1.Introduction

Since China entered the global business market in 1970s, numerous enterprises from various countries have come and set their branches in China. “Foreign company employees” (FCEs), referring to local Chinese employees who work for the Chinese branches of foreign-invested enterprises, has become a considerable social community in modern China [1]. One salient characteristic of this group is the language they use: the mixed language of Chinese and English [2]. The author of this research personally experienced and used this “mixed language of Chinese (here specifically refers to Mandarin) and English” during nearly 3 years’ working experience in a Chinese subsidiary of an American company, which offered this research a solid observation and deep understanding of this code-mixing phenomenon, and also the trigger to interrogate this code-switching strategy.

2.Literature Review

Code-switching (CS), defined as the “use of more than one language in the course of a single communicative episode” [3], has become a broadly discussed research area. Previous study on CS can be roughly divided into two groups. One group looks into CS within a linguistic ontological perspective: Scholars are studying on the syntactic or morphosyntactic constraints on language alternation [4,5]. The other group studies CS in a broader area of sociolinguistics, psycholinguistics and anthropological linguistics [3,6,7]. Code-mixing (CM) is defined as the intra-sentential code-switching within sentence boundary. The other type of CS is inter-sentential code-switching, referring to switch between different languages beyond the sentence boundary [8]. Among Chinese local FCEs in mainland China, the everyday most frequently used conversational pattern is the insertion of English lexical or phrasal elements into the matrix language of Mandarin, which is a typical example of intra-sentential CS. This article will use CM to refer to this integral Chinese-English mixed communicative pattern, while use CS to describe a specific case where a certain element is switched to English.

Though this code-mixing of FCEs has been noticed by Zhang’s work in 2005 [2], her research focused on Chinese yuppies’ phonological variation of Mandarin instead of the “mixed language” itself. No previous research of FCEs’ Chinese-English code-mixing has been found so far. Although there are some researches about the Cantonese-English CM in Hong Kong in a broader social settings [9], findings from these researches clearly cannot be generalized to cover the case of FCE community. These two situations are under significantly different social and cultural conditions, since CM in Hong Kong is more of a language contact-induced language change after Hong Kong’s British colonial period [10]. Areas used to be colonies is a classic situation where CM might occur, which is also the place where the majority of other worldwide CM researches are conducted, such as Bangladesh and Indonesia [11-14]. On the other hand, language-related researches that has been conducted in working places are mostly landed in the topics of inter-sentential CS among employees when facing different situations and top-down company language policy or language management [15]. None research focused on CM in multinational companies in China has been found. Thus, this present research might open up a new domain for future investigation.

3.Research Questions and Methods

The speech community where this research is conducted is a Chinese subsidiary belonging to an American company who has set its branches in more than 80 countries worldwide, and the Chinese subsidiary was set in late 1980s. With more than 8,000 local Chinese employees (LCEs) accounting for more than 95% of the total amount of employees working for its branch in mainland China, the conversations happening in this working place are mostly among Chinese (here equals to Mandarin) native speakers. And the reason why this research only studies oral discourses is because employees are often required to use only English when it comes to writing. The CS happening in this discourse is intra-sentential code-mixing (insertion of English lexical or phrasal elements into the matrix language of Chinese) instead of inter-sentential code-switching. Therefore, it is clear that the discourse this research wants to study has three constraints: 1) oral context; 2) with all participants being Chinese native; 3) intra-sentential code-mixing. This research is designed to answer two specific questions:

RQ1: What certain types of English lexical/phrasal items are switched in this discourse?

RQ2: What are the main pragmatic motivations of this code-mixing?

The data serving for this present research is mainly collected through four methods listed below in a chronological order. Findings of the present research could be proved more credible if each data source provides corroborative evidence to verify information obtained by other methods with triangulation:

a.Observation: Examples and notes are recorded immediately after related cases were observed;

b.Corpus analysis: 10 hours’ audio recordings of conversations that meets the 3 constraints clarified above have been taken and transcribed into text;

c.Semi-structured interview: Interview questions are designed based on the hypothesis (based on observation and corpus analysis) of each RQ. Then follow-up interviews have been conducted one by one with each participant for 45 to 60 minutes via phone call or face-to-face talk. Interviews are conducted in Chinese and recorded as audio files, transcribed by the software Iflyrec, and then double checked, corrected and translated into English by the author;

d.Questionnaire: typical statements collected from the participants’ expressions in interviews, and the preliminary findings drew from first three methods constitute the majority of question items in the format of seven-level Likert scale with “1” standing for “strongly disagree” and “7” meaning “strongly agree”. Therefore, present research hopes to double check whether the preliminary conclusions drew from the previous three methods with limited participants can be generalized to a larger population. 35 effective answer sheets have been collected in the end;

The participants involved in the interviews are deliberately selected with a balanced allocation in gender and oversea experience with similar age. All participants are not only the author’s colleagues but also familiar friends, which can help to create a comfortable atmosphere when conducting the interview, so it can encourage the participants to be more straightforward and open-up to talk. Overall information of interview participants can be checked as below in Table 1.

Table 1: Overview of The Interview Participants.

5.Classification of Code-switched Elements

A total number of 334 inserted English words or phrases has been discerned from the 10 hours’ audio recordings, among which most phrasal units actually also function as lexical elements since they are strongly conventionalized to be used together to present an integral meaning. Thus, in this classification, phrasal elements are divided into different groups together with all the other lexical units based on the grammatical meanings they carry in the structure of matrix language.These insertions allocate over all grammatical types of words, with nouns absolutely accounting for the majority, taking 58% of the total number. Therefore, nouns has been further divided into sub-categories as: proper noun (created to indicate a certain unique concept existing in the company’s repertoire, which might be unable to understand by people out of the company), noun-business terminologies and noun-others, with examples that can be checked in Table 2. The overall allocation of the code-switching elements can be check in Figure 1 below.

Figure 1: Allocation of Code-switched Elements.

Table 2: Detailed Classification of Code-switched Elements.

It is easy to explain why code-switching occurs when it comes to proper-nouns, since this type of elements were originally created in English without Chinese equivalents. Code-switching is conducted to fill the lexical gap [8]. Some of the business terminologies can also be explained in this way. Nevertheless, it can be seen in Table 2 that, other common nouns that are neither proper nouns nor terminologies actually account for a larger proportion. This type of code-switched elements normally has Chinese equivalents. Why code-switching happens in this situation is exactly where further investigation is needed. Based on the observation and corpus analysis, present research found that the original meanings of this type of nouns in English will change when they are used as insertions in the CM discourses. This kind of change of word’s meaning also follows the same two paths of the words’ meaning evolution in a monolingual environment: meaning narrowing and meaning transfer.

5.1.Meaning narrowing

In this case, when an English word gets inserted, only one of its various original meanings remains active in the CM context. Examples are as (1) and (2):

(1)Filed-de tóngshì-men kěnéng méi nàme duō-de exposure.

Filed’s colleagues might not have so much exposure

‘Colleagues in the field might not have so much exposure.’

(2)LM huì yāoqiú wǒmen zuò gè case chūlái.

LM will ask us make a case out

‘LM will ask us to make a case.’

The original English word field carries multiple meanings. But in sentence (1), it can only refer to “the working location that is out of the main offices of the company but in the field”. Another typical example is the word case in sentence (2), where it carries the specific meaning of “a well-performed project that can be submitted to bosses in the form of written report via email to gain exposure and notice from the bosses and reputation among counterparts, which can eventually help one get promoted”. This meaning is very strictly limited here, but when hearing this single word, everyone belonging to this speech community will get all of the information of the very specific standards a case must meet within a minute. This is also expressed by our participant E5 in his interview:

“For example, when we say make a case, we all know what that case refers to, and we know that this case is unique, like the case in our company is probably also different from the case in other domestic companies, but I would know exactly what kind of case I am supposed to make. ”

E5 also implied here that this type of code-switching can serve as an access to meet speakers and interlocutors’ pragmatic demands: By code-switching, both speakers and other interlocutors can express or understand a certain concept in a more accurate and efficient way, avoiding the effort on clarifying. This motivation will be discussed more in the following session.

5.2.Meaning transfer

This path is distinguished from the meaning narrowing in the way that, the active meaning working in the sentence are either an expansion of the original meaning in English, or a new derivative conventionalized in this speech community, with example of the word firm in sentence (3):

(3)Míngtiān yīqǐ kàn xiàgè cáinián-de firm.

Tomorrow together check next fiscal year firm

‘(Let’s) check the firm of next fiscal year together tomorrow.’

The word firm here carried an extended meaning from the original one as “a solid and massive business develop plan for a whole fiscal year”. Similar to the path of meaning narrowing, when meaning transfer happened in CM, the transferred meaning becomes the only one remaining active in the context, and members of this speech community have the shared knowledge of this specific meaning. It is safe to assume that by meaning transfer or meaning narrowing, the code-switched noun becomes a new or updated code that is different from the original code either in matrix language or embedded language system, which is agreed and conventionalized among the members of this speech community to express a certain concept existing in this setting via a more efficient and accurate way.

Although data shown above aligns with the previous research claiming that the most frequent code-switched elements are nouns and discourse markers [8,16], it can be seen in Figure 1 that, verbs, adjectives and adverbs also take a considerable proportion of 35%. These types of words together with discourse markers are carrying more salient pragmatic functions which will be further discussed in the next section.

6.Pragmatic Motivations

6.1.Accuracy

This motivation occurs mostly in the code-switching of nouns discussed in the previous section. CS can support speakers to express a certain concept more accurately when there is no corresponding word in the matrix language, or the meaning carried by the code-switched embedded language elements has already been narrowed or transferred, so that it can avoid misinterpretation or misconstruing among interlocutors. One of the interview participants E4 also supported this pragmatic motivation by stating: “I know what I am trying to express, but I don’t know its exact Chinese corresponding word… I don’t know how I can express it in Chinese in the most accurate way”. Data collected form questionnaires also proved that this is one of the motivations strongly agreed by participants with the highest average score of 6.16 (out of 7) towards the statement “The reason why certain words are converted into English expressions is because the meaning of these words is clearer/more accurate in English”.

6.2.Efficiency

Efficiency is the motivation mentioned most frequently by participants in interviews. How CS of nouns can provide a more efficient way to communicate has already been mentioned in the previous session, and this session will have a more focused discussion on proper nouns and acronyms.

This motivation is also well noticed by speakers as E3 stated in the interview: “(If without code-switching) it will be more difficult for me to express, or I may use a more complicated way to explain what it is…but in the culture of this company, if you use the acronyms or proper nouns, it can actually be understood by everyone immediately within a second, and this can improve efficiency.”

Acronym takes 10% out of the total English insertions collected from the corpus, allocating broadly in categories of proper noun and business terminology. The number of 10% cannot show how frequently speakers use acronyms because the times each acronym appeared in corpus was not counted in this research. But a very high frequency of the usage of acronyms is demonstrated both in observation and corpus collected, since these concepts are closely related to employees’ work. Acronym can offer an easier and less time-consuming way for speakers to express a certain concept, especially the concept has a high frequency in the discourse. It’s very common that speakers may not even know the full English words of the acronyms they use because there is no need for them to master the full words. Acronyms are already able to offer a more efficient code to express the specific concept. Some examples of acronyms in our corpus can be checked in Table 3 below:

Table 3: Examples of Acronyms.

Most proper nouns and business terminologies including acronyms are originally created in English, so it is easy to understand this type of code-switching. Participant E4 added that these proper nouns and business terminologies are very commonly used not only in Chinese branches, but also among other oversea colleagues coming from different language background. To keep a same conventionalized code system to express these frequently used concepts also improves the efficiency when communicating with foreign colleagues.

In terms of other nouns that are neither proper noun nor business terminology, previous section has proposed that via meaning transfer or meaning narrowing, CS also offers a more efficient way for members of this speech community to express concepts. And these nouns after CS actually also function similarly as proper nouns do.

The motivation to achieve a more efficient communication is also double clicked by the data collected from questionnaires, with a high average score of 6.03 (out of 7) indicating participants’ high agreement level to the statement “The reason why some words are converted into English is because it is more convenient, less time-consuming and more efficient to speak”.

6.3.Emphasis

Code-switching can also function as an emphasis and this motivation is especially salient when there are insertions of verbs and discourse markers. In the working place where present research looks into, code-switched verbs usually occur when speaker wants an action from others. Examples can be checked in sentence (4) and (5):

(4)Zhègè wèntí nǐ follow yīxià.

This issue you follow (up)

‘(I want) you to follow up on this issue.’

(5)618-de shǔ nǐ check-le ma?

618’s data you check-ed (interrogative particle)

‘Have you checked 618’s data?’

The verbs follow, check are switched into English to emphasize that the speaker wants others to take the action. Participant E5 compared this code-switching to “an alarm clock” that when he hears “the language is changed”, he would immediately be aware that here comes his “assignment”, and this is the information he needs to pay attention. But he wouldn’t feel the same way if the word is maintained in Chinese. Switched verb here helps the speaker to gain interlocutor’s attention with the emphasis created by the change to another language.

Similar function is also carried by the CS of discourse markers. In monolingual contexts, discourse marker usually signals a relationship between the following segment and the prior segment [17]. And when discourse marker is switched into another code, it also carries a function of emphasis [18], which is proved in this research as well that when discourse marker is switched into English, it alerts the interlocutors that there is going to be a change of the ongoing topic. Present research also found that CS of discourse markers can also be applied when the interlocutor wants to interrupt the speech of someone else. E2 stated in the interview as: “When I feel someone is talking too much, I may just say anyway (in English to cut in) ... I can’t find a better Chinese word, which is same as anyway (in English) …it is shorter. One can quickly end the topic. (Another example is) by the way, when someone is talking about one thing endlessly, if I want to interrupt... sometimes I may want to be tougher (by saying this)”. Conclusion can be drawn here that CS of discourse markers can be driven by a motivation to emphasis the switch of the topic of one’s own speech or to interrupt other’s speech.

6.4.Emotion Strengthening/Softening

Speakers tend to switch into English to say some adjectives and adverbs when they want to strengthen or soften the emotion carried in these words. Examples can be seen in sentence (6), (7) as below:

(6)Wǒ juéde zhègè shìqíng shì unfair-de.

I think this thing is unfair

‘I think this is unfair.’

(7)Zhègè yāoqiú yǒu yīdiǎn ridiculous.

This requirement is a little bit ridiculous.

Code-switching of the word unfair is conducted to emphasize the speaker’s emotion of disagreeing and being irritated, while the CS of the word ridiculous is meant to be more euphemistic. Whether the CS is conducted to soften or strengthen the emotional expression is based on the intonation of the speaker and the context of the discourse. Participant E2 stated in the interview that, if we use the Chinese word for ridiculous, it might be too direct and aggressive. CS into another language which is not the mother tongue of the interlocutor functions as a filter on the emotional expression. This motivation is also highly agreed by questionnaire participants with an average score of 5.19.

While code-mixing among Chinese local employees working for foreign companies in China have gained a lot attention and discussion from the society, and also have been noticed by previous scholars, few research has been conducted in this area. Thus, this research aims to have a brief overview of this code-mixing case and also to find the pragmatic motivations of this CM. Main findings of this study and verification of triangulation can be check in Table 4 as below:

Table 4: Matrix of Findings and Sources for Data Triangulation.

Note. Sources of data: O = observation, C = corpus, I = interview, Q = questionnaire

7.Conclusion

Limitations have to be admitted to exist in this research due to the limited time and effort: 1) The limited samples can’t support a more solid data analysis; 2) Corpus analysis was done merely within the lexical level, more interesting phenomena found in the observation and corpus are left to be studied with morphology/syntax/phonology perspectives; 3) Present research didn’t put enough efforts to take the diversity of different forms of discourses into consideration, such as informal talk with colleagues vs. formal business meetings and townhalls with bosses.

With these limitations, this research hopes to shed a light for the future researches from more perspectives by offering the possible avenue for the future studies in this area: 1) How matrix language constrains the insertion of embedded language; 2) Corpus analysis in the aspect of morphology/syntax/phonology; 3) The difference of code-switching strategies used in formal and informal context; 4) Investigation in the correlation between CM strategy and the variables like gender, years of service, language policy, language ideology with adequate samples collected.

CS/CM in multinational companies is an area worth more attention and can offer with a new aspect to look back at the previous theories and frameworks of CS. And this is also a place where resource can be offered for more valuable findinds to be drawn in areas like sociolinguistics, anthropology, psycholinguistics and etc.