1.Introduction

In most communicative situations involving interlocutors, discourse not only expresses knowledge but also conveys a stance on the topic at hand. In other words, we speak to convey propositional content and, most importantly, to channel a range of subjective meanings related to our value systems, identity, and even confidence (Hunston and Thompson 2000Hunston, Susan, and Geoffrey Thompson 2000 “Evaluation: An Introduction.” In Evaluation in Text: Authorial Stance and the Construction of Discourse, ed. by S. Hunston and G. Thompson, 1–26. Oxford: Oxford University Press.; Bucholtz and Hall 2005Bucholtz, Mary, and Kira Hall 2005 “Identity and Interaction: A Sociocultural Linguistic Approach.” Discourse Studies 7(4–5): 585–614. ). Such understanding of interaction has sparkled research investigation into stance expression, particularly epistemic-stance, in which “a speaker signals their relationship towards the talk they are producing” (Kiesling 2011Kiesling, Scott 2011 “Stance in context: Affect, Alignment and Investment in the Analysis of Stancetaking.” Paper presented at the iMean Conference, The University of the West of England, 15 April., 4). Previous research on epistemic-stance expression has largely focused on written language. For instance, Aijmer (2002)Aijmer, Karin 2002 English Discourse Particles. Evidence From a Corpus. Amsterdam/Philadelphia: John Benjamins. compares the range and frequency of certain key modal words in the English writing of advanced-level university students. McEnery and Kifle (2002)McEnery, Tony, and Nazareth Amselom Kifle 2002 “Epistemic Modality in Argumentative Essays of Second-Language Writers.” In Academic Discourse, ed. by John Flowerdew, 182–195. London: Longman. examine the ways in which learners of English qualify their claims in argumentative essay writing. Despite the contributions made by these studies to understanding of expressions of stance, it is also necessary to investigate the negotiation of position between and among language users in spoken or even computer-mediated discourses, particularly as these modes of interaction allow for limited editing and may demand the immediate processing of utterances (Bygate 2009Bygate, Martin 2009 “Teaching and Testing Speaking.” In The Handbook of Language Teaching, ed. by Michael H. Long and Catherine J. Doughty, 412–440. UK: Wiley-Blackwell . ).

Among various stance-markers, I think is one that has attracted great attention from applied linguists, for its prevalence across genres in spoken discourse (e.g. Baumgarten and House 2010Baumgarten, Nicole, and Juliane House 2010 “ I think and I don’t know in English as Lingua Franca and Native English Discourse.” Journal of Pragmatics 42(5): 1184–1200. ; Kaltenböck 2009Kaltenböck, Gunther 2009 “English Comment Clauses: Position, Prosody, and Scope.” Arbeiten aus Anglistik und Amerikanistik 34 (1): 49–75.). As Scheibman (2001)Scheibman, Joanne 2001 “Local Patterns of Subjectivity in Person and Verb Type in American English Conversation.” In Frequency and the Emergence of Linguistic Structure, ed. by Joan Bybee and Paul Hopper, 61–89. Amsterdam: Benjamins. suggests, I think allows speakers to personalize their utterances through passing judgment and expressing attitude by “automatically introduc[ing] an explicit argumentative perspective to the discourse” (Baumgarten and House 2010Baumgarten, Nicole, and Juliane House 2010 “ I think and I don’t know in English as Lingua Franca and Native English Discourse.” Journal of Pragmatics 42(5): 1184–1200. , 1185). Some researchers have revealed that nonnative speakers use this grammatical construction differently from native speakers. For instance, Zhang and Zabet (2014)Zhang, Grace Qiao, and Peyman G. P. Sabet 2014 “Elastic ‘I Think’: Stretching Over L1 and L2.” Applied Linguistics 35(3): 1–21. observed that I think as a turn-taking device is used more often by nonnative speakers than by native speakers. Baumgarten and House (2010)Baumgarten, Nicole, and Juliane House 2010 “ I think and I don’t know in English as Lingua Franca and Native English Discourse.” Journal of Pragmatics 42(5): 1184–1200. also found that nonnative speakers use I think mainly to express either certainty or a lack of knowledge, whereas native speakers use this and other epistemic expressions to organize discourse and construct a verbal routine. Additionally, an investigation of English interviews with L1 Turkish speakers revealed that nonnative English speakers prefer to use personal stance expressions such as I think to convey messages while employing these expressions as discourse organizers less often (Sahin and Kilimci 2014Şahin Kýzýl, Aysel, and Kilimci, Abdurrahman 2014 “Recurrent Phrases in Turkish EFL Learners’ Spoken Interlanguage: A Corpus-Driven Structural and Functional Analysis.” Journal of Language and Linguistic Studies. 10(1): 195–210.). Another study highlighted the preference for evaluative I think by lower-level learners, which was attributed to linguistic constraints in nonnative speakers (Salsbury and Bardovi-Harlig 2000Salsbury, Tom, and Bardovi-Harlig, Kathleen 2000 “Oppositional Talk and the Acquisition of Modality in L2 English.” In Social and Cognitive Factors in Second Language Acquisition, ed. by B. Swierzbin, F. Morris, M. E. Anderson, C. A. Klee, and E. Tarone, 57–76. Somerville, MA: Cascadilla Press.).

The studies mentioned above only focus on differences between native and nonnative I think. Limited systematic analyses appear to have been conducted of the use of I think in communication between nonnative speakers. Baumgarten and House (2010)Baumgarten, Nicole, and Juliane House 2010 “ I think and I don’t know in English as Lingua Franca and Native English Discourse.” Journal of Pragmatics 42(5): 1184–1200. , among others, highlight the importance of such investigation, remarking that “[i]n ELF communication, speakers of different L1s with potentially differing conventionalized patterns of stance-marking and stance-taking and differently diversified L2 varieties of English interact” (p. 1185).

Taken together, this study explores the pragmatic functions of I think as used by Taiwanese and Japanese undergraduate students in computer-mediated cross-cultural communication. The aim of the investigation is to identify the distribution of I think functions and the possible factors affecting their frequency of use. Specifically, the research questions this paper intends to address are:

1. How frequently do Taiwanese and Japanese university students use I think in a multilingual and multimodal corpus of synchronous, computer-mediated cross-cultural communication?

2. Which functions of I think are used by Taiwanese and Japanese university students in the corpus?

3. Do the frequencies and functions of I think resemble or differ between the two groups of students?

4. What are the possible contributing factors to the observed similarities and differences?

2.Pragmatic functions of I think

This study substantiates the claims made by Kaltenböck (2010)Kaltenböck, Gunther 2010 “Pragmatic Functions of Parenthetical I Think.” In New Approaches to Hedging, ed. by Gunther Kaltenböck, Wiltrud Mihatsch, and Stefan Schneider, 237–266. Bingley UK: Emerald. regarding the basic role of I think and the development of its other functions. In his view, I think has an inherent core meaning, which is activated in various ways depending on the ‘concrete contextual realization’ (Kaltenböck et al. 2010Kaltenböck, Gunther, Wiltrud Mihatsch, and Stefan Schneider 2010 “Introduction.” In New Approaches to Hedging, ed. by Gunther Kaltenböck, Wiltrud Mihatsch and Stefan Schneider, 1–13. Bingley UK: Emerald. , 11). The stance marker has a basic function (evaluative) that can be intensified (emphatic) or de-intensified (tentative and mitigating), and can even manifest an evolved function (discursive). In other words, Kaltenböck (2010)Kaltenböck, Gunther 2010 “Pragmatic Functions of Parenthetical I Think.” In New Approaches to Hedging, ed. by Gunther Kaltenböck, Wiltrud Mihatsch, and Stefan Schneider, 237–266. Bingley UK: Emerald. contends that the basic evaluative function of I think conveying the central meaning of ‘in my opinion’ serves as a foundation for the particular marker in its acquisition of other pragmatic functions. Supporting Kaltenböck, Zhang (2014)Zhang, Grace Qiao 2014 “The Elasticity of I Think: Stretching Its Pragmatic Functions.” Intercultural Pragmatics 11(2): 225–257. posits that in a broader scope, I think manifests elasticity in its four functions, tentative, mitigating, emphatic, and discursive, which respectively depict fluidity, overlap, correlation, and the co-existence of roles. The tentative I think is used by speakers to express uncertainty and approximation (Jucker 1986Jucker, Andreas 1986 News Interviews. Amsterdam: John Benjamins. ). In her study of political interviews, Zhang (2014)Zhang, Grace Qiao 2014 “The Elasticity of I Think: Stretching Its Pragmatic Functions.” Intercultural Pragmatics 11(2): 225–257. observes that I think as an epistemic modal marker conveys vagueness and an avoidance of commitment to save social face (Aijmer 1997Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. ). This tentative function is particularly evident when I think is positioned at the end of a clause to express an epistemic afterthought (Conrad and Biber 2000Conrad, Susan, and Douglas Biber 2000 “Adverbial Marking of Stance in Speech and Writing.” In Evaluation in Text: Authorial Stance and the Construction of Discourse, ed. by Susan Hunston and Geoff Thompson, 56–73. Oxford: Oxford University Press.; Kaltenböck 2013Kaltenböck, Gunther 2013 “Development of Comment Clauses.” In The Verb Phrase in English: Investigating Recent Change with Corpora, ed. by Bas Aarts, Joanne Close, Geoffrey Leech, and Sean Wallis, 286–317. Cambridge: Cambridge University Press. ; Zhang 2014Zhang, Grace Qiao 2014 “The Elasticity of I Think: Stretching Its Pragmatic Functions.” Intercultural Pragmatics 11(2): 225–257. ). As I think in the final position conveys doubt and suggests a lack of authority, it is, for example, not found in this position in political discourse, as this would defeat the communicative goals of a political speaker (Simon-Vanderbergen 2000Simon-Vandenbergen, Anne-Marie 2000 “The Functions of I Think in Political Discourse.” International Journal of Applied Linguistics 10(1): 41–63. ).

In contrast with the expression of uncertainty in the tentative I think, the mitigating I think is used to soften or tone down the assertiveness of a message. Termed a subjectivizer (Blum-Kulka, House, and Kasper 1989Blum-Kulka, Shoshana, Juliane House, and Gabriele Kasper 1989 Cross-Cultural Pragmatics: Requests and Apologies. Norwood, N.J.: Albex.), this phrasal downgrader not only softens directives but also supports a speaker’s face-saving attempts (Ruzaitė 2007Ruzaitė, Jūratė 2007 “Vague References to Quantities as a Face-Saving Strategy in Teacher-Student Interaction.” Lodz Papers in Pragmatics 3: 157–178. ) and mitigates face threat (Aijmer 1997Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. ). Occurring almost as frequently as tag questions, the mitigating I think serves a hedge-like function particularly in emphatic statements (Preisler 1986Preisler, Bent 1986 Linguistic Sex Roles in Conversation: Social Variation in the Expression of Tentativeness in English. Berlin: Mouton de Gruyter. ). It softens a blunt assertion or functions as a negative politeness marker expressing primarily affective meaning (Holmes 1990Holmes, Janet 1990 “Hedges and Boosters in Women’s and Men’s Speech.” Language and Communication 10(3): 185–205. ; 1995Holmes, Janet 1995 Women, Men and Politeness. London: Longman.). Additionally, the mitigating I think qualifies commitment, indicating the speaker’s uncertainty about the truth of a proposition (Kärkkäinen 2010Kärkkäinen, Elise 2010 “Position and Scope of Epistemic Phrases in Planned and Unplanned American English.” In New Approaches to Hedging, ed. by Gunther Kaltenböck, Wiltrud Mihatsch and Stefan Schneider, 207–241. Bingley UK: Emerald. ).

The third function, the emphatic I think, also known as the deliberative I think, conveys the speaker’s certainty in and reassurance of the validity of the interpretation of the facts stated (Holmes 1984Holmes, Janet 1984 “Modifying Illocutionary Force.” Journal of Pragmatics 8(3): 345–365. ; Holmes 1990Holmes, Janet 1990 “Hedges and Boosters in Women’s and Men’s Speech.” Language and Communication 10(3): 185–205. ). Using I think adds weight, emphasis, and confidence to a statement. Additionally, the emphatic I think is considered deliberative, as it appears first in the utterance and is prosodically prominent (Aijmer 1997Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. ). In its fourth, discursive function, I think is a structural device (Aijmer 1997Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. ; Kärkkäinen 2003Kärkkäinen, Elise 2003 Epistemic Stance in English Conversation. A Description of Its Interactional Functions, with a Focus on I Think. Amsterdam: John Benjamins. ; Kaltenböck 2010Kaltenböck, Gunther 2010 “Pragmatic Functions of Parenthetical I Think.” In New Approaches to Hedging, ed. by Gunther Kaltenböck, Wiltrud Mihatsch, and Stefan Schneider, 237–266. Bingley UK: Emerald. ). For instance, it can function as a stalling or filling device in the event of a gap in communication (Stenström 1995Stenström, Anna-Brita 1995 “Some Remarks on Comment Clauses.” In The Verb in Contemporary English, ed. by Bas Aarts and Charles F. Meyer, 290–301. Cambridge: Cambridge University Press.). I think in its discursive role thus plays an important role in online processing, hesitant speech, word searches, and repair (Mullan 2010Mullan, Kerry 2010 Expressing Opinions in French and Australian English Discourse: A Semantic and Interactional Analysis. Amsterdam: John Benjamins. ). In its fifth, evaluative function, I think is used in utterances to convey speakers’ evaluation of the topic discussed. It thereby serves as “a marker of the speaker’s degree of belief, opinion or subjective evaluation of the proposition” (Baumgarten and House 2010Baumgarten, Nicole, and Juliane House 2010 “ I think and I don’t know in English as Lingua Franca and Native English Discourse.” Journal of Pragmatics 42(5): 1184–1200. ; 1197). As suggested by Aijmer (1997)Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. and supported by Ifantidou (2001)Ifantidou, Elly 2001 Evidentials and Relevance. Amsterdam: John Benjamins. , I think in this capacity conveys the central meaning of “I’m expressing my opinion.” The evaluative I think neither softens nor boosts a statement, but instead expresses a propositional attitude, revealing the speaker’s assessment and judgment of the truth of the statement.

3.Methodology

To investigate the functions and frequency of the use of I think, data were extracted from the Cross-Cultural Distance Learning (CCDL) corpus. Focusing on computer-mediated international communication, the CCDL corpus is a collaborative project undertaken by Waseda University (WU) in Tokyo, Japan and Yuan Ze University (YZU) in Taoyuan, Taiwan. To enrich cultural exchange and language learning, a two-way interactive online chat-room system (via the Live On software interface) was established to allow students to converse synchronously face to face. The interactions were primarily spoken, yet this interface also enabled the students to use multiple modalities such as audio, video, and text and to share images, website links, and other multimedia files. The multilingual CCDL corpus contains 112,498 words, comprising 68,954 English, 37,171 Japanese, and 6,373 Chinese words.

The data drawn from the CCDL corpus were provided by 46 students (see Table 1), comprising 23 Taiwanese students from YZU and 23 Japanese students from WU. All were nonnative English speakers with proficiency levels ranging from B1 to B2 according to the Common European Framework of Reference for Languages. In terms of exposure to other languages, the Taiwanese participants had majored in both English and Japanese in their department, while the Japanese participants understood some Chinese characters because they used Kanji to write Japanese. As the CCDL corpus is multilingual, the participants used English, Japanese, and Chinese to facilitate their interactions.

The data were collected in two periods: fall 2013 and spring 2014. First, the students volunteered to participate in the CCDL project, which was then followed by an orientation session explaining the processes of logging in to the online chat-room system and using the recording software. After the meeting, a group assignment was held in which the 46 participants were assigned to 11 groups according to availability. Each of the groups was provided with a list of weekly topics determined by both their Japanese and their Taiwanese professors (see Appendix A). The groups engaged in weekly 50-minute sessions for five weeks to maintain ongoing cross-cultural communication in cyberspace. The online interactions were mainly spoken through the two-way audio and video mode, and students resorted to text and multimedia to further aid their communication. The groups from fall 2013 were labeled ‘Lavender’ and those from spring 2014 were marked ‘Yellow’. Their recorded interactions, which were transcribed verbatim in the abovementioned periods, were collected and analyzed in this study. The flow of communication afforded by the Live On platform is structured in a way that facilitates multimodal interactions. As seen in Figure 1, the spoken interactions with the presence of the interlocutors’ faces are akin to the typical face-to-face interactions, with features such as turn-taking, overlapping, discourse markers and back-channels. Topic development may be initiated around objects introduced visually such as the glass of bubble tea in Figure 1. Served primarily as a supplementary role, typed texts, as those in other text messaging platforms, are arranged by interlocutors, one after another. All these modes of interactions work with each other to facilitate meaningful interactions among parties involved.

Microsoft Excel Find function was run on the transcribed data to search for I think occurrences. Although the interactions were mainly spoken, an indicator was provided to note utterances in text format. To determine clusters and discourse context, occurrences were manually identified, analyzed, and coded from the transcripts. To ensure credibility, another rater checked 20% of the data with 90 % agreement (item to item). Additionally, two other researchers were also consulted in relation to data coding.

The data analysis was mainly qualitative, supplemented by quantitative analysis to strengthen the robustness of the research. The qualitative part of the analysis was conducted at discourse level, due to the significant role of correlated utterances in implicitly or explicitly prompting a stance (Kärkkäinen 2003Kärkkäinen, Elise 2003 Epistemic Stance in English Conversation. A Description of Its Interactional Functions, with a Focus on I Think. Amsterdam: John Benjamins. ; 2006Kärkkäinen, Elise 2006 “Stance Taking in Conversation: From Subjectivity to Intersubjectivity.” Text and Talk – An Interdisciplinary Journal of Language, Discourse Communication Studies 26(6): 699–731. ). It is essential to methodically associate the structure and function of the correlated utterances; in Kärkkäinen’s (2012Kärkkäinen, Elise 2012 “ I Thought It Was Very Interesting: Conversational Formats for Taking a Stance.” Journal of Pragmatics 44(15): 2194–2210. , 2194) words, “viewing the utterance as a whole as performing some discourse function or constituting some action”. The descriptive data reveal the speakers’ maneuvers and turn taking in their use of I think during their interactions.

Criteria for analyzing I think

Insights into the functions of I think as an expression of epistemic stance, such as expressing opinions, maintaining relationships between interlocutors, and organizing discourse, have motivated researchers to propose different sets of criteria to help determine the functions of I think. For instance, Holmes (1990)Holmes, Janet 1990 “Hedges and Boosters in Women’s and Men’s Speech.” Language and Communication 10(3): 185–205. points out that context, which specifies the relationships between the participants, the topic, and the level of formality of the interaction, is key to determining I think functions. Aijmer (1997)Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. also observes that prosodic, grammatical, and positional criteria must be considered when identifying the functions of I think, particularly to determine whether it takes a deliberative or a tentative function. Several other researchers integrate the aforementioned criteria to argue that semantic criteria, syntactic position, prosodic features, and the linguistic and situational context of occurrence must be analyzed to determine the functions of I think in spoken discourse (Aijmer 1997Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. ; Simon-Vanderbergen 2000Simon-Vandenbergen, Anne-Marie 2000 “The Functions of I Think in Political Discourse.” International Journal of Applied Linguistics 10(1): 41–63. ; Kaltenböck 2009Kaltenböck, Gunther 2009 “English Comment Clauses: Position, Prosody, and Scope.” Arbeiten aus Anglistik und Amerikanistik 34 (1): 49–75.).

The current study adopted the specifications made by Aijmer (1997)Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. as well as Kadar and Haugh (2013)Kádár, Dániel Z., and Haugh, Michael 2013 Understanding Politeness. Cambridge: Cambridge University Press. , specifically with respect to collocation, the situational context of occurrence and sequentiality. Although prosodic features and syntactic positions are included in Aijmer’s criteria, this study focuses on the three aforementioned criteria since they are conveniently found in the data and take interactional view into account:

Collocation patterns

One important criterion used in analyzing I think functions in the data is clustering. I think clusters with conjunctions such as and, because, but, and so; such stance marker mainly performs evaluative or emphatic functions (Zhang 2014Zhang, Grace Qiao 2014 “The Elasticity of I Think: Stretching Its Pragmatic Functions.” Intercultural Pragmatics 11(2): 225–257. ). Maybe added before I think can be used to express tentativeness and mitigation, and clustering the epistemic phrase with fillers such as uh or um serves a discursive function (Wu et al. 2010Wu, Yong, Wang Jingli, and Cai Zhou 2010 “The Use of I Think by Chinese EFL Learners: A Study Revisited.” Chinese Journal of Applied Linguistics 33(1): 3–23.). Meanwhile, I think followed by that typically expresses evaluation and emphasis (Thompson and Mulac 1991Thompson, Sandra, and Anthony Mulac 1991 “A Quantitative Perspective on the Grammaticalization of Epistemic Parentheticals in English.” In Approaches to Grammaticalization, ed. by Elizabeth Closs Traugott and Bernd Heine, 313–339. Amsterdam and Philadelphia: John Benjamins. ; Aijmer 1997Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. ). According to Aijmer (1997)Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. , repeated I think or I think followed by maybe can convey tentativeness or fulfill a discursive function. Besides identifying the first word before and after I think, the study considers other neighboring words in the utterance that may help to determine the pragmatic role of the marker. For example, I think clustered with nonadjacent neighboring expressions such as should, quite, and almost can emphasize a statement, convey uncertainty, or soften a proposition.

Discourse contexts

In addition to collocation, the broader context of occurrence is a useful criterion for identifying the function of I think. The study adapts the list of discourse contexts that may contain I think as presented by Baumgarten and House (2010)Baumgarten, Nicole, and Juliane House 2010 “ I think and I don’t know in English as Lingua Franca and Native English Discourse.” Journal of Pragmatics 42(5): 1184–1200. : answering a question, corroborating another’s utterance, introducing a new or related topic, expressing agreement, drawing a conclusion, expressing a contrasting view, displaying knowledge, elaborating on a previous utterance, sharing personal experience, rejecting another’s contribution, reinforcing one’s own preceding claim, simple explanation, and making a suggestion.

Sequentiality

As proposed by Kadar and Haugh (2013)Kádár, Dániel Z., and Haugh, Michael 2013 Understanding Politeness. Cambridge: Cambridge University Press. , sequentiality “refers to the way in which current turns or utterances are always understood relative to the prior and subsequent talk, particularly talk that is continuous” (p. 112). Similarly, Heritage (2012)Heritage, John 2012 “Epistemics in Action: Action Formation and Territories of Knowledge.” Research on Language and Social Interaction 45(1):1–29. argues that consideration of the relative epistemic statuses of the speaker and hearer are a fundamental element in the construction of social action. This criterion is crucial as we take into account the interactional perspective when analyzing the functions of I think, i.e. the immediate co-text or initiating and responding utterances.

4.Findings

4.1Overall frequency of I think

This investigation of the functions of I think as used by Japanese and Taiwanese university students in their online cross-cultural communication reveals 316 occurrences of the stance marker from a total of 112,498 words used in the corpus (see Table 2). The prevalence of this subject-predicate combination in the CCDL corpus was also found in Baumgarten and House’s (2010)Baumgarten, Nicole, and Juliane House 2010 “ I think and I don’t know in English as Lingua Franca and Native English Discourse.” Journal of Pragmatics 42(5): 1184–1200. study of COCA and BNC, notably the spoken component. Such significant number of occurrences also supports the previous claim that speakers generally speak more often about themselves than others in conversation, sharing their feelings about, attitudes toward, and views of the world (Scheibman, 2001Scheibman, Joanne 2001 “Local Patterns of Subjectivity in Person and Verb Type in American English Conversation.” In Frequency and the Emergence of Linguistic Structure, ed. by Joan Bybee and Paul Hopper, 61–89. Amsterdam: Benjamins. ).

Table 2.Frequency of I think functions in the corpus

I think function	Raw figure	Percentage
Evaluative	101	32.0%
Emphatic	94	29.7%
Tentative	65	20.6%
Discursive	32	10.1%
Mitigating	24	7.6%
Total	316	100%

The data also indicate that the aforementioned nonnative speakers used 32% of I think in its typical role: to mark their stance, conveying the sense of ‘in my opinion’. Such finding is somewhat contrary to the claim that speakers use I think mainly to express either certainty or lack of knowledge (Baumgarten and House 2010Baumgarten, Nicole, and Juliane House 2010 “ I think and I don’t know in English as Lingua Franca and Native English Discourse.” Journal of Pragmatics 42(5): 1184–1200. ). Another noticeable result was the high frequency of use of the emphatic I think (29.7%) and the tentative I think (20.6%) by these nonnative speakers of English, who preferred to use I think to emphasize their epistemic claims than to use the mitigating I think (7.6%) to tone down strong claims. In other words, these speakers preferred to add strength (Holmes 1984Holmes, Janet 1984 “Modifying Illocutionary Force.” Journal of Pragmatics 8(3): 345–365. ) and weight (Aijmer 1997Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. ) to their assertions by using the emphatic I think as a booster to express certainty, authority, and reassurance (Brown 1980Brown, Penelope 1980 “How and Why Are Women More Polite: Some Evidence from a Mayan Community.” In Women and Language in Literature and Society, ed. by Sally McConnell-Ginet, Ruth Borker, and Nelly Furman, 111–136. New York: Praeger.; Holmes 1990Holmes, Janet 1990 “Hedges and Boosters in Women’s and Men’s Speech.” Language and Communication 10(3): 185–205. ). These findings contrast with the previous claim that the use of I think to indicate certainty is infrequently used (Aijmer 1997Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. ; Kaltenböck 2010Kaltenböck, Gunther 2010 “Pragmatic Functions of Parenthetical I Think.” In New Approaches to Hedging, ed. by Gunther Kaltenböck, Wiltrud Mihatsch, and Stefan Schneider, 237–266. Bingley UK: Emerald. ). In addition to the emphatic I think, however, the tentative I think was regularly used by these speakers, conveying approximation and uncertainty. They might have used I think in this way to express a lack of knowledge (Simon-Vandenbergen 2000Simon-Vandenbergen, Anne-Marie 2000 “The Functions of I Think in Political Discourse.” International Journal of Applied Linguistics 10(1): 41–63. ) and a high level of subjectivity. It is also noteworthy that the discursive I think, commonly used by native speakers to carry out “routinized work in conversation organization” (Kärkkäinen 2003Kärkkäinen, Elise 2003 Epistemic Stance in English Conversation. A Description of Its Interactional Functions, with a Focus on I Think. Amsterdam: John Benjamins. , 145), was also used by the Japanese and Taiwanese students, albeit at a minimal frequency (10.1%). In sum, the results indicate that I think had retained its epistemic functions of evaluation and certainty in the discourse of the Japanese and Taiwanese students recorded, suggesting a transition in the function of I think from epistemic to interpersonal and then to discourse-organizational (Kaltenböck 2013Kaltenböck, Gunther 2013 “Development of Comment Clauses.” In The Verb Phrase in English: Investigating Recent Change with Corpora, ed. by Bas Aarts, Joanne Close, Geoffrey Leech, and Sean Wallis, 286–317. Cambridge: Cambridge University Press. ; Zhang 2014Zhang, Grace Qiao 2014 “The Elasticity of I Think: Stretching Its Pragmatic Functions.” Intercultural Pragmatics 11(2): 225–257. ).

4.2Comparison of distribution of I think

The distribution of I think functions in the Japanese students’ discourse is illustrated in Table 3. Of the five main functions of I think, the evaluative I think was used at the highest frequency (41.9%), followed by the emphatic I think (33.8%). In addition to using the marker to express stance in a typical way, conveying the sense of ‘in my opinion’, the Japanese students regularly used the emphatic I think, despite expressing minimal assessment and contributing less to the conversation than their Taiwanese counterparts.

Table 3.Distribution of I think functions in Japanese and Taiwanese students’ discourses

I think function	Japanese students			Taiwanese students
	Raw figure	Percentage		Raw figure	Percentage
Evaluative	31	41.9%		70	28.9%
Emphatic	25	33.8%		69	28.5%
Tentative	10	13.5%		55	22.7%
Discursive	5	6.8%		27	11.2%
Mitigating	3	4.1%		21	8.7%
Total	74	100%		242	100.0%

The Taiwanese students most frequently used the evaluative I think (28.9%) and the emphatic I think (28.5%) in their discourses. They also used the tentative I think (22.7%), which made up almost a quarter of their I think utterances, in contrast with the claim that the use of I think to indicate certainty is scarcely used (Aijmer 1997Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. ; Kaltenböck 2010Kaltenböck, Gunther 2010 “Pragmatic Functions of Parenthetical I Think.” In New Approaches to Hedging, ed. by Gunther Kaltenböck, Wiltrud Mihatsch, and Stefan Schneider, 237–266. Bingley UK: Emerald. ). A possible reason for their significant use of I think is the informal and conversational nature of the data. As the Taiwanese speakers sought to elicit responses and encourage their Japanese counterparts to participate more, they might have preferred to convey their stance in a less authoritative way to sound less threatening. The Taiwanese students also used I think to perform routine organizational tasks (the discursive I think) and to tone down assertiveness (the mitigating I think), albeit at lower frequencies.

Comparison of the use of I think between the Taiwanese and Japanese students revealed that both sets of nonnative speakers of English used the evaluative I think, followed by I think in its emphatic, tentative, discursive, and mitigating functions, respectively. The data reveal that both sets of speakers made multifunctional use of I think. The evaluative I think serves as a baseline, which when strengthened becomes the emphatic I think and when weakened becomes the tentative and mitigating I think (Zhang 2014Zhang, Grace Qiao 2014 “The Elasticity of I Think: Stretching Its Pragmatic Functions.” Intercultural Pragmatics 11(2): 225–257. ). With further evolution, the discourse marker develops a new function: to perform organizational tasks during interaction.

A difference was found in the total frequency of use, as the Taiwanese students more frequently expressed their epistemic stance than the Japanese students. This difference in frequency may be attributed to the different roles played by the Japanese and Taiwanese students in their computer-mediated communication. The Japanese students were more passive: merely responding or reacting to the utterances of their Taiwanese counterparts. Without the latter’s encouragement, the former mostly remained silent and unresponsive. Additionally, the Japanese students used silence to stall for time, to gain an adequate grasp of the subject matter, or to prepare for their next utterances (Nakane 2006Nakane, Ikuko 2006 “Silence and Politeness in Intercultural Communication in University Seminars.” Journal of Pragmatics 38(11): 1811–1835. ), which as previously mentioned were mostly responses to questions asked or comments made by the Taiwanese speakers. These occurrences reflected the typical dynamics of Japanese conversation, which is analogous to “the game of bowling” (Sakamoto and Naotsuka 1982Sakamoto, Nancy, and Reiko Naotsuka 1982 Polite Fictions: Why Japanese and Americans Seem Rude to Each Other. Tokyo: Kinseido.). Japanese interlocutors project reserved behavior, wait for their turn to speak, express their ideas carefully, and allow for appropriate pauses between turns, making them appear unresponsive and unspontaneous (Kato 2001Kato, Kumi 2001 “Overseas-Educated Teachers in Australian Classrooms.” Babel 36(2): 30–36.; Ogasawara 1995Ogasawara, Linju 1995 “Native Cultural Interference in Japanese English.” In Georgetown University Roundtable on Languages and Linguistics, ed. by J. Alatis, C. Straehle, B. Gallenberger and M. Ronkin, 105–111. Washington DC: Georgetown University Press.; Tanaka 1986Tanaka, Noriko 1986 “Politeness in Japanese and English.” Babel 21(2): 25–28.; Tsuda, Shigemitsu, and Murata 2007Tsuda, Sanae, Yuka Shigemitsu, and Kazuyo Murata 2007 “Cultural Awareness in Teaching English: Analysing Intercultural Communication and Teaching Positive Politeness Strategies.” The Journal of Asia TEFL 4(3): 161–185.). The Taiwanese students played a comparatively dominant role in the interaction. They mostly initiated topics, introduced new ideas, and asked questions of the other interlocutors. They seemed to carry the burden of maintaining the flow of the conversation and avoiding dead air, which reflects the Taiwanese students’ less tolerance of silence, as is evidenced by their interaction dynamics in the corpus. Such verbal behaviors displayed by the Taiwanese students are identical to that of the native English speakers in their cross-cultural communication with the Japanese speakers (Kato 2001Kato, Kumi 2001 “Overseas-Educated Teachers in Australian Classrooms.” Babel 36(2): 30–36.; Ogasawara 1995Ogasawara, Linju 1995 “Native Cultural Interference in Japanese English.” In Georgetown University Roundtable on Languages and Linguistics, ed. by J. Alatis, C. Straehle, B. Gallenberger and M. Ronkin, 105–111. Washington DC: Georgetown University Press.; Tanaka 1986Tanaka, Noriko 1986 “Politeness in Japanese and English.” Babel 21(2): 25–28.; Tsuda, Shigemitsu, and Murata 2007Tsuda, Sanae, Yuka Shigemitsu, and Kazuyo Murata 2007 “Cultural Awareness in Teaching English: Analysing Intercultural Communication and Teaching Positive Politeness Strategies.” The Journal of Asia TEFL 4(3): 161–185.).

In the following sections, selected examples will be used to illustrate each of the five pragmatic functions of I think.

4.2.1Evaluative ‘I think’

As a marker of epistemic stance, the evaluative I think is used to “present speaker comments on the status of information in a proposition” (Biber et al. 2010Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad, and Edward Finegan 2010 (8th edition). Longman Grammar of Spoken and Written English. London: Longman., 972). The Taiwanese university students (70 instances in total) used the evaluative I think almost twice as frequently than their Japanese counterparts (31 instances). The gap between these sets of two nonnative speakers of English in their use of the evaluative I think may suggest that Taiwanese students are more inclined to speak about their own thoughts, attitudes, and opinions than Japanese students. On the other hand, Japanese students’ reluctance to express their opinions may be attributed to their perceived potential of making “the interlocutors lose face” (Blight and Stephens 2006Blight, Richard, and Meredith Stephens 2006 “Discourse Strategies Involving Topic Nomination in English and Japanese Language Usage.” English Education Center Journal 4: 41–53., 43).

In Extract 1, a Taiwanese student elaborates on her previous utterance in relation to studying English in Taiwan. As the Japanese speakers respond with some surprise to her description of her early English training in her country (signified with the use of ‘!?’), she later provides further details on this preference for early language training. Her explanation results in a Japanese student’s apparent expression of surprise, “Wow,” which prompts the original speaker to express an evaluative I think utterance: “But I think that is a little bit pressure” a way of providing her opinion with regard to the negative consequence of this educational practice. All of the interlocutors then respond by laughing, and the same Japanese speaker replies, “Yes, it’s hard for children in kindergarten.” This remark prompts the Taiwanese speaker to agree with the former’s statement and provides additional reasons for agreeing, and then finally to end her utterance by stating her evaluation of the situation as in “I think that is difficult.”

Extract 1.(Lavender 7, Week 5) (TW = Taiwanese student; JP = Japanese student)

1   TW1    But, uh, now… now many children have study their English in kindergarten.
2   JP2    Kindergarten?
3   JP1    Kindergarten!? Wow!
4   TW1    Yeah! Because their parents want to they be more competitive so they think they
5          need, uh train Chinese and English at same time.
6   JP1    Wow!
7   TW1    But I think that is a little bit pressure.
8   All    Haha.
9   JP1    Yes, it’s hard for children in kindergarten.
10  TW1    Yeah! Because they don’t know letter but they have to say the English sentence. I
11         think that is difficult.

An example of the use of the evaluative I think by the Japanese speakers is provided in Extract 2. After introducing the Japanese term for the month of May, a Japanese speaker elaborates on her previous utterance by further explaining the word. Toward the middle of her discourse, she offers an evaluative statement: “I think it is a beautiful word in Japan.” I think in this utterance is evaluative as it carries the meaning of “in my opinion” and focuses on the epistemic assessment of the Japanese speaker. Additionally, the stance marker is clustered neither with a softener to convey tentativeness nor a booster to strengthen the claim or convey authority.

Extract 2.(Lavender 7, Week 5)

1   JP1    So in Japan, uh, May is called さつき
                                          sa tsu ki
                                          ‘May’
2   TW1    さつき?
           sa tsu ki
           ‘May’
3   JP1    This is traditional name of May. [Texting: さつき]
                                                      sa tsu ki
                                                      ‘May’
4   TW1    [Texting: oh]
5   JP1    I think it is a beautiful word in Japan. [Texting: さつき]
                                                              sa tsu ki
                                                              ‘May’
6   TW1    [Texting: さつき?]
           sa tsu ki
           ‘May’
7   JP1    Yes, this is a traditional name of May in Japan.
8          えー、さつきって漢字出せるかな？
           e ~ satsuki tte kanji daseru kana?
           ‘Well, can I get you to type the Chinese character for satsuki?’

4.2.2Emphatic ‘I think’

Although both the emphatic and evaluative I think are used to express evaluative stance, they differ in that the emphatic I think conveys the assertiveness and authority of the speaker while the evaluative I think centers on epistemic assessment (Zhang 2014Zhang, Grace Qiao 2014 “The Elasticity of I Think: Stretching Its Pragmatic Functions.” Intercultural Pragmatics 11(2): 225–257. ). In this study, the Taiwanese students (69 instances) used emphatic I think approximately three times more than that of their Japanese counterparts (25 instances). As the Japanese students only minimally expressed their stance and contributed less than the Taiwanese speakers to the conversation, it is unsurprising that they used the emphatic I think less frequently in their discourses. One of the possible explanations for the Japanese’ minimal use of emphatic construction is their tendency to “behave politely” (Shigemitsu 2012Shigemitsu, Yuka 2012 “Question-Answer Sequences in English Conversation and Japanese Conversation: Suggestion for English Teaching.” Academic Report 35(2): 1–15., 2)

An example of emphatic I think located in the corpus is shown in Extract 3. To describe the discourse context, the Japanese student said he read the book being discussed in the conversation. To further ask about his opinion, the Taiwanese student asked if he likes it and as a response, the Japanese student replied, “Yeah, and I also see the movie.” Having heard his response to her question, the Taiwanese student supported his claim by saying, “Yeah, and I think it is really a sad story.” In this extract, I think is clustered with the confirmation marker “yeah” and the booster “really”, which further strengthens and intensifies the Taiwanese student’s claim regarding the book being discussed.

Extract 3.(Lavender 8, Week 5)

1   JP1           I read this book.
2   TW1           Umm, so do you like this?
3   JP1           Yeah, and I also see the movie.
4   TW1           Yeah, and I think it is really a sad story.
5   JP1           Umm.
6   TW1           Yeah. OK, so that’s all about me.
7   JP1 and JP2   Thank you.

Extract 4 gives an illustration of how Japanese students use emphatic I think in their discourses. While the Taiwanese student was looking for the logo of a Taiwanese beer, the Japanese student responded, “Yes please.” The Taiwanese student requested for the Japanese interlocutors to wait as he was searching for it. The Japanese student responded, “Okay” followed by the utterance of another Japanese student: “I think Jump is very bad.” The Taiwanese student wanted the previous speaker to clarify what she said and so she responded, “You like woman and alcohol. Haha” Such I think utterance expresses the Japanese student’s opinion of the Taiwanese student in its clause-initial position clustered with a booster, “very”, which adds weight to the claim.

Extract 4.Lavender 12, Week 2

1   JP1    Yes please.
2   TW1    Wait me a moment.
3   JP1    Okay.
4   JP2    I think Jump is very bad.
5   TW1    What? What?
6   JP2    You like woman and alcohol. Haha.
7   TW1    It’s the logo.

4.2.3Tentative ‘I think’

In addition to its function as a booster “strengthening the speaker’s commitment,” I think can also perform as a hedging device, “generally weakening the speaker’s commitment” (Ifantidou 2001Ifantidou, Elly 2001 Evidentials and Relevance. Amsterdam: John Benjamins. , 155). Accordingly, the stance marker is used in utterances as an epistemic qualifier or quality hedge, expressing a minimal level of certainty and indicating a noncommittal stance on the truth of the utterance (Aijmer 1997Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. ), thereby performing a tentative function. The Taiwanese student participants (55 instances) used the tentative I think five times more than the Japanese students (10 instances). These data suggest that the Taiwanese speakers might have preferred to convey their stance with approximation and uncertainty to appear less authoritative and establish rapport with their Japanese counterparts.

In Extract 5, recording the participants’ discussion of the sizes of television screens in their respective countries, the Taiwanese student expresses uncertainty when elaborating on her previous utterance regarding the ‘42 size’ of a television. She starts with a description of a typical television size in a Taiwanese household, clustering it with ‘maybe’, which signals tentativeness and uncertainty (Aijmer 1997Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. ), and ‘like’, which suggests approximation. Additionally, I think in this utterance is in the clause-final position, as if an afterthought (Conrad and Biber 2000Conrad, Susan, and Douglas Biber 2000 “Adverbial Marking of Stance in Speech and Writing.” In Evaluation in Text: Authorial Stance and the Construction of Discourse, ed. by Susan Hunston and Geoff Thompson, 56–73. Oxford: Oxford University Press.; Kaltenböck 2013Kaltenböck, Gunther 2013 “Development of Comment Clauses.” In The Verb Phrase in English: Investigating Recent Change with Corpora, ed. by Bas Aarts, Joanne Close, Geoffrey Leech, and Sean Wallis, 286–317. Cambridge: Cambridge University Press. ), expressing tentativeness (Aijmer 1997Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. ; Simon-Vandenbergen 2000Simon-Vandenbergen, Anne-Marie 2000 “The Functions of I Think in Political Discourse.” International Journal of Applied Linguistics 10(1): 41–63. ). In this excerpt, the tentative I think functions as “a marker of the speaker’s lack of knowledge” (Simon-Vandenbergen 2000Simon-Vandenbergen, Anne-Marie 2000 “The Functions of I Think in Political Discourse.” International Journal of Applied Linguistics 10(1): 41–63. , 51) of television size.

Extract 5.(Lavender 8, Week 5)

1   TW1    Do you know 42 size? 42
2   JP1    42 size?
3   TW1    For example, in the normal family, the TV’s size, the screen is maybe like 30,
4          30 I think. But maybe in then カラオ the size may be 42 to 60.
                                         karaoke
                                         ‘karaoke’
5   JP1    Oh.
6   TW1    About 60.

Extract 6 records an exchange between two Japanese students and a Taiwanese student in which the latter starts a new topic by asking a question pertaining to a particular food. Attempting to elicit a response, she asks if the Japanese speakers know about the two kinds of the food. One of the Japanese speakers replies, “Maybe in Japan かき is only the hard one. I think so. How about you?” This utterance conveys uncertainty and a lack of knowledge, as it includes the word “maybe.” Although the speaker specifies “only the hard one,” which expresses some degree of certainty, this utterance is succeeded by “I think so,” conveying the sense of an afterthought, before the speaker projects the question to the other speakers.

Extract 6.(Lavender 7, Week 3)

1   TW1    Do you know there are two kind of かき?
                                             ka ki
                                             ‘oysters’
2   JP1    Two kinds of?
3   TW1    The tough and the soft one. You know?
4   JP1    Maybe in Japan かき is only the hard one. I think so. How about you?
                          ka ki
                          ‘oysters’
5   JP2    I think so, too.
6   JP1    Maybe we have not eaten the soft one. かき
                                                 ka ki
                                                 ‘oysters’
7   TW1    Oh.

4.2.4Discursive ‘I think’

As mentioned earlier, I think can also be used to organize information in conversation (Kärkkäinen 2003Kärkkäinen, Elise 2003 Epistemic Stance in English Conversation. A Description of Its Interactional Functions, with a Focus on I Think. Amsterdam: John Benjamins. ). In this textual and interactional capacity, the marker is labeled the discursive I think. A significant difference was observed in the use of this function between the two sets of speakers. Approximately five times more of the uses of I think by the Taiwanese students (27 instances) functioned as the discursive I think, compared with those of the Japanese students (five instances). This difference in frequency strongly suggests that the Taiwanese students were more aware of the flexibility available in positioning the I think marker, particularly in placing it outside the grammatical structure of the utterance. The versatility of the positioning of I think in the utterances of the Taiwanese students may indicate a process of language evolution (Traugott 1995Traugott, Elizabeth 1995 “Subjectification in Grammaticalisation.” In Subjectivity and Subjectivisation: Linguistic Perspectives, ed. by Dieter Stein and Susan Wright, 31–54. Cambridge University Press: Cambridge. ) that is evident not only to native speakers but even to nonnative speakers. According to Traugott (1995)Traugott, Elizabeth 1995 “Subjectification in Grammaticalisation.” In Subjectivity and Subjectivisation: Linguistic Perspectives, ed. by Dieter Stein and Susan Wright, 31–54. Cambridge University Press: Cambridge. , such progression involves a linguistic form originating as a lexical item that loses its referential meaning as it develops but concurrently acquires meanings at the textual, pragmatic, and subjective levels. The data on the uses of I think by the Taiwanese students suggest that the speakers used the linguistic form in different syntactic positions, reflecting its versatility, its vagueness, and the context-dependence of its meaning. As a result, I think became significantly less dependent on the syntax of the utterance and could thus be used in various structures, each performing different roles in the discourse.

Extract 7 illustrates the use of the discursive I think by a Taiwanese student. The speaker refers to a ‘buffet’, but the other speakers do not immediately comprehend the meaning of the word. This gap in communication results in word repetition, which the Japanese speakers eventually understand. Next, the same Taiwanese student introduces a new topic, “ カラオケ ” ‘karaoke’. She begins her description with “In Taiwan カラオケ is umm I think is the umm…” In this case, I think is used for online planning: the speaker attempts to buy time and stall the conversation until she has identified an appropriate word to describe a music room or karaoke.

Extract 7.(Lavender 8, Week 5)

1   TW1    Umm, maybe because maybe we will have the buffet.
2   JP1    Buffet.
3   TW1    Yes, buffet. If you pay more money, you can eat.
4   JP1    Oh.
5   TW1    Buffet B--f-f- buffet.
6   JP1    Oh oh oh buffet.
7   TW1    Yes.
8   JP1    Oh.
9   TW1    In Taiwan カラオケis umm I think is the umm…in the special room.
                     karaoke
                     ‘karaoke’

A Japanese student’s use of the discursive I think is recorded in Extract 8. A Taiwanese student initially expresses his thoughts on being acquainted or not with others. To confirm that he has accurately understood what has been said, a Japanese student says, “Feeling, acquainted, each other. Er…me, I think,” which is affirmed by the Taiwanese speaker with “yap.” As he begins answering the question and expressing his opinion, he says, “Er…I think my, ha ha.” The I think occurrence in this extract shows no clear syntactic position, and is clustered with “er” and “haha.” Therefore, this use of I think has a discursive role, as it functions as a filling or stalling device, providing the speaker with time to gather his thoughts. Furthermore, although the laughter toward the end of the utterance may manifest face-saving, suggesting that I think has a mitigating function, no clear opinion is stated by the Japanese speaker, which may disqualify the phrase from performing such a role.

Extract 8.(Yellow 7, Week 1)

1   TW1    Sometimes, because I thought, I think you are not the, you are not the, you
2          are not very acquainted with each other, would you think about, would you
3          have a feeling about that? not acquainted, not acquainted with each other.
4   JP1    Feeling, acquainted, each other. Er…me, I think…
5   TW1    Yap.
6   JP1    Er…I think my, ha ha.
7   TW1    Yeah your teammate.
8   JP1    My teammate.
9   TW1    Is also, is all fantastic?

4.2.5Mitigating ‘I think’

Functioning to weaken the intensity of an utterance, the mitigating I think conveys softness and politeness (Jucker 1986Jucker, Andreas 1986 News Interviews. Amsterdam: John Benjamins. ; Trappes-Lomax 2007Trappes-Lomax, Hugh 2007 “Vague Language as a Means of Self-Protective Avoidance: Tension Management in Conference Talks.” In Vague Language Explored, ed. by Joan Cutting, 117–137. Hampshire: Palgrave Macmillan. ). Compared with the tentative I think, which is rarely used on purpose, as in the case of a lack of knowledge, the mitigating I think is used by speakers intentionally and strategically (Zhang 2014Zhang, Grace Qiao 2014 “The Elasticity of I Think: Stretching Its Pragmatic Functions.” Intercultural Pragmatics 11(2): 225–257. ). In our data, the Taiwanese students (21 instances) used the mitigating I think seven times more frequently than their Japanese counterparts (three instances). These findings indicate that the Taiwanese students might have attempted to reduce the intensity of their expression of stance, probably motivated by their interest in engaging the Japanese speakers in the interaction by sounding less threatening.

Extract 9 illustrates the use of the mitigating I think by a Taiwanese student. She responds in the affirmative to a Japanese student’s inquiry regarding a Japanese television program, and elaborates on her answer by mentioning the many singers who appear on the show. As a final remark, she says, “Very worth to see it, I think.” This utterance initially conveys evaluation in an assertive tone, particularly in the use of “very.” However, toward the end of the statement, the Taiwanese speaker says, “I think.” Therefore, this occurrence of the marker achieves a mitigating function, as it occupies the clause-final position and performs a hedge-like role within an emphatic statement.

Extract 9.(Lavender 7, Week 1)

1   JP1    So have you ever seen this TV program?
2   TW1    Yes.
3   JP1    Kohaku.
4   TW1    Yes, because there are many famous singer in Japan to come to show the
           different show in program, very worth to see it, I think.
5   JP1    Oh…so…uh…what are you good doing? What are you good at doing?
6   TW1    Is our strength?

An example of the use of the mitigating I think by a Japanese student is shown in Extract 10. Here, a Taiwanese student solicits suggestions from the other speakers for improving the CCDL class. A Japanese student replies, “Umm, actually your English is good, I think.” This utterance at first appears emphatic, with the use of “actually” to convey intensity. However, the Japanese student uses I think toward the end of the utterance, softening and toning down his claim. Here, therefore, the use of I think has a mitigating role, serving as a hedge to the emphatic statement “actually your English is good.”

Extract 10.(Lavender 7, Week 5)

1   TW1    I makes you. And so…right now we have…
2   JP1    And there…right now we have all learned CCDL class.
3   TW2    Do you find anything we can improve next time? Or some advice to
4          give us, or you need to…
5   JP1    Umm, actually your English is good, I think. And Japanese are also
6          good. I was surprised.
7   TW2    Yeah.
8   TW1    And I think your English is also good, is more fluent than other people.
9   TW2    Yeah.

5.Discussion

The results indicate that in their computer-mediated, cross-cultural communication, both Taiwanese and Japanese speakers used the evaluative I think more frequently than I think in its other functions, and to use the mitigating I think the least. In addition, both the Taiwanese and Japanese students seemed to show interest in establishing rapport in their cross-cultural communication, but used different approaches. For example, the Japanese students tended to use the emphatic I think to convey assurance when agreeing with another speaker, whereas the Taiwanese students were more likely to use the tentative I think to convey approximation when expressing evaluative statements.

In terms of differences, the findings indicate that the Taiwanese students displayed more awareness of the different functions of I think in discourse, as suggested by their high frequency of use of I think utterances performing various roles. These nonnative speakers used the stance marker to convey opinion, i.e., a propositional attitude (the evaluative I think); to hedge (the tentative or mitigating I think); as a booster (the emphatic I think); or in its evolved function, as a structural device. The Taiwanese students played a comparatively dominant role in the interaction by initiating topics, introducing new ideas, asking questions and establishing rapport with the presence of the various functions of I think. They seemed to carry the burden of maintaining the flow of conversation and avoiding dead air.

On the other hand, the Japanese students used the five functions of I think much infrequently than their Taiwanese counterparts in their online cross-cultural communication. This suggests that the Japanese speakers are less aware of the multiple functions of I think, and that their conversational styles may be a significant factor determining the ways in which such nonnative speakers express subjective meanings and take stances in their interactions. As Davidson (1998)Davidson, Bruce 1998 “Comments on Dwight Atkinson’s ‘A Critical Approach to Critical Thinking in TESOL’: A Case for Critical Thinking in the English Language Classroom.” TESOL Quarterly 32(1): 119–123. explains, Japanese discourse involves low risk taking, which suggests that Japanese speakers delay their expressions of stance to allow them to consider others’ viewpoints and to be certain of their evaluation. They thus avoid engaging in behavior that puts them at risk of being negatively evaluated (Brown 2004Brown, Robert 2004 “Learning Consequences of Fear of Negative Evaluation and Modesty for Japanese EFL students.” The Language Teacher 28(1): 15–17. ) or making “the interlocutors lose face” (Blight and Stephens 2006Blight, Richard, and Meredith Stephens 2006 “Discourse Strategies Involving Topic Nomination in English and Japanese Language Usage.” English Education Center Journal 4: 41–53.: 43). Such reserved behavior makes them appear unresponsive and unspontaneous (Kato 2001Kato, Kumi 2001 “Overseas-Educated Teachers in Australian Classrooms.” Babel 36(2): 30–36.; Tsuda, Shigemitsu, and Murata 2007Tsuda, Sanae, Yuka Shigemitsu, and Kazuyo Murata 2007 “Cultural Awareness in Teaching English: Analysing Intercultural Communication and Teaching Positive Politeness Strategies.” The Journal of Asia TEFL 4(3): 161–185.). After delaying expressions of stance to give ample time for careful deliberation, the Japanese speakers involved in this study tended to use the emphatic I think to convey a final evaluative statement with certainty and commitment. This apparent preference for the emphatic I think seems to be in contrast with the findings of a study suggesting that Japanese students use self-qualification to avoid committing to a stance (Mori 1999Mori, Junko 1999 “ Well I may Be Exaggerating But…: Self-Qualifying Clauses in Negotiating of Opinions Among Japanese Speakers.” Human Studies 22(2–4): 447–473. ). A possible explanation of this difference is that the Japanese students involved in the current study delayed the expression of their opinion for long enough to gain a high degree of certainty, hence their decision to use the emphatic I think to express certainty, authority, and reassurance.

6.Conclusion

This study provides a detailed analysis of the use of I think in computer-mediated cross-cultural communication and explores the distribution of this discourse marker in the CCDL corpus. The five functions of I think – evaluative, emphatic, tentative, mitigating, and discursive – are the focus of the study. To identify these functions, the study adopted the list of criteria used by Aijmer (1997)Aijmer, Karin 1997 “ I think–an English Modal Particle.” In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter. as well as Kadar and Haugh (2013)Kádár, Dániel Z., and Haugh, Michael 2013 Understanding Politeness. Cambridge: Cambridge University Press. , specifically with respect to collocation, the situational context of occurrence and sequentiality.

This study raises awareness of the various roles by I think can perform – roles that functioned to express speaker stance in computer-mediated cross-cultural communication between Japanese and Taiwanese students. As online intercultural communication appears to be a popular means for English learners to practice the target language, considerable attention should be paid to the expression of opinion, assessment, and attitude by these language learners. Language teachers can design curricula and develop materials to hone intercultural communicative and strategic competence, particularly in conveying stance. The incorporation of effective stance marking into language learning can build learner confidence in expressing evaluative statements and further aid students as they engage in cross-cultural interaction. Additionally, as the findings indicate that the groups to which the Japanese and Taiwanese students were assigned significantly influenced the use of I think in their cross-cultural communication, it is important for teachers to allocate learners to groups with due care and attention. Degree of familiarity tends to influence the expression of opinion, particularly Taiwanese students, and it is crucial for language teachers to be conscious of this factor as they plan lessons so as to avoid causing anxiety. Rather, teachers should aim to encourage students to express their thoughts in English.