dade.
gave
'She has given ten dollars.'
(English-Farsi;Mahootian 1993c:102, (21))
b. Only small prizes morattane.
got you know
'We got only small prizes, you know.'
(English-Japanese;Nishimura 1986:128, (3))
Violations of the Word-GrammarIntegrity Corollary involving adjectives are attested as
well, as shown in (13).
(13) a. Adjectivefrom N-Adj language, nounfrom Adj-N language
I got a lotta blanquitofriends.
whitey
(English-Spanish;Poplack 1980:600, (16b))
b. Nounfrom N-Adj language, adjectivefrom Adj-N language
p6sta le cailin Brazilian
marriedwith girl
'marriedto a Brazilian girl'
(Irish-English;Stenson 1990:171, (7b))
c. Adjectivefrom Adj-N language, nounfrom N-Adj language
Ma ci stanno dei smart italiani.
but there are of-the Italians
'But there are smartItalians.'
(Italian-English;Di Sciullo, Muysken, and Singh 1986:15, (40a))
d. Nounfrom Adj-N language, adjectivefrom N-Adj language
He presenteda paper exceptionnel.
exceptional
(English-French;Bokamba 1989:282, (16a))
In conclusion, since the Word-GrammarIntegrityCorollaryrequiresthe placementof each
single word of a language to be consistent with that language's grammar,it reduces in effect to
the well-known Equivalence Constraint,accordingto which "the orderof sentence constituents
immediatelyadjacentto and on both sides of the switch point must be grammaticalwith respect
to both languagesinvolved simultaneously"(SankoffandPoplack 1981:5). As a result,the Word-

' We use the traditionalterms VX and XV, setting aside the proposal in Kayne 1994 to derive XV orders from a
universalVX base.

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions
470 REMARKS AND REPLIES

GrammarIntegrityCorollary,like the EquivalenceConstraint,is incompatiblewith the largebody


of evidence that has been accumulatedagainstthe latter(Sridharand Sridhar1980, Nartey 1982,
Bentahilaand Davies 1983, Nishimura1985, 1991, Berk-Seligson 1986, Pandit 1986, 1990, NaYt-
M'barekand Sankoff 1988, Bokamba1988, 1989, Stenson 1990, 1991, Mahootian1993c, Myers-
Scotton 1993, and Poplack and Meechan 1995; see also Poplack 1980:600).

1.3 ConceptualConsiderations
Having demonstratedthe empiricalinadequacyof the FunctionalHead Constraintand the Word-
GrammarIntegrityCorollary,we turnnow to two conceptualshortcomingsof Belazi, Rubin, and
Toribio's analysis. First, the notion of language feature,upon which the FunctionalHead Con-
straintrelies, is suspect in any version of a principles-and-parameters approachto syntax from
Chomsky 1981 onward. In contrastto the pretheoreticalnotion of E-language, I-language, the
object of concernto linguists in such an approach(Chomsky 1986), is not a theoreticalprimitive,
but the space of possible combinationsof linguistic parameters.Since the particularparameter
combinationshave no theoreticallyprivilegedstatusof theirown, referenceto them in the form of
a languagefeatureamountsto a significantweakeningof the principles-and-parameters approach.
Second, the notion of language featureviolates conceptualeconomy quite independentlyof
whethera principles-and-parameters approachto syntaxis correct.The FunctionalHeadConstraint
requiresfunctionalheads andtheircomplementsin code-switchingsentencesto agree with respect
to languagefeature.Belazi, Rubin, and Toribio arguethat such agreementis trivially satisfied in
monolingualsentences and conclude that their analysis succeeds in eliminatingmechanismsspe-
cific to code switching. But the notion of language featureis not itself independentlymotivated
(Mahootian1993c:39-40).
In conclusion, then, Belazi, Rubin, and Toriblo's proposalfalls short of their stated goal of
providingan analysis accordingto which "patternsof code switching are ... constrainedsolely
by Universal Grammar"(1994:234).

2 Code Switching and the Complement/Adjunct Distinction


In this section we outlinean alternativeanalysisthatrelies on generalprinciplesof phrasestructure
ratherthan on constraintsthat are specific to code switching, such as the FunctionalHead Con-
straintor the Word-Grammar IntegrityCorollary(see also Mahootian1993a,c, to appear,Mahoo-
tian and Santorini1994, Santoriniand Mahootian1995). The basic insight of the analysis is given
in (14) (see also Finer 1990:918-922, Lee 1986, 1990 (cited in Finer 1990), Pandit 1990:43).
(14) Heads determinethe syntacticpropertiesof their complementsin code-switchingand
monolingualcontexts alike.
Sections 2.1 and 2.2 illustratethe importof (14) for functionaland lexical heads in turn.Section
2.3 providesan analysis of code switchingin modificationstructures.Finally, section 2.4 suggests
an explanationfor the discrepancybetween Belazi, Rubin, and Toribio's data and the data upon
which our own analysis is based.

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions
REMARKS AND REPLIES 471

2.1 Switches Involving Functional Heads


An analysisbased on (14) straightforwardlyaccountsfor switches involving functionalheads like
the one in (15).
(15) E wo green dress ko.
he/she PAST tone wear ART
'(S)he wore a green dress.'
(Adaime-English; Nartey 1982:187, (6))
First, it allows such switches in principle, contraryto the FunctionalHead Constraint.Second,
the orderof the determinerand its complementfollows from (14): since the Adarme determiner
ko is postnominal,its English complementgreen dress is requiredto precedeit, just as an Addrjme
complementwould be requiredto in monolingualAdaime discourse.By contrast,(14) disallows
the hypotheticalswitch in (16), a consequencethatis consistentwith the absenceof such examples
from the code-switching literature.
(16) *E wo ko green dress.
he/she PAST tone wear ART
'(S)he wore a green dress.'
Plainly, all the switches between functionalheads and their complementsreviewed in section 1.1
are consistentwith (14), as are the switches that led Belazi, Rubin, and Toribio (1994) to propose
the FunctionalHead Constraint.(As noted above, we propose an explanationfor the discrepancy
between their data sets and ours in section 2.4.)

2.2 Switches InvolvingLexical Heads


In contrast to what happens under Belazi, Rubin, and Toribio's analysis, under our analysis
lexical heads behave no differentlyin code-switchingcontextsthanfunctionalheads. For instance,
consider code switching between a VX language like English and an XV language like Farsi,
Hindi, Japanese, or Tamil. With such a combinationof languages, there are four hypothetical
switches between a verb and its object: a verb from a VX language could precede or follow an
object from an XV language, and a verb from an XV languagecould precede or follow an object
froma VX language.Of these fourhypotheticalswitches, only those areattestedthatareconsistent
with the verb's requirementsregarding the position of its complement. A verb from an XV
language can therefore follow a complement from a VX language, as already shown in (12).
Conversely, a verb from a VX language can precede a complementfrom an XV language, as in
(17).
(17) a. You'll buy xune- ye jaedid.
house Poss new
'You'll buy a new house.'
(English-Farsi;Mahootian 1993c:152, (106))

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions
472 REMARKS AND REPLIES

b. Nisei no jidai ni wa we never knew annakoto nanka.


Nisei Poss days in TOP such things FOC
'In the days of the Nisei (second-generationJapanese), we never knew such a
thing.'
(Japanese-English;Nishimura 1985:76, (74))
Further,in what might be called lenient XV languages-namely, ones that allow complements
to follow verbs in monolingualdiscourse-a verb from the XV language can precede a comple-
ment from a VX language, as in (18).
(18) Avanga vantu keeTuTu iruntaangaquestions.
they filler asking were
'They were asking questions.'
(Tamil-English;Sankoff, Poplack, and Vanniarajan1990:79, (2))
However, (14) correctlyrules out sequences consisting of a complementfrom an XV language
preceding a verb from a VX language, as in (19), which to our knowledge are not attested.
(19) *You'll xune- ye jaedid buy.
house Poss new
'You'll buy a new house.'
Although we follow Belazi, Rubin, and Toribio (1994) and much of the code-switching
literaturein focusing on phrasestructureposition in this article,the syntacticpropertiesreferredto
in (14) includeothergrammaticalfeaturessuch as syntacticcategoryandfiniteness.Forinstance,if
a verb from one language subcategorizesfor an infinitival complement, our analysis correctly
prohibitsit from taking a finite complementfrom anotherlanguage in code-switching contexts
(Bentahilaand Davies 1983:322-324).
In summary,we would like to emphasize that our analysis does not allow code switching
at every phrasestructurenode. Rather,heads play a pivotal role, imposing their syntacticrequire-
ments and therebydeterminingthe phrasestructureposition, syntacticcategory, and featurecon-
tent of their complements.

2.3 ModificationStructures
We turn now to the analysis of modification structuresinvolving adnominaladjectives,9which
has long representedone of the most vexed issues in the study of code switching (Stenson 1990:
186, 189-190). A number of previous studies have proposed a head-complementanalysis of

9 To the best of our knowledge,the datafor adjunctsotherthanadnominaladjectivesareconsistentwith our analysis.


The reason we focus on adnominaladjectives is that the adjective-nouncase contrastsparticularlyclearly with the head-
complementcase. AdverbialVP adjunctsenjoy significantlygreaterword orderfreedom than adnominaladjectives,and
their placement is thereforeoften consistent with a head analysis of them. A test case of considerableinterest for our
analysis concerns switches between relative clauses and the nouns they modify in languages with pre- and postnominal
relative clauses, but we have found no relevantexamples in the literature.

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions
REMARKS AND REPLIES 473

nouns modified by adjectives (or more precisely, adjective phrases-see (21d)). Both logical
possibilities are discussed in the literature:according to Aguirre (1976), Wentz and McClure
(1976), Wentz (1977), and Bentahilaand Davies (1983), such modificationstructuresare headed
by the adjective, whereas according to Di Sciullo, Muysken, and Singh (1986:9) and Pandit
(1990), they are headed by the noun. As in the case of heads and complements, there are four
hypotheticalswitches between adnominaladjectives and nouns: an adjectivefrom an Adj-N lan-
guage like Englishl could precedeor follow a noun from an N-Adj languagelike Irish,Romance,
or Swahili, and an adjective from an N-Adj language could precede or follow a noun from an
Adj-N language. But contraryto the head-complementcase, all four potential sequences are
attested,as alreadyshown in (13) in connectionwith our critiqueof the Word-Grammar Integrity
Corollary. If adjective-noun sequences were headed by adjectives, (13a-b) shouldn't occur,
whereas if they were headed by nouns, (13c-d) shouldn't.
We conclude from the availabilityof all four switches in (13) that modificationcannot be
forced into the mold of the head-complementrelation, but is better representedby adjunction
structures.'1For instance,the mixed Irish-Englishnoun phrasein (13b), repeatedhere, is derived
by Chomsky-adjoiningthe AdjP node dominatingBrazilian to the NP node dominatingcailin
'girl'.
(20) posta le cailin Brazilian
marriedwith girl
'marriedto a Braziliangirl'
(Irish-English;Stenson 1990:171, (7b))
Since there is no head in this case to unilaterallydeterminethe order of the modifying and the
modified phrases, we conclude that the direction of adjunctioncan be determinedby either of
the grammarsof the languagesinvolved in code switching.In (20) it is the grammarof Irish that
determinesthe order of the resulting NP cailin Brazilian 'Brazilian girl', which then serves as
the complementof an Irish (null) determiner.
Since we place no constraintson Chomsky-adjunctionthat are specific to code switching,
adjectives and nouns from one language are free to appear in an order unique to the other.
This strikingconsequence, which is completely unexpected given the Word-GrammarIntegrity
Corollary,is borne out by the data, as illustratedin (21).
(21) a. E he house red O.
he/she PAST tone buy ART
'(S)he bought the red house.'
(Adaime-English; Nartey 1982:187, (5))

10
As is usual in the code-switchingliterature,we abstractaway from the availabilityof postnominaladjectivephrases
in English when the adjective phrase is sufficiently heavy.
1 We restrictour remarkshere to ordinarymodification.As is well known, certain adjectives-or more typically,
certaininterpretationsof such adjectives-exhibit the syntacticbehaviorof heads (see Bernstein 1992, 1993 for indepen-
dent motivationfor distinguishingadjectival heads from ordinaryadjectives, which are phrasaladjuncts).As we argue
in Santoriniand Mahootian 1995, the behavior of adjectivalheads in code-switching contexts is consistent with (14).

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions
474 REMARKS AND REPLIES

b. ar nos dha'chat ar roof galvanized


like two cats on a
'like two cats on a galvanized roof'
(Irish-English;Stenson 1991:571, (1ld))
c. Hao wanataka timingproper.
those they want
'Those [people] want propertiming.'
(Swahili-English;Myers-Scotton1993:87, (9))
d. Nikaona i- na taste lousy sana.
I thought it with very
'I thoughtit had a very lousy taste.'
(Swahili-English;Myers-Scotton1993:29, (10))
To our knowledge, correspondingexamples involving heads and complements are not attested.
The asymmetryin this regardbetween heads and complementson the one hand and adjectives
and nouns on the other thereforeprovides strong supportfor the analysis that we have presented
here.
In conclusion, the emphasis that we place on the complement/adjunctdistinctionis unique
in the code-switching literature(though see Woolford 1983:530 for mention of the distinction).
A convenient summaryof our analysis is that "heads dictate what goes in head-complement
structures,but anythinggoes elsewhere."

2.4 Naturally Occurringversus Elicited Data


The analysisjust presentedallows all of the examples that Belazi, Rubin, and Toribio's analysis
does. In addition, it allows all but one of the examples in their article that they rule out as
impossible.12 In our view, the discrepancybetween the two analyses stems from the fact that the
dataupon which we rely are naturallyoccurringexamples, whereasBelazi, Rubin, and Toribio's
data"were collected by eliciting grammaticalityjudgmentsand ... furthercorroboratedby record-
ings of naturalconversation" (1994:222, fn. 2). Like Belazi, Rubin, and Toribio, Bentahilaand
Davies (1983) also used a combinationof naturallyoccurringdata and elicited judgmentsas the
basis of their study of code switching, but they adopteda more nuancedmethodology (p. 307).

We feel thereforethatone should be cautiousin using informationobtainedin this way [by elicitation]
to draw conclusions about syntactic constraintson code-switching. While respondents'judgements
that an utteranceinvolving a switch sounds perfectly naturalmay be consideredto suggest that there
are no syntactic constraintsagainst such switches, we should not be so quick to assume that their
rejection of other utterancesas odd is evidence for the existence of some syntactic constraint.

Precisely the same methodological concerns regardingthe reliability of elicited judgments of


12
Underthe standardassumptionthatclitics are heads, Belazi, Rubin,and Toribio's analysis and ours agree in ruling
out the English-Spanishswitch *I saw lo 'I saw him/it' (= their (8b)) by the Word-GrammarIntegrityCorollaryand
the principlein (14), respectively.

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions
REMARKS AND REPLIES 475

unacceptabilityare voiced in connection with the study of nonstandardmonolingual speech by


Henry (1992:281-282, 1995:12); see also Rickford 1975. The social unacceptabilityof code
switching for many bilingual speakers,togetherwith the difficulty of providinga motivationfor
switches in artificially constructedexamples, has the consequence that elicitation introducesa
restrictive bias into the data. Indeed, speakers are reportedto "reject as impossible the very
sentences they have been recordedusing" (Pfaff 1979:301, fn. 7); for similar observations,see
Mahootian 1993c:7-8, Pandit 1990:63, fn. 21, and Sankoff, Poplack, and Vanniarajan1990:75.
Given Belazi, Rubin, and Toribio's methodologicalbias, the broaderempiricalcoverage of our
analysis is thereforeexpected.

3 Conclusion
In this response to Belazi, Rubin, and Toribio (1994), we have shown that the two universal
constraintson code switching that they propose-the FunctionalHead Constraintand the Word-
GrammarIntegrity Corollary-are empirically inadequate.We have argued furtherthat these
constraintsare inconsistentwith a principles-and-parametersapproachto syntax as well as with
the theory-neutralaim of deriving attested code-switching patternsfrom general principles of
phrase structure.After presenting an empirically and conceptually more satisfactory analysis,
accordingto which there are no phrase structureconstraintsspecific to code switching, we pro-
posed an independently motivated sociolinguistic explanation for the discrepancies between
Belazi, Rubin, and Toribio's analysis and our own.

Appendix: The Constraint on Closed-Class Items


In what follows, we show that although there is reason to reject Joshi's (1985) Constrainton
Closed-ClassItems on empiricalgrounds,the particularcritiqueof it in Belazi, Rubin,andToribio
1994 is unfounded.Joshi takes the two languages involved in code switching to be of unequal
status: one language is the matrix language, and the other is the embedded language. Switches
are asymmetrical:they are allowed from matrix to embedded language, but not in the opposite
direction.Joshi furtherproposes (22) (= his (5), his italics).
(22) The Constrainton Closed-Class Items
Closed-classitems (e.g., determiners,quantifiers,prepositions,possessive, Aux, Tense,
helping verbs, etc.) cannot be switched.
As Joshi notes (1985:194), he intends the constraintin (22) to rule out switches of closed-class
items only when they are not partsof longer sequences in the same language as the closed-class
item.
Belazi, Rubin, and Toribio (1994) reject the Constrainton Closed-Class Items on the basis
of the examples in (23).
(23) a. Safae:t ni-tkalmuSal l'anemie.
sometimes we-speak about the anemia
'Sometimes we speak about anemia.'
(TunisianArabic-French;Belazi, Rubin, and Toribio 1994:227, (15a))

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions
476 REMARKS AND REPLIES

b. J'ai joue avec il-ku:ra.


I have played with the-ball
'I played with the ball.'
(French-TunisianArabic;Belazi, Rubin, and Toribio 1994:227, (15b))
In both cases they take the switch between the prepositionand its complementto be problematic.
However, since the prepositions are in the matrix language and not switched from it, neither
example actually violates Joshi's constraint.13
Whereasthe examples in (23) turnout not to bear on the Constrainton Closed-ClassItems,
a numberof authorshave observedthatthe constraintis indeedempiricallyinadequate(Nishimura
1985, Pandit 1990, Mahootian1993c, Myers-Scotton1993). Some counterexamplesare given in
(24); see also (5a). 4
(24) a. Complementizers
Anyway, I figured ke if I worked hard enough, I'd finish in the summer.
that
(English-Farsi;Mahootian1993c:32, (22))
b. Conjunctions
Look, a Jew was: a doctor, a lawyer, oder a businessman.
or
(English-Yiddish;Prince and Pintzuk 1984:8, (16g))
c. Determiners
Where are they, los language things?
the
(English-Spanish;Poplack 1981:175)
d. Prepositions
Maen throughtaemam-e ina raeftaem.
I all Poss these I went
'I've gone throughall of this.'
(Farsi-English;Mahootian 1993c:111, (38))
It is worthnoting that in view of the extensive overlapbetween closed-class items and functional
heads, counterexamplesto the Constrainton Closed-ClassItems are generallyalso counterexam-
ples to the FunctionalHead Constraint(though not conversely).

13
We assume that the matrix language is the language of the tensed verb. In the opposite case, Joshi's analysis
would indeed rule out the examples in (23), but by his assumptionthat switching is asymmetrical,not by the Constraint
on Closed-Class Items.
14 Attempting to salvage the Constraint on Closed-Class Items by deriving examples like (24a-d) with two
switches-a higher switch of the constituentimmediately dominatingthe closed-class item, and a lower switch of the
closed-class item's sister-is inconsistent with Joshi's assumption that switching is asymmetrical.Depending on the
choice of matrix and embedded language, it is the higher or the lower switch that is problematic (Nishimura 1985,
Mahootian 1993c).

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions
REMARKS AND REPLIES 477

References
Abney, Steven. 1987. The Englishnounphrasein its sententialaspect.Doctoraldissertation,MIT,Cambridge,
Mass.
Aguirre,Adalberto,Jr. 1976. Acceptabilityjudgementsof code-switchingphrasesby Chicanos:Some prelim-
inary findings. ERIC ED 129 122.
Belazi, Hedi M., EdwardJ. Rubin,andJacquelineAlmeidaToribio. 1994. Code switchingand X-bartheory:
The FunctionalHead Constraint.LinguisticInquiry25:221-237.
Bentahila,Abdelali, and Eirlys E. Davies. 1983. The syntax of Arabic-Frenchcode-switching.Lingua 59:
301-330.
Bentahila, Abdelali, and Eirlys E. Davies. 1992. Code-switching and language dominance. In Cognitive
processing in bilinguals, ed. RichardJackson Harris,443-458. Amsterdam:North-Holland.
Berk-Seligson, Susan. 1986. Linguistic constraintson intrasententialcode-switching:A study of Spanish-
Hebrew bilingualism.Language in Society 15:313-348.
Bernstein, Judy. 1992. On the syntactic status of adjectives in Romance. In CUNY Forum 17, 105-122.
Departmentof Linguistics, City University of New York.
Bernstein, Judy. 1993. Topics in the syntax of nominal structureacross Romance. Doctoral dissertation,
City University of New York.
Bokamba,EyambaG. 1988. Code-mixing, language variation,and linguistic theory:Evidence from Bantu
languages.Lingua 76:21-62.
Bokamba,EyambaG. 1989. Are there syntactic constraintson code-mixing? WorldEnglishes 8:277-293.
Brown, Becky. 1986. Cajun/Englishcode-switching:A test of formal models. In Diversity and diachrony,
ed. David Sankoff, 399-406. Amsterdam:Benjamins.
Chomsky, Noam. 1981. Lectures on governmentand binding. Dordrecht:Foris.
Chomsky, Noam. 1986. Knowledge of language: Its nature, origin, and use. New York: Praeger.
Chomsky, Noam. 1993. A minimalistprogramfor linguistic theory. In The viewfrom Building 20: Essays
in linguistics in honor of Sylvain Bromberger,ed. Kenneth Hale and Samuel Jay Keyser, 1-52.
Cambridge,Mass.: MIT Press.
Clyne, Michael. 1967. Transferenceand triggering. The Hague: Nijhoff.
Clyne, Michael. 1987. Constraintson code-switching:How universalare they? Linguistics 25:739-764.
Di Sciullo, Anne-Marie,PieterMuysken,and RajendraSingh. 1986. Governmentand code-switching.Jour-
nal of Linguistics 22:1-24.
Eliasson, S. 1989. English-Maorilanguage contact:Code-switchingand the Free-MorphemeConstraint.In
Reportsfrom Uppsala UniversityDepartmentof Linguistics 18, 1-28. Departmentof Linguistics,
Uppsala University.
Finer,Daniel. 1990. Modularityand lexical parameterizationin the adultgrammar.Linguistics28:905-927.
Grevisse, Maurice. 1986. Le bon usage: Grammairefranfaise. 12th ed., revised by Andre Goosse. Paris:
Duculot.
Henry, Alison. 1992. Infinitives in a for-to dialect. Natural Language & Linguistic Theory 10:279-301.
Henry, Alison. 1995. Belfast English and standardEnglish: Dialect variation and parametersetting. New
York: Oxford University Press.
Joshi, AravindK. 1985. Processing of sentences with intra-sententialcode-switching. In Natural language
processing: Psycholinguistic, computationaland theoreticalperspectives, ed. David Dowty, Lauri
Karttunen,and Arnold Zwicky, 190-205. New York: CambridgeUniversity Press.
Karimi,Ezat. 1990. Persian-Englishcode-switching. Ms., University of Texas, Austin.
Kayne, RichardS. 1994. The antisymmetryof syntax. Cambridge,Mass.: MIT Press.
Lee, Mee-Hwa. 1986. Code-mixingand binding. Paperpresentedat the Second LanguageResearchForum,
University of SouthernCalifornia.

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions
478 REMARKS AND REPLIES

Lee, Mee-Hwa. 1990. A parametricapproachto code-mixing. Doctoraldissertation,StateUniversityof New


York, Stony Brook.
Levelt, Willem J. M. 1989. Speaking:From intentionto articulation.Cambridge,Mass.: MIT Press.
Mahootian, Shahrzad. 1993a. Bilingual codeswitching: A null theory. Paper presented at NWAVE 22,
University of Ottawa.
Mahootian,Shahrzad.1993b. Farsi/Englishcode-switching corpus. Unpublished.
Mahootian,Shahrzad.1993c. A null theoryof codeswitching.Doctoraldissertation,NorthwesternUniversity,
Evanston, Ill.
Mahootian,Shahrzad.To appear.A competencemodel of codeswitching.In Sociolinguisticvariation:Data,
theory,and analysis. Selectedpapersfrom NWAVE23 at StanfordUniversity.Stanford,Calif.: CSLI
Publications.
Mahootian,Shahrzad,and Beatrice Santorini.1994. Adnominaladjectives, codeswitching and lexicalized
TAG. In 3e colloque internationalsur les grammairesd'arbresadjoints(TAG+ 3), ed. Anne Abeille,
Sophie Aslanides, and Owen Rambow, 73-76. Technical reportTALANA-RT-94-01. TALANA,
Paris.
Myers-Scotton,Carol. 1993. Duelling languages: Grammaticalstructurein codeswitching.Oxford:Oxford
University Press.
Nait-M'barek,M., and David Sankoff. 1988. Le discours mixte arabe/fran9ais:Des empruntsou des alter-
nances de langue? CanadianJournal of Linguistics/Revuecanadienne de linguistique33:143-154.
Nartey, Jonas. 1982. Code-switching,interferenceor faddism: Language use among educated Ghanaians.
AnthropologicalLinguistics 24:183-193.
Nishimura,Miwa. 1985. Intrasententialcode-switchingin Japaneseand English. Doctoraldissertation,Uni-
versity of Pennsylvania,Philadelphia.
Nishimura, Miwa. 1986. Intrasententialcode-switching: The case of language assignment. In Language
processing in bilinguals: Psycholinguistic and neuropsychologicalperspectives, ed. Jyotsna Vaid,
123-143. Hillsdale, N.J.: LawrenceErlbaum.
Nishimura,Miwa. 1991. Varieties of Japanese/Englishbilingual speech: Implicationsfor theories of code-
switching and borrowing.Ms., GeorgetownUniversity, Washington,D.C.
Pandit,Ira. 1986. Hindi English code switching:Mixed Hindi English. Delhi: Datta Book Centre.
Pandit, Ira. 1990. Grammaticalityin code switching. In Codeswitchingas a worldwidephenomenon,ed.
Rodolfo Jacobson, 33-69. New York: Peter Lang.
Pfaff, Carol. 1979. Constraintson language mixing. Language 55:291-318.
Poplack,Shana. 1980. "Sometimes I'll starta sentencein Spanishy terminoen espaniol":Towarda typology
of code-switching.Linguistics 18:581-618.
Poplack, Shana. 1981. Syntactic structureand social function of code-switching. In Latino language and
communicativebehavior, ed. RichardP. Duran, 169-184. Norwood, N.J.: Ablex.
Poplack, Shana, and MarjoryMeechan. 1995. Patternsof language mixture:Nominal structurein Wolof-
French and Fongbe-Frenchbilingual discourse. In One speaker, two languages: Cross-disciplinary
perspectives on code-switching,ed. Lesley Milroy and PieterMuysken, 199-232. Cambridge:Cam-
bridge University Press.
Prince, Ellen F., and Susan Pintzuk. 1984. Bilingual code-switching and the open/closed class distinction.
Ms., University of Pennsylvania,Philadelphia.
Rickford, John. 1975. Carryingthe new wave into syntax: The case of Black English BIN. In Analyzing
variationin language,ed. RalphW. Fasold andRogerW. Shuy, 162-183. Washington,D.C.: George-
town University Press.
Rubin, Edward J., and Almeida JacquelineToribio. 1995. Feature checking and the syntax of language
contact. In Contemporaryresearch in Romance linguistics, ed. Jon Amastae, GrantGoodall, Mario
Montalbetti,and MariannePhinney, 177-185. Amsterdam:Benjamins.

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions
REMARKS AND REPLIES 479

Sankoff, David, and ShanaPoplack. 1981. A formal grammarfor code-switching.Papers in Linguistics 14:
3-45.
Sankoff,David, ShanaPoplack,andSwathiVanniarajan.1990. The case of the nonce loan in Tamil.Language
Variationand Change 2:71-101.
Santorini,Beatrice, and ShahrzadMahootian. 1995. Codeswitching and the syntactic status of adnominal
adjectives.Lingua 95:1-27.
Sridhar,S. N., and Kamal Sridhar.1980. The syntax and psycholinguisticsof bilingualcode-mixing. Cana-
dian Journal of PsychologylRevuecanadienne de psychologie 34:407-416.
Stenson, Nancy. 1990. Phrasestructurecongruence,government,and Irish-Englishcode switching. In The
syntaxof the modernCeltic languages, ed. RandallHendrick,167-197. San Diego, Calif.: Academic
Press.
Stenson, Nancy. 1991. Code-switchingvs. borrowingin modern Irish. In Language contact in the British
Isles: Proceedings of the EighthInternationalSymposiumon Language Contactin Europe,Douglas,
Isle of Man, 1988, ed. P. StureUrelandand George Broderick,559-579. Tiibingen:Max Niemeyer.
Toribio, Almeida Jacqueline,and EdwardJ. Rubin. To appear.Code-switchingin generative grammar.In
Spanish in contact, ed. Ana Roca and John Jensen. Somerville, Mass.: CascadillaPress.
Wentz, Jim. 1977. Some considerationsin the development of a syntactic descriptionof code-switching.
Doctoral dissertation,University of Illinois at Urbana-Champaign.
Wentz, Jim, and Erica McClure. 1976. Ellipsis in bilingual discourse. In Papers from the 12th Regional
Meeting, Chicago Linguistic Society, 656-665. Chicago Linguistic Society, University of Chicago,
Chicago, Ill.
Wilmet, M. 1981. La place de l'epithete qualificative en frangais contemporain:Etude grammaticaleet
stylistique.Revue de linguistiqueromane 45:17-73.
Woolford, Ellen. 1983. Bilingual code-switching and syntactic theory. LinguisticInquiry 14:520-536.

(Mahootian)
Departmentof Linguistics
NortheasternIllinois University
5500 North St. Louis Avenue
Chicago, Illinois 60625-4699
s-mahootian@neiu.edu

(Santorini)
Departmentof Linguistics
NorthwesternUniversity
2016 SheridanRoad
Evanston,Illinois 60208-4080
b-santorini@nwu.edu

This content downloaded from 91.229.248.152 on Sat, 14 Jun 2014 11:53:23 AM


All use subject to JSTOR Terms and Conditions

You might also like