Chapter 2
 
 CODE INTERACTIONS BETWEEN TWO ORAL LANGUAGES
 
  This chapter and the next review literature 
concerning three phenomena occurring in language contact 
situations; specifically, code- changing within the speech of a 
bilingual, the din in the head phenomenon which occurs when a 
second or foreign language is learned, and the social situation of 
diglossia. Chapter Two focuses on the literature dealing with two 
oral languages. Chapter Three concentrates on these same phenomena 
when one of the languages is signed. The three phenomena discussed 
in these chapters are those language contact phenomena most closely 
related to bimodal communication. One of the goals of this 
dissertation is to compare and contrast BC with these phenomena. 
Chapter Two begins with an extensive discussion of code-switching 
and borrowing, which is followed by a brief section on language 
acquisition phenomena such as the din in the head phenomenon. The 
chapter concludes with a very brief discussion of diglossia.
 
2.1 Code-Switching and Borrowing 
 The literature on code-switching and borrowing presents many 
conflicting definitions and contradictory studies. Section 2.1.1 
discusses the previously proposed definitions of these and related terms. The 
terminology used in the remainder of this dissertation is presented at the end 
of this section. Following the definitions are discussions on the 
major areas of interest in the code-switching literature; 
specifically, the number of grammars involved (section 2.1.2), the 
syntactic constraints on code-switching (section 2.1.3), and the 
functional uses of code- switching and borrowing (section 2.1.4). 
 
 2.1.1 Definitions 
 There are many different definitions of code-
switching. Robert Di Pietro (1977:3) gives a very inclusive 
definition of this phenomenon: "[C]ode-switching is the use of more 
than one language by communicants in the execution of a speech 
act." 
 Poplack (1987:97), however, has a more restrictive 
definition of code-switching. She defines code-switches as, 
"multi-word sequences, which remain lexically, syntactically, and 
morphologically unadapted to recipient language patterns."
 Blom and Gumperz (1971) describe three different types of code- 
switches. If there are two languages used in a community, each of 
which is typically used in a specifically defined situation, then 
a situational code-switch is one in which a change in situations 
results in a change in the code used. A metaphorical code-switch 
occurs when: a) the situation stays the same, b) the topic changes 
to one more typically found in the other situation, and c) the code 
is switched to correspond with the code used in the other 
situation. An example of such a code-switch would be the use of a 
church-language to discuss a church-related topic when one is not 
physically in a church. Finally, conversational code-switching may 
occur without a change in either the topic or the situation. 
Conversational code-switches may be intra-sentential.
 Gumperz (1982:59) defines code-switching itself as "the 
juxtaposition within the same speech exchange of passages of speech 
belonging to two different grammatical systems or sub-
systems." 
 Other classifications of code-switching exist. 
For example, Poplack (1982) differentiates intimate and emblematic 
code-switching. Intimate code-switching occurs frequently, and is 
often intra-sentential. The phrase of intimate code-switching 
refers to the form of the code-switch, not to the relationship 
between the conversants. An English-Spanish example of this type 
of code-switching, taken from Sankoff and Poplack (1981: 5), is "So 
you todava haven't decided lo que vas a hacer next week." 
Emblematic code-switching, which some linguists consider to be 
borrowing, rather than true code-switching, occurs only 
occasionally. It involves mostly tag questions and single nouns. 
A Spanish-English example of emblematic code-switching, taken from 
Sankoff and Poplack (1981: 33) is, "Yo estaba aburrecido, 
murindome, you know?"
 Scotton's (1987) original definition of code-switching is very 
different from most. She relies upon frequency of occurrence of a given 
item in a passage to determine whether it is an instance of borrowing or 
code-switching. Her definition of code-switching is thus very broad, and 
includes many things which others would label as borrowing. In a later study, 
(Scotton, 1988:157), she uses a different criterion, and defined 
code-switching as
 
the use of two or more linguistic varieties in the same 
conversation, without prominent phonological assimilation of one 
variety to the other...switching may be either intra- or extra-
sentential and often (but not necessarily) involves stretches of 
more than one word.
 
 Phonological assimilation is the primary criterion according to 
this definition. However, Scotton cautions that the distinguishing 
of code- switching from borrowing solely by structural criteria 
poses many problems. For example, she points out that there are 
intermediate degrees of phonological assimilation; she states that 
assimilation is on a continuum. Consequently, according to her, 
code-switching must be distinguished from borrowing by means of 
social context. She asserts that if a word (or words) from one 
language is (are) incorporated into another language in a way which 
carries social significance, a code-switch has occurred.  If there 
is no social significance, a borrowing has occurred.
 Finally, Messing (1990:463) defines code-switching as a 
sequential transition from one language to another within a `speech' act. 
(Speech is in quotes to emphasize that signing may constitute a 
`speech' act, even though no speech is involved.) 
 
The word sequential is included in this definition to differentiate 
code- switching from bimodal and simultaneous communication, in 
both of which an individual signs and speaks 
simultaneously.
 The choice of terminology is further complicated when only 
one word is used from one of the languages, and the rest of the sentence is in 
the other (matrix) language. As we have seen, some individuals would consider 
this an example of code-switching. Others use special terminology to 
describe the individual word. The word is considered to be borrowed from the 
source language. The terms calque or loan translation, and relexification 
are often used. Calques preserve the morphological structure of one 
language while they borrow lexemes from another language.
 Borrowings and items which have been relexified have sometimes 
been contrasted by saying that relexified items are more permanent; borrowings 
may be nonce occurrences. Two factors are often used to differentiate 
borrowings and code-switches. First, only a borrowing may be phonetically 
integrated into the matrix language. Second, monolinguals are capable of using 
borrowings, but they are not able to code-switch. In order to avoid 
the difficulty of distinguishing between the two terms, Pfaff 
(1982) introduced the term mixing to refer to code-switching and 
borrowing collectively.
 McClure and McClure (1975) use code-mixing to refer to a 
single word in one language integrated into an utterance of another 
language because the other language's word is not readily available at 
the time of the utterance. This lack of availability can be due either to 
the speaker's not knowing the word in one of the languages, or because of 
an inability to retrieve the word from memory rapidly enough. Code-changing, 
according to them, is a switch from one language to another which 
occurs at constituent boundaries. Code-changes are used 
stylistically to indicate a change in affect, addressee, mode, etc. 
They use the term code-switching to refer to both code-mixing and 
code-changing.
 The above discussion reveals the current 
terminological confusion in the existing literature. It is 
therefore especially important that all writing which uses these 
terms include definitions of them in order to make clear to the 
reader the writer's intended use of the terms. The current study 
assumes the following definitions for these words:
 
1. Code-shifting occurs when a speaker changes languages between
 utterances. An utterance is a stretch of speech preceded and
 followed by silence or a change of speaker. Scotton (1987)
 provides an example of a code-shift between Swahili and Luyia.
 After a security guard has asked a visitor in Swahili whom she 
 wishes to see, and after she responds in Swahili that she
 wishes to see Solomon Inyama, the following dialogue occurs:
 
  Guard (Swahili): Unamjua kweli? Tunaye
   Solomon Amuhaya--nadhani ndio yule.
   (Translation:Do you really know him? We have
   a Solomon Amuhaya--I think that's the one you mean.)
   Visitor (Swahili): Yule anayetoka Tiriki--yaani Mluyia.
  (Translation: This one comes from tiriki--that is, a Luyia person.)
   Guard (switches to Luyian): Solomon mwenuyu
   wakhumanya vulahi? (Translation: Will Solomon know you?)
 
2. Code-switching occurs when two languages are used within an
  utterance such that the utterance contains grammatical
 elements from both languages. Even if only one word of one of
 the languages is present, it is enough to have a code-switch
 if that word contains grammatical elements, such as
  inflectional endings which agree with other elements in the
 sentence. Code-switches can be intra-sentential or inter-sentential.
 An example of a code-switch between English and Spanish, taken 
 from Sankoff and Poplack (1981: 5), is "So you todava haven't
 decided lo que vas a hacer next week." Code-switches can 
 be either sequential or simultaneous. Code-switches between 
 two oral languages will always be sequential. Simultaneous 
 code-switches would be instances of either bimodal 
 communication or simultaneous communication.
 
3. Borrowing occurs when a word or words from one (source)
 language is introduced into an utterance which is otherwise in
 another (matrix) language in such a way that the utterance
 contains no grammatical elements unique to the source
 language. Because the languages involved in this study are
 articulated in different modes simultaneously, a criterion of
 phonological integration to distinguish borrowings from code-
 switches is irrelevant. Therefore, syntactic and
 morphological criteria are being used. The term borrowing will
 include both what have been called calques and what have been
 referred to as instances of relexification. An example of a
 borrowing from English into Spanish, taken from Sankoff and
 Poplack (1981: 18) is "Lleg yesterday la mama ma." 
 Borrowings can be either sequential or simultaneous.
 Borrowings between two oral languages will always be
 sequential. In a sequential borrowing, only one language at a
 time is articulated. Cross-modal borrowing can be either
 simultaneous or sequential.
 
4. The term code-changing is a new term which will be employed in
 this dissertation to refer to code-shifting, sequential code-
 switching, and sequential borrowing collectively. This term
 will include only sequential changes in codes; hence, it will
 exclude bimodal communication.
 
 Note that with these definitions, a word or 
phrase which does not contain grammatical elements unique to one 
of the languages in question will be classified as a borrowing, 
rather than a code-switch. A French- English example of this taken 
from Poplack (1987:96) is, "il y avait une grosse beach." 
 
2.1.2 The Number of Grammars 
 There are a few issues which are 
discussed, either implicitly or explicitly, in almost every paper 
concerning the syntax of code-switching. The first concerns the 
number of grammars involved. There is assumed to be a separate 
grammar for each of the languages concerned (Lederberg and Morales, 
1985, Petersen, 1988, Poplack, 1979). However, there is debate 
over whether or not there is also an entire additional grammar for 
code- switching. Such a third grammar would contain selected 
elements from each of the first two grammars. Poplack (1979) cites 
as evidence for a third grammar the lack of disruption of speech 
rhythm, and the lack of repetitions, hesitations, or corrections 
at the point of the code-switch. The ability of speakers to give 
judgments, and basically to agree with each other on these 
judgments, on the wellformedness of sentences containing code-
switches has also been cited as evidence for a third 
grammar.
 Lederberg and Morales (1985) and Petersen (1988) 
give as evidence against a third grammar the fact that people do 
not need extensive exposure to samples of code-switching in order 
to make grammaticality judgments which are similar to those made 
by people who have been code- switching for years. A 
related issue concerns whether or not there is a single matrix 
language for each sentence, and how to determine which language is 
the matrix one for a given sentence. Petersen (1988) considers an 
asymmetry in the ability of her subject to code-switch from one 
language to another in a given syntactic structure to be evidence 
in favor of the existence of a matrix language for her subject. 
However, evidence for such an asymmetry in other bilinguals is at 
best mixed. Furthermore, the asymmetry may be due to non-syntactic 
factors or to syntactic constraints on code-switching without it 
being the result of the existence of a matrix language. 
People who assume the existence of a matrix language have different 
ways of determining which language is the matrix one, and which is 
the embedded. Klavans (1985), for example, provides evidence from 
Spanish-English code-switching that the matrix language is the one 
which contains INFL.
 Not every potentially possible 
combination of elements from two languages is permissible. 
Proficient bilinguals will deem certain theoretically possible 
code-switches to be ungrammatical. These judgments might be due 
to a third grammar which the bilinguals have, or they might be due 
to specific constraints which exist on the interactions of the two 
languages. The latter possibility is discussed in the next 
section. 
 
2.1.3 Syntactic Constraints on Code-switching
  Most of the literature concerning syntactic constraints on 
code-switching have examined code-switching between Spanish and English. 
Consequently, one must take most statistical observations on code- 
switching with a grain of salt. Many of the tendencies noted below 
may be due to the specific languages involved. With this caveat, 
let us proceed to a description of some of the statistical 
tendencies. Most switches occur at sentence boundaries. There are 
more extra-sentential switches than there are intra-sentential 
ones. Extra-sentential switches include, among other things, 
switches involving whole sentences, tag questions, exclamations, 
and fillers. Within a sentence, single nouns are the most 
frequently switched category. Subject noun phrases are more likely 
to occur in the same language as the verb than object noun phrases 
are. Timm (1975) notes many construction specific tendencies, such 
as the impossibility of a switch between an auxiliary and a verb 
in Spanish. 
 There has been an effort to move away from 
construction specific tendencies to more general constraints. 
Lance's (1975) proposal that there are no syntactic constraints on 
code-switching is currently in disfavor. Several, often 
conflicting, constraints have been proposed. Perhaps the two most 
well known constraints are Sankoff and Poplack's (1981) free 
morpheme and equivalence constraints. The free morpheme constraint 
states that "a switch may not occur between a bound morpheme and a 
lexical form unless the latter has been phonologically integrated 
into the language of the bound morpheme." (Sankoff and Poplack, 
1981:5). The equivalence constraint indicates that 
 
the order of sentence constituents immediately adjacent to and on 
both sides of the switch point must be grammatical with respect to 
both languages involved simultaneously...[T]he local co-
grammaticality or equivalence of the two languages in the vicinity 
of the switch holds as long as the order of any two sentence 
elements, one before and one after the switch point, is not 
excluded in either language. (Sankoff and Poplack, 1981:5-6)
 
Poplack (1981:183) argues that
 
[t]he equivalence constraint suggests that the code-switching mode 
proceeds from that area of the bilingual's grammar where the 
surface structures of L1 and L2 overlap, whereas the outer areas 
where there is no equivalence will tend to be reserved for 
monolingual segments of discourse.
 
Woolford (1983) agrees with this assessment. She also agrees that 
the grammars for the two languages are not disjoint; instead, they 
overlap. Code-switching occurs within this overlapping area. It 
would be logical to assume that bimodal communication (BC) would 
also arise within this intersection. If this is so, then the 
greater the area of overlap, the more opportunities there would be 
for BC to be manifested. Consequently, the more English-based a 
person's signing is, the more BC that person will use. This is the 
sixth hypothesis presented in Chapter One. 
 These two hypothesized constraints have received mixed empirical 
support. Poplack (1982) verifies that even non-fluent Spanish-English 
bilinguals obey both constraints. On the other hand, Scotton 
(1988) provides evidence that East African bilinguals violate the 
free morpheme constraint in their code-switching between Swahili 
and English. Poplack (1988) examines two bilingual communities. 
She finds that the code- switching between Spanish and English done 
by Puerto Ricans in New York City is used as an identity marker. 
It obeys the free morpheme and the equivalence constraints. In 
contrast to this, Canadians in the Ottawa-Hull code-switch between 
French and English for stylistic purposes. Their code- switching 
often violates both constraints. Poplack concluded that different 
code-switching constraints can operate in different speech 
communities. Nartey (1982) notes that Ghanians do not use the 
free- morpheme constraint when they code-switch between Ghanian 
languages and English. He proposes that, although Poplack's 
constraints might not be universal, they may nevertheless hold 
universally within a given socio- cultural environment. He feels 
that other constraints might hold within other socio-cultural 
environments. Kendall (1980) examines the code- switching done by 
French speakers learning Bambara (a postpositional language). 
These speakers originally switched codes within prepositional/
postpositional phrases (PP), thus violating the equivalence 
constraint. As they learned more Bambara, however, they ceased to 
switch within the PP. Therefore, Kendall concludes that more 
proficient bilinguals will use more constrained code-switching. 
All of these studies support the hypothesis that all proficient 
bilinguals within a speech community will follow certain 
constraints on code-switching. The specific constraints may vary 
across communities, and non-proficient bilinguals will not 
necessarily follow the constraints.
 Lederberg and Morales (1985) also provide experimental evidence 
in support of the free morpheme constraint and the equivalence constraint 
for Spanish-English bilinguals. They asked adults and children to give 
acceptability judgments on sample code-switches, some of which 
violated one of these constraints. Adults had both constraints. 
Children had only the equivalence constraint. The researchers feel 
that this is due to the children's stage of general cognitive 
development, rather than to a lack of knowledge concerning code-
switching per se. They also note a tendency for switches of content 
words to be deemed more acceptable than switches of function words.
 Pfaff (1982) proposes a constraint similar to the equivalence 
constraint, and a tense/aspect constraint. The latter constraint 
requires tense/aspect to be overtly marked somewhere in each 
sentence, such as by an affix indicating tense (e.g., English -ed).
 A final unresolved issue concerns whether constraints are 
absolute or probabilistic. A violation of an absolute constraint would 
result in a completely ungrammatical construction. A violation of 
a probabilistic constraint may only result in a construction which 
is not heard frequently, or which sounds awkward, but is not 
impossible. Lederberg and Morales (1985:133) posit probabilistic 
constraints such as, "It is more acceptable to code-switch a 
content word than a function word." Sankoff and Poplack (1981:20) 
also suggest introducing a probabilistic element into a grammar in 
order to, "account for the many striking quantitative patterns 
evident in the discourse." The study of bimodal 
communication might shed some light on the number of grammars and 
the linguistic constraints involved in code- changing. By studying 
BC, we can determine which of the code-changing constraints 
heretofore discovered are due solely to the sequentiality imposed 
by unimodal oral communication.
 
2.1.4 Functional Uses of Code-changing  
 Many different uses of code-changing in various societies have been 
described by a large number of researchers. The six most commonly mentioned 
functions, along with a few of the researchers who described them, are listed 
below.
 
 Strategic Negotiation: Code-changing can be used
 strategically in negotiating one's relationship with one's
 conversational partner. (Di Pietro, 1977, Dabene and Billiez,
 1986, Grosjean, 1982, Heller, 1988, McClure, 1981, Scotton,
 1987, Valdes, 1981). 
 
 Identity Marker: Code-changing can be used as an identity
 marker, to indicate such things as one's ethnicity, or amount
 of education. (Di Pietro, 1976, Grosjean, 1982, McClure,
 1981, Poplack, 1982, 1988, Scotton, 1987, Taylor and Royer, 
 1980, Titone, 1987).
 
  Domain Marker: If each of the languages is typically used in a 
 certain place, or discussing certain topics (i.e., in a
 certain domain), code-changing can be used when the domain
 changes or when more than one domain is involved. (Blom and
 Gumperz, 1971, Gibbons, 1987, Jacobson, 1982, Meloni, 1982,
 Titone, 1987).
 
  Compensation: Code-changing can be used when the speaker in
 the conversation either does not have the ability to discuss a
 topic in one of the languages, or simply can not access the
 necessary vocabulary rapidly enough on a given occasion. 
 (Auer, 1984, Grosjean, 1982, Jacobson, 1982, Meloni, 1982,
 Titone, 1987).
 
  Accommodation: Code-changing can be used by speakers to
 accommodate conversational partners whom they believe do not
 have the ability to converse comfortably in at least one of
 the languages. This becomes especially likely if there is no
 one language in which all of the conversants are fluent.
 (Auer, 1984, Genishi, 1981, McClure, 1981, Platt, 1980, Taylor
 and Royer, 1980, Titone, 1987).
 
  Stylistic Effect: Finally, code-changing can be used for
 stylistic effect, either to employ the mot juste, or as
 emphasis. (Auer, 1984, Grosjean, 1982, Jacobson, 1982,
 McClure, 1981, Meloni, 1982, Titone, 1987).
 
  It seems reasonable to hypothesize that BC, a phenomenon 
which differs from code-changing only by the simultaneity of the 
articulation of the languages involved, serves one or more of these 
functions. Chapter Six includes an examination of the corpus of 
data collected in the current research in an attempt to ascertain 
the function(s) of BC.
 
2.2 Language Acquisition Phenomena
 
 Many language acquisition phenomena, when plotted 
against language proficiency, yield an inverse-U graph. Two 
examples of such phenomena will be discussed in this 
section. 
 The first language acquisition phenomenon is the 
number of environments in which code-switching can occur. As 
previously mentioned, Kendall (1980) found that as French speakers 
became proficient in Bambara, they ceased to code-switch within 
PPs; i.e., they began to obey the equivalence constraint. 
Monolingual French speakers could not code-switch at all. 
Completely fluent French-Bambara bilinguals would obey the 
equivalence constraint. Therefore, non-proficient French speakers 
of Bambara, who violate the equivalence constraint, code-switch in 
more environments than either monolinguals or proficient 
bilinguals. Thus, a plot of the number of syntactic environments 
in which speakers code-switch, plotted against the second language 
proficiency of the speakers, will be an inverse-U.
 The din in the head is another language acquisition phenomenon 
which has an inverse-U graph. Barber (1980:30) coined the term din in 
the head to describe an experience she had had during her stay in 
Russia. As she explained, 
 
[b]y the third day the linguist in me was noticing a rising din of 
Russian in my head: words, sounds, intonations, phrases, all 
swimming about in the voices of the people I talked with...The 
sounds in my head became so intense after five days that I found 
myself mindlessly chewing on them, like so much linguistic 
cud...The constant rehearsal of these phrases of course was making 
it easier and easier to speak quickly and fluently; things popped 
out as prefabricated chunks. But I had no control over what my 
subconscious fed into my "chewer" each day.
 
 Krashen (1983) read her description and at once 
identified her experience with the din in the head as being similar 
to his own. In his article, he hypothesizes that this is a common 
occurrence which results when the Language Acquisition Device (LAD) 
is stimulated. The LAD is proposed to be an innate set of language 
constraints, rules, and language acquisition principles. Krashen 
says that it is the LAD which permits children to acquire a language 
or languages in an apparently easy, natural, fashion. The LAD is 
triggered by comprehensible input which contains sufficient 
quantities of the learner's i + 1 structures. This means that 
people naturally acquire a language when they are exposed to 
samples of the language which they can understand, but which 
contain some structures which they have not yet mastered. Krashen 
(1983) states that a graph of second language ability vs. frequency 
of experiencing the din in the head will be in the shape of an 
inverse-U. This relationship would be due to the brain processing 
a language which was not fully acquired.
 If a graph of BC as a function of signing ability is also in the 
shape of an inverse-U, then BC might also be a byproduct of the language 
acquisition process. This is the third part of the second 
hypothesis described in Chapter One. 
 
2.3 Diglossia
  Ferguson (1959: 325) defines diglossia as a language 
situation in which speakers in a speech community use "two or more 
varieties of the same language...under different conditions." One 
of the varieties, called the high code, is typically a standardized 
code used in formal settings. The high code might be used, for 
example, in religious contexts, and in conducting business. The 
other varieties are called low codes. They are typically regional 
dialects used in informal settings. Low codes might be used, for 
example, in conversations with family and friends. While the high 
code is typically the most prestigious code, the low codes are the 
ones used to promote solidarity among speakers. 
 Ferguson (1959) cites the following examples of diglossia:
 
1. Haiti uses French as a high code, and Haitian Creole as a low code. 
 
2. Arabic speakers use Classical Arabic as a high code and Egyptian as a 
 low code.
 
3. Swiss German speakers use standard German as a  high code and Swiss as 
 a low code.
 
  There is a broader usage of the term diglossia which is 
also extensively employed. Fishman's (1971) definition of 
diglossia permits any two types of codes to be in a diglossic 
relationship.
 This general discussion on diglossia provides the background 
necessary to understand the section on sign language diglossia in the next chapter, 
which also contains a more extensive discussion on this topic. It is important 
to understand the relationships among signed and spoken languages in order to 
understand the social milieu in which BC occurs. 
 
 2.4 Summary
  Bilingual individuals can integrate elements from 
both of their languages into a single conversation. They can do 
this by code-shifting between utterances, by code-switching within 
an utterance, or by borrowing a word or phrase from one language 
into the other one. This dissertation refers to code-shifting, 
sequential code-switching, and sequential borrowing collectively 
as code-changing.
 The different code-changing phenomena 
have been of interest to linguists for the light they may shed both 
on the storage and retrieval of languages in the mind of the knower 
of a language, and on the social uses of language. This chapter 
has discussed research on code-changing between two oral languages. 
The next chapter will discuss code-changing and other phenomena 
which occur between an oral and a signed language. 
 Two language acquisition phenomena were discussed in order to 
show that a plot of such phenomena against second language proficiency 
results in an inverse-U. If BC has such a plot, it may also be a 
language acquisition phenomenon. 
 The brief description of diglossia included in this chapter was 
provided as a background for the next chapter's discussion of the different types of 
signing.
 
 
 
 NOTES
 
 

 
 
 
 
 
Chapter 3: Code Interactions Involving Sign 




Back to my dissertation page