DESIGNING FOR CROSS-CULTURAL GROUP CHAT

ABSTRACT

Cross-cultural communication is taking place everywhere as the world is getting flat. Research conducted across many disciplines (e.g. education, psycholinguistics, sociology, HCI and CSCW) shows that language fluency plays an important role in communication. Non-native speakers have been suffering from communication problems caused by language fluency issues both in face-to-face (FTF) communication and distributed online communication (Gonzalez, 2003; Neeley, Hinds, & Cramton, 2009). Building upon past studies, I investigated the communication process and outcomes in online cross-cultural group chat, with a particular focus on issues related to language fluency. Based on findings from this work, I proposed and designed a lightweight tool supporting secondary conversation stream. I evaluated and confirmed with a comparison study that group satisfaction and perceived control of the conversation are enhanced for the participants who used the tool. I also discovered new patterns of annotation use and design implications for group chat tools to support cross-cultural communication.

 

Chapter 1

 

Communication Problems in Cross-cultural Groups

Cross-cultural communication and collaboration has become more and more common in organizations and education systems, as multinational corporations hire local employees in their overseas branches, American companies outsource services to foreign states, and international students take online courses at U.S. universities. Compared to intra-cultural communication, communicating with people from other countries requires considerable extra effort, such as more confirmation checks, conversational remedies for misunderstandings, extended time for grounding and so on. Even when extra attention and extended times are allotted for cross-cultural communication, the communication outcomes are often inferior to intra-cultural exchanges. Culture is such a complex construct that it varies among multiple dimensions of language, value, social norm, belief and knowledge. It requires adequate and sometimes sophisticated strategies to reconcile those differences in a successful cross-cultural communication. As such, cross-cultural communication problems continue to be experienced and are a challenge worthy of attention from multiple research disciplines.

The Study of Cross-cultural Communication

Samovar & Porter (Samovar & Porter, 1997) point out that as “cultures differ from one another, the communication practices and behaviors of people will inevitably vary as a result of their different perceptions of the world.” Cross-cultural communication is defined as the study of communication between people whose “cultural perceptions and symbol systems are distinct enough” to alter their communication. Some research shows that messages conveyed by a person in one culture can be interpreted differently by a person in another culture. For example, studies have frequently found that non-native speakers rely heavily on their first language skills in interpreting English sentences (Koda, 1993). Unfortunately, as Koda points out, the strategies used in one language do not always cross to other languages, because “linguistic features that are essential to sentence comprehension and production vary from one language to another, and, importantly, the specific skills and strategies involved in language processing are developed to capitalize on the essential linguistic information” (Koda, 1994, p. 3). When it is impossible to overcome the incongruities between the two languages, non-native speakers must adjust their strategies, which slows comprehension processes even more.

In addition to varying linguistic experiences, people from different cultures also differ in their knowledge about the world, such as social norms, values, and beliefs. These affect one’s experiences in a society, which in turn impact how words and sentences are interpreted in context. When the context is similar to one of their past experiences, they can quickly retrieve that experience and use the knowledge of that context to understand the current text. When the context is unfamiliar to them, they need to build their understanding purely from imagination or logical inference, which can lead to poor retention or even misunderstanding of the text. For example, Steffensen et al. (Steffensen, Joag-Dev, & Anderson, 1979) found that in a reading task containing a passage of an American wedding and a passage of an Indian wedding, the American participants read the American wedding passage faster and recalled the contents better, while Indian participants read the Indian wedding passage faster and recalled the contents better.

The study of social norms, values and beliefs in cross-cultural communication

Hofsted’s cultural dimensions

Hofstede’s (Hofstede, 1980) cultural dimensions – Individualism-Collectivism, Uncertainty Avoidance, Power Distance, and Masculinity-Femininity – suggested some potential sources of conflict and problems in cross-cultural communication. For example, in individualistic cultures, the needs, values, and goals of the individual precede that of the group. In collectivistic cultures, the needs, values, and goals of the group precede that of the individual (Triandis, 1995). When there are conflicts between individual goals and group goals, collectivists will emphasize the attainment of group outcomes and subordinate their personal interests to ensure that group outcomes are attained. They also will perceive themselves as an indispensable part in a group’s success. However, individualists will seek to maximize their self-gain, because they derive satisfaction from performance based on their own achievements (Earley, 1989). Many crosscultural studies have used individualism-collectivism to explain group work performance. For example, Wagner found that individualists (American) are less apt to engage in cooperative group projects than collectivists (Chinese) (Wagner, 1995).

With respect to cross-cultural communication, the cultural factors directly affect communication styles and behaviors. For example, Gudykunst and Ting-Toomey (Gudykunst, Ting-Toomey, & Chua, 1988) argued that Individualism-Collectivism affected the use of low- and high-context communication, which I will review in detail in the next paragraph. In another large-scale national survey study (P. B. Smith, 2004), Smith found that people from high uncertainty avoidance felt more anxious about interacting with others and had less positive expectations concerning others. Brockner et al. (Brockner et al., 2001) studied the influence of power distance on the magnitude of voice effects. They found that the tendency for people to respond less favorably to lower levels of voice was greater in low power distance cultures (United States) than in high power distance cultures (China).

Hall’s emphasis on context

Hall (Hall, 1976) classified communication patterns in different cultures into low-context and high-context communication. Speakers in low-context cultures tend to express themselves verbally and explicitly. Speakers in high-context cultures are more sensitive to context information such as non-verbal cues and interpersonal relationships. Research has shown that low-context and high-context communication is predominant in individualistic and collectivistic cultures, respectively. People from individualistic cultures are more inclined to talk (Gaetz, Klopf, & Ishii, 1990) and more concerned with clarity in conversations than people from collectivistic cultures (M. S. Kim, 1994).

There is also research showing that people from a high-context culture are more likely to participate in group chat if they are allowed to see each other while chatting (Elizabeth, Judith, Gary, & Xiaolan, 1999; Leslie, Susan, & Christine, 2004). However, the effect of communication patterns is still unclear. There are controversial results showing that people from high context culture are less talkative when they are able to see each other (Wang, Fussell, & Setlock, 2009). Admittedly, there are other confounding factors in this result. For example, in this study the participants from the high-context culture in the study were not from English-speaking countries. Being anonymous and avoiding eye contact while chatting effectively may alleviate their tension when communicating in a foreign language.

Negotiation in Chinese language practices

The negotiation style of Chinese communication partners may be influenced by Chinese cultural values and language usage. Many studies that try to find particular cultural factors related to Chinese negotiation behavior suggest four concepts: power distance, holism, face and guanxi (Graham & Lam, 2003; Kirkbride, Tang, & Westwood, 1991). These four concepts are considered important to Chinese people and therefore have been used widely to explain Chinese negotiation style. Guanxi is the personal connections between people. In a collectivist society, building and maintaining guanxi is very important. In business activities, a good guanxi is a sign of success of two parties’ collaboration. Face is related to guanxi. Chinese’s adoption of mutualface maintenance is a way to maintain guanxi. Power distance is a mediator of guanxi. In situations where older people or those of higher social status are involved, the other party will take greater attempts to maintain guanxi and protect face. Based on the above discussion, one would expect less open argumentation and debate with Chinese in negotiations. Also, Chinese are reported to be happy to remain at the level of general principles to avoid confrontation related to specific issues, and they tend not to engage in extreme verbal posturing and aggressive positiontaking. (Kirkbride et al., 1991) This negotiation style is also considered to be related to Chinese holistic thought processes, which suggests that people will think issues as a whole and are sensitive to context. (Nisbett, Peng, Choi, & Norenzayan, 2001; Redding, 1980)

To the contrary, American negotiation style is known as competitive and confrontational style. Stella Ting-Tommey (Ting-Toomey, 1988) finds that Americans emphasize their self-face maintenance more than their mutual-face maintenance. Therefore, they tend to be aggressive and confrontational when negotiating specific issues. Americans’ negotiation style is analytic, which is the opposite of Chinese negotiation style. Americans usually decompose a negotiation task into a series of logical points or issues and settle the issues one at a time (Graham & Herberger, 1983).

The study of language in cross-cultural communication

People may think language differences will not affect English-based communication as long as all speakers have some working knowledge of English. However, in a survey

(Govindarajan & Gupta, 2001) of 58 executives from multinational organizations, the barrier of language differences is rated as the second biggest challenge for effectiveness of global business teams. A qualitative study (Neeley et al., 2009) has shown that uneven English language proficiency in global organizations has degraded meaning making, group collaboration, and individual relationships.

These language problems stem from the fact that cross-cultural communication typically takes place in one language (i.e. English in the context here), thereby requiring that some participants communicate in a non-native language. This generates communication problems in many ways. Research in second language learning has shown that low language proficiency may lead people to feel overwhelmed and even lost in fast-paced conversations (Mynard, 2002). However, most cross-cultural communication studies neglect this language difference. For example, Chinese participants in cross-cultural studies are typically described as fluent or nearly fluent in English, for example, “The Chinese participants were all fluent or nearly fluent in English. Although they were all currently studying at a U.S. university, the majority had been in the U.S. less than 2 years” (Wang et al., 2009) and “Sixteen were nationals of the United States and spoke English as their first language, and 14 were visiting students from the People’s Republic of China and had been in the United States for fewer than 2 years and were fluent or nearly fluent in English” (Stewart, Setlock, & Fussell, 2007). The special selection of fluent nonnative speakers may mitigate or disguise the impacts of language proficiency differences, but no one can deny language proficiency as a pervasive factor in non-native speakers’ communication problems. The next section will explain in detail why language proficiency matters.

Language Proficiency Matters

“Do you know any jokes?” “…Timing isn’t my strong suit.” King George VI embarrassingly stammered out his problem in the movie “The King’s Speech”. Yes, being able to control the rhythm is important when telling a joke. Uttering the same words at different speeds and with different pauses can create a very different listening experience for an audience. In some extreme cases, the utterance could be totally incomprehensible even if the audience hears every individual word clearly. This can be seen in the first public speech by King George VI in the film, where his severe stammer made his message excruciatingly difficult to understand.

Clearly speaking proficiency matters to the sense-making process. But problems in listening proficiency can also incur misunderstandings of a discourse, as can variations in reading and writing proficiency. Many people have such experiences when watching foreign movies; they often feel the dialogue is going by too fast, even if they know the words in that language. While they are trying to work out what the actors just said, or read what the subtitles just displayed, the dialogue is already moving to something else, and they begin to lose their place in the conversation. In general, language proficiency in all four modalities is a common source of problems for non-native speakers, problems that can haunt them in every aspect of their social lives. Therefore understanding language proficiency differences between native and non-native speakers as well as the psycholinguistic processes underlying language proficiency are important to the HCI scholars and designers focusing on non-native speakers as a user group. In the rest of this section, I will overview some common concepts of language fluency across modalities in order to prepare a theoretical common ground for later discussions.

Proficiency or fluency (we will use these two words interchangeably throughout this work) is an often-used term in linguistic research to describe someone’s high-level operational command of a language. It has a number of meanings that vary according to the different facets of a language. First of all, proficiency can vary by modalities: it can be used as a descriptor for speaking proficiency, listening proficiency, reading proficiency, or writing proficiency. The meaning of fluency as a summary term can also vary by speaker population, for example the operationalization of “fluent native speaker” can be very different from that for “fluent non-native speaker”. For native speakers, fluency implies a fluid speed in language use, and more than that a manifestation of the proficient command of the language, such as demonstrations of control over coherence, complexity and aesthetic functions of the language (Schmidt, 1992). Whereas for nonnative speakers, the meaning of fluency does not have the same richness as it does for native speakers and there is much less consistency in how it is used.

Fluency is often used as a rough synonym for global ability (Lennon, 1990). For example when people say someone speaks four languages fluently, they only mean that the person has the ability to communicate well using these different languages, without differentiating the many ways “well” can be operationalized. Given people’s various understandings of perceived fluency, there has never been a unified definition of non-native speakers’ language proficiency. Different researchers take on different definitions according to the need of their studies. For example, Lennon defines speaking proficiency in terms of its temporal aspect, “speech at the tempo of native speakers, unimpeded by silent pauses and hesitations, filled pauses… self-corrections, repetitions, false starts and the like” (Lennon, 1990, p. 390). Brumfit instead emphasizes the communicative outcomes of proficiency, where proficiency is defined as “the maximally effective operation of the language system so far acquired by the [ESL] student” (Brumfit, 1984,

1. 543). In his book, he explains fluency in contrast to accuracy. Under his definition, we might say that “He writes fluently, but with many grammatical errors.” or “He speaks fluently, but on occasion chooses odd vocabulary or expressions.” A shared emphasis of Lennon and Brumfit’s definition of proficiency is the speed and flow of language use. It defines proficiency from an information processing perspective that reflects the ongoing status of the brain’s language processing. I will use this definition of proficiency throughout my proposal, except otherwise noted. In fact, native and non-native speakers’ difference in language proficiency reflects a difference in their language processing and constructing process. This contrasting relationship is also true for adult and adolescent language users, or for that matter, any people with high versus low language skills.

We can think of a language user with high proficiency and a language user with low proficiency as a chess master and an amateur. A chess master is faster than an amateur in terms of planning a move or remembering a mid-game composition. The processes underlying these actions are “controlled” in the amateur’s mind in a way that he is aware of the rules of moving, the consequence of a move, and a particular position of a chess, etc. For the chess master, these processes have been automated so that he can free his mind for some higher-order processes, such as planning for several steps and reconstructing a mid-game composition. Psychologists have been studying this phenomenon for decades, and have described its relationship to working memory capacity (Baddeley, 1986). Working memory is described as “immediate memory processes involved in the simultaneous storage and processing of information in real-time” (Harrington & Sawyer, 1992, p. 25). Chess masters are found to have more efficient processing skills (the automated processing ability), thus allowing more capacity to be devoted to the storage of other tasks (Degroot, 1965). For example in a think aloud reading experiment (Horiba, 1990), researchers found that the non-native speakers paid heavy attention to language mechanics but mostly failed to infer meanings from their knowledge. This suggested that their minds were busy recognizing words and syntax, while higher-level comprehension was left untouched or addressed only to a limited degree. In contrast the native speakers mostly ignored language mechanics and devoted considerable attention to inferring the meaning of a text. This suggested that native speakers were able to automatically extract linguistic information from a text, and thus were more efficient in elaborating and comprehending the semantic content it conveyed.

The notion that a high working memory capacity leads to greater memory storage space in turn rests on the scientific consensus that human memory is limited. This is well documented in Miller’s (1969) book chapter, “The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information.” Seven is approximately the number of items (whether numbers, words representing complex concepts, or any partial products of a process) that you can hold in your conscious mind at one time before the information begins to dissipate. Chess masters are able to connect mid-game chess positions in a current game with similar matches they have played in the past; therefore they only need to remember that past game as one item along with the relevant chess positions from the past game. The amateurs, not having a lot of game experience, can only remember approximately seven chess positions.

Language users with high proficiency have an ability similar to chess masters, but with respect to language use. They use their proficient experience to process basic information quickly, so that their processing capacity is left free for thinking about what a text means, how a paragraph can be organized, or what implications are suggested by content they have just read or heard. To the contrary, language users with low proficiency do not have similar processing of the language, leaving their minds more occupied by words instead of phrases or concepts. Because of having to focus on spelling, or meaning, or sounds of these words, their cognitive processes give lower priority to comprehending and organizing words and phrases at a sentence level. Often, the processes is too slow to be maintained within the limit of seven items; this means that content will disappear from their mind before they have a chance to make sense of a sentence, draw connections between sentences, or synthesize the meaning of a group of sentences. This is why language proficiency is important.

During group text chats, non-native speakers receive and understand information more slowly than native speakers; as a result, they need more time to organize expressions and they frequently make grammatical mistakes. The imbalanced language proficiency between non-native and native speakers could affect the dynamic of group conversation. For example, it can lead to asymmetrical relationships. When a native speaker perceives that a non-native speaker has language deficiencies that interfere with communication, the native speaker may assume control of the conversation (Gaies, 1982). Native speakers also demonstrate more cognitive and metacognitive skills for comprehending conversations, due to their higher language proficiency

(Horiba, 1990).

Because of the prevalence of communication problems in cross-cultural groups, professional communication research has a body of work to develop strategies for these problems. For example, some research has suggested that we should increase native-speaking employees’ cultural awareness and train them to adapt their communication strategies. Communication adaptation strategies include speaking slowly, avoiding jargon, writing clear and structured text, allowing wait time, and so on (for a full list, refer to (Anawati & Craig, 2006)).

Following the cultural and language studies, there is a growing body of research investigating how cultural and language factors influence people’s use of computer-mediated communication (CMC) tools, which in turn influence their computer-mediated cross-cultural communication experiences and outcomes. Chapter 2 will review relevant literature considering

the roles and impacts of CMC in cross-cultural communication.

 

DESIGNING FOR CROSS-CULTURAL GROUP CHAT

Sharing is caring!