Light Verb Constructions (LVCs; Grimshaw and Mester, 1988; Rosen, 1989; Butt, 1995; etc.) are basically constituted by a semantically bleached verb (e.g., make, have, get, give) and the action nominal complement (e.g., answer, advice, help), most typicssally denoting a motion event or state as in (1a). Its meaning is equivalent to the counterpart synthetic verb derived from the nominal complement as in (1b). In addition, LVCs are analogous to the ditransitive clause in syntactic form as in (1c).

This construction exhibits a cross-linguistic feature, such as Mandarin Chinese (Huang et al., 2014), Japanese (Grimshaw and Mester, 1988), Indo-European languages (e.g., Butt, 1995; Golshaie, 2016; Sundquist, 2018), etc. The common features mainly include: (1) verbo-nominal combination; (2) semantically semi-compositional; (3) ‘light’ verbs; and (4) ‘heavy’ nominal complements.

Yet cross-linguistic studies suggest that the licensing conditions of LVCs vary from language to language. Nagy et al. (2019) focus on the automatic detection of LVCs in four languages, namely, English, German, Spanish and Hungarian; they generalize both common and specific linguistic features from a typological perspective distinguished by five categories, namely, statistical, lexical, morphological, syntactic, and orthographic. Furthermore, the difference can be detected across regional varieties of English. Ronan and Schneider (2015) carried out an automated parser-based study to detect LVCs by inter-variety comparison in the Great Britain component and Ireland component in International Corpora of English. In their study, light verbs with higher frequency are more often used in British English, whereas light verbs with lower frequency tend to be more actively used in Irish English (Ronan and Schneider, 2015). Similar cross-regional investigations of Asia English versions have been carried out by Hoffmann et al. (2011) and Mehl (2017), but show different results in onomasiological preferences. These previous studies illustrate that exploring the common and specific features of LVCs across languages or varieties is weighted toward LVC detection in machine translation and L2 acquisition.

Similar to English, Chinese possesses a robust distribution of LVCs (Huang, 2009), especially in registers, such as official public speeches, legislation and science and technology texts (Wang and Zhang, 2014). As two typologically different languages, English and Chinese have fundamental differences that are embodied in cross-linguistic features of LVC usage, such as the presence/absence of inflectional markers, flexibility of modifications, position of PPs, fixedness of the combinations, etc. These discrepancies increase the difficulties in acquisition for Chinese EFL learners, especially evident in training their interpreting skills, which require the intellectual capacity to instantly transform idioms, colloquialisms and collocations into the equivalent information in the target language. Available Chinese–English comparative studies in LVCs are mainly carried out by Chinese researchers in the field of comparative linguistics, such as semantic properties (Chou, 2019), formal features applied in NLP (Wang and Zhang, 2014; Bai and Xue, 2015); syntactic formation in relation to argument structures (Zhu, 2019), etc. These studies compare five properties of LVCs from different aspects in the two linguistic systems. However, despite their prevalence in both languages, LVCs have not received much attention as formulaic sequences in EFL learners’ translations. The fixity of LVCs is presented as a gradient ranging from rigid to free, depending on the syntactic variability and the degree of lexical opacity. Most of the LVCs are not fixed enough to qualify as idioms, but the combinations of the components and modification used are constrained. The semantic complexities of light verbs show “greater cross-linguistic variability than nominal one” (Foley, 2010, p. 84). Such variability across the two languages might be the main cause of difficulties faced by Chinese EFL learners, which attracts interest in the inappropriate usage of LVCs and motivates explorations of corresponding strategies. Light verbs and other high frequency verbs are thought to be a significant barrier for EFL learners because of their limited knowledge (Altenberg and Granger, 2001). However, such knowledge of LVCs is limited in terms of Chinese–English translation strategies in the context of consecutive interpreting performance. Opting for appropriate strategies of the interpretation during learning can save processing and production efforts when retrieving the translation equivalents (McDonald and Carpenter, 1981; Gile, 1995), thus effortlessly bringing about better performance. In addition, locating LVC segments in the production of interpreting text-based sources may truly reflect LVCs usage in a linguistic context. Considering the gaps in the literature and necessity of the research in L2 language acquisition, this study aims to investigate effective translation strategies fit for Chinese EFL learners based on a Chinese–English consecutive interpreting (CECI) test. The study addresses the following three questions:

The answers to these questions can help Chinese EFL learners better understand LVCs in both languages and improve their performances in interpreting.

In this study, the literature review primarily concentrates on comparative studies on LVCs between Chinese and English because translation strategies of LVCs between these two languages have not been addressed in previous studies. This review can serve as the foundation for the discussion of translation strategies by identifying the typical syntacto-semantic features of LVCs in the two target languages.

Studies on English light verbs can be traced back to nearly a century ago. Poutsma (1929) described complex predicate constructions as ‘group verbs’, including but not limited to light verb constructions, which since then has begun to be noticed for its syntacto-semantic idiosyncrasy. The term ‘light verb’ was first coined by Jespersen (1954, p. 117) to denote semantically low-content verbs, and ‘light verb construction’ has since become a commonly accepted term for a bipartite complex predicate (e.g., Grimshaw and Mester, 1988; Rosen, 1989; Butt, 1995, 2003, 2010). Major disputes rest on whether the verb must be pertinent to an isomorphic (zero-derived) form (e.g., drink in to have a drink), a derivative (e.g., decide (v.)—decision (n.) in to make a decision), or a verbal noun (e.g., effort in to make an effort; e.g., Wierzbicka, 1982; Quirk et al., 1985; Algeo, 1995; Allerton, 2002; Dixon, 2005), and whether the direct object in the construction must be analysed as verbs or as nouns (e.g., Hoffmann et al., 2011).

Studies on Chinese LVCs occurred much later and has reached no consensus on terminology and classification of the light verb. Some use ‘formal verb’ to highlight its purely formal functions, as the meaning of the light verb is impoverished with no semantic contribution to the clause (e.g., Lv, 1980; Fan, 1981). Others (e.g., Yuan and Xia, 1984; Zhu, 1985) proposed ‘delexical verb’ or ‘quasi-predicate verb’, holding that the nominal complement collocating with a light verb must be a two-syllable verb-derived noun or modifier-head combination. Recent Chinese LVC studies mainly center on its syntactic representation and semantic features within Chomskyian generative linguistic sphere (Wen and Cheng, 2007; Zhang, 2013).

A limited number of Chinese–English comparative studies by Chinese researchers concentrate on delineating formal features of LVCs between the two languages (Wang and Zhang, 2014; Bai and Xue, 2015; Zhu, 2019). Under the guidance of Hierarchical Network of Concepts Theory (Huang, 1998), Wang and Zhang (2014) targeted Chinese–English machine translation of LVCs, taking the typical Chinese light verb jìnxíng ‘make’ type as an example. Three linguistic factors have been found to modulate the choice of syntactical structure in Chinese–English translation as follows: the semantic category of the light verb’s nominal complement, the syntactical form of LVCs in the target language, and the function of LVCs in the clause where they are located. Three translating rules have been put forward, but their effectiveness and applicability need to further testing. Bai and Xue (2015) attempted to distinguish predicative verbs with vague meaning from the true light verbs by analyzing syntactic and semantic features and argument assignments of the latter. This approach is analogous to Kearns (2002), in which the true light verb (e.g., give a groan) and the vague action verb (e.g., give a demonstration) are differentiated by testing their passivisation, WH-movement, and pronominalisation, etc. These two studies imply that both languages are involved in the issue of delimitation of the light verb, and syntactic and semantic functions of the nominal complement are the key determining factors to distinguish these two kinds of verbs.

The issue of delimitation of the light verb is aligned with grouping their different types by means of shared semantic attributes. Bai and Xue (2015, p. 11) rank the degree of ‘lightness’ of five major types of Chinese predicative verbs as shown in (2) (from light to heavy), and thus must be treated differently when annotated.

In the same vein, Feng (2016, p. 141) attempted to sequence five types of Chinese light verbs in terms of the degree of grammaticalization as in (3).

Though the two orderings partly overlap, they differ in the research materials, naming of the light verb groups, testing methods, and semantic classification of major types; consequently, the results are different. Similarly, Zhu (2019, p. 155–156) classifies light verbs into four groups, namely, DO, CAUSE, CONSIDER, PREP (a provisional term, indicating a type of covert light verbs that can be used with prepositional phrases), but without sequencing likewise. Furthermore, the naming method that mixes syntactic and semantic attributes obscures classification.

Notably, Feng’s (2016) finding mainly comes from his substantial diachronic analysis of old Chinese (OC) compared with modern Chinese (MC). Interestingly, compared with LVCs in MC, those in OC are more similar to those in English (Feng, 2016, p. 114). For example, the light verbs in both OC and English are silent (without phonetic realization), which trigger syntactic shifts and give rise to semantic changes via denominalisation or causativisation, as shown in (4–5) (Huang, 2009, p. 2–3):

However, in MC, the silent light verb becomes overt and fills the hypothesized position originally occupied with a silent category (Huang, 2009, p. 2). Thus, the examples in (4–5) can be instantiated by a light verb dǎ ‘hit’ in MC as in (6).

As the two languages evolve, their typological differences emerge. For example, the analytic/synthetic account reveals that English has a large number of bound morphemes to denote the word property. Several action nominal complements in English LVCs are derived from verbs with bound morphemes, such as contribute (v.) converted into (make a) contribution (n.), or investigate (v.) into (give an) investigation (n.), while Chinese uses more free morphemes and has no inflectional markers to distinguish the parts of speech. As such, the nominal complements in Chinese LVCs are often in isomorphic (zero-derived) form. For example, diàochá (n.) ‘investigation’ in (2a) and diàochá (v.) ‘investigate’ in (7b) are identical in form.

In addition, paratactic/hypotactic difference indicates that the relationship between the components or clauses is loose and flexible in the Chinese language but conforms to strict order with connectives in English. In Example (8), the theme argument introduced by the Chinese prepositional case marker duì ‘to’ can be placed either before (8a) or after (8b) the subject. However, the syntactic structure in (8b) is not acceptable in English. Such difference can be explained by topic−/subject-dominant account, that topic plays a fundamental role in Chinese clausal constructions. In addition to the canonical syntactic structure SVO like that in English, SOV and OSV are also prevalent in Chinese.

Such flexibility is similarly exemplified by ‘separable LVCs’, i.e., a verbo-nominal combination is at times embedded with other attributive components, showing greater syntactical flexibility.

In (9), quánmiàn ‘comprehensive’ separates the light verb zuò ‘make’ and its nominal complement zǒngjié ‘summary’. This term can also be substituted by more modifiers, either simple adjectives or complex attributive clauses. However, such property varies in degree of flexibility and accessibility among different types of light verbs. Kuo and Ting (2007) divided the light verbs into two types, namely, MAKE group (e.g., jìnxíng ‘proceed’, zuò ‘do’) and GIVE group (e.g., jiāyǐ ‘give’, yùyǐ ‘give’, gěiyǔ ‘give’). It is claimed that only the insertion of modification by the MAKE group is allowed, and not with the GIVE group. In general, the common type of modification– such as articles, quantifiers, possessors, or adjectives–is placed before the nominal complement in Chinese LVCs. By contrast, Claridge (2000) dichotomizes English LVCs into LV + NP pairings and LV + NP + PP, such as to run the risk of. Unlike English, Chinese has no post modifier PP like that in English. The similar component in Chinese LVC sentences often functions as a theme argument projected by the deverbal noun, introduced with or without a preposition. For example,

In (10a), the PP duì zhège ànjiàn ‘to this case’ comes before the complex predicate. By comparison, in (10b), the theme argument zhège ànjiàn ‘this case’, as a pre-nominal modifier, is placed directly before the nominal complement diàochá ‘investigation’.

While Chinese LVCs allow more diversified modifiers, the article usage, modification, and pluralization in English LVCs tend to be more fossilized (Brinton, 2008). Such phenomena can be traced back to the Middle English period, during which adjectival modifications were confined to a small range of adjectives (Matsumoto, 1999, p. 83). For several combinations, no modification has been identified, i.e., LVCs in this situation are lexicalized into fixed expressions, such as take effect, lose sight of, give rise to (Claridge, 2000, p. 157–158). However, Claridge (2000, p. 158) also indicated that, though rather rare, modification is found in well-established units such as take full place, or find so much fault, because the noun is salient and independent enough to be modified.

In addition, syntactic operation such as passivisation can render the light verb and its nominal complement separable and inverted, and is sometimes treated as a testing method to distinguish true light verb from vague action verb as mentioned earlier (Saito and Hoshi, 2000; Kearns, 2002; Kuo and Ting, 2007). Notably, the light verb and its nominal complement in a true light verb construction cannot be passivized. Such property is observed in both Chinese and English.

Another issue relates to the semantic difference between an LVC and its simplex predicate verb. It is generally agreed that LVCs are ‘semantically more lightweight than the same word would have been in a normal context’. (Allerton, 2002, p. 172) It is the same case in Chinese as is shown in example (11).

In (11a), Quánmiàn-de ‘comprehensive’ means that an overall summing-up has been completed while quánmiàn ‘comprehensively’ in (11b) denotes one typical property of the action ‘summarize’. For the record, not all nominal complements can be converted into the counterpart synthetic verb. For example, the change of effort into to make an effort cannot be used in the form *to effort. In this case, make is a light verb and effort denotes an abstract event. Chinese also has a group of specialized event nouns that cannot be converted into the counterpart synthetic verbs, such as zhànzhēng ‘warfare’, yíshì ‘ceremony’, shoǔshù ‘surgery’ (Lu, 2012).

In summary, from a broader perspective, the canonical order of LVCs basically shared by Chinese and English is ‘the light verb + nominal complement’, with or without modifiers such as articles, quantifiers, possessors, adjectives inserted in between. Besides, both languages are characterized by the categorization of true light verbs and vague action verbs, classification of event nouns, semantic differentiation between an LVC and its simplex predicate verb. The differences are reflected mainly in the presence or absence of inflectional markers, flexibility of modifications, position of PPs, fixedness of the combination, etc. These differences pose difficulties in the comprehension of LVCs and lead to various problems in conversion across languages. The present investigation attempts to be carried out with this line of research to delve into translation strategies of LVCs in the CECI test.

In response to the three research questions, three sets of data are reported: (1) distribution of translation strategies (including the proportion of frequency and entropy values) used by the participants for the 12 LVC segments; (2) the appropriate rates of the 12 LVC segments and the nine translation strategies, as well as their interrelation with the semantic attributes of light verbs; and (3) the correlation between vocabulary knowledge of LVCs and the interpreting score of the 12 target LVC segments.

In general, literal translation type (A) takes up the highest proportion in selection (54.92%), followed by Types C (10.61%), B (6.44%) and G (6.31%; see Table 2). In terms of entropy value by LVC segments (Figure 1), the scores of L8, 9, 10, 12 are relatively low, which implies that when they are translated into English, the translation strategies used by the participants are relatively consistent and mostly centralized in Type A. These LVC segments mainly pertain to BE and BECOME grouped by semantic attributes of the light verb. By contrast, L1 (DO group) shows the highest entropy (H = 0.643) indicating high variability in strategy selection distributed in almost all strategies except Types E and F. For entropy value by strategies (Figure 2), the scores of Types H, I and E are relatively low, implying that those strategies are applied to a limited number of LVC segments. By contrast, Type A shows the highest entropy (H = 1.02) indicating its wide application in various LVC segments.

The appropriate rates of the 12 LVC segments (Figure 3) and of the nine translation strategies (Figure 4) are unevenly distributed. A relatively higher rate is achieved in L3, 4, 5, 7, 10 and in Types A, B, C, respectively, (given that only one case is translated as Type H, the result can be ignored).

To obtain a closer look at the appropriate rate by semantic attribute groups, we compute the grand means of appropriate rate using the three semantic attribute groups, as shown in Figure 5. The BECOME group has the highest rate, and the DO group has the lowest.

We also compute the grand means of entropy value by semantic attribute groups in a similar manner, and the results are given in the reverse order than those of the appropriate rates (Mean_DO = 0.643; Mean_BE = 0.497; Mean_BECOME = 0.366). These results indicate that as the light verb becomes more abstract, the translation appropriate rate decreases and the variability increases (or consistency decreases) in strategy selection. In summary, the degree of lightness of the light verbs exerts a reverse effect on the appropriate rate and consistency of strategy selection.

Furthermore, a correlation test is carried out between the English vocabulary knowledge related to the 12 target LVCs and the appropriate rates of the 12 target LVC segments in the interpreting test. Given that both of the testing variables are normally distributed, and the assumption of linearity is not markedly violated, Pearson correlations are computed to examine the two variables. The result shows a significant correlation (r = 0.272, p < 0.05). According to Cohen (1988), the vocabulary knowledge and the appropriate rates of the 12 target LVC segments have a positive correlation, which is considered as a close medium effect size (r = 0.3). This finding means that students with a good command of target topic-related lexical knowledge are likely to have high LVC scores in the interpreting test.

In this section, under the background of typological differences and syntactic and semantic properties of the construction, we mainly discuss the achieved results from three motivating factors proposed by the research questions: a. preferences of strategy selection; b. structural patterns in LVC translation; and c. relations of lexical knowledge and appropriate use of LVCs. Thereupon, the felicitous conditions of the nine translation strategies are summarized along with related pedagogical implications.

This study explores the appropriateness and variability of translation strategies in dealing with the 12 target LVC segments by L2 learners to capture effective translation strategies fit for Chinese learners of English in this regard. The results show that the frequency of use and the appropriate rates of nine types of translation strategies are distributed unevenly; as such, preference of strategy selection conforms to the economy principle, i.e., saving the processing effort to allocate more attention to wherever necessary. Moreover, the consistency between the appropriate rates and of strategy selection grouped by semantic attributes and the order of the degree of grammaticalization implies that the lightness of the light verb affects the appropriateness and strategy selection of translation, i.e., the degree of lightness of the light verbs exerts a reverse effect on the appropriate rates and consistency of strategy selection. Meanwhile, the positive correlation between the after-test questionnaire and interpreting score shows that a good command of target topic-related lexical knowledge helps to improve the interpreting performance. The teacher needs to determine whether the L2 learners have or have not appropriately acquired the target expressions to determine whether it is the strategy itself or unfamiliarity with the target expression that leads to failure in bilingual translation. The felicitous conditions of the nine translation strategies are thus delineated. The findings support that cultivating the awareness of formulaicity and acquiring translation strategies help L2 learners develop a set of matching system across languages, and improve fluency and accuracy of C-E interpreting.

As an attempt to explore the interpreting strategies of LVC, the current study only offers data observed from Chinese to English translation, and not the other way around. A full picture can be obtained if bidirectional interpreting tasks can be designed. We hope that future studies will further probe into the issue by providing more comprehensive experimental support.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Xiamen University Tan Kah Kee College. The patients/participants provided their written informed consent to participate in this study.

DW: the conception and design of the work, the acquisition, analysis, and interpretation of data for the work, drafting the original version. GJ: the conception and design of the work, revising the draft critically for important intellectual content. YZ: data collection, revising the draft critically. All authors contributed to the article and approved the submitted version.

This work was supported by the Humanities and Social Science General Program sponsored by the Ministry of Education of the People’s Republic of China (grant number 18YJA740047).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.