LINGUIST List 12.1550

Tue Jun 12 2001

Review: Muysken, Bilingual Speech (2nd review)

Editor for this issue: Terence Langendoen <terrylinguistlist.org>


What follows is another discussion note contributed to our Book Discussion Forum. We expect these discussions to be informal and interactive; and the author of the book discussed is cordially invited to join in. If you are interested in leading a book discussion, look for books announced on LINGUIST as "available for discussion." (This means that the publisher has sent us a review copy.) Then contact Simin Karimi at siminlinguistlist.org or Terry Langendoen at terrylinguistlist.org.

Directory

  1. Nkonko Kamwangamalu, Re: Book Review: Muysken (2000)

Message 1: Re: Book Review: Muysken (2000)

Date: Sat, 02 Jun 2001 13:33:27 +0200
From: Nkonko Kamwangamalu <KAMWANGAnu.ac.za>
Subject: Re: Book Review: Muysken (2000)

Muysken, Pieter (2000) Bilingual Speech: A Typology of Code-mixing,
Cambridge University Press, xvi+306 pp., ISBN 0521771684 Hardback.


Reviewed by: Nkonko M Kamwangamalu, University of Natal, Durban, South
Africa

BILINGUAL SPEECH: A TYPOLOGY OF CODE-MIXING, is perhaps the latest
addition to publications in the ever growing field of codeswitching.
The striking feature of this book is the extensive amount of data it
covers and the depth of analysis it offers. The central question
addressed in the book is "how can a bilingual speaker combine elements
from two languages when processing mixed sentences (p. 1). Muysken
addresses this question by drawing on various case studies, hence the
subtitle "A typology of code-mixing", from recent research on
codeswitching. In so doing, he aims to provide a general account, not a
single model, of codemixing across languages and cultures. Three types
of code-mixing are distinguished, insertion, alternation, and
congruent lexicalization. The bilingual's choosing of one type of code-
mixing or another is said to depend on the grammatical structures of
the languages involved as well as on sociolinguistic and
psycholinguistic factors. The book consists of nine chapters, the last
of which is followed by an extensive list of references and an author
index and a subject index. In what follows I provide a brief summary of
each of the chapters. This will be followed by a general evaluation of
the book.

Chapter 1, 'The study of code-mixing', presents the general aim of the
book, namely, to explain 'how bilingual speakers switch from one
language to another in the course of conversation". In an attempt to
address this question, Muysken distinguishes between the three types of
codemixing patterns already mentioned: (a) insertion, (b) alternation,
and (c) congruent lexicalization. Both 'insertion' and 'alternation'
focus on structural constraints on mixing. The former views constraints
in terms of the structural properties of one language, the matrix
language; the latter [alternation] views constraints in terms of
structural equivalence between the languages involved (i.e., switching
is possible only where it does not violate the structural integrity of
either of the participating languages (Poplack 1980)). 'Insertion' is
said to be akin to spontaneous lexical borrowing and, depending on
languages, it may consist of single bare nouns, bare noun phrases (p.
95) or adverbial phrases (pp. 3, 5). 'Alternation' entails a true
switch from one language to the other, and involves both grammar and
lexicon. 'Congruent lexicalization' refers to a 'a situation where the
participating two languages share a grammatical structure which can be
filled lexically with elements from either language (p. 6). It is said
to resemble style shifting and variation within a language. The second
half of this chapter presents a typology of constraints on
codeswitching, with a focus on Poplack's (1980) Equivalence and
Morpheme constraints; Myers-Scotton's (1993) Matrix Language Frame
Model, and Di Sciullo, Muysken, and Singh's (1986) government
constraint. Also, an attempt is made to bring all these constraints
under one roof, or what Muysken calls 'Unification and the escape hatch
model', which explains how mixing occurs in the three-way codemixing
patterns discussed: insertional codemixing, alternational codemixing,
and congruent lexicalization. The chapter concludes with a description
of the structure and scope of the book.

Chapter 2, Differences and similarities between languages, attempts to
address questions such as the following, among others: Is language a
'fortress', with clearly defined boundaries both socially and
cognitively? If languages are thought of as fortresses, what keep these
fortresses intact? How does code-mixing break into these fortresses?
And how strong are the fences around a given language, both those
deriving from language as an identity carrier and from it being
embedded in a processing system? Muysken addresses these issues against
the background of the distinction Chomsky (1986) makes between I-
language (internal) and E-language (external). He then hypothesizes
that E-language and I-language are perhaps the two factors that
conspire in the building of fortresses. Viewed from this perspective,
the E-language leads to separateness of outer form (defined by
vocabulary, pronunciation and morphology), and the I-language to a
clearly distinct inner form (marked in the syntax and semantics of the
language). However, Muysken presents evidence that I-language does not
always correspond to E-language. The evidence, discussed on pp.43-45,
consists of the following cases: (a) several E-languages correspond to
a relatively coherent I-language (e.g. Hindi and Urdu have one common
I-language); (2) several I-languages correspond to something perceived
as one E-language (e.g. patois, dialect, Quechua); (c) dialect
continuum carved up into E-languages not directly corresponding to I-
language borders (e.g., West Germanic continuum); (d) a bilingual E-
language (i.e. a combination of modules from different languages)
corresponds to components from various I-languages. The chapter then
presents a sketch of the dimensions of typological differentiation
between languages (e.g. word order, word class, syntactic categories,
etc.), and of the history of the generative approach to language
differences. It concludes that codemixing patterns, whether insertion,
alternation or congruent lexicalization have to do with the linguistic
typology of the languages involved.

Chapter 3 discusses 'Insertion', with a focus on its grammatical
dimensions. Two main claims are made at the beginning of this chapter.
The first claim is that phenomena such as borrowing, nonce-borrowing,
and constituent insertion belong together and so are subject to the
same conditions. The second claim is that insertional codemixing can be
appropriately accounted for within the government model proposed in
DiSciullo, Muysken, and Singh (1986). Before Muysken justifies these
claims, he outlines the characteristic features of insertional
codemixing, drawing on Spanish-Quechua codemixing. Also, he raises two
questions: How is a matrix language defined? What is the relationship
between lexical borrowing and insertion? These two questions have been
an object of on-going debate in current research on CM (see, for
instance, Codeswitching List) and so they are indeed worth raising.
With respect to the features of insertional codemixing, Muysken says
that insertions tend to be (a) content words rather than function
words, (b) morphologically integrated constituents, (c) selected
elements (e.g., objects or complements) rather than adjuncts, (d)
nested (i.e., the fragment preceding the insertion and the fragment
following are grammatically related), (e) single, unique,
constituents. The occurrence of either of the features of insertional
codemixing is said to be determined by the matrix language. But how is
the matrix language determined? Muysken lists a number of approaches
that have been proposed in the literature. Some of these include (a)
the language of conversation (the question remains of how we go about
determining the language in question); (b) left-to-righ parsing, which
says that the first word or set of words in the sentence determines the
base language; (c) morpheme-counting, i.e. the language with more
morpheme in the discourse is the ML. Other possibilities listed include
defining the ML from a psycholinguistic or a structural perspective;
with the former referring to the most activated language for the
speaker, and the latter to the language of INFL. Like many CM
researchers before him (e.g. Poplack 1980), Muysken rejects a purely
structural definition of the ML because it leads to circularity. It is
circular to identify a language as a matrix language and then invoke it
to explain the origin of system morphemes such as verbal inflections. I
shall return to this point in the final section of this review,
critical evaluation.

Chapter 4, 'Alternation', presents a typology of alternational code-
mixing patterns based largely on data from French/Dutch mixing in
Brussels (Treffers-Daller, 1994). The patterns identified include,
among others, non-nested mixing, long switches, emblematic or tag-
switching, peripheral switches (e.g. coordination, clefting, left-
dislocation, etc.). Besides these patterns, it is noted that
alternational code-mixing does also have other properties, among them
embedding in discourse, doubling (repetition of a CM structure in both
languages in mixed clauses) and dummy insertion. The first of these
three patterns is illustrated with data from Chinese/English, and the
second and the third with data from Finnish/English. Halfway into the
chapter functional elements such as discourse markers, conjunctions and
adpositions are discussed with reference to codemixing involving
Spanish and an Amerindian language: the Mexican Otomanguean languages
Popoloca and Otomi, and the Andean language Quechua. This is followed
by a discussion of discourse markers in Moroccan Arabic/Dutch
codemixing and, briefly, in French/Swahili codemixing. Drawing on de
Rooij (1996), it is explained that discourse markers occur in
alternational codemixing because they are 'highly salient within the
discourse which they help structure' (p.114). The chapter concludes
with a discussion of word order in alternational codemixing.

Chapter 5 discusses the third major type of codemixing patterns,
'Congruent lexicalization', with a focus on the following issues: (1)
What are the main features of 'congruent lexicalization'? (2) How
different is codemixing from language variation and style shifting? Can
these phenomena be studied in the same theoretical framework?
Additional issues raised in this chapter have to do with language
convergence. In particular, (3) does convergence facilitate codemixing
or does codemixing pave the way for convergence? Muysken first
addresses the question whether there is a difference between variation
and codemixing. He argues that there is no difference between the two
phenomena and so they must be collapsed. The justification for the
proposed collapse, which I will not go into here, is given on pp. 124-
126. As to congruent lexicalization, Muysken notes that it precludes
the notion of 'matrix language', that is, the languages involved share
the grammatical structure of the sentence; any category can be
switched. Against this background, Muysken (pp. 128-134) lists the
following as the main features of 'congruent lexicalization', drawing
on studies of language variation and style shifting: (a) linear and
structural equivalence, (b) mutli-constituent code-mixing, (c) non-
constituent or 'ragged' mixing (Poplack 1980), (d) non-nested a b c
structures, (e) switching of function words (and any other category),
(f) switching of selected elements (e.g. PPs), (g) bi-directional code-
mixing, (h) back-and-forth switches, (i) homophonous diamorphs, and
(j) morphological integration, (k) triggering (of CM by words from the
other language), and (l) mixed collocations and idioms. These features
are illustrated with examples from codemixing in several settings.

Chapter 6 , 'Function words', focuses on the distinction between
lexical and grammatical categories or what Muysken calls 'functional
elements', and the role of the latter in codemixing. The discussion of
function words departs from the assumption that differences between
languages derive from different characteristics of functional elements
(e.g. agreement, tense, modals, auxiliaries, discourse markers, etc.).
In codemixing the functional elements generally 'come' from one
language, the matrix language. Two hypotheses are tested concerning the
role of the functional elements in codemixing. These run as follows:
(a) The functional element effect derives from the special status of
functional elements within the mental lexicon and speech production, as
argued by Myers-Scotton (1993) and Azuma (1993) on the basis of
assumptions made by some psycholinguists. (b) The functional element
effect derives from the lack of equivalence of functional elements
across different languages.

Muysken makes the case for the second hypothesis, noting that the
evidence he offers for it is not conclusive. The discussion of
functional elements begins with the issue of how the elements
themselves can be defined. Several criteria are presented, among them
'open' vs. 'closed' class distinction (e.g., functional elements belong
to closed class; whereas lexical elements (e.g. nouns and verbs) belong
to open class); bound vs. free morpheme distinction (e.g., bound
morphemes are functional elements, whereas free morphemes are not);
autosemantic vs syn-semantic element distinction (the former [e.g.
content words] have concrete meaning and so can be modified, but the
latter [function words] have no independent meaning and so cannot be
modified) (pp. 157), to list just a few. Muysken then surveys a number
of areas in which he says the distinction between lexical and
functional elements plays an important role. These areas include
language mixing, speech production, agrammatic speech, language
development, foreign talk, language change, creolization, and lexical
borrowing. In surveying these areas, Muysken distinguishes between
three different types of functional elements: shifters (e.g., pronouns,
demonstratives, quantifiers, question words), functional categories
proper (e.g. articles, tense markers, agreement, auxiliaries, modals),
and linkers (prepositions, conjunctions, complementizers, connectives).
The discussion then moves on, disjunctively, to the difference between
content and system morphemes, with a focus on plural markers, past and
present participles, and pronouns. The main claim here is that the
distinction between content and system morphemes (i.e. between lexical
and functional elements) should be made in terms of equivalence. (p.
172).

Chapter 7, 'Bilingual Verbs', examines the difference between
alternation, insertion, and congruent lexicalization, with a focus on
the verb. Four types of bilingual complex verbs are distinguished, (a)
inserted verbs: a guest verb is inserted into a position ordinarily
reserved for a host, native, verb; (b) nominalized verbs in a compound:
a guest verb is a nominalized complement to a causative helping host
verb; (c) adjoined verbs: a guest verb is adjoined to a helping verb;
and (d) infinitive verbs: a guest verb is an infinitive and the
complement of a host auxiliary. Data is drawn from several language
pairs to provide evidence for the proposed typology of bilingual
complex verbs. The author points out that no single analysis can
account for all the bilingual complex verb constructions. The chapter
concludes by raising the question whether there exists a 'bilingual
grammar', a question that has been raised in earlier studies of
codemixing (Poplack 1980, Myers-Scotton 1993, Kamwangamalu 1997).

Chapter 8, Variation in mixing patterns, summarizes what is known about
the relation between mixing patterns and extralinguistic factors. The
summary focuses on three key issues: (a) What are the relevant
grammatical and extralinguistic factors influencing the choice of a
mixing pattern? (b) Is it possible to differentiate between bilingual
corpora in terms of the proposed three-way distinction between
insertion, congruent lexicalization, and alternation? (c) What factors
explain best why an individual speech community shows one pattern of
codemixing rather than another one? In order to address these
questions, the author says, and I agree with him (see last section:
critical evaluation), that one must take into account the various
settings in which codemixing takes place: (i) the social definition of
the bilingual situation on the macro-level (e.g., colonial
language/dominated language, migrant communities, bilingualism of
native elites, etc.); (ii) the sociolinguistic difference between
communities on the meso-level (e.g. attitudes towards bilingualism and
codemixing, structure of linguistic domination (transplanted or
endogenous bilingual community), etc.); (iii) duration of language
contact; and (iv) the interactional setting on the micro-level (e.g.
classroom, marketplace, etc). With respect to attitudes, it is
suggested that insertion and congruent lexicalization occur mostly in
communities that are invariably favorable to language mixing. As to
bilingual proficiency, it is noted that more fluent bilinguals tend to
display congruent lexicalization and more complex insertional
(intrasentential) codemixing; while alternation is said to be
characteristic of less fluent bilingual. In terms of dominance in use,
it is suggested that in migrant communities, styles of codemixing will
develop from insertion to alternation to congruent lexicalization. In
this case, there is possibility for shift from one matrix language to
another across generations. And in terms of structural factors,
codemixing involving typological similar languages will tend to lead to
congruent lexicalization; whereas in the reverse case (i.e. dissimilar
languages) insertional patterns will occur. A key point made in this
chapter is that of directionality in codemixing: insertional mixing
tends to be unidirectional (from the host language to an alien one),
alternational mixing tends to be bi-directional (see page 236).

Chapter 9, 'Codemixing, bilingual speech, language change, discusses
codemixing and language change, the relationship between codemixing and
other language contact phenomena (e.g., borrowing, interference, etc.),
and bilingual language use/processing. With respect to bilingual
processing, the author rejects theories that analyze codemixing
(understood as clause internal switching) as an on/off phenomenon; that
is, that the participating languages are accessed sequentially, when
one language is on the other is off and vice versa (e.g. Poplack 1980,
Myers-Scotton 1993). He cites evidence from psycholinguistic research
(e.g. Grosjean, 1996) that calls for the 'bilingual mode' of speech
production during codemixing, in which both languages are active
simultaneously. He stresses modularity, that is, both languages are
accessed at the same time, but different modules of each. The author
presents data from several language pairs in support of the 'bilingual
mode' approach to codemixing processing. In the data he examines
phenomena such as idioms and collocations in bilingual compound verbs,
pro-drop, and delayed lexicalization. He concludes this section on
bilingual processing by explaining how simultaneous access to both
languages may be realized: (a) both languages are fully present; (b)
any one module can be selected from either language, but not from both.
He seems to favor the second option, which he likens to Grosjean's idea
of 'chamber orchestra', "where different languages play different
instruments, as it were, in the chamber orchestra of sentence-
production" (p. 262). The idea of chamber orchestra of sentence-
production entails division of labor for different modules (syntax,
function words, phonology, lexicon, semantics). The author next
discusses language contact and language change, followed by a sketch of
the relation between codemixing and other language contact phenomena.
With respect to language change, Muysken focuses on the notion of
asymmetry between the languages involved in codemixing. Following
Thomason and Kaufmann (1988), he explains that in contact-induced
language change, there is interference from the dominant language
within the subordinate language (i.e., borrowing under maintenance of
L1), or retention of features of the subordinate language while
shifting to the dominant language (i.e. interference under shift to
L2). The chapter concludes with a survey of diachronic aspects of
language contact research, with a focus on relexification, language
genesis, lexical borrowing, second language learning and substrate
influence, convergence due to prolonged existence, and calquing and
imitation of the prestige patterns (pp. 266-274). This is followed by
suggestions for further research in the area of bilingual processing.

CRITICAL EVALUATION:
"Bilingual Speech" provides codemixing researchers with a wealth of
data on codemixing patterns across languages. It is, to my knowledge,
the second study that offers such a comprehensive typology of
codemixing, the first one (again to my knowledge) being Kamwangamalu's
(1989) doctoral dissertation entitled "Codemixing across languages:
structure, functions, and constraints." I would like to comment on one
issue and then raise another. My comment concerns the issue of
determining the matrix language in codemixing. I agree with Muysken
that a structural approach to the matrix language is inappropriate
because it leads to circularity, as explained earlier. However, it is
not clear what the author's own position is concerning this issue.
Muysken concludes the discussion of the matrix language on a rather
ambiguous note: "a generally valid criterion for defining the ML ... is
hard to find, but researchers have no trouble identifying it ... There
is much evidence that CM is asymmetrical and involves a dominant, base
or matrix language." In a recent study, motivated by the debate on
this issue in the Linguist List (Vol. 10.1759, 30 November 1999),
Kamwangamalu (2000) explains that studies that attempt to define the
matrix language in codemixing do not take into account the
sociolinguistic context in which codemixing takes place. He proposes
that in codemixing in a diglossic context (e.g. multilingual
communities in Africa), the language identified as 'Low' (e.g., an
African language) tends to be the matrix language, and the one
identified as High (e.g., English, French or Portuguese) the embedded
language. It seems that diglossia is a useful theoretical construct in
attempts to define the matrix language: the distinction between High
language and Low language parallels the distinction between the matrix
language and the embedded language. In codemixing involving English,
French or Portuguese with an African language the latter is usually the
matrix language, and the former the embedded language (for details, see
Kamwangamalu 2000: 200-204).

The question I would like to raise concerns the three-way distinction
that Muysken makes between 'insertional codemixing', 'alternational
codemixing', and 'congruent lexicalization'. Is there really a need for
introducing these 'new' concepts? To what extent does the introduction
of these concepts advance our knowledge in the field of codeswitching
(a concept that I am using as cover term for both intrasentential as
well as intersentential switching). How different is the three-way
distinction between 'insertion', 'alternation', and 'congruent
lexicalization' from the classical distinction between 'codemixing',
'codeswitching', and 'lexical borrowing'? Muysken is not the first
codeswitching scholar seeking to introduce new terms into the study of
codeswitching/codemixing. For instance, in a study of Mexican-American
children's codeswitching, McClure (1977: 97) distinguishes between two
dimensions of codeswitching, namely, 'codemixing' and 'code-changing'.
In particular, McClure writes that "The children's codeswitching
appears to reflect the operation of two linguistic devices: 'code-
changing' and 'code-mixing'. 'Code-changing', generally motivated by
situational and stylistic factors, is the alternation of languages at
the level of the major constituent (e.g. NP, VP, S). The code-change is
a complete shift to another language system". She then defines 'code-
mixing' as "the individual's use of opposite language elements which
cannot be considered to be borrowed by the community" (McClure, 1977:
98). Rather than shed light on children's codeswitching, McClure's and
related studies added to the complexity of an already complex
phenomenon, codeswitching. In this regard, Baker (1980:1), in an
article aptly titled 'Categories of codeswitching in Hispanic
Communities: untangling the terminology', observes that "the research
which has been done on 'codeswitching' in Hispanic communities of the
United States is confusing and difficult to understand, not only
because of different samples and methods but also because of the
overlapping terminologies employed to describe codeswitching
categories". Baker's point, made twenty years ago with regard to
codeswitching in American Hispanic communities, applies equally to
current studies of codeswitching, including "Bilingual Speech". In my
view, there seems to be no need to rename existing codeswitching
categories. The distinction between 'insertion' and 'alternation' is
parallel to the classical distinction between 'code-mixing' and
'codeswitching', where the former refers to intrasentential switching,
and the latter to 'intersentential switching'. Likewise, 'congruent
lexicalization' corresponds to 'borrowing' (since one of the features
of 'congruent lexicalization' is 'morphological integration'). In this
regard, Muysken writes: "a frequent feature of 'congruent
lexicalization' is the incidence of MORPHOLOGICAL INTEGRATION" (p. 134,
emphasis added). Now, compare this statement with the following, which
Muysken makes about 'insertion': ".. another diagnostic feature of
insertion [is] MORPHOLOGICAL INTEGRATION*" (p. 63, emphasis added).
"Insertions tend to be single* often MORPHOLOGICALLY INTEGRATED
CONSTITUENTS (p. 64, emphasis added). If 'insertion' and 'congruent
lexicalization' involve morphological integration, why is it necessary
to distinguish between the two in the first place? There are several
instances in "Bilingual Speech" that show that the book simply recycles
existing codeswitching categories, namely intrasentential and
intersentential codeswitching.. Consider, for instance, the following:
"Bilinguals dispose of two grammars and lexicons, and the lexicons can
be viewed as one large collection that consists of several subsets.
THUS LEXICAL BORROWING COULD BE TERMED LEXICAL SHARING (p. 69, emphasis
added). Note that it is this 'sharing' that constitutes the basis for
one of the proposed three 'new' categories of codeswitching, namely,
'congruent lexicalization'. Further, consider the distinction between
the other two categories, 'insertion' and 'alternation'. In this
regard, Muysken writes, "Insertion is mostly a form of unidirectional
language influence, while alternation often goes both ways. INSERTION
IS CONSTITUENT-INTERNAL, ALTERNATION IS PHRASE- OR CLAUSE-PERIPHERAL"
(p. 75, emphasis added). "'Insertion' is akin to (spontaneous) lexical
borrowing, which is limited to one lexical unit." "In the case of
alternation, there is a TRUE SWITCH FROM ONE LANGUAGE TO THE OTHER,
involving both grammar and lexicon" (p. 5, emphasis added). This
definition of 'alternation' echoes McClure's (1977) definition of what
she terms 'code-changing', which also entails a complete shift/switch
from one language to another. Another point that concerns me is that,
Muysken appears to lump style-shifting and codemixing together. The
author writes: "The phenomenon of style shifting can be seen as one
subtype of code-mixing, namely congruent lexicalization". My question
is: doesn't codemixing presuppose competence in at least two languages?
If so, in what ways can codemixing be equated with style-shifting.
Also, the two terms, style shifting and codemixing, make totally
different claims about the competence of the individual speaker: 'style
shifting' can occur in both monolingual and bilingual speech, but as
Pfaff (1979:295) rightly points out, codemixing is necessarily a
product of bilingual competence. In spite of my skepticism about the
usefulness of the three-way distinction Muysken makes between
'insertion, alternation, and congruent lexicalization', I must say that
his book, "Bilingual Speech", is a welcome addition to publications on
codemixing, especially in terms of the amount and variety of data it
makes available to professionals in this field.


References

Azuma, Shoji (1993) The frame-content hypothesis in speech production:
evidence from intrasentential codeswitching. Linguistics 31: 1071-94.

Baker, Opal R. (1980) Categories of codeswitching in Hispanic
communities: Untangling the terminology. Working Papers in
Sociolinguistics, 74-80.

Chomsky, Noam (1986) Knowledge of language: its nature, origin, and
use. New York: Praeger.

DiSciullo, Anne Marie, Muysken, Pieter and Singh, Rajendra (1986)
Codemixing and government. Journal of Linguistics 22: 1-24.

Grosjean, Fran´┐Żois (1996) Processing mixed language: issues, findings,
models. In Annet de Groot and Judith Kroll (eds.), Tutorials in
bilingualism: psycholinguistic perspectives, 225-54. Hillsdale, N.J.:
Lawrence Erlbaum.

Kamwangamalu, Nkonko M. (1989) Codemixing across languages: Structure,
Functions and Constraints. Doctoral dissertation. University of
Illinois at Urbana-Champaign.

Kamwangamalu, Nkonko M. (1996) Sociolinguistic aspects of siSwati-
English bilingualism. World Englishes 15, 3: 295-306.

Kamwangamalu, Nkonko M. (1997) Language contact, codeswitching, and I-
languages: evidence from Africa. South African Journal of Linguistics
15, 2: 45-51.

Kamwangamalu, Nkonko (2000) INFL as a marker of matrix language in
codeswitching in a diglossic context. In Arika Okrent and John P. Boyle
(eds.) The Proceedings from the Main Session of the Chicago Linguistic
Society's Thirty-sixth Meeting, Vol. 36, 1: 197-207.

Myers-Scotton, Carol (1993) Duelling languages: grammatical structure
in codeswitching. Oxford: Clarendon.

Pfaff, Carol W. (1979) Constraints on language mixing. Language 55:
291-318.

Poplack, Shana (1980) Sometimes I'll start a sentence in Spanish Y
TERMINO EN ESPANOL. Linguitics 18: 581-618.

Rooij, Vincent de (1996) Cohesion through contrast. French-Swahili
codeswitching and Swahili style shifting in Shaba Swahili. Doctoral
dissertation, Universiteit van Amsterdam/IFOTT.

Sankoff, David and Poplack, Shana (1984) Borrowing: the synchrony of
integration. Linguistics 22: 99-136.

Thomason, Sarah G. and Kaufmann, Terence (1988) Language contact,
creolization and genetic linguistics. Berkeley: University of
California Press.

Traffers-Daller, Jeanine (1994) Mixing two languages: French-Dutch
contact in a comparative perspective. Berlin: Mouton.


About the reviewer:
Nkonko M Kamwangamalu holds a PhD in linguistics from the University of
Illinois at Urbana-Champaign, Illinois. He has taught English language
and linguistics at the National University of Singapore and the
University of Swaziland and is currently Associate Professor of
Linguistics at the University of Natal in Durban, South Africa. His
main areas of research interests include multilingualism, codeswitching
and codemixing, language policy and planning, new Englishes, and
African linguistics.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue