John Benjamins Publishing Company

John Benjamins Publishing Company

This is a contribution from International Journal of Corpus Linguistics 18:3 © 2013. John Benjamins Publishing Company This electronic file may not be altered in any way. The author(s) of this article is/are permitted to use this PDF file to generate printed copies to be used by way of offprints, for their personal use only. Permission is granted by the publishers to post this file on a closed server which is accessible to members (students and staff) only of the author’s/s’ institute, it is not permitted to post this PDF on the open internet. For any other use of this material prior written permission should be obtained from the publishers or through the Copyright Clearance Center (for USA: www.copyright.com). Please contact [email protected] or consult our website: www.benjamins.com Tables of Contents, abstracts and guidelines are available at www.benjamins.com

Overuse or underuse A corpus study of English phrasal verb use by Chinese, British and American university students* Meilin Chen

City University of Hong Kong

This study explores Chinese university students’ use of phrasal verbs in comparison with their American and British counterparts by utilizing a corpus of learner English and four native novice corpora of two English varieties and two genres (argumentative and academic writing). The results show that it is difficult to state whether the Chinese learners of English over- or underuse phrasal verbs in writing because a more striking difference emerges between the British and American students. American students tend to use many more phrasal verbs in both genres than British students and they also use a greater variety of phrasal verbs. Notwithstanding the differences, both American and British students tend to use fewer phrasal verbs in academic writing than in argumentative writing. The learners do not show a fundamental difference from the British students regarding overall frequencies of phrasal verbs; however, the learner-native writer gap does exist between the Chinese and American students. Keywords: phrasal verbs, leaner corpus studies, native novice corpora

1. Introduction: Challenges of finding comparable reference corpora in learner corpus studies Since the 1960s when the first modern English corpus, i.e. the Brown corpus (Kucěra & Francis 1967), became available to the academic community, corpus linguistics has developed rapidly and made considerable contributions to different areas of linguistics. Its contribution to second language acquisition (SLA) research is also widely acknowledged; for instance, numerous learner corpora, such as the International Corpus of Learner English (ICLE, Granger et al. 2009), have been emerging since the early 1990s (see Xiao 2008). Different native English corpora have also become available to researchers, such as the British N ational International Journal of Corpus Linguistics 18:3 (2013), 418–442. doi 10.1075/ijcl.18.3.07che issn 1384–6655 / e-issn 1569–9811 © John Benjamins Publishing Company

Overuse or underuse: A corpus study of English phrasal verb use 419

orpus (BNC 2007), the British Academic Written English Corpus (BAWE C undated, see Gardner & Nesi 2012) and the Michigan Corpus of Upper-Level Student Papers (MICUSP 2009). Notwithstanding the growing number of available corpora, one of the challenges in applying corpora to language learning and teaching is that very few available corpora could be “exactly the kind of English corpora that will give frequency information relevant to learners” (Leech 2011: 25). The same challenge exists in learner corpus studies when one searches for a native comparable corpus. Also the importance of establishing a shared ground prior to any analysis has been long acknowledged and repeatedly discussed in contrastive linguistics (see James 1980; Krzeszowski 1984, 1990; Connor & Moreno 2005), “since it is only against a background of sameness that differences are significant” (James 1980: 169). This shared ground or common platform is often referred to as ‘equivalences’ or ‘tertium comparationis’ by contrastive linguists (James 1980, Krzeszowski 1984, Connor & Moreno 2005). Connor & Moreno (2005) demonstrate the importance of variables such as the mode, the author, the setting, the topic, the genre, etc., in the selection of the appropriate texts for comparison, and how ‘tertium comparationis’ can be created at different levels through controlling these variables. In learner corpus studies, however, little discussion has been devoted to the importance of ‘tertium comparationis’ in selecting native corpora and the possible influences of the above-mentioned variables on the results. Among the very few researchers who have addressed this issue, Ädel (2008) suggests that the differences observed in her study between the learners and the native writers regarding the use of metadiscourse may be due to extralinguistic factors such as intertextuality (i.e. the native writers had secondary texts on which they could base their argument) and task setting (i.e. the native essays were untimed while the learner essays were written within a short time limit). Kaszubski (1998) also suggests that “it may be psycholinguistically more appropriate to compare EFL learner corpora not with ideal ‘expert performances’ in the target language but (more realistically) with the attainable performances of native learners of a comparable, preferably slightly lower, age and experience” (Kaszubski 1998: 25). In addition to the extralinguistic factors discussed above, another factor that cannot be ignored in selecting a native corpus is the variety of L1 English. Differences between varieties of English, American and British in particular, have been widely acknowledged and well documented in a large body of literature (e.g. Algeo 2006, Bauer 1988, Potts & Baker 2012, Rohdenburg & Schlüter 2009). In learner corpus studies, discrepancies between novice writers of different L1 varieties in the use of certain vocabulary items have also been identified. In a study on learners’ use of metadiscourse, Ädel (2008) breaks down the Louvain Corpus of Native English Essays (LOCNESS, Granger et al. 2009), which includes essays © 2013. John Benjamins Publishing Company All rights reserved

420 Meilin Chen

by American and British students, into two sub-corpora of American and British English novice writing. The results show that American novice writers use significantly more personal metadiscourse than their British counterparts. Surveys of ESL/EFL learners’ attitudes towards English varieties show that learners are aware of the differences between American and British English, and they show marked preference for a certain variety or varieties (Jarvella et al. 2001, Virtanen & Lindgrén 1998, Zhang & Hu 2008). Furthermore, based on the Swedish and Finnish sub-corpora of ICLE, Virtanen & Lindgrén (1998) found that the majority of the Swedish and Finnish learners in those corpora do not mix American and British spellings and vocabulary in their writing. Although performance of the Swedish and Finnish learners does not represent English learners of other L1s, the survey findings mentioned above alone bring to light the necessity of distinguishing between different English varieties in English teaching and interlanguage assessment. The present study, in order to take this issue into account, uses four different native novice corpora to facilitate the investigation of phrasal verb use in learner writing. The reasons for choosing these corpora are given in Section 3. 2. Phrasal verbs in English language and SLA Phrasal verbs, like other phraseological units, are omnipresent in all fields of language use. Taking the 100-million-word BNC, Gardner & Davies (2007) find that phrasal verbs occur approximately “every 192 words, that is, almost two phrasal verbs per page of written text on average” (Gardner & Davies 207: 347). Mastery of a large repertoire of phrasal verbs, together with other types of phraseological units by native English writers, is considered an important difference between their writing and learner writing. Despite their frequency, phrasal verbs are perceived as notoriously difficult for ESL/EFL learners because they are semantically non-compositional, very often polysemous, and syntactically more flexible than other types of phraseological units (e.g. variation of particle positions and pronoun or noun insertions are allowed in phrasal verbs).1 Previous empirical studies have repeatedly found that learners tend to avoid using phrasal verbs when there is a single-verb counterpart available (Dagut & Laufer 1985, Hulstijn & Marchena 1989, Laufer & Eliasson 1993, Liao & Fukuya 2004, Siyanova & Schmitt 2007). The evidence that accounts for this avoidance behaviour by ESL/EFL learners, however, is not conclusive in these studies. Laufer and her colleagues (Dagut & Laufer 1985, Laufer & Eliasson 1993) found that Hebrew-speaking learners tend to avoid using phrasal verbs due to the absence of such a construction in their L1. Other researchers (for Dutch learners of English © 2013. John Benjamins Publishing Company All rights reserved


see Hulstijn & Marchena 1989, for Chinese EFL learners see Liao & Fukuya 2004), in contrast, find that the L1 does not have any great impact on the use and avoidance of phrasal verbs. Liao & Fukuya (2004) find that overall English proficiency is an important parameter in learners’ use of phrasal verbs, while Siyanova & Schmitt (2007) discover that a lengthy stay (at least 12 months) in an Englishspeaking environment leads to an increase in the use of phrasal verbs. Another factor that has an influence on the use of phrasal verbs by ESL/EFL learners is identified by Schmitt & Redwood (2011). They find that the frequency of phrasal verbs has a significant correlation with learners’ productive knowledge of this type of construction. Among the 50 high frequency phrasal verbs in the BNC taken from Gardner & Davis (2007) and 10 less frequent phrasal verbs taken from student course books and grammar reference books, more high frequency phrasal verbs are used by most of the learners in a productive test than the less frequent ones. Learner corpus studies also reveal a mixed picture of phrasal-verb use by ESL/ EFL learners. For example, Waibel (2007) find that while the frequency of phrasal verbs in many sub-corpora of ICLE (e.g. French, Spanish, Italian, Russian, etc.) is lower than that in LOCNESS (a comparison native corpus), phrasal verbs in the Dutch and Polish sub-corpora do not show a significant numerical difference from those in LOCNESS and phrasal verbs in the German sub-corpus outnumber those in LOCNESS. Based on LOCNESS and the Louvain Corpus of Native English Conversation (LOCNEC, Gilquin et al. 2010), Gilquin (2011) carries out a case study of phrasal verbs with up in written and spoken interlanguage in comparison with those in native English. The results show that the learners tend to use phrasal verbs more frequently in writing in comparison with native writers, yet their use of phrasal verbs in speech is significantly less frequent than that of native speakers. Both Gilquin (2011) and Waibel (2007) attribute the discrepancies between the learners to L1, i.e. learners whose L1 (e.g. a Germanic language) possesses the phrasal verb construction use phrasal verbs more frequently than those whose L1 does not (e.g. a Romance language). The present study explores phrasal verb use by Chinese EFL learners, whose L1 lacks the construction, in comparison with phrasal verb use by American and British novice writers. 3. The corpora The 188,628-word learner corpus used in this study was compiled by the author. It is comprised of 780 argumentative essays contributed by 130 English majors (six essays by each student) at a Chinese university in their first three years of undergraduate studies. The essays were collected at the end of each semester under © 2013. John Benjamins Publishing Company All rights reserved

422 Meilin Chen

e xamination conditions where the students were required to finish an essay within 30 minutes with no access to reference tools. When the data were collected, the Chinese EFL learners passed the Chinese national English proficiency exam for English majors Test for English Majors-4 (TEM-4).2 This exam, which students are required to pass for the completion of their four-year undergraduate studies, requires mastery of a 6,000-word vocabulary (which includes phraseological units such as idioms, multi-word combinations, phrasal verbs, etc.). A pass score from this exam is roughly equivalent to a 6.5 score for IELTS or 550 for paperbased TOEFL. The learners, therefore, could be considered upper intermediate EFL learners. The programme in which these students were enrolled emphasises equal input of British and American English with a complementary introduction to other varieties of English such as Australian and Canadian English. For instance, the learners listened to news broadcasts from both the British Broadcasting Corporation and the Voice of America in their listening courses. Both American and British literary works were required for the reading module of their programme. The phrasal verb use in the learner corpus (CH) was compared to that in four native corpora. The first two corpora include argumentative essays written by American and British novice writers taken from: the LOCNESS corpus (LOCNESS-US) and the General Studies corpus (GS-UK) (Milton 2001).3 The second two corpora consist of academic essays taken from the MICUSP and the BAWE. The reason for utilizing the two native argumentative corpora is evident, while the reason for using the two academic corpora is as follows. At the initial stage of this study, the author compared argumentative essays from the LOCNESS corpus and the General Studies corpus and found a striking difference in phrasal verb use: considerably more phrasal verbs were found in the LOCNESS corpus than in the General Studies corpus. After dividing the LOCNESS corpus into two sub-corpora, i.e. one including only American essays and the other only British essays, the results showed that over 60% of the phrasal verbs in the LOCNESS corpus were from the American essays. In order to determine whether the difference was triggered by the English variety factor, two native novice corpora of a different genre, i.e. academic writing, were included in this study (MICSUP and BAWE). If more phrasal verbs were also found in the American academic corpus, it could be confirmed that American novice writers tend to use more phrasal verbs in writing in general. To make the four native novice corpora as comparable as possible to the Chinese learner corpus, many undesired texts were excluded. From LOCNESS, all the British essays were taken out so that the abridged version, i.e. LOCNESS-US, could represent American argumentative writing in contrast to the GS-UK corpus of British argumentative writing. Second, essays in LOCNESS-US that were © 2013. John Benjamins Publishing Company All rights reserved


not written by native English speakers and those on literary topics were also ruled out. As for GS-UK, texts which contained only short answers to exam questions and those on literary or specialised subject topics (e.g. on biology topics) were excluded. To make a comparable native corpus out of BAWE, papers written by students at level 4 (graduate students) and by non-native English writers were first excluded from the corpus. Second, texts labelled other than essay were eliminated. Third, texts labelled as essays were narrowed down to those written by students of the following four disciplines: English, linguistics, anthropology and sociology. Finally, the author went through the topic of every essay and eliminated those that did not appear to be argumentative essays (see Appendix 1 for a list of BAWE essay topics chosen for this study). After filtering out those “illegitimate” essays, there were altogether 99 essays from the BAWE corpus containing 211,929 tokens. The same procedure was carried out with the MICUSP corpus, and 40 essays written by undergraduate students from the following disciplines were chosen for this study: English, linguistics, education and sociology. Table 1 presents the profiles of the five corpora used in the study. Table 1. Profiles of the five corpora used in the study Tokens

CH

LOCNESS-US

GS-UK

MICUSP

BAWE

188,628

227,972

398,279

88,217

211,929

AmE

BrE

AmE

L1

Chinese

Genre

argumentative argumentative essay essay

argumentative academic essay essay

academic essay

No. of topics

6

84

29

40

85

No. of essays

780

318

590

40

99

BrE

4. Methodology For the data analysis, phrasal verbs were extracted from the corpora and determined through the following procedure. The first step was to annotate the five novice corpora with part-of-speech (POS) tags by using the online free CLAWS POS tagger (C7 tagset) provided by Lancaster University.4 Next, all co-occurring lexical verbs and adverbial particles were extracted from the corpus using the Concord function of WordSmith Tools 5.0 (Scott 2008). For the purpose of this study, phrasal verbs were defined as any two-part verbs consisting of a lexical verb followed (continuously or discontinuously) by an adverbial particle, which “behaves to some extent either lexically or syntactically as a single verb” (Quirk et al. 1985: 1150). However, the concordances automatically © 2013. John Benjamins Publishing Company All rights reserved

424 Meilin Chen

generated by WordSmith Tools included not only phrasal verbs but also verb + particle free combinations and verb + prepositional phrases since many particles function as both adverbs and prepositions. For example, act out is a phrasal verb in These children acted out their anger in the previous game., but it is a verb + a prepositional phrase in [...] and they acted out of concern for anyone who may [...]. Therefore, the second step involved a manual check in order to rule out verb + prepositional phrases. The third step was to check all the results in phrasal verb dictionaries in order to determine which meanings made them phrasal verbs or verb + particle free combinations, using: Collins COBUILD Dictionary of Phrasal Verbs 2nd Edition, Longman Dictionary of Phrasal Verbs (Longman hereafter) and Oxford Phrasal Verbs Dictionary for Learners of English. After going through these procedures, the analysis was carried out to explore answers to the following questions regarding the Chinese EFL learners’ use of phrasal verbs: i. Do the Chinese EFL learners use phrasal verbs less frequently than their native counterparts? In other words, do they avoid using phrasal verbs as observed in previous empirical studies (e.g. Dagut & Laufer 1985, Hulstijn & Marchena 1989, Liao & Fukuya 2004)? ii. Does the individual factor show a significant impact on the overall results of phrasal verbs in the learner corpus? In other words, are phrasal verbs distributed evenly across individual novice writers? This question also applies to native novice corpora. iii. Does phrasal verb frequency have a positive association with the learners’ production of phrasal verbs as suggested in the previous empirical study (cf. Schmitt & Redwood 2011)? Apart from the questions concerning phrasal verb use in the learner corpus, the present study aims to explore an important methodological issue in learner corpus studies: iv. Is the L1 variety factor an important parameter in the selection of native comparison corpora in learner corpus studies? In other words, do the American and British corpora show any difference in terms of phrasal verb use? For the first research question, the overall frequencies of phrasal verbs in the learner corpus were calculated in comparison with those in the native novice corpora. The significance of the differences between the corpora (if there are any differences) was then determined by using Chi-square tests. To further ensure the reliability of the differences observed between the corpora with respect to the overall results, an adjusted frequency test was carried out using Gries’s (2008, © 2013. John Benjamins Publishing Company All rights reserved


2009, 2012: 2) measure for dispersion: DP (for ‘deviation of proportions’). The purpose of conducting this test was to decide whether the overall frequency of phrasal verbs in each corpus was affected by highly frequent use of this construction by certain individual writers and/or by text length. The distribution results thus provide answers to the second research question. To address the third question, the highly frequent phrasal verbs in the BNC and the COCA from Gardner & Davies (2007) and Liu (2011) respectively were used as a yardstick to determine the frequency of certain phrasal verbs in English. The analysis first involves comparing the frequency rankings of phrasal verbs in the expert corpora (BNC and COCA) with those in the five novice corpora. Pearson correlation coefficients5 were used to determine the association between the frequency rankings in the expert corpora and the novice corpora. In addition to the results from the complete BNC and COCA, the 50 most frequent phrasal verbs in the academic corpora of the BNC and the COCA were also used (see Appendix 2). The reason for including the results from the academic sub-corpora is that phrasal verbs behave very differently in academic genres than in other genres. The results may not be reliable if the phrasal verbs in the native novice academic corpora were compared to those in the whole expert corpora. During the exploration for answers to the first three research questions, comparisons were carried out not only between the learner corpus and the native novice corpora but also between the native novice corpora themselves. The similarities as well as differences between the native novice corpora were recorded and discussed, which leads to answers to the fourth research question. The following section presents the results of analysis. 5. Results 5.1

Overall frequencies of phrasal verbs

Analysis of the overall frequencies of phrasal verbs in the five corpora first reveals that the Chinese EFL learners’ use of phrasal verbs resembles that of the British students in argumentative writing most, but shows a striking difference from that by the American students. Table 2 shows that the relative frequency of phrasal verb types measured in parts per million words (ppm) in the learner corpus (891 ppm) is slightly higher than that in GS-UK (861 ppm), and the relative frequency of phrasal verb tokens in the learner corpus (3,054 ppm) is also very similar to that in GS-UK (2,998 ppm). However, the relative frequencies of both phrasal verb types and tokens in the learner corpus are much lower than those in LOCNESS-US (types: 1,838 ppm, tokens: 4,188 ppm). © 2013. John Benjamins Publishing Company All rights reserved

426 Meilin Chen

Table 2. Phrasal verb frequencies in the five corpora* Corpus CH LOCNESS-US GS-UK MICUSP BAWE

Corpus size 188,628 144,709 398,279 88,217 211,929

PHV types

PHV tokens

Abs.

Rel.

Abs.

Rel.

168 266 343 125 171

891 1838 861 1417 807

576 606 1194 229 448

3054 4188 2998 2596 2114

* Abs. = absolute frequency. Rel. = relative frequency. The relative frequency is acquired by normalising the absolute frequency to a one million token basis.

The comparison between the learner corpus and the two native novice academic corpora (MICUSP and BAWE) shows that the Chinese students use phrasal verbs more frequently in argumentative writing than their American and British counterparts do in academic writing. The relative frequency of phrasal verb tokens in the learner corpus (3,054 ppm) is much higher than in both academic corpora (MICUSP: 2,596 ppm, BAWE: 2,114 ppm). Less frequent use of phrasal verbs by the native novice writers of both L1 varieties in academic writing reflects the convention of English academic writing, i.e. phrasal verbs are more informal or colloquial in style and should be avoided in formal writing (Swales & Feak 2004). Chi-square test results (applied to absolute frequencies of phrasal verb tokens) show that the difference between the learner corpus and GS-UK is not significant (x2 = 0.13, p > .5), but phrasal verbs in the learner corpus are significantly more frequent than those in both native novice academic writing (CH vs. MICSUP: x2 = 4.34, p < .05; CH vs. BAWE: x2 = 34.56, p < .01). Although the phrasal verb tokens in the learner corpus are fewer than those in LOCNESS-US by a statistically significant margin (x2 = 29.81, p < .01), the evidence that phrasal verb tokens in the learner corpus slightly outnumber those in GS-UK suggests a positive answer to the first research question, i.e. the learners are capable of producing a sufficient number of phrasal verbs in writing and the avoidance of this type of construction observed in previous empirical studies was not found in the present study. The third finding lies between the native novice corpora of both varieties. The total number of both phrasal verb types and tokens in the two American novice corpora is much larger than those in the British ones. The relative frequency of phrasal verb types in the American argumentative novice corpus (LOCNESS-US: 1,838 ppm) is almost 1,000 ppm more than in its British counterpart (GS-UK: 861 ppm), and the phrasal verb types in the American academic corpus (MICUSP: 1,417 ppm) outnumber those in its British counterpart (BAWE: 807 ppm) by more than 600 ppm. Chi-square tests show that the differences are significant © 2013. John Benjamins Publishing Company All rights reserved


(LOCNESS-US vs. GS-UK: x2 = 45.48, p < .01; MICUSP vs. BAWE: x2 = 6.43, p .05; ^^ p > .1

Similarly, the strengths of the correlation coefficients for LOCNESS-US are positive (LOCNESS-US – COCA: r = 0.42, p < .01; LOCNESS-US – BNC: r = 0.5, p < .01). 18.2% (r2) of the variance in phrasal verb frequency rankings in LOCNESSUS is attributable to that in American English (COCA) while 25.1% to that in British English (BNC). As for the British argumentative corpus, the strength of the association between phrasal verb frequency rankings in GS-UK and those in the COCA is small (r = 0.27, p < .05), while the strength of the association between those in GS-UK and the BNC is moderate (r = 0.36, p < .01). Only 7.8% of the variance in phrasal verb frequency rankings in GS-UK is correlated with that in the COCA while 13.2% with that in the BNC. The phrasal verb frequency rankings in both native novice argumentative corpora (LOCNESS-US and GS-UK) show a closer association with those in the BNC than in the COCA. On the other hand, there is no significant correlation between the phrasal verb frequency rankings in the native novice academic corpora (MICUSP and BAWE) and those in the general English corpora (r < 0.3, p > .05 for both MICUSP and BAWE). The lack of association between phrasal verb frequency rankings in the native novice academic corpora and those in the general English corpora does not indicate that the native novice writers do not use high-frequency phrasal verbs commonly in academic writing. A more reasonable explanation for the lack of association is as follows. Phrasal verbs are more commonly used in spoken or informal registers than in written or formal ones (Biber et al. 1999). Given that texts from spoken and fiction registers make up a large proportion of the BNC (over 25%) and the COCA (around 40%), the 50 most frequent phrasal verbs from the complete BNC and COCA may consist of many phrasal verbs that are more frequently used in spoken or informal registers than in written or formal ones. These spoken or informal phrasal verbs were avoided by the native novice writers, which leads to the lack of association between frequency rankings in the general English corpora and those in the native novice academic corpora. © 2013. John Benjamins Publishing Company All rights reserved


To test this hypothesis, the 50 most frequent phrasal verbs in the academic sub-corpus of the COCA (COCA-AC) were first drawn from Liu (2011). As Gardner & Davis (2007) do not explore the frequencies of the top 100 phrasal verbs in different BNC sub-corpora, we searched the top 100 phrasal verbs from Gardner & Davis (2007) in the academic sub-corpus of the BNC (BNCweb CQP-Edition) to obtain the 50 most frequent phrasal verbs in the BNC-AC (see Appendix 2). The frequency rankings of the top 50 phrasal verbs in COCA-AC and BNC-AC were then compared with their rankings in the five novice corpora. Table 6 details the results in comparison with those based on the complete BNC and COCA presented previously in Table 5. Table 6. A comparison between the correlation results based on the complete COCA and BNC and their academic sub-corpora Corpus CH LOCNESS-US GS-UK MICUSP BAWE

COCA

COCA-AC

r

r2

0.41** 0.42** 0.27* 0.13^^ 0.09^^

17.1 18.2 7.8 1.8 0.9

(%)

BNC

r

r2

0.38** 0.45** 0.38** 0.49** 0.35*

14.7 20.5 14.5 23.5 12.0

(%)

BNC-AC

r

r2

0.34* 0.50** 0.36** 0.20^^ 0.25^

11.9 25.1 13.1 4.2 6.3

(%)

r

r2 (%)

0.40** 0.43** 0.36* 0.38** 0.45**

16.2 18.2 13.2 14.6 20.2

* p < .05; ** p < .01; ^ p > .05; ^^ p > .1

Pearson correlation coefficients in Table 6 show that the correlations between phrasal verb frequency rankings in the learner corpus and BNC-AC and COCAAC are similar to the complete BNC and COCA results. Ranging from 0.38 to 0.4, the strengths of the associations between phrasal verb frequency rankings in the learner corpus and those in the BNC-AC and COCA-AC are positive but moderate (0.34 to 0.41 for the complete BNC and COCA). In short, the phrasal verb frequency rankings in any of these four corpora do not predict the frequency rankings in the learner corpus a great deal better than the remaining three (from 12 to 17%). This suggests that the difference between phrasal verb frequency in general English and academic English or that between phrasal verb frequency in American and British English seems not to greatly affect the Chinese learners’ use of phrasal verbs. This corroborates Schmitt & Redwood’s (2011) finding. They make a distinction between the phrasal verb frequencies in the written and the spoken sub-corpora of the BNC but discover no significant difference in the learners’ phrasal verb knowledge regarding the distinction. The correlations between phrasal verb frequency rankings in the native novice argumentative corpora and those in the academic corpus of their own L1 v arieties

© 2013. John Benjamins Publishing Company All rights reserved

432 Meilin Chen

(LOCNESS-US: COCA-AC: r = 0.45, p < .01; GS-UK: BNC-AC: r = 0.42, p < .01) are also similar to the general corpora results. Phrasal verb frequency rankings in COCA and COCA-AC predict those in LOCNESS-US at the 18–21% covariance level. Similarly, the correlation between phrasal verb frequency rankings in GSUS and those in BNC-AC (r = 0.36, p < .01) shows virtually no difference from the complete BNC result (r = 0.36, p < .05). Phrasal verb frequency rankings in BNC and BNC-AC predict those in the GS-UK at the 13% covariance level. However, the correlations between phrasal verb frequency rankings in the native novice academic corpora and the COCA and BNC academic sub-corpora suggest a considerable difference from the complete COCA and BNC results. Table 6 shows there is no significant correlation between phrasal verb frequency rankings in the native novice argumentative corpora and those in the BNC and the COCA. The correlation between those in the native novice academic corpora and in the BNC and COCA academic sub-corpora is, nevertheless, positive (r = 0.35–0.49, p < .05), and the association is closer between the corpora of the same variety (BAWE and BNC-AC, MICUSP and COCA-AC). 20.2% of the variance in phrasal verb frequency rankings in BAWE is correlated with that in BNC-AC while 12% with that in COCA-AC. On the other hand, 23.5% of the variance in phrasal verb frequency in MICUSP is attributable to that in COCA-AC but only 14.6% to that in BNC-AC. The comparisons of phrasal verb frequency rankings show that phrasal verb frequency has a positive association with the learners’ production of phrasal verbs. Such association was also found with the native novice argumentative corpora. The distinction between phrasal verb frequencies in general English (BNC and COCA) and academic written English (BNC-AC and COCA-AC) reveal no noticeable difference in phrasal verb frequencies in both the learner corpus and the native novice argumentative corpora. On the other hand, phrasal verb frequency rankings in the native novice academic corpora are only related to those in the academic sub-corpora of the BNC and COCA. This indicates that the genre factor does not have a great influence on phrasal verb frequency in both learner and native novice argumentative writing, while it is important in native novice academic writing. 6. Discussion Based on the learner corpus of Chinese university students and the native novice corpora of American and British students, this study has identified similarities and differences not only between the learners and the native novice writers but also between the native novice writers of two English varieties. In response to the © 2013. John Benjamins Publishing Company All rights reserved


four research questions, we now discuss the findings and their implications for learner corpus studies and for English phrasal verb learning and teaching. 6.1

Do the Chinese students underuse or avoid using phrasal verbs?

The analysis of the overall frequencies in the learner corpus suggests that in general the learners do not show a numerical difference from the native novice writers in phrasal verb use. This finding is not congruent with the findings in Waibel (2007) and Gilquin (2011) that learners whose L1 possesses the phrasal verb structure tend to use it more frequently than those whose L1 does not: while the phrasal verb structure is missing in the Chinese language, Chinese learners of English are capable of producing a sufficient number of phrasal verbs in writing. The L1 factor, therefore, does not necessarily have a considerable impact on phrasal verb acquisition and production. An alternative explanation for more frequent use of phrasal verbs by German EFL learners than EFL learners of other L1s observed by Waibel (2007) and Gilquin (2011) could be the difference between the English proficiency of the learners. In Thewissen’s (2013) study on L2 accuracy developmental patterns, texts from the German sub-corpus of ICLE are rated as advanced or near-native according to the Common European Framework of Reference for Languages, while texts from the French and the Spanish sub-corpora are rated upper intermediate to advanced and lower to upper intermediate respectively. German EFL learners’ highly frequent use of phrasal verbs observed by Waibel (2007) and Gilquin (2011) is possibly due to their mastery of English in general. Findings from previous experimental studies show that Chinese students, similar to EFL learners of other L1s, tend to avoid phrasal verbs in experimental tasks when they are given both the phrasal verbs and their single-word verb counterparts (see Section 1). The overall results in the present study, however, indicate that learners’ avoidance behaviour observed under experimental conditions does not correlate with considerable underuse of phrasal verbs in actual writing. The phrasal verbs in the learner writing are as frequent as those in at least two native novice corpora. However, this does not mean that the learners in this study preferred phrasal verbs to single-word verbs. It is possible that the Chinese EFL learners in the present study tend to use verbs very frequently in general, and that only a small proportion of all lexical verbs in the learner corpus co-occur with adverbial particles in phrasal verbs. To test this hypothesis, a quick search for lexical verbs in the five novice corpora was carried out and the results show that verb tokens in the learner corpus (112,147 ppm) do outnumber those in the four native novice corpora (111,659 ppm in LOCNESS-US and even fewer in the other three). Therefore, Chinese


434 Meilin Chen

students’ ability to produce a sufficient number of phrasal verbs does not necessarily mean that their knowledge of phrasal verbs has reached native level. Their frequent use of phrasal verbs is due to their highly frequent use of verbs in general. Moreover, there is still a considerable frequency gap between the learner corpus and LOCNESS-US; the gap between the Chinese EFL learners and their American counterparts still exists. 6.2 Are individual differences important in learner corpus analysis? The even (although not perfectly even) distribution of phrasal verbs in the learner corpus displays a balanced picture of phrasal verb use across writers, which reflects consistent progress in phrasal verb acquisition. In other words, there are no students who “lagged behind” in the acquisition and production of phrasal verbs. The even distribution of phrasal verbs observed in the GS-UK corpus and in BAWE indicates that they are potentially very useful reference corpora for English phrasal verbs research and teaching as they present the most balanced picture of phrasal verb use in native novice writing. However, this does not mean that the two American novice corpora are inferior, as those used in this study are abridged versions. The original LOCNESS and MICUSP are most likely more balanced. The results suggest that in addition to overall frequencies, distribution of the searched item in the corpus is equally important in data analysis. In learner corpus studies, overall frequencies present a general picture of group performance. However, they may be raised to a false high level due to considerably frequent use of the item by a few individual learners. This may lead to a distorted image of the group performance and hence undermine the reliability of the results. The importance of exploring the distribution of the searched item across writers also applies to the native comparison corpus used in learner corpus studies. 6.3

Does the frequency factor matter in phrasal verb acquisition and production?

The comparisons of phrasal verb frequency rankings in the five novice corpora and those in expert corpora (BNC and COCA) reveal that phrasal verb frequency has a positive association with the production of phrasal verbs by the Chinese learners. Around 12–17% of the variance in phrasal verb frequency rankings in the learner corpus can be predicted by that in expert corpora. This finding ties in well with Schmitt & Redwood’s (2011) observation of the “general trend of high frequency leading to a greater chance of learning phrasal verbs to a productive degree of mastery” (Schmitt & Redwood 2011: 184). © 2013. John Benjamins Publishing Company All rights reserved


Schmitt & Redwood (2011) also find that the frequencies of phrasal verbs in spoken and written discourse do not show significant difference in the prediction of learners’ phrasal verb knowledge, and they suggest that “it is probably sufficient to use overall corpus frequency results when thinking about the likely acquisition of phrasal verbs, as there seems to be no real advantage to distinguishing between spoken and written frequencies” (Schmitt & Redwood 2011: 185). However, the results from the present study suggest otherwise. Even though the distinction between phrasal verb frequencies in general English and in academic written English does not reveal much difference in phrasal verb frequency in both the learner and the native novice argumentative corpora, it does show a significant difference in both native novice academic corpora. Therefore, whether to distinguish results from different discourses is determined by different factors such as genre. 6.4 Is the L1 variety factor important in the selection of native comparison corpora in learner corpus studies? Based on the four native novice corpora, this study has identified differences and similarities between native novice writers of American and British English in phrasal verb use. The American novice writers use a larger variety of phrasal verbs and tend to use them much more frequently in both argumentative and academic writing than their British counterparts. The results also suggest that the two American corpora present a less even distribution of phrasal verbs than their British counterparts. The findings bring to light an important methodological issue in learner corpus studies, i.e. that the choice of native reference corpora in learner corpus studies may affect the results considerably. In the present study, the Chinese students would be considered to use far fewer phrasal verbs if they were compared to the American novice writers (LOCNESS-US and MICUSP), but they would not display a fundamental difference from British writers (GS-UK and BAWE). The findings suggest that it is better to use native corpora with texts from different L1 varieties separated; utilizing a native corpus with texts from different L1 varieties may lead to distorted results. Students’ learning experiences also point to the importance of the L1 variety factor in the selection of native comparison corpora. If the learners studied were only exposed to one English variety (British English, for instance), it would not make any sense if their English performance were evaluated against the American English standard or a standard with a mix of different English varieties.


436 Meilin Chen

7. Limitations of the study The present study investigates the Chinese EFL learners’ use of phrasal verbs taking a quantitative approach. It does not provide any insight into the appropriateness of phrasal verbs in the learner corpus. This limitation will be overcome in a follow-up study which will analyse the learners’ use of phrasal verbs from a qualitative perspective. The second limitation of the study lies in the small size of the corpora. In order to make the native corpora more comparable, many texts were removed, which may affect the representativeness and balance of the corpora used. This might partly be the cause of the high DP value with respect to the distribution of phrasal verbs in the American novice corpora. Third, it would be ideal if academic writing by the Chinese students could be included in the present study. This would show how the learners use phrasal verbs in academic writing and provide a more comprehensive picture of their knowledge of phrasal verbs. Unfortunately, when this study was carried out, the author had no access to any academic writing by the students. Finally, a comparative analysis of the phrasal verbs and their single-word counterparts in the corpora could be carried out. Such an approach would provide more solid evidence for the counter-avoidance observation. However, such a task might not be easy to undertake, as not every phrasal verb has a single-word equivalent and vice versa. 8. Conclusion and implications The Chinese EFL learners in this study were capable of producing a sufficient number of phrasal verbs in writing despite the repeated assumptions that phrasal verbs are more difficult for EFL learners whose L1 does not possess this type of construction. Clearly, the L1 factor is not the most influential parameter in phrasal verb acquisition; other factors may be of more importance. Phrasal verbs in the learner and the British corpora are evenly distributed across writers. There is no considerable individual difference in terms of phrasal verb production among the Chinese learners. However, the uneven distribution across writers in the American corpora suggests that factors such as individual differences and text length are of great importance in learner corpus analysis and may distort the final results. The positive association between phrasal verb frequency in the learner corpus and expert corpora shows that high frequency of occurrences does lead to the learning and eventual production of phrasal verbs by EFL learners. This finding can be beneficial for English teachers as phrasal verb frequency can serve as a criterion to determine which phrasal verbs should be taught and in what order. © 2013. John Benjamins Publishing Company All rights reserved


The present study has illustrated how the choice of native corpora would affect the overall results in learner corpus studies. Differences within the native novice corpora were identified across English varieties and genres. The finding shows that the L1 variety factor is just as important as other factors and should not be ignored in the selection of native comparison corpora in learner corpus studies. Despite the differences, novice writers of both English varieties tend to use fewer phrasal verbs in academic writing than in argumentative writing. The consistency observed in the native corpora (e.g. the variation of phrasal verb use in different genres) is of significant pedagogical value in the teaching and learning of phrasal verbs, showing that stylistic features of phrasal verbs are an indispensable part of phrasal verb knowledge.

Notes * The author would like to thank Professor John Flowerdew, Professor Peter Kelly, editor Victoria Hasko and the two anonymous reviewers for their invaluable comments and suggestions. She would also like to acknowledge the editorial help of Lorenzo Mastropierro. 1. See Schmitt & Redwood (2011) and White (2012) for detailed discussion of the intralinguistic factors that contribute to the difficulty of phrasal verbs. 2. English majors at Chinese universities are required to take another national proficiency exam, i.e. Test for English Majors-8 (TEM-8) in the second semester of their final year, but the results of this exam do not affect the completion of their programme. 3. I would also like to express my sincere gratitude to John Milton for kindly letting me use his corpus. 4. The online free CLAWS POS tagger is available at http://ucrel.lancs.ac.uk/claws/ (accessed January 2012). 5. The formula to calculate the correlation coefficient used in the present study is as follows:

where n is the total number of samples, xi (x1, x2, ... , xn) are the x values and yi are the y values.


438 Meilin Chen

References Ädel, A. 2008. “Metadiscourse across three varieties of English: American, British, and advanced learner English”. In U. Connor, E. Nagelhout & W. Rozycki (Eds.), Contrastive Rhetoric. Amsterdam: John Benjamins, 45–62. Algeo, J. 2006. British or American English? A Handbook of Word and Grammar Patterns. Cambridge: Cambridge University Press. Bauer, L. 1988. An Introduction to International Varieties of English. Edinburgh: Edinburgh University Press. Biber, D., Johansson, S., Leech, G., Conrad, S. & Finegan, E. 1999. The Longman Grammar of Spoken and Written English. London: Longman. The British National Corpus, version 3 (BNCweb CQP-Edition). 2007. Distributed by Oxford University Computing Services on behalf of the BNC Consortium. Available at: http:// www.natcorp.ox.ac.uk/ (accessed June 2013). COBUILD Dictionary of Phrasal Verbs (2nd Edition). 2002. Sinclair, J., P. Hanks, G. Fox, R. Moon & P. Stock (Eds.). Edinburgh: HarperCollins. Connor, U. M. & Moreno, A. I. 2005. “Tertium Comparationis: A vital component in contrastive research methodology”. In P. Bruthiaux, D. Atkinson, W. G. Eggington, W. Grabe & V. Ramanathan (Eds.), Directions in Applied Linguistics: Essays in Honor of Robert B. Kaplan. Clevedon: Multilingual Matters, 153–164. Dagut, M. & Laufer, B. 1985. “Avoidance of phrasal verbs: A case for contrastive analysis”. Studies in Second Language Acquisition, 7 (1), 73–79. Gardner, D. & Davies, M. 2007. “Pointing out frequent phrasal verbs: A corpus-based analysis”. TESOL Quarterly, 41 (2), 339–360. Gardner, S. & Nesi, H. 2012. “A classification of genre families in university student writing”. Applied Linguistics, 34 (1) 1–29. Gilquin, G. 2011: online. “Corpus linguistics to bridge the gap between World Englishes and Learner Englishes”. 12th International Symposium on Social Communication, Santiago de Cuba, January, 2011. Available at: http://hdl.handle.net/2078.1/112509 (accessed August 2012). Gilquin, G., De Cock, S. & Granger, S. (Eds.) 2010. The Louvain International Database of Spoken English Interlanguage. Handbook and CD-ROM. Louvain-la-Neuve: Presses Universitaires de Louvain. Granger, S., Dagneaux, E., Meunier, F. & Paquot, M. (Eds.) 2009. The International Corpus of Learner English: Handbook and CD-ROM V2. Louvain-la-Neuve: Presses Universitaires de Louvain. Gries, T. S. 2008. “Dispersions and adjusted frequencies in corpora”. International Journal of Corpus Linguistics, 13 (4), 403–437. Gries, T. S. 2009. “Dispersions and adjusted frequencies in corpora: Further explorations”. In T. S. Gries, S. Wulff & M. Davies (Eds.), Corpus Linguistic Applications: Current Studies, New Directions. Amsterdam: Rodopi, 197–212. Gries, T. S. 2012. “Correction to ‘Dispersions and adjusted frequencies in corpora’”. International Journal of Corpus Linguistics, 17 (1), 147–149. Hulstijn, J. H. & Marchena, E. 1989. “Avoidance: Grammatical or semantic causes?” Studies in Second Language Acquisition, 11, 241–255. James, C. 1980. Contrastive Analysis. London: Longman.



Jarvella, R. J., Bang, E., Jakobsen, A. L. & Mees, I. M. 2001. “Of mouths and men: Non-native listeners’ identiﬁcation and evaluation of varieties of English”. International Journal of Applied Linguistics, 11 (1), 37–56. Kaszubski, P. 1998. “Learner corpora: The cross-roads of linguistic norm”. In C. Stephens (Ed.), TALC98 Proceedings. Oxford: Humanities Computing Unit, Oxford University, 24–27. Krzeszowski, T. P. 1984. “Tertium comparationis”. In J. Fisiak (Ed.), Contrastive Linguistics: Prospects and Problems. The Hague: Mouton, 301–312. Krzeszowski, T. P. 1990. Contrasting Languages: The Scope of Contrastive Linguistics. Berlin: Mouton. Kucěra, H. & Francis, W. 1967. Computational Analysis of Present-day English. Providence: Brown University Press. Laufer, B. & Eliasson, S. 1993. “What causes avoidance in L2 learning: L1-L2 difference, L1-L2 similarity, or L2 complexity?” Studies in Second Language Acquisition, 13 (1), 35–48. Leech, G. 2011. “Frequency, corpora and language learning.” In F. Meunier, S. De Cock, G. Gilquin & M. Paquot (Eds.), A Taste for Corpora. In Honour of Sylviane Granger. Amsterdam: John Benjamins, 7–32. Liao, Y. & Fukuya, J. Y. 2004. “Avoidance of phrasal verbs: The case of Chinese learners of English”. Language Learning, 54 (2), 193–226. Liu, D. 2011. “The most-frequently used English phrasal verbs in American and British English: A multi-corpus examination”. TESOL Quarterly, 45 (4), 661–688. Longman Dictionary of Phrasal Verbs. 1983. R. Courtney (Ed.). London: Longman. Michigan Corpus of Upper-level Student Papers. 2009. Ann Arbor, MI: The Regents of the University of Michigan. Available at: http://micusp.elicorpora.info/ (accessed June 2013). Milton, J. 2001. Research Reports V2: Elements of a Written Interlanguage: A Computational and Corpus-based Study of Institutional Influences on the Acquisition of English by Hong Kong Chinese Students. Hong Kong: Hong Kong University of Science & Technology Press. Oxford Phrasal Verbs Dictionary for Learners of English. 2001. D. Parkinson (Ed.). Oxford: Oxford University Press. Potts, A. & Baker, P. 2012. “Does semantic tagging identify cultural change in British and American English?” International Journal of Corpus Linguistics, 17 (3), 295–324. Quirk, R., Greenbaum, S., Leech, G. & Svartvik, J. 1985. A Comprehensive Grammar of the English Language. Harlow: Longman. Rohdenburg, G. & Schlüter, J. (Eds.) 2009. One Language, Two Grammars? Differences between British and American English. Cambridge: Cambridge University Press. Scott, M. 2008. WordSmith Tools version 5. Liverpool: Lexical Analysis Software. Schmitt, N. & Redwood, S. 2011. “Learner knowledge of phrasal verbs: A corpus-informed study”. In F. Meunier, S. De Cock, G. Gilquin & M. Paquot (Eds.), A Taste for Corpora. A Tribute to Professor Sylviane Granger. Amsterdam: John Benjamins, 173–208. Siyanova, A. & Schmitt, N. 2007. “Native and nonnative use of multi-word vs. one-word verbs”. International Review of Applied Linguistics, 452, 119–139. Swales, J. & Feak, C. 2004. Academic Writing for Graduate Students: Essential Tasks and Skills (2nd Ed). Ann Arbor: University of Michigan Press. Thewissen, J. 2013. “Capturing L2 accuracy developmental patterns: Insights from an errortagged EFL learner corpus”. Modern Language Journal, 97 (Suppl. 1), 77–101. Virtanen, T. & Lindgrén, S-A. 1998. “British or American English? Investigating what EFL students say and what they do”. In H. Lindquist, S. Klintborg, M. Levin & M. Estling (Eds.), The Major Varieties of English. Papers from MAVEN 97. Vaxjo: Acta Wexionensia, 273–282.


440 Meilin Chen

Waibel B. 2007. Phrasal Verbs in Learner English: A Corpus-based Study of German and Italian Students. PhD thesis. Freiburg: Albert-Ludwigs-Universität Freiburg. White, B. J. 2012. “A conceptual approach to the instruction of phrasal verbs”. The Modern Language Journal, 96 (3), 419–438. Xiao, R. 2008. “Well-known and influential corpora”. In A. Lüdeling & M. Kyto (Eds.), Corpus Linguistics: An International Handbook Volume 1. Berlin: Walter de Gruyter, 383–457. Zhang, W. & Hu, G. 2008. “Second language learners’ attitudes towards English varieties”. Language Awareness, 17 (4), 342–347.

Appendix 1. Examples of essay topics in BAWE chosen for the present study 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

All varieties of English are equal. Discuss. Are all forms of art part of the ‘culture industry’? Can sociologists make use of official statistics or should they seek to generate their own? Chimpanzees can use signs, but do they have language? Consider the view that gender should be included as one of the core analytical categories in social anthropology. Critically evaluate the assertion that sporting practices and their associated subcultures bear no relation whatsoever to the social construction of masculine ideals. Critically explore the relationship between ethnicity and mental illness. Did the Death Camps introduce a new principle into modern political life? Discuss and evaluate, with the use of examples, the extent to which social control has become a prime function of modern medicine. Discuss the importance of working-class pressure in the development of state policies to overcome mass unemployment in the 1930s. European women both challenged and supported the Empire-building project. Discuss. Every morning I see a Neanderthal in my shaving mirror (Milford Wolpoff). How closely, if at all, do you think Neanderthals are related to ourselves? Explore the limitations and uses of official statistics, using the internal and radical approaches, and conclude should sociologists use official statistics? Gender identities assignment. Discuss. How would you describe the relationship between science and society in a democracy: compatible or oppositional? Is motherhood a woman’s destiny or the source of her subordination? Is ‘race’ just another term for ethnicity? Discuss. Is the social division between man as ‘Breadwinner’ and woman as ‘Homemaker’ a product of modern society? Private/independent and state schools are completely different (and separate) entities. Critically discuss. Religion is sometimes a difficult concept to apply to beliefs and practices. Discuss. The social construction of black masculinity necessarily manifests contradictory experiences. Critically discuss. There are no parallels in life to the concentration camp. Its horror can never be fully embraced by the imagination for the very reason it stands outside of life and death. Discuss. To what extent is our view of child-directed speech and its effects culturally biased?



What accounts for the differences in offending between girls and boys? What are the consequences of ‘eating badly’ and whose responsibility is it if people do? What arguments do you think may be given for or against democracy? What considerations frame debates around delinquency among girls? What groups and individuals are most at risk from unemployment and what explains their vulnerability? 29. Who are the ‘winners’ and the ‘losers’ in the modern, global food system? 30. Write an essay using the ethnography or other sources about a particular people, to show how ideas about dirt and cleanliness, pollution and/or taboo may be related to systems of classification. 24. 25. 26. 27. 28.

Appendix 2. The 50 most frequent phrasal verbs in the BNC and the COCA and their academic sub-corpora Rank of frequency

COCA

BNC

COCA-AC

BNC-AC

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

go on pick up come back come up go back find out come out go out point out grow up set up turn out get out come in take on give up make up end up get back look up figure out sit down

go on set up pick up go back come back go out point out find out come up make up take over come out come on come in go down work out set out take up get back sit down turn out take on

point out carry out go on take on make up set up turn out bring about give up pick up grow up take up find out end up go back come up break down set out open up take over work out figure out

make up take up set out carry out point out bring out go on set up give in take on take over bring about turn out find out take in break down come in give up go back go through pick up go out

23 24 25 26 27

get up take out come on go down show up

give up carry out get up look up carry on

come back bring in come out build up come in

put in bring in come out come about bring up


442 Meilin Chen

Rank of frequency

COCA

BNC

COCA-AC

BNC-AC

28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

take off work out stand up come down go ahead go up look back wake up carry out take over hold up pull out turn a/round take up look down put up bring back bring up look out bring in open up check out move on

go up get out take out come down put down put up turn up get on bring up bring in look back look down bring back break down take off go off bring about go in set off put out look out take back hold up

go out sum up bring back stand out show up carry on rule out catch up lay out take out play out bring up look back reach out cut off keep up hold up clean up bring out follow up get out move on come down

move in move on get out hold on get up go in take out look back get on break up come back make out pick out put up bring back hold out go up break out go down come down move out carry on look up

Author’s address Meilin Chen Department of English City University of Hong Kong Tat Chee Avenue, Kowloon Hong Kong SAR [email protected]
