Table of Contents
3 Language Mixing
4 Study - Phoneme triggered lexical decision task
4.1 Study - Guest words
4.2 Effects of guest word recognition
4.3 The TRACE model
For the interest to researchers of Bilingualism the code-switching (CS) received the highest interest in the past few years. CS can be defined as the ability that bilinguals can switch easily between two or more languages. Also it is normal, that bilinguals integrate their two languages to a given degree. Further the act of CS can be studied as a reflection of social constructs and of the cognitive mechanism that control language switching. With other words this phenomena is known as on-line processing, which describes for instance operations that take place during speech comprehension and production.1 For the comprehension this is the moment where a utterance is heard or seen by a receptor and then accordingly processed and stored in the brain. In speech production this applies to the operations taking place between the moment of the intention to communicate something and the actual utterance.2
While the first view of CS is the psycholinguistic one it is important to mention also the sociolinguistic view on CS. It is important to understand that there is a different behavior of CS between women vs. men, middle-class vs. working-class and other social groups. It is claimed that the language manner and use are related to the speakers' social identity. From the sociolinguistic point of view it can be argued that sociolinguists can help us to understand CS in a different way than mentioned in the preceded passage.3
All in all CS comprises a wide range of contact phenomena. First there are insertions of single words or larger segments of a speech. Second, when bilinguals produce differing degrees of proficiency who reside in various types of language contact. For these phenomena, there are many reasons: filling linguistic gaps, expressing ethnic identity, achieving particular discursive etc.4
The following Essay shows first some general aspects of Bilingualism and the difference kinds of CS. After this short explanation there are two studies presented and reasoned model will be shown, which explains reasons for CS leaned on the studies before.
For this Essay are two language modes important, the monolingual language mode and the bilingual language mode. Therefore monolinguals try to deactivate their second language as good as they can, but it causes speaker-specific deviation from language at all levels: phonological, lexical, syntactic, semantic, pragmatic. These deviations are called static interferences and dynamic interferences. Static interferences are permanent traces of one language on the other, like permanent accent, the meaning extensions of particular words, specific syntactic structures etc. Dynamic interferences mean the short intrusion of the second language, like accidental slips on the stress pattern, the momentary use of syntactic structure from the other language etc. On the other hand bilinguals have both of their languages activated, so the language can shift from one to the other language, which can be caused by the situation, the topic, the interlocutor and the function of interaction.5
The question, which psycholinguists ask is, if a bilingual possess one or two internal lexicons which gets us to three different theories. So antagonists of the one-lexicon view claim that all is stored in one single semantic system, but every word is linked to the language it belongs to (interdependent storage view). Researchers of the independent storage view think that the information from one language is only available through a translation process. A third theory says that bilinguals have three stores, one conceptual store for the bilingual's knowledge of the world and one language store for each language he/she knows.6
3 Language Mixing
In the bilingual language mode the speaker chooses one language which he speaks (language choice), but we can see there different Phenomena's. Language mixing is the kind of Phenomena, where for example aspects of a native language are transferred into a second/subsequent language. Furthermore there is a hiring of forms and structures from one language into another. Language mixing is often made up of two different processes: CS and borrowing. CS is also distincted into two different types, the intersentential and intrasentenial. Intersentential means that the switch occurs between sentences. For example, when the topic or the participant changes. The second kind of CS occurs when the switching of languages is within sentence boundaries. In other words it means that both languages are in use at the same time. Both switching types are often seen in informal situations, where bilingual communities talk a lot (conversational CS). Borrowing is called the use of a word from one language in the usage of another. Every kind of word class can be borrowed. For example the word "tortilla" is a borrowed word from the Spanish language and describes a round type of flat bread used in Mexican food dishes. Languages can also borrow meanings (without the form). Calques are a good example for this kind of borrowing, where meaning is borrowed and represented morpheme by morpheme in a native form, like the English "Superman" which is borrowed from the German "Übermensch".7
4 Study - Phoneme triggered lexical decision task
In the study of Soares and Grosjean investigated the lexical access of base-language words and code-switched words by means of the Phoneme Triggered Lexical Decision task. They did this, because word recognition in bilinguals has received much less attention, especially in the point of code-switches and borrowings.
In the experimental set-up English-Portuguese bilingual subjects got sentences and they were asked to listen to words or nonwords, which began with a pre-specified phoneme. If a probant found this word/nonword, he/she had to decide as quickly as possible whether the item was a real word or not. English monolingual subjects became only the English sets of sentences. So the bilingual probants got tested on three sets of sentences: English, Portuguese and code-switched and before each set the appropriate language mode was induced.8
Two main findings were followed from the study. The first finding was, that bilinguals accessed real words just as the English monolinguals, but they were slower in responding to nonwords. Because Soares and Grosjean thought that this finding is linked up to the residual activation of the other language, when the bilingual is in monolingual language mode. They hypothesized that a nonword appeal to a complete search of the base-language lexicon (or an activation of the lexicon, depending on the access theory one espouses). Besides it is followed by a search of the other lexicon. In this way both researchers explain the longer reaction time. The second finding was that bilinguals took longer to access code-switched words in the bilingual language mode than they did base-language words in the monolingual language mode. Like in the first finding they suggested, those bilinguals always search the base-language lexicon before they activate the less used lexicon, but in a later paper both claimed that much more factors can affect this delay: Cross-language coarticulation, delay or absence in the speakers’ production, base-language context and the listener's tendency to assimilate ambiguous items during the perception of code-switches.9
4.1 Study - Guest words
In the studies of CS a guest language is the one from which words or structures are taken and embedded into another host language. So all words, which are integrated phonologically and morphologically into the host language, can be called guest words. Consequently this means that this words are pronounced as if they were native words and they receive prefixes/suffixes as if they were native words.10
In a more recent study Grosjean wants to point out, why it takes more time for a bilingual to recognize guest words. He assumed that there are two lexicons which are interconnected in some way. He thinks that guest words are stored in the less activated lexicon and just have to be accessed during bilingual speech (his hypothesis).
In his Experiment he had two variables: (1) a structural or word type variable and an (2) output or language phonetics variable. Based on the first variable there are three questions which should be answered by the experiment:
1. Would guest words that are marked phonotactically be recognized sooner and with more ease than words not marked in this way?
2. Would guest words that belong solely to the guest-language lexicon be easier to process than words that do not belong to just one lexicon?
3. Would guest words that have near homophones in the base language be recognized with more difficulty than other guest-language words?
For the second variable he asked, if "code-switched words, which normally retain a phonetic cue as to the lexicon they are a part of, be accessed more easily than borrowings which are usually integrated into the base language, and hence have lost some of the cues pertaining to their lexicon of origin?"11
The Gating-experiment was realized by French-English bilinguals. A guest word (English verb) was embedded into a French phrase (Il faudrait qu'on...) and followed by a noun phrase. The bilinguals listened to the accordings, which were recorded by a French-English bilingual with no accent, and subjects were asked to guess the word being presented.
In the relation to the three questions, which Grosjean pointed out before there are answers for every question after this study:
1. Phonotactically marked words were identified sooner than words that are not marked in this way.
2. Words that solely belong to guest lexicon were recognized sooner.
3. Words in the guest-language lexicon that have close homophones in the base language were processed with more difficulty.
Besides it can be said that the hypothesis is proved correct. There is a higher correlation between word frequency and isolation point.12
For the second variable he discovered that during the selection phase the relationship of guest-language candidates was greater for code-switchers than for borrowings. He also figured out that strong phonetic cues activated both lexicons, depending on the phonetics of the guest word and it affected the language the candidates used.13
1 Grosjean P. 259f
2 Bullok P.1ff
3 Bullok P. 97f
4 Bullok P.2
5 Grosjean P.259 ff
6 Grosjean P.260
7 Field P.103/30/25
8 Grosjean P.265
9 Grosjean P.265f
10 Field P. 79
11 Grosjean P. 266
12 Grosjean P. 267
13 Grosjean P. 268