Bilingual Research Journal

ISSN: 1523-5882 (Print) 1523-5890 (Online) Journal homepage: https://www.tandfonline.com/loi/ubrj20

Functions of Code Switching in Schoolchildren's

Conversations

Iliana Reyes

To cite this article: Iliana Reyes (2004) Functions of Code Switching in Schoolchildren's
Conversations, Bilingual Research Journal, 28:1, 77-98, DOI: 10.1080/15235882.2004.10162613

To link to this article: https://doi.org/10.1080/15235882.2004.10162613

Published online: 23 Nov 2010.

Submit your article to this journal

Article views: 1857

View related articles

Citing articles: 66 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=ubrj20
 Functions of Code Switching in
Schoolchildren’s Conversations

Iliana Reyes
University of Arizona

Abstract
This study examined the code-switching patterns in the speech of
immigrant Spanish-speaking children. Seven- and 10-year-old boys
and girls from bilingual classrooms were each paired with a mutually
selected friend, and their speech was collected in two contexts:
while the children waited for an expected science experiment and
when they worked together to follow an instruction worksheet
about hands-on magnetic materials. This study presents data on the
discourse characteristics of children’s code switching, and the
functions that Spanish and English have according to context. In
addition, the data are described in relation to children’s language
competence and preference. The study found that code switching
occurred both within and across turns. The older children’s switches
were more frequent and were deployed for a wider variety of
functions than the younger children’s. The results challenge the
negative view that code switching by children who are learning two
languages is due to lack of proficiency, and instead support the view
that it is used as a strategy to extend their communicative competence
during peer interaction.

Introduction
Conversation constitutes a major part of human interaction. It is through
language and conversation that children begin to communicate information
through a variety of speech registers and style switching (Ervin-Tripp, 2001).
Of particular interest to sociolinguistic and developmental researchers is the
impressive ability of bilingual speakers to switch with ease at different points
in conversation. From the sociolinguist’s point of view, code switching (CS) is
studied to understand why people who are competent in two languages
alternate languages in a particular conversation or situation. The concept of

Functions of Code Switching in Schoolchildren’s Conversations 77

CS, as defined by Gumperz (1973), refers to the alternate use of two or more
languages in the same utterance or conversation.
Most of the early research on CS has looked at adult–adult interaction
(e.g., Clyne, 1967; Gumperz, 1967, 1973; Weinreich, 1953). However, the early
1980s witnessed increased interest in studying children’s CS (Álvarez, 1979;
Genishi, 1976; Zentella, 1982, 1997). These studies have shed some light on
how bilingual children use different languages according to addressee and
context. However, these studies do not describe how children develop CS
over the years, and how CS is used to extend communicative competence for
achieving conversational goals during peer interaction. In the classic CS study
by Poplack (1980) with Puerto Rican American bilinguals, she analyzed adults’
conversations in natural settings and speech during a sociolinguistic interview
to learn about bilinguals’ linguistic competence and their use of CS. Poplack’s
findings pointed out that CS was used by those individuals whose language
skills in both languages were balanced. While CS might be an indicator of
bilingual ability in adults (Poplack, 1980), it might indicate development of
bilingual communicative competence in children who are still learning a second
language (L2) (Reyes, 2001).
The present study analyzed the functions of CS in the speech of peers,
for the primary purpose of exploring the relation between CS functions and
development of children’s bilingual communicative competence. In this study,
communicative competence refers to the speaker’s knowledge of when to
speak and when not to, and what to talk about with whom, when, where, and
in what manner (Hymes, 1974). Based on previous research, we know that
monolingual speakers have the capacity to switch between a variety of
linguistic registers (e.g., Gumperz, 1967; Ervin-Tripp, 1964, 2001). Similarly,
bilinguals also have these communicative registers that monolinguals have,
but in addition they have two linguistic codes that they can alternate as a
resource for sociolinguistic strategies during conversation.

Literature Review: Code Switching in Children

Early research in this area has identified the types of CS and the factors
affecting it (e.g., Clyne, 1967; Ervin-Tripp, 1964; Fishman, 1965; Gumperz,
1964, 1967; Gumperz & Hernández-Chavez, 1975). More specifically, CS varies
according to the situation (situational code switching) and within a
conversation (metaphorical code switching). Under the metaphorical category,
CS varies according to discourse function (e.g., to include or exclude someone
from a conversation, to convey intimacy, or to emphasize a message).
Research on children’s CS has shown that simultaneous bilinguals develop
knowledge on how and when to use their two languages depending on the
addressee, the topic of the conversation, and the situation (Fantini, 1985;
Genishi, 1976; Halmari & Smith, 1994; McClure, 1981; Zentella, 1982, 1997).

78 Bilingual Research Journal, 28: 1 Spring 2004

More recently, Genesee and his colleagues (Genesee, 2002; Genesee, Boivin,
& Nicoladis, 1996) have found that French–English bilinguals as young as 2
years of age develop the ability to use and adjust each of their languages
differentially and appropriately with parents and an unfamiliar interlocutor as
part of their communicative competence. Genesee (2002) goes further to state
that “true bilingual communicative competence entails the ability to adapt
one’s language use on-line in accordance with relevant characteristics of the
situation, including the preferred or more proficient language of one’s
interlocutor” (p. 174).
Scientific evidence from studies with young Spanish–English bilinguals
(McClure, 1981; Zentella, 1997) report that younger children show more lexical-
item CS than older children. A common assumption to explain this finding is
that children code switch when they do not know the word in one language.
Hence, they draw on the other language. It is incorrect, however, to assume
that all cases of this type of code switch are the result of incomplete knowledge
of one of the languages. In some cases, children might be momentarily unable
to access a word for a concept in the language in use, but can access one more
readily in the other language. On the other hand, older children seem to
manipulate their linguistic codes for a wider variety of stylistic purposes and
situational demands than younger children do (Zentella, 1997). In a study
with peer Turkish–Danish bilinguals, Jørgensen (1998) found that 7- to 10-
year-olds could manipulate their two languages for power-wielding purposes.
The children in his study strove to gain control of the conversation when
involved in problem-solving tasks. Research findings also indicate that older
children code switch when they learn that elements of the other language
convey the meaning of the intended idea more accurately (Halmari & Smith,
1994; Zentella, 1997). However, we do not know well how “developing bilingual
children” develop the ability to code switch over the years and how they
make use of this strategy during peer interaction.

Theoretical Framework
This study was guided by a number of theoretical assumptions related to
bilingualism and communicative competence. First, in terms of children’s use
of CS, it was expected that older children would use code switches at a higher
rate and for a greater range of sociolinguistic functions than younger children.
This expectation is based on the assumption that older children would have
developed a higher level of proficiency in their L2 and greater knowledge of
the grammatical systems of both their first and second languages—factors
that have been found necessary for CS (Poplack, 1980). In addition, other
studies report that as children get older, they develop and use more
sophisticated discourse strategies (e.g., Ervin-Tripp & Mitchell-Kernan, 1977).
The nature of language development in these developing bilinguals must be
understood in relation to their development of bilingual communicative
competence (Genesee, 2002; Reyes, 2001). In other words, as bilingual children

Functions of Code Switching in Schoolchildren’s Conversations 79

become older, their exposure to different social and linguistic experiences
increases, and these experiences in turn affect and enlarge their knowledge
and ability to use their different languages and to deploy CS for sociolinguistic
purposes. Second, children’s patterns of using language and CS often mirror
the ways in which language is used in their communities (Lanza, 1997;
Schieffelin & Ochs, 1986). In addition, research by Poplack (1980) with adults
and by Zentella (1982, 1997) with children indicate that those speakers who
are more fluent and balanced in both languages tend to be better at alternating
and switching languages. Based on these previous findings, a relation between
CS and the number of years children have been exposed to their L2 was
expected. The working hypothesis from a developmental perspective is that
CS increases in bilingual children as they have more exposure to the L2, and it
is used as a resource to extend communicative competence. A sociolinguistic
theory of CS acquisition by bilingual children is necessary to understand the
development of communicative competence when learning an L2 and becoming
bilingual.
The focus in this article is to investigate how schoolchildren from two
age groups (7- and 10-year-olds) switch between two different languages,
English and Spanish, and draw on features of each language to accomplish
their conversational goals. In particular, peer interaction without an adult
present was analyzed to explore how children make use of their native language
(Spanish) and their L2 (English) in a social and school-related context. The
research questions were as follows:
1. How do children use CS during peer interactions?
2. What are the sociolinguistic functions of the code switches used by
school-age children from different age groups?
3. Is there an effect of context on the type of CS and language used by
schoolchildren (e.g., social setting vs. school-related setting)?

Methodology
Participants
Demographics
A total of 20 self-selected friend dyads participated in this study. These
children were in second- and fifth-grade classrooms in an elementary school
in Oakland, California. Ten dyads from each grade were recruited. Most of the
children were of Mexican background, with the exception of two, who were
Central Americans. The mean ages of the children were 10 years, 8 months and
7 years, 9 months for the fifth and second graders, respectively. Children were
given a small gift (e.g., stickers and pencils) for their participation.
The school where this study took place is part of a working-class Latino
and African American neighborhood. The children learned and continue to
use Spanish at home but hear and learn English in school and the community
around them (Reyes, 2001). These children participate in a transitional bilingual

80 Bilingual Research Journal, 28: 1 Spring 2004

program, but the school is not bilingual. In general, teachers supported the
use and development of both Spanish and English. Teachers and children
used both languages outside the classroom as well as during classroom
interaction.
Language background
Information regarding each child’s language background was obtained
from the parents and the teachers. The parents reported the language they
thought their child spoke the best, and they gave information about the age at
which the child learned English. The teachers were asked the same questions;
in addition, teachers used school test scores to report on each child’s language
proficiency. In a brief interview, the children were asked about their language
preference at home with parents and siblings, as well as in school with teachers,
staff, and friends.
The parents’ and teachers’ reports were compared for agreement. In most
cases, parents and teachers agreed about the relative proficiency of the
children. However, in those cases where their reports did not agree, we looked
at the students’ school English placement scores to determine their relative
fluency in English. In the few cases in which parents and teachers disagreed,
the parents underestimated the children’s fluency in English. This could be
explained by the fact that most of the parents see their children in a monolingual
Spanish environment. In contrast, teachers see the children at school, where
children are learning English and speak both languages throughout the day.
Children’s relative language knowledge was determined for each language by
looking at the information about each child’s language abilities as observed
by the parents, teachers, and researchers. Table 1 shows students’ relative
proficiency in each language and their language preference (data for all
children are included, except for two who were absent the day of the interview).
Table 1
Children’s Language Background: Proficiency and Preference by
Age Group

Students' language proficiency Students' language preference

7- year- olds 10- year- olds 7- year- olds 10- year- olds

Spanish 69% 25% 33% 20%

English 6% 12% 47% 47%

Both 25% 63% 20% 33%

Functions of Code Switching in Schoolchildren’s Conversations 81

 The results in Table 1 show that on average, the younger children were
mostly dominant in Spanish (69%), followed by 25% who were considered
balanced in both languages, and only 6% were considered English dominant.
By contrast, we see that on average, 63% of the older children were considered
balanced bilinguals (at least at the oral level). A very different picture emerged
when we looked at language preference reported by the children. A total of
47% of the younger children reported that they preferred to speak English,
33% preferred Spanish, and 20% said that they had no preference. For the
older children, a similar pattern was observed, with a 47% preference for
English. However, for Spanish preference and absence of preference, there
was an inverse pattern: 33% of 10-year-olds had no language preference,
while only 20% reported that they preferred to speak Spanish. The language
preference pattern contradicts data regarding younger children’s relative
proficiency. While nearly 50% of the children in each group said they preferred
English, only 25% of the younger children were considered proficient in both
English and Spanish. Regardless of their language proficiency, children from
a young age already can grasp the dominant perception that English is the
language of higher prestige in U.S. society.

Design and Procedure

About 10 hours of conversations were audiotaped over a period of 8
weeks during the children’s lunch break. Participants were told that the
researchers were interested in children’s conversations during a science project
and that they would be participating in a science activity. The children were
asked to nominate a friend with whom they wanted to participate in the science
activity.
Each dyad participated in one session. First, children were taken to the
cafeteria to eat lunch, and after that they were taken to the music room in the
school’s auditorium to be tested. The children’s conversations were recorded
as soon as the microphones were put on them. They were instructed to have
a seat and just talk while they waited for the two researchers to prepare the
science activity. Both researchers present were bilingual and gave instructions
to the children, alternating between English and Spanish. The researchers left
the room, and they monitored and recorded the children’s conversation from
an adjacent room. After 15 minutes, the researchers returned and explained
the science activity to the children, alternating between English and Spanish.
The researchers code switched among themselves and to the children, in
order to set the stage for CS. This way, the researchers modeled CS to let the
children know that it was acceptable to speak both languages. Various
researchers (e.g., Bayley & Zapata, 1993; Zentella, 1997) have noted that,
historically, there has been a pejorative connotation to “Spanglish” and CS.
To avoid negative perceptions that some children might have had about CS, it
was important to let them to know that CS was valued as a discourse style by

82 Bilingual Research Journal, 28: 1 Spring 2004

the researcher and student assistants in the project. After another 15 minutes,
the researchers returned and inquired about the dyad’s science findings.

Materials
A wireless microphone was connected to a transmitter inside a belt pack
that was strapped to each child’s waist, each with a different radio frequency.
Each microphone was clipped to the child’s shirt. Radio receivers and a cassette
recorder were set up outside the room where the study was being conducted
in order to receive and record the children’s conversations on separate channels
for each voice.
The materials used for the science activity included a compartmented
box, with labels in both languages and containing magnets of different forms
and colors, a compass, various magnetic rocks, and miscellaneous metal and
nonmetallic objects (e.g., paper clips, thumbtacks, keys), and a clear plastic
box containing iron filings. A worksheet had eight questions, in both English
and Spanish, regarding the items in the box and how they interacted with one
another (see Appendix A). This worksheet was used to prompt conversation
regarding the science activity and minimize silent play. Children had the option
of discussing the questions in English or Spanish.

Data Analysis
The conversations were transcribed for analysis using Gumperz and
Berenz’s (1993) transcription conventions (see Appendix B). Each of the
conversations lasted about 30 minutes. Each speaker’s turn, between the
turns of speaking by the other child, was taken as a unit of analysis. For the
frequency analysis, counts of all code switches were made at two levels. First,
all the code switches from English to Spanish and Spanish to English were
counted. Second, the number of turns per conversational dyad and per
language were computed separately for each setting: the social-talk context
and task-related conversation within the science activity. During the analysis,
it was determined that some of the talk in the science setting was off task and
social in nature (e.g., children talked about their favorite television show while
answering the science questions). This intermittent social talk within the
science activity was analyzed separately. For comparison purposes,
proportions were computed by dividing the number of code switches by the
number of turns for each child to yield a measure of CS probability per turn in
each context.
For the sociolinguistic analysis, coders were trained to use a specific
coding frame. It is important to clarify that only code switches that constitute
more than one lexical item were included in the analysis. Many investigators
in the field of language CS do not consider single switches (also called
borrowings or tag switches) “true switches,” therefore, only those longer
switches that clearly indicated a syntactic switch into the other language

Functions of Code Switching in Schoolchildren’s Conversations 83

have been included in the present analysis (see Romaine, 1995, for further
discussion on this issue). The following examples illustrate these two types
of code switches1 (see Appendix B for transcription annotations):
Example 1: e: se agarran asi ira shake your booty
t: they do it like this look
In the next example, the word lonchera illustrates a one-word switch or
lexical item. Overall, this type of one-word switch has been phonologically
assimilated into the speech community, and often it will be used in the speech
of monolingual speakers as well.
Example 2: e: i’m going to the classroom to get my lonchera
t: lunch box
The code switches were analyzed in terms of content and function in
relation to location in the conversation. To determine the extent of use of each
language, the total number of turns in only English or only Spanish, and in
both English and Spanish, were counted separately. The following are the
three categories used to code each child’s turns:
1. Spanish only: All phrases in one turn were Spanish, with no CS to English.
2. English only: All phrases in one turn were English, with no CS to Spanish.
3. Mixed language: Both languages were used within the same turn (one
word or more).
For the analysis of conversational functions, the different types of code
switches that were most prevalent in the children’s conversations were first
identified through analysis of the transcripts and then compared with the CS
typologies used in previous frameworks by Gumperz (1967), McClure (1981),
and Zentella (1997). In the end, combining new data-based and older typologies
facilitated identification of the categories most relevant to the code switches
produced by the children in this study. The study used 12 categories in the
analysis for conversational CS. These categories were used to indicate the
conversational function of the code switch and were always determined in
relation to the contextual information at a specific point of the ongoing
conversation.2
1. Representation of speech: CS employed to represent talk.
t: all right I’m calm I’m calm … a lo major están diciendo estan chulos
e: . . . may be they are saying that we are good looking
2. Imitation quotation: CS involving imitation and change in tone of voice to
play a particular character.
t: y luego le hace si {[robot voice] i’m hungry}
e: and then he says yes
3. Turn accommodation: CS occurring between speakers’ turns.
Claudia: t: y luego se va salir a las cuatro y media from school

84 Bilingual Research Journal, 28: 1 Spring 2004

 e: and then she’s getting out at four thirty from school

Jimena: t: de la project school?

e: from the school project?
Claudia: t: summer school she don’t need to go to summer school
4. Topic shift: CS occurring due to a change of topic in conversation.
t: we finished all the books … thank you mira mis calzones se me andan
cayendo
e: . . . look at my underwear they are falling
5. Situation switch (on/off topic in academic work): CS marking a switch
between science talk and non-science talk.
t: mira mira los magnets … what was he saying during recess
e: look look the magnets
6. Insistence (non-command): CS indicating a child’s persistence in a specific
idea. The child usually repeated the same utterance in both languages.
t: a ver . . . let me see
e: let me see . . .
7. Emphasis (command): CS used to put emphasis on a specific command.
t: loud leelo loud!
e: loud read it loud!
8. Clarification or persuasion: CS giving more information to clarify an idea
or message.
t: andale pues esa cosa we don’t need this no more
e: yeah that thing
9. Person specification: CS occurring when children referred to
another person during their conversation.
t: you should ask A#given name# si quiere comer nieve después de la
escuela
e: if she wants to eat ice cream after school
10. Question shift: CS indicating a switch in language when children had a
question.
t: let me see cómo le hiciste?
e: how did you do it?
11. Discourse marker: Discourse markers are linguistic elements that do not
necessarily add to the content of the utterance but act as markers of the
context in which the utterance is taking place (Escalera, 2002).
t: okay asi ira
e: like this look
12. Other: This last category was used to code instances in which the function
of the CS could not be identified.
t: a mi me gusta big teeth
e: i like

Functions of Code Switching in Schoolchildren’s Conversations 85

 Results
Language Use
During the social talk, both the 7-year-olds and 10-year-olds used more
Spanish than English. This finding was expected, considering the children’s
linguistic proficiency in Spanish and English. Unexpectedly, a similar pattern
was observed for language use during the science activity. Developmentally,
it was anticipated that the 7-year-old children would use their native language
(Spanish) with a higher frequency than the 10-year-olds because the latter
have developed higher degrees of bilingual proficiency. Table 2 shows the
frequency for each language used by age group across the two contexts: the
social setting and the science activity setting.3
As can be seen in Table 2, on average, the 7-year-olds used Spanish in
81% of their total conversational turns in the social setting and used English
for only 7% of their conversational turns in that segment. For the 10-year-
olds, there was a decrease in the use of Spanish, hence an increase in English,
as well as a decrease in code-switched turns. Both groups of children used
their native language, Spanish, as the base language during their
conversations. However, we see that the 10-year-olds used English and mixed
English and Spanish in about 33% of their turns, in comparison with the 7-
year-olds, who only used English and mixed English and Spanish for about
19% of their turns.
Table 2
Average Percentage Distribution of Total Conversational Turns by
Language Choice

C ontext- language choice in social talk

Participants Spanish English Both

7- year- olds 81% 7% 12%

10- year- olds 67% 17% 16%

Language choice in science activity (on- task and off- task talk included)

7- year- olds 77% 9% 14%

10- year- olds 68% 15% 17%

During the science activity, surprisingly, both age groups used Spanish
again as the base language. The 7-year-olds used Spanish for more than 75%
of their total turns during the science activity. Hence, the 7-year-olds showed
a slight increase in use of English during the science activity, compared with
the social setting, but they still used Spanish as the primary language for

86 Bilingual Research Journal, 28: 1 Spring 2004

discussing social and academic topics during peer interaction. The 10-year-
old children showed a similar pattern of language use in both contexts (social
talk and on-task science talk).

Code Switching
The results indicate that the younger children code switched less than
the older children. The 7-year-old children used Spanish as the base language
during the social talk, code switched to English, and then code switched back
to their base language. The 10-year-old children code switched with greater
frequency than the younger children (see Table 3). Furthermore, the analysis
reveals that on average, the 7-year-olds used three types of CS, in comparison
with the 10-year-olds, who used an average of five types of CS during their
conversations. Older children seemed to use a wider range of CS to fulfill
different sociolinguistic functions. This finding seems to indicate a
relationship between the number of years children have been exposed to their
L2 and the influence on their bilingual communicative skills.

Code-Switching Frequency
A comparison of CS during the two contexts—social talk and the science
activity (on-task talk)—found that 7-year-olds showed a similar pattern in the
frequency of code switches across contexts: an average of 13% and 10% for
social talk and science activity (on-task talk), respectively (see Table 3). Only
a small decrease (3%) of code switches was observed in looking at social talk
as opposed to science talk for the 7-year-old children. An identical pattern
was observed for the 10-year-old children in the social talk and science activity
(on-task talk) contexts. The average frequency of code switches in their
conversational turns—in both contexts—was about 28%.
Table 3
Average Percentage of Turns With Code
Switches Used by Children in Different Contexts
Age group Social talk Science activity
(on- task talk)

7- year- olds 13% 10%

10- year- olds 28% 28%

The results indicate that 10-year-old children code switched significantly

more in all contexts when compared to 7-year-old children. An analysis of
variance (ANOVA) with age as the between-subjects variable and context as
the within-subjects variable confirmed these important findings. First, the
difference in frequency of CS between the two age groups was statistically
significant (F [1, 31] = 11.80, p < .01), showing a higher frequency of CS by the

Functions of Code Switching in Schoolchildren’s Conversations 87

older children (10-year-olds). However, the difference in frequency of CS
according to context was not significant. That is, both groups of children
code switched with the same frequency during the social context as they did
during the science activity context.
In addition, an analysis was done to determine if there was a relation
between the frequency of CS and the number of years a child had been exposed
to the L2. In order to determine if these two variables were correlated, the
researcher assigned the 10-year-old children to two different groups. The first
group included those children who had been exposed to English (the L2) for
2 years or less. In the second group, only those children who had been exposed
to the L2 for at least 3 years were included. The percentage of conversational
turns in which the two groups of children code switched is illustrated in
Figure 1.

35

30
Percentage of total code-switches

25

20

15

10

0
1 or 2 years 3 or more years
Years of exposure to second language (English)

Figure 1. Percentage of 10-year-olds’ conversational turns that used

code switching, by number of years exposed to English.

The 10-year-old children who had been exposed to English for 3 or more
years code switched in 33% of their total conversational turns, in comparison
with the 10-year-olds who had been exposed to English for 1 to 2 years, who
only code switched in 15% of their total turns. This result shows a trend in the
relationship between the number of years children have been exposed to an
L2 and the frequency and use of CS in their conversations (this will be further
addressed in the Discussion and Conclusion section).

Code-Switching Functions
For the sociolinguistic analysis, children showed variation in the types
of CS functions they used according to context. A summary of the most
commonly used types of CS by age group is given in Table 4.

88 Bilingual Research Journal, 28: 1 Spring 2004

Table 4
Sociolinguistic Functions of Code Switches by Activity, in Order of Frequency

Age group Social talk Science activity Science activity

on- ask talk off- task talk

7- year- olds 1. Topic shift 1. Clarifcation 1. Situation switch

2. Emphasis 2. Emphasis 2. Topic shift
3. Clarification 3. Question shift 3. Emphasis

10- year- olds 1. Topic shift 1. Clarification 1. Situation switch

2. Clarification 2. Topic shift 2. Topic shift
3. Accommodation 3. Question shift 3. Accommodation

Across age groups, for the social context, the four most commonly used
types of functions were topic shift, clarification, emphasis, and turn
accommodation. The topic shift type of code switch was the most frequently
used by both groups of children during the social setting (24% of total code
switches among 7-year-olds and 19% of switches among 10-year-olds). Children
frequently code switched while changing topics. This type of CS function
seems to be particularly important in social talk, because children spent a
great deal of time talking about social events and teasing each other (see
example below):
Fernando: t: ira lo que voy hacer
e: look what i’m going to do #child is playing with Play-Doh#

Cesar: t: que son? son Power Rangers?

e: what are they? are they Power Rangers?

t: go go Power Ranger… man, is this the activity? es la migra

blood
e: they are the immigration
#child is referring to researchers as immigration officers#

[both children laugh]

In this example, we see how Fernando and Cesar, 10-year-old boys, switch
languages from Spanish to English when talking about popular characters
(Power Rangers) in a television show, but switch back to Spanish when
changing topics to refer to “immigration officers.” These children were teasing
each other while code switching to Spanish to refer to the researchers (who
were in the next room) as the migra [immigration officers]. This particular
dyad continues to code switch throughout the conversation in both the social
and science activity contexts.

Functions of Code Switching in Schoolchildren’s Conversations 89

 The second most used category by the young children was emphasis,
followed by clarification. For the 10-year-olds, clarification was the second
most used sociolinguistic function, followed by turn accommodation.
Specifically, older children accommodated their speech by CS according to
their peers’ linguistic knowledge, a strategy that draws from implicit
metalinguistic knowledge4 to monitor speech. These findings showed that
older children, who have developed bilingual communicate competence, were
aware of their listeners’ linguistic abilities, as exemplified in the following
conversation of two fifth graders:
Alberto: t: hay *vas mas o menos con el *inglés
e: you’re getting english more or less

Francisco: t: yo invento *cosas /

e: i invent things

Alberto: t: *eh?
e: ‘huh?’

Francisco: t: *yo invento *cosas/

e: i invent things

Alberto: t: pero *porque buey?

e: but why dude?

Francisco: t: *asi como el *inglés

e: like in english
t: le *hace...le hace mi *hermana..l-..(es)tá...-ta todavia
-ta *chiquita y le hace...lo *hace/
e: she goes…my sister goes..l-..my sister goes l-...she’s she’s
still she’s small and she goes . . . like this’

Francisco: t:{[with a deep Spanish accent] oh tu espeak

*espanish..*no/
no *inglish..no *comprende}/
e: oh you speak Spanish...no...no english..you don’t
understand
t: she say oh you speak spanish no.. no english..don’t
understand
e: she (says) oh you speak Spanish . . . no . . . you don’t
understand english

Alberto: t: oh/ good/

In the example above, Alberto (bilingual) and Francisco (Spanish
dominant) are talking about how Francisco is getting better at speaking English.
Alberto refrained from CS and carried out a conversation with Francisco mostly

90 Bilingual Research Journal, 28: 1 Spring 2004

in Spanish. In this example, we also observed one of the two code switches
produced by Francisco during their conversation. This example supports the
claim that those children who can code switch to accommodate peers during
interaction demonstrate higher bilingual communicative skills.
During the science activity (on-task talk), the four most common types of
code switches used across age groups were clarification, emphasis, topic
shift, and question shift. In terms of code-switch variation during the science
activity, the only difference across age groups was the use of the emphasis
type of CS for the young children and the topic shift type of CS for the older
children. The pattern of code switches children used during the science activity
when they were off task is more similar to that of the social-talk context. The 7-
year-olds used situation switch, topic shift, and emphasis more frequently
than any other type of CS. In a similar way, 10-year-olds used situation switch,
topic shift, and accommodation shift within their conversations.
This slight variation in the type of code switches observed across age
groups suggests a developmental trait. For example, the use of the emphasis
type of CS by the 7-year-olds during the science activity (on-task talk) reflects
the use of more commands, hence, more command-related emphasis through
CS among 7-year-olds to get through the science activity questions. This
finding corroborates with Jørgensen’s (1998) and Bain’s (2000) earlier findings
that document young children’s use of CS for power, control, and dominance.
Therefore, the variation in the use of sociolinguistic functions for CS reflects
particular means to achieve discourse goals during peer interactions.

Discussion and Conclusion

An unexpected finding was the consistent use of Spanish (the children’s
native language) as the base language in both types of conversations and
contexts. A possible explanation to these findings is that the children might
have felt some pressure in carrying out the science activity and answering the
questions, in order to report back to the adult researcher. Therefore, the children
were more comfortable using Spanish to solve the science activity. Because
the activity was a cognitive task, the children might have switched to their
native language, Spanish, as a result of wanting to get through all the science
activity questions. A similar finding was reported by Glaessner (1995), who
noticed that her bilingual students used two languages strategically to engage
in private speech and collaborative speech when cognitively challenged.
Moreover, Benjamin (1996) also documented schoolchildren’s use of their
native language, Spanish, to communicate their knowledge to carry out
complex academic tasks while negotiating participation in ongoing
conversations. Similarly, the children in the present study used their native
language (Spanish), in addition to CS, to negotiate conversational involvement
while seeking explanations during the science activity, as exemplified below:

Functions of Code Switching in Schoolchildren’s Conversations 91

 Araceli: t: [ reading] ¿qué es lo que los im::anes hacen al compass
*brújula brújula?
e: what do magnets do to a compass compass?’
Mari: t: um…
Araceli: t: compás?
e: compass?
Mari: t: estás segura?
e: are you sure?
Araceli: t: (you) know
Mari: t: ~um-huh
Araceli: t: do you understand one?
Mari: t: can you use it?
Araceli: t: entiendes número uno?
e: do you understand number one?
Mari: t: huh?
Araceli: t: le entiendes?
e: do you get it?
Mari: t: yeah, it goes north

In the example above, two fifth-grade girls are switching back and forth
between Spanish and English to check with each other their understanding of
one of the questions during the science activity. In this particular example, the
children use CS as a clarification function successfully to answer the questions
on the worksheet.
In terms of the frequency patterns of CS, the 10-year-olds showed a
higher frequency than the 7-year-olds of code switches in all contexts—
social talk, science activity on-task talk, and science activity off-task talk.
Interestingly, both age groups showed about the same percentage of code
switches across contexts. In other words, context, at least in this specific
situation, did not seem to affect the frequency of CS among the children.
Previous research by McClure (1981) also reported that setting was a secondary
factor for language choices by Mexican American children.
In sum, there was some variation in the sociolinguistic functions of code
switches used with respect to the type of context. During the social talk, the
most systematic function for CS among children was the topic of the
conversation. An important difference found between the two age groups
was the function of accommodation CS to the other speaker. In the present
study, children during peer interaction seemed to be monitoring and
accommodating their peers’ linguistic abilities (see example above). These
findings showed that older children, who have developed bilingual competence

92 Bilingual Research Journal, 28: 1 Spring 2004

(including metalinguistic knowledge of two languages), were aware of their
listeners’ linguistic abilities and accommodated to their peers’ skills. Lanza
(1997) has pointed out that metalinguistic awareness can be evidenced in
different ways. For the bilingual child, CS is an index of bilingual competence
development. For the science activity (on task), switching to add emphasis to
a statement and to clarify information given to the listener was important
when a child had a question regarding the activity. Children manipulated their
languages in order to make sure the other child understood what they were
trying to communicate. Results from this study confirmed McClure’s (1981)
previous finding that using CS to clarify meaning through translation seems
to be learned at quite a young age. Of particular interest in this study is the
finding that CS to mark topic shift seems to be a type of code switch that is
learned at an earlier age than other, more sophisticated types of code switches
(e.g., turn accommodation, imitation quotation).
In conclusion, these results show that children during peer interaction
use the language with which they both feel most comfortable and have greater
competence. The results also indicate that those speakers with the greatest
degree of bilingual communicative competence are the ones who most
frequently use CS as a strategy to meet their conversational goals and to
communicate with their peers. This finding suggests that a positive
relationship exists between bilingual CS and language proficiency, and that
“the number of instances of ‘code-switching’ can be interpreted to reflect the
child’s developing communicative competence” (Genesee, 2002, p. 190). This
is an important finding that challenges the view of CS as a sign of
communicative incompetence. Instead, it calls for attention from teachers,
parents, and educators to draw on children’s language resources (Ruiz, 1984)
to help students achieve academic and communicative competence. In recent
work in educational sociolinguistics, Gumperz, Cook-Gumperz, and Szymanski
(1999) have identified CS as one of the skills bilingual children used during
cooperative learning environments. They argue that CS is part of children’s
linguistic knowledge, and that it must be recognized as conveying essential
information that affects their understanding of others.
Children, as they get older, seem to learn to code switch for more
sophisticated purposes. One explanation could be that they become more
sensitive to their peers’ linguistic abilities, consequently becoming better at
CS to accommodate their linguistic demands. That is, they know roughly the
level of bilingual fluency that their friends and adults possess, and at the
same time this helps them to be better communicators by taking advantage of
their bilingual skills. In a recent article, Ervin-Tripp and Reyes (in press) have
described this skill as part of children’s developmental bilingual pragmatics.
Therefore, as bilingual children’s exposure to the L2 increases, their bilingual
communicative skills advance and they are able to switch languages with
ease, knowing when it is appropriate to mix them and when they should keep
the languages separate.

Functions of Code Switching in Schoolchildren’s Conversations 93

 This study has attempted to explore the different patterns of CS functions
Spanish–English bilingual children use during peer conversation. For future
studies, it will be necessary to do a comparative analysis including other
languages and children from different ages to see how the use of their native
language and CS during peer interaction extends the linguistic repertoire of
bilingual language learners for greater pragmatic and academic competence.
These findings are particularly important for classroom teachers, because CS
should be understood as a tool for cognitive development and a skill children
use to achieve communicative goals. Moreover, the findings from this study
indicate that CS is a complex skill children develop as part of communicative
competence. Educators, teachers, and parents should not consider CS a sign
of cognitive confusion, but instead an indicator of children’s bilingual
competence. This research should help educators of second language learners
facilitate discourse practices in the classroom according to students’ first
language and second language linguistic skills.

References
Álvarez, A. I. (1979). Code-switching among bilingual children.
Unpublished doctoral dissertation, University of Massachusetts,
Amherst.
Bain, B. (2000). Toward a fuller appreciation of codeswitching. Journal of
Pragmatics, 32, 1405–1411.
Bayley, R., & Zapata, J. (1993). Prefiero Español because I’m used to it:
Code-switching and language norms in south Texas. Working
Papers, 02. San Antonio: University of Texas at San Antonio,
Hispanic Research Center.
Benjamin, R. (1996). The functions of Spanish in the school lives of Mexicano
bilingual children. Bilingual Research Journal, 20(1), 135–164.
Clyne, M. G. (1967). Transference and triggering. The Hague, Netherlands:
Martinus Nijhoff.
Ervin-Tripp, S. (1964). An analysis of the interaction of language, topic &
listener. American Anthropologist, 66(6, Part 2), 86–102.
Ervin-Tripp, S. (2001). Variety, style-switching, and ideology. In P. Eckert &
J. Rickford (Eds.), Style and variation (pp. 44–56). New York: Cambridge
University Press.
Ervin-Tripp, S., & Reyes, I. (In press). From child code-switching to adult
content contrast. International Journal of Bilingualism.
Escalera, E. (2002). The role of context in children’s use of discourse
markers. Unpublished doctoral dissertation, University of California,
Berkeley.
Fantini, A. E. (1985). Language acquisition of a bilingual child: A
sociolinguistic perspective. San Diego, CA: College-Hill Press.

94 Bilingual Research Journal, 28: 1 Spring 2004

Fishman, J. A. (1965). Who speaks what language to whom and when? La
Linguistique, 2, 67–88.
Genesee, F. (2002). Portrait of the bilingual child. In V. Cook (Ed.),
Perspectives on the L2 user (pp. 170-196). Clevedon, England:
Multilingual Matters.
Genesee, F., Boivin, I., & Nicoladis, E. (1996). Talking with strangers: A
study of bilingual children’s communicative competence. Applied
Psycholinguistics, 17, 427–442.
Genishi, C. (1976). Rules for code-switching in young Spanish-English
speakers: An exploratory study of language socialization.
Unpublished doctoral dissertation, University of California, Berkeley.
Glaessner, B. E. (1995). Bilingual children in school: The negotiation of
language use in a first-grade classroom. Unpublished doctoral
dissertation, University of Chicago.
Gumperz, J. J. (1964). Linguistic and social interaction in two communities.
American Anthropologist, 66(6, Part 2), 137–154.
Gumperz, J. J. (1967). On the linguistic markers of bilingual communication.
Journal of Social Issues, 23, 48–57.
Gumperz, J. J. (1973, April). The communicative competence of bilinguals:
Some hypotheses and suggestions for research. Language in Society,
2(1), 143–154.
Gumperz, J. J., & Berenz, N. (1993). Transcribing conversational exchanges.
In J. A. Edwards & M. D. Lampert (Eds.), Talking data (pp. 91–122).
Hillsdale, NJ: Lawrence Erlbaum Associates.
Gumperz, J. J ., Cook-Gumperz, J., & Szymanski, M. (1999). Collaborative
practices in bilingual cooperative learning classrooms: Research Report
7. Santa Cruz, CA: Center for Research on Education, Diversity, &
Excellence.
Gumperz, J. J., & Hernández-Chavez, E. (1975). Bilingualism, bidialectalism,
and classroom interaction. In C. Cazden, V. P. John, & D. Hymes
(Eds.), Functions of language in the classroom (pp. 84–107). New
York: Teachers College Press.
Halmari, H., & Smith, W. (1994). Code-switching and register shift: Evidence
from Finnish–English child bilingual conversation. Journal of
Pragmatics, 21(4), 427–445.
Hymes, D. H. (1974). Foundations in sociolinguistics: An ethnographic
approach. Philadelphia: University of Pennsylvania Press.
Jørgensen, J. (1998). Children’s acquisition of code-switching for power
wielding. In Peter Auer (Ed.), Code-switching in conversation:
Language, interaction and identity (pp. 237–261). London:
Routledge.

Functions of Code Switching in Schoolchildren’s Conversations 95

Lanza, E. (1997). Language mixing in infant Bilingualism. New York:
Oxford University Press.
McClure, E. (1981). Formal and functional aspects of the code switched
discourse of bilingual children. In R. P. Duran (Ed.), Latino language
and communicative behavior (pp. 69–84). Norwood, NJ: Ablex.
Poplack, S. (1980). Sometimes I’ll start a sentence in English y termino en
español. Linguistics, 18(7&8), 581–618.
Reyes, I. (2001). The development of grammatical and communicative
competence in bilingual Spanish speakers. Unpublished doctoral
dissertation, University of California, Berkeley.
Romaine, S. (1995). Bilingualism (2nd ed.). Oxford, England: Blackwell.
Ruiz, R. (1984). Orientations in language planning. Bilingual Research
Journal, 8(2), 15–34.
Saunders, F. (1975). Code switching in bilingual children. Unpublished
doctoral dissertation, University of California, Santa Barbara.
Schieffelin, B., & Ochs, E. (1986), Language socialization across cultures.
Cambridge, England: Cambridge University Press.
Weinreich, U. (1953). Languages in contact. The Hague, Netherlands:
Mouton.
Zentella, A. C. (1982). Code-switching and interaction among Puerto Rican
children. In J. Amastae & L. Ellas-Olivares (Eds.), Spanish in the
United States: Sociolinguistic aspects (pp. 354–385). Cambridge,
England: Cambridge University Press.
Zentella, A. C. (1997). Growing up bilingual. Maiden, MA: Blackwell.

Acknowledgments
Support for this research was provided by a Social Science grant by the
University of California, Berkeley. I am grateful to Susan Ervin-Tripp for
comments and discussion on earlier drafts of this manuscript. Please address
correspondence to Iliana Reyes at Department of Language, Reading and
Culture, College of Education, University of Arizona, Tucson, AZ 85721-0069.
E-mail can be sent to: [email protected]

Endnotes
1.
The first line, t:, is the transcript literally from the tape. Line e: is a free translation
into English to give the meaning. The * indicates accent and normal prominence.
2.
Pseudonyms have been used to protect children’s identity.
3.
For comparison purposes, proportions were computed by dividing the number of
code switches by the number of speaking turns for each child to yield a measure of CS
probability per turn in each context.
4.
The term metalinguistic knowledge refers to the ability to reflect and think about the
nature and functions of language (Lanza, 1997).

96 Bilingual Research Journal, 28: 1 Spring 2004

 Appendix A

Science Activity Questions

Cuestionario de ciencias naturales

1. ¿A que lado apunta la brújula?

Which way does the compass point?

2. What do magnets do to a compass?

¿Qué es lo que los imanes hacen a la brújula?

3. What do the magnets do to the other things?

¿Qué hacen los imanes a los otros objectos?
Find all the different things the magnets do. Encuentren todas las cosas
que se pueden hacer con los imanes y los diferentes objetos.

4. Usen los imanes con la caja que contiene el polvo con hierro, y volteenlo
al revés ¿Qué es lo que pasa? ¿porqué? Put a magnet on the box with iron
filings and turn it upside down. What happens? Why?

5. How can you tell which objects contain iron?

¿Cómo puedes saber que objetos contienen hierro?

6. Traten de darle vuelta a la piedra“hematite” (negra/lisa) con uno de los

imanes. ¿Porqué da vueltas? Try to spin the hematite with the bar magnet.
Why does it spin?

7. The N and S on the magnets mean north and south. What happens when
you put two norths or two souths together? How about a north and a
south? La N y S en los imanes significan norte y sur. ¿Qué pasa cuando
tratas de poner dos N o dos S juntas, y que pasa cuando pones la N y la S
juntas?

8. ¿Cómo puedes saber cual es Norte o Sur en los imanes redondos? How can
you tell which side of the circle magnets is north or south?

Functions of Code Switching in Schoolchildren’s Conversations 97

 Appendix B

Transcription Notation

(from Gumperz & Berenz, 1993)

Symbol Significance

* Accent; normal prominence

** Extra prominence
~ Fluctuating intonation over one word
? Rising intonation at end of intonation contour
/ Falling intonation at end of intonation contour
, Holding intonation at end of intonation countour
// Turn-final intonation
- Truncation (e.g., what ti- what time is it/)
:: Lengthened segments (e.g., wha::t)
.. Pauses of less than .5 second
... Pauses greater than .5 second (unless precisely timed)
<3> Precise units of time (= 3 second pause)
( ) Unintelligible speech
(they) A good guess at an unclear word
(“ “) Regularization (e.g., i’m gonna (“going to”) come soon/)
# # Use hatchmarks when extratextual information needs to be
included within the text (e.g., R: did you ask E #surname# to
come?)
[ ] Nonlexical phenomena which interrupts the lexical stretch
(e.g., text[laugh] text/ )
{[ ]} Nonlexical phenomena which overlays the lexical stretch
(e.g., {[laugh] text/} text/)
= To indicate overlap and latching of speakers’ utterances
(e.g. L: so you understand = the requirements =
G: = yeah, I under = stand them/)

Note. The use of uppercase letters is restricted to proper names to identify a person
or place. All regular orthographic annotations have a specific meaning for this
transcription notation.

98 Bilingual Research Journal, 28: 1 Spring 2004