Introduction

The NP de VP construction in modern Chinese has long been investigated in academia. It has been discussed no less frequently than other Chinese expressions such as taishang zuo zhe zhuxituan ‘on the platform sits the presidium’ and wangmian qisui si le fuqin ‘Wang Mian’s father died when he was seven’. The employment of the concept “construction” in this research underlies Goldberg’s definition that construction is “a form-meaning pair that some aspect of the form or the meaning is not strictly predictable from the construction’s component parts” (Goldberg, 1995, p. 4); and constructions themselves are meaningful (Goldberg, 1995, 2006). The NP de VP construction is a grammatical pattern in which the NP in the first slot represents a nominal phrase and the VP in the third slot represents a verbal phrase; the two types of phrases are threaded by the possessive particle de. This construction could be typically instantiated by mubiao de shixianFootnote 1 ‘realization of target’ in example (1), in which mubiao ‘target’ is a nominal phrase or NP and shixian ‘realize’ is a verbal phrase or VP. The construction itself is generally regarded as a nominal phrase as a whole (e.g., Guo, 2000; Lu, 2003; Jin, 2019; Yang and Xiong, 2021; Li, 2021, etc.).

(1)法学教育以人权为教育理念, 无疑有助于法治目标的实现。Footnote 2

faxue jiaoyu yi renquan wei jiaoyu linian, wuyi

law education on human rights be education idea undoubtedly

youzhuyu fazhi mubiao de shixian

helpful legal target de realization

‘It is undoubtedly helpful to realize the legal targets that legal education takes human rights as its educational idea.’

Previous studies have generally centered on discussions of word classes of the VP (e.g., Wang, 1954; Zhang, 1957; Li and Liu, 1960; Zhu et al., 1961; Zhu, 1982; Wu, 2006; Shen, 2009, 2015; Yuan, 2010; Zhou, 2012; Yang and Xiong, 2021, etc.), heads of the NP de VP construction (cf. Lu, 1985, 1993; Jin, 1987, 1992; Hu and Fan, 1994; Cheng, 1999; Li, 2002; Lu, 2003; Si, 2002, 2004; Chen, 2009; Shen, 2009; Wang and Yang, 2017), and semantic relationships between the NP and the VP (cf. Chen, 1987; Zhang, 1993; Wang, 2002, 2010; Wu and Guo, 2018; Fang and Liu, 2021). There is also research investigating the meaning patterns of the construction in terms of the VP slot (cf. Zhan, 1998; Wang, 2002) and the NP slot (cf. Shen and Wang, 2000). Although the aforementioned research have uncovered different aspects of the NP de VP construction, e.g., the grammatical category of the VP, the head of the NP de VP construction, the semantic relationship between the NP and the VP in the construction, and even the meaning patterns that could enter the NP slot and the VP slot, respectively, there is no study, according to my knowledge, that considers the meaning patterns of the VP and the NP in the NP de VP construction drawing on significantly attracted instances to this construction by means of statistic measures such as the statistical significance and/or the association strength based on the real data from a large corpus. It is still unknown what typical meanings of the NP slot and the VP slot denote respectively, leading further to the uncertainty of the meaning patterns of the NP de VP construction because construction itself is meaningful (cf. Goldberg, 1995, 2006). Consequently, this paper, drawing on the covarying collexeme analysis (cf. Stefanowitsch and Gries, 2005; Gries and Stefanowitsch, 2010; Hilpert, 2014) and the hierarchical cluster analysis (cf. Gries and Stefanowitsch, 2010; Desagulier, 2017; Zhou, 2021), aims at uncovering instances that are significantly attracted to the NP de VP construction so as to identify typical meanings of the construction and meaning patterns of the VP and the NP in this construction. Accordingly, the following research questions are proposed.

1)What are the typical meanings of the NP de VP construction?

2)What are the meaning patterns of the VP in the NP de VP construction?

3)What are the meaning patterns of the NP in the NP de VP construction?

Literature review

Elements in different slots of the NP de VP construction were examined by previous research, for example, lexical items in the VP slot, the semantic relationship of lexical items between the NP slot and the VP slot, and meaning patterns of both the VP slot and the NP slot of the construction. Therefore, this section briefly reviews these respects.

In relation to word classes of the VP in the NP de VP construction, there are generally two theoretical hypotheses. One hypothesis assumes that the VP in the NP de VP construction has been nominalized (cf. Lu, 1985; Chen, 1987; Ye, 2020 for nominalization from the word rank; and Guo, 2000; Xiong, 2001, 2006; Shi, 2008; Lu abd Pan, 2013 for nominalization from the phrase rank). The argument of nominalization of the VP is in accordance with the overall grammatical class of the construction which is universally accepted as a nominal phrase; it also sufficiently explains the linguistic phenomenon that the VP is grammatically pre-modified by the possessive particle de. However, the argument of nominalization with respect to the VP in the construction could not persuasively expound the grammaticality and acceptability of the expression mubiao de chichi bu shixian ‘delayed realization of target’, in which the VP shixian ‘realize’ still keeps its verbal syntactic feature of being modified by adverbials such as chichi ‘with delays’ and negative markers such as bu ‘not’. This leads to the other theoretical hypothesis that word classes of the VP in the NP de VP construction are still verbs (cf. Zhu et al., 1961; Zhu, 1982; Wu, 2006). Verbs pre-modified by adverbials are intrinsic natures in the language system, which sufficiently expounds the grammaticality of the expression mubiao de chichi bu shixian ‘delayed realization of target’. Both arguments are to some extent reasonable in that the former complies with the category agreement between the nominal nature of the VP and that of the construction, and the latter complies with the simplicity of language systems (Jin, 2019). The VP in the NP de VP construction is simply regarded as verbs in this research in that they are in line with the tagging system of the BCC corpus. As a matter of fact, treating the VP as nominalized items or verbs does not influence this research to achieve its purposes because we specifically highlight the meaning patterns of the VP and the NP in the NP de VP construction.

Chen (1987), Zhang (1993), Wang (2002, 2010), and Wu and Guo (2018) have examined the semantic relationship between the NP and the VP in the NP de VP construction. They argued that the NP could be either the agent or the patient of the VP. Concerning mubiao de shixian ‘realization of target’ in example (1), the NP mubiao ‘target’ functions as the patient of the VP shixian ‘realize’. There are also cases that the NP functions as the agent of the VP in the NP de VP construction; this could be exemplified by lingdao de tiba ‘promotion by leaders’ (Wang, 2002, p. 62), in which the NP lingdao ‘leaders’ functions as the agent of the VP tiba ‘promote’. Significantly attracted instances to the NP de VP construction demonstrate that most instances belong to the former kind while only a rather small proportion of these instances belong to the latter kind. Although researchers as such have uncovered the semantic relationship between the NP and the VP in the construction, they do not count on the typical meanings that these elements could denote or how these meanings could be patterned.

There is also research investigating the meaning patterns of the construction that could enter the VP slot (cf. Zhan, 1998; Wang, 2002) and the NP slot (cf. Shen and Wang, 2000). Zhan (1998) argued that the VP in Chinese could denote the senses of either strong “action” and weak “event” (e.g., pao ‘run’), or weak “action” and strong “event” (e.g., shixian ‘realize’ and zhaokai ‘convene’), or weak “action” and weak “event” (e.g., baokuo ‘include’); however, only lexical items that denote weak “action” and strong “event” are capable of entering the VP slot of the NP de VP construction. Zhan’s (1998) argument is further corroborated by Wang (2002) who summarized that the weaker the transitivity of a verb possesses, the more likely the verb enters the VP slot of the construction; in other words, the VP in NP de VP construction usually denotes a comparatively weak sense of “action”. However, Zhan’s (1998) and Wang’s (2002) conclusions underlie the examples that are not based on large corpora and thus need further testified by examples sourced from a large corpus such as BCC. Concerning the meaning patterns of the NP in the NP de VP construction, Shen and Wang (2000) argued that the NP in this construction must have a sense of prominence, which combines two factors, i.e., informativity that refers to “the prompting information” (Shen and Wang, 2000, p. 29) and accessibility that refers to “the difficulty of a relevant concept being retrieved from the memory or the environment” (Shen and Wang, 2000, p. 30). Precisely, the NP with high informativity and accessibility are extremely likely to enter the NP slot of the construction. Nevertheless, Shen and Wang’s (2000) argument is not frequency- and/or statistical significance-based, hence it also needs further testification in that their conclusion may underlie peripheral instances which do not represent typical meanings of these NPs.

Methodology

Corpus

This research retrieved data from the corpus of the Beijing Language and Culture University Corpus Center or BCC. Reasons for considering the BCC are three-fold. First, it is presently one of the largest online corpora free of charge, covering ~9.5 billion tokens from various disciplines. Second, it is part-of-speech or POS tagged, which caters to the purpose of this research very well to highlight occurrences of nouns and verbs in the NP slot and the VP slot of the NP de VP construction. Third, data in the BCC are taken from different genres, incorporating Newspapers (2 billion), Literature (3 billion), Micro-Blog and film (0.6 billion), Classical Chinese (2 billion), and Miscellaneous (1.9 billion).

Data collection

The search query “../n 的../v”, which reads as a construction in the sequence of a 2-character noun, a possessive particle de, and a 2-character verb, is implemented to retrieve sufficiently relevant hits of the construction at issue. The purpose of restricting instances of the NP slot and VP slot within 2-character lexical items is to preclude those lexical items with 1-character in that a large proportion of 1-character items retrieved are not typical cases of NP de VP construction by means of a pilot study, although some legitimate hits will be wrongly precluded from the analysis such as (这本)书的出版(zheben) shu de chuban ‘publication of this book’; in addition, those items with more than two characters will also be successfully retrieved. Another purpose is motivated by the fact that meaning patterns of the NP de VP construction considered in this research are based on association strengths between specific lexical items and constructions, which correlates significantly positively with the observed frequencies of these lexical items (cf. Desagulier, 2017; Zhou, 2021). Precisely, although attested hits of this construction with low frequencies in the BCC enter the computation of covarying collexeme analysis (cf. subsection “Approaches”), they will be regarded as repelled ones by this approach and the produced results will not influence the findings of this research because what is highlighted is the typical meaning pattern that the construction possibly denotes.

Approaches

Two approaches are employed to facilitate the fulfillment of the purposes in this study: the covarying collexeme analysis and the hierarchical cluster analysis. Drawing on the covarying collexeme analysis, we could statistically identify lexical items that are capable of entering the VP slot and the NP slot in the NP de VP construction, respectively; drawing on the hierarchical cluster analysis, we could objectively cluster these significantly attracted lexical items according to their shared semantic meanings. Therefore, their operationalizations and motivations for selection are further expounded.

Covarying collexeme analysis is one of the three members in collostructional analysis, the other two being the collexeme analysis (cf. Stefanowitsch and Gries, 2003; Hilpert, 2014; Desagulier, 2017; Zhou, 2021) and the distinctive collexeme analysis (cf. Gries and Stefanowitsch, 2004, 2005; Hilpert, 2014; Zhou, 2021). The covarying collexeme analysis identifies “the association strength between pairs of lexical items occurring in two different slots of the same construction” (Stefanowitsch and Gries, 2005, p. 9) or investigates which lexical items in one slot covary with those in another slot. Specifically, this determines which potential lexical items in slot 2 cooccur with each potential lexical item occurring in slot 1 significantly more often than expected or vice versa. Its operationalization could be illustrated by the NP de VP construction demonstrated in the contingency Table 1. Table 1 shows that, for the purpose of computing the association strength between the lexical item (L) in the slot of NP and that (M) in the slot of VP in the NP de VP construction, four numbers are required; that is, the frequency of cooccurrences of L in the slot of NP and M in the slot of VP (i.e., freq1), the frequency of cooccurrences of other lexical items rather than L in the slot of NP and M in the slot of VP (i.e., freq2), the frequency of cooccurrences of L in the slot of NP and other lexical items rather than M in the slot of VP (i.e., freq3), and the frequency of cooccurrences of other lexical items rather than L in the slot of NP and other lexical items rather than M in the slot of VP (i.e., freq4).

Table 1 Contingency table of the covarying collexeme analysis of the NP de VP construction.
Full size table

Computation of the association strength between the VP and the NP in the NP de VP construction is finished by performing the function coll.analysis scripted by Gries (2014) in R language. In this study, by drawing on the Fisher–Yates exact testFootnote 3, results of the association strength between the NP and the VP in NP de VP construction are transformed by means of negative base-10 logarithm of their p valuesFootnote 4. Values of association strengths that are larger than 1.301 indicate the statistical significance at the level of p < 0.05; 2 at the level of p < 0.01; 3 at the level of p < 0.001; and Inf indicates that the association strength is infinite and could not be measured by numbers.

Selection of covarying collexeme analysis is directly motivated by the fact that it is an approach that testifies the probability of mutual prediction between the NP and the VP in the NP de VP construction. By so doing, we could easily identify instances that are significantly attracted to the NP slot and the VP slot in the construction. Drawing on these significantly attracted instances that could enter both the NP and the VP slots in the NP de VP construction, it is possible for us to further pattern the lexical items that are similar in meanings by means of the hierarchical cluster analysis.

Hierarchical cluster analysis is a member of cluster analysisFootnote 5 which is an umbrella term for a number of related agglomerating analyses. It depicts a series of multifactorial methods to examine the structure in data for the purpose of identifying clusters of similar objects (cf. Everitt et al., 2011; Desagulier, 2014). It agglomerates individuals based on their distances which rely on the parameters that characterize them. This approach has been widely used in linguistics, particularly in corpus linguistics (cf. Divjak and Gries, 2006, 2008; Gries and Stefanowitsch, 2010; Divjak and Fieller, 2014; Desagulier, 2017). Hierarchical cluster analysis requires the data be a table T of I observations or individuals and K variables. Numbers of the dependent variable could be continuous (real numbers), discrete (integers), or both. It converts the table into a distance object by implementing an amalgamation rule (e.g., Ward’s method which, by employing an analysis of variance, evaluates the distances between clusters) which determines in what ways elements in the distance object could be clustered into groups. This method is performed by the function hclust or pvclust in R language.

Hierarchical cluster analysis is chosen in this study in that it could cluster lexical items that are significantly attracted to the VP slot and the NP slot in the NP de VP construction according to their meanings. In addition, the clustering of lexical items in both slots underlies one another’s covarying collexemes; precisely, the clustering of lexical items in the VP slot generally relies on those that are significantly attracted to the NP slot of the NP de VP construction, and the clustering of lexical items in the NP slot chiefly relies on those that are significantly attracted to the VP slot of the construction. This means of clustering, i.e., relying on covarying collexemes, is initially proposed by Gries and Stefanowitsch (2010) who argued that it is more objective and precise than previous clustering that relies only on a linear context or certain word classes within a specific window around the node word. In so doing, it will “shed light on the most typical senses of a construction as well as sub-senses instantiated by coherent semantic classes of words occurring in it” (Gries and Stefanowitsch, 2010, p. 73).

Covarying collexeme analysis of the NP de VP construction

Implementation of the covarying collexeme analysis produced the results that are presented in Table 2. Instances of the NP de VP construction are ranked according to their association strengths between lexical items and the construction, which are presented in the first column; column name ‘NP de VP’ profiles typical instances of this construction; column name ‘obs. freq.’ stands for observed frequencies of the construction in the corpus; column name ‘exp. freq.’ means the expected frequency that a certain instance of the construction should occur in the corpus; column name ‘relation’ demonstrates whether a certain instance is attracted or repelled to the NP de VP construction; column name ‘Coll.s’ shows values of the association strength.

Table 2 Results of covarying collexeme analysis of the NP de VP constructionFootnote

It should be noted that relying on the observed frequency is sometimes misleading because it does not consider the occurrence of a certain lexical item in the corpus. For instance, shigu de fasheng ‘occurrence of accidents’ ranks No. 14 in Table 2, while it should rank No. 2 if its observed occurrence (1893) is considered.

.
Full size table

Table 2 shows that there are 515 instances that are significantly attracted to the NP de VP construction, the 515th one being the shuiping de zengjia ‘increasement of levels’ (Coll.s = 1.71 > 1.303, p < 0.05). Among the 515 significantly attracted instances to the construction, there are 71 instances whose association strengths reach the infinite level (i.e., Coll.s = Inf). Those instances are typical cases that are capable of representing meanings of the NP de VP construction and thus are chiefly employed to analyze the meaning patterns of the VP slot and the NP slot in sections 5 and 6, respectively. Scrutinization of those typical cases demonstrates that typical meanings of the NP de VP construction briefly include the pairing of “regulations” in the NP slot and “implementation” in the VP slot, the pairing of “systems” in the NP slot and “establishment” in the VP slot, and the pairing of “results” in the NP slot and “achievement” in VP slot.

The pairing of “regulations” in the NP slot and “implementation” in the VP slot

The first typical meaning pattern of the NP de VP construction pairs the lexical items in the NP slot that denote the sense of “regulation” and those in the VP slot that denote the sense of “implementation”. The former is briefly realized by such lexical items as zhengce ‘policy’, zhidu ‘regulation’, gongneng ‘function’, and zuoyong ‘function’, and the latter is generally realized by verbs such as zhiding ‘enact’, guanche ‘implement’, lvxing ‘perform’, and xingshi ‘perform’. This pairing could be exemplified by the sentence in (2), in which zhengce ‘policy’ which expresses the sense of “regulations” significantly covaries with zhiding ‘enact’ which expresses the sense of “implementation”.

(2)公共行政就是公共政策的制定。

gonggongxingzhengjiushigonggongzhengcedezhiding

publicadministrationbepublicpolicydeenact

‘Public administration is the enactment of public policy.’

The pairing of “systems” in the NP slot and “establishment” in the VP slot

The second typical meaning pattern of the NP de VP construction pairs the sense of “systems” in the NP slot with the sense of “establishment” in the VP slot. Lexical items that realize the former sense include jigou ‘organization’, tizhi ‘regulation’, tixi ‘system’, and jizhi ‘mechanism’, and those that realize the latter include jianli ‘establish’, sheli ‘set up’, and kaishe ‘set up’ (cf. rows 4 and 5 in Table 2). This pairing is evidenced by the example (3), in which the typical meaning regarding the pairing of “systems” and “establishment” in both slots of the NP de VP construction is realized by the significant covarying between tizhi ‘regulation’ and jianli ‘establish’.

(3)这必然阻碍市场体制的建立。

zhebiranzhu’aishichangtizhidejianli

thismusthindermarketregulationdeestablish

‘This will definitely hinder the establishment of the market regulations.’

The pairing of “results” in the NP slot and “achievement” in the VP slot

The third typical meaning pattern of the construction under consideration pertains to the pairing of lexical items that denote the sense of “results” in the NP slot and those that denote the sense of “achievement” in the VP slot. The sense of “results” in the NP slot of the construction is briefly realized by such lexical items as mubiao ‘targets’, chengji ‘results’ and chengjiu ‘achievements’, and that of “achievement” in the VP slot is generally realized by items such as qude ‘achieve’, shixian ‘realize’ and wancheng ‘accomplish’ (cf. rows 1 and 13 in Table 2). Considering the instantiated clause in (1), shixian ‘realize’ in the VP slot significantly covaries with mubiao ‘targets’ in the NP slot of the NP de VP construction, and therefore the third typical meaning pattern is formulated.

A cursory survey of those top cases in Table 2 will simply demonstrate that lexical items in the VP slot by and large denote senses of achieving something (e.g., shixian ‘realize’, qude ‘achieve’, etc.), augmenting something (e.g., jiakuai ‘speed up’, jiada ‘expand’, and zengqiang ‘enhance’), etc. Lexical items in the NP slot generally denote senses of organization or system (e.g., tizhi ‘regulation’, tixi ‘system’, etc.), personal traits (e.g., zeren ‘responsibility’, zhenxin ‘sincerity’), etc. However, this way of semantically patterning the lexical items in both slots of the NP de VP construction will unavoidably impose researchers’ subjectivity into the meaning patterns; furthermore, the results are usually problematic and unsuccessful at length (cf. Gries and Stefanowitsch, 2010). The most appropriate way to avoid researchers’ subjectivity and to achieve objective and successful meaning patterns is to cluster these lexical items by implementing a hierarchical cluster analysis.

Meaning patterns of the VP in the NP de VP construction

Subsequent manual processing of those overlapping lexical items (e.g., jianli ‘establish’ cooccurs with both tizhi ‘regulation’ and tixi ‘systems’ in rows 4 and 5 of Table 2, respectively) yields 40 types of items in the VP slot. Taking their covarying collexemes in the NP slot of the construction into consideration, we finally formed a 40 (types of verbs in the VP slot) × 211 (types of covarying collexemes in the NP slot) contingency table. Inputting the.txt formatted contingency table and subsequently performing the function hclust in R language produced a cluster dendrogram as shown in Fig. 1. According to Fig. 1, verbs that fall into the scope of 9 clusters are extremely likely to be particularly preferred in the VP slot of the construction. Specifically, these clusters, except for clusters 1 and 9 could not capture a common sense, generally denote senses of “cognition”, “augmentation”, “implementation”, “achievement”, “establishment”, and “report”, which are to be detailed in the following paragraphs of this section.

Fig. 1
figure 1

Dendrogram of meaning patterns of the VP in the NP de VP construction.

Full size image

Meaning pattern of “cognition”

The first typical meaning pattern denoted by the typical senses of the VP in the NP de VP construction pertains to the sense of “cognition”, which is typically realized by mental verbs juede ‘think’ and xihuan ‘like’. They are clustered in that they generally cooccur with lexical items in the NP slot such as zhenxin ‘sincerity’ and neixin ‘heart’ that denote people’s internal traits. The finding that such mental verbs could enter the VP slot is also in accordance with Qi, et al.’s (2004, pp. 219-220) argument that verbs of the conjoining type could enter the VP slot of the construction such as lijie ‘understand’ and sikao ‘think’ that are typical of metal verbs.

Meaning pattern of “augmentation”

The second meaning pattern that could uncover the intrinsic nature of the VP in the construction pertains to the sense of “augmentation”. Members of this cluster chiefly incorporate kuoda ‘expand’, zengjia ‘augment’, zengqiang ‘enhance’, tigao ‘improve’, jiakuai ‘speed up’, jiada ‘increase’, and zengda ‘magnify’. All these verbs denote more or less a sense of augmentation of degrees and/or amounts. This is partially exemplified by sentences in (4), in which zengjiang ‘enhance’ in (4a) denotes the addition of degrees in terms of economic power and zengjia ‘augment’ in (4b) denotes the addition of amounts in terms of information channels in question.

(4) a. 今后, 随着国家经济实力的增强, 教育经费会逐年增加。

jinhou suizhe guojia jingji shili de zengqiang jiaoyu jingfei

in the future with nation’s economy power de enhance education fund

hui zhunian zengjia

will year-by-year increase

‘In the future, education funds will increase year by year with the enhancement of national economic strength.’

b. 信息渠道的增加原本可以淡化校园网站的重要性。

xinxi qudao de zengjia yuanben keyi danhua xiaoyuan

information channel de augment originally might dilute campus

wangzhan de zhongyaoxing

internet de importance

‘The augmentation of information channels might originally dilute the importance of campus internet.’

Meaning pattern of “implementation”

The third meaning pattern is about the sense of “implementation”, which is typically realized by such verbs as zhiding ‘enact’, guanche ‘implement’, kaizhan ‘carry on’, qianshu ‘sign’, lvxing ‘perform’, and xingshi ‘perform’. They are clustered in that their covarying collexemes in the NP slot are generally biaozhun ‘standard’, zhengce ‘regulation’, and quanli ‘rights’. Consider examples in (5), in which zhengce ‘policy’ is a covarying collexeme of both guanche ‘implement’ in (5a) and zhiding ‘enact’ in (5b) which is rewritten from example (2).

(5) a. 产业政策的贯彻、实施无法有力执行。

chanye zhengce de guanche shishi wufa youli zhixing

industry policy de implement perform cannot effectively carry out

‘The implementation of industrial policy cannot be carried out effectively.’

b. 公共行政就是公共政策的制定。

gonggong xingzheng jiushi gonggong zhengce de zhiding

public administration be public policy de enact

‘Public administration is the enactment of public policy.’

Meaning pattern of “achievement”

Verbs in the VP slot of the construction also denote a sense of “achievement”, indicating reaching specific results with efforts. These verbs generally include qude ‘achieve’, jieshu ‘finish’, jiejue ‘resolve’, shixian ‘realize’, zhangwo ‘command’, and wancheng ‘accomplish’. Their covarying collexemes chiefly pertain to positive targets such as mubiao ‘target’, chengji ‘result’, chengjiu ‘achievement’, and jiazhi ‘value’. Example (6) will demonstrate the sense of “achievement” that is realized by one of these verbs.

(6) 它们从不同的角度来保障公正价值的实现。

tamen cong butongde jiaodu lai baozhang gongzheng jiazhi de

they from different angles come guarantee justice value de shixian

realize

‘They guarantee the realization of justice value from different angles.’

Meaning pattern of “establishment”

The meaning pattern of “establishment” in the VP slot of the construction is denoted by such verbs as jianli ‘establish’, sheli ‘set up’, and kaishe ‘set up’. The clustering of these verbs is primarily attributed to the covarying collexemes (e.g., organization names and regulations) in the NP slot they cooccur with. Consider examples in (7), in which caigou zhongxin ‘purchasing center’ in the NP slot is the significant cooccurrence of sheli ‘set up’ in the VP slot as shown in (7a) and dongbao fagui ‘regulations of animal protection’ is that of jianli ‘establish’ as shown in (7b). caigou zhongxin ‘purchasing center’ in (7a) represents the name of an organization and dongbao fagui ‘regulations of animal protection’ in (7b) is the name of a regulation.

(7) a. 跨国采购中心的设立, 拉动了深圳的出口增长。

kuaguo caigou zhongxin de sheli ladong le Shenzhen

international purchase center de set up promote tense part name

de chukuo zengzhang

de export increase

‘The setting up of international purchasing center promoted the increment of export in Shenzhen’.

b. 中国动保法规的建立, 不只依靠任何组织或团体。

zhongguo dongbao fagui de jianli buzhi yikao

China’s animal protection rules de establish not only rely upon

renhe zuzhi huo tuanti

any organization or groups

‘The establishment of regulations of China’s animal protection relies not solely on any organization or group.’

Meaning patterns of “report” and “others”

Another important prototypical meaning pattern with respect to the VP in the NP de VP construction is concerned with the reporting verbs. In other words, reporting verbs are also likely to enter the VP slot of the construction. Typical reporting verbs of this type include baodao ‘report’, tichu ‘propose’, and biaoda ‘express’. The contents that are reported are generally about news media, issues or questions, opinions, which constitute the covarying collexemes in the NP slot to attribute the meaning pattern of “report”. This will be instantiated by examples in (8), in which contents of the reporting verbs baodao ‘report’ and tichu ‘propose’ are zhongdong wenti ‘middle east issues’ as shown in (8a) and wenti ‘question’ as shown in (8b) respectively.

(8) a. 西方媒体对中东问题的报道最清楚不过地表明了这一点。

xifang meiti dui zhongdong wenti de baodao zui qingchu buguo

west media to middle east issue de report most obvious part

di biaoda le zheyidian

adv. marker express tense marker this point

‘Nowhere is clearer than the report delivered by west media concerning middle east issues.’

b. 一个好的问题的提出, 具有更加重要的意义。

yige haode wenti de tichu juyou gengjia zhongyaode yiyi

a good question de propose have more important meaning

‘The proposing of a good question has more important meanings.’

In addition to the abovementioned meaning patterns that verbs in the VP slot of the construction could denote, there are also some other verbs that could not be patterned into a single meaning group. However, they are also equally important to represent the typical meanings of lexical items in the VP slot. Those verbs mainly include zhaokai ‘convene’, chengdan ‘undertake’, yingxiang ‘influence’, etc.

In summary, the typical meanings that could be patterned in the VP slot of the NP de VP construction generally incorporate senses of “cognition”, “augmentation”, “implementation”, “achievement”, “establishment”, and “report”, which are abstracted by referring to their covarying collexemes in the NP slot of the construction. In the same vein, the meaning patterns of NP in the NP de VP construction could also be clustered by referring to their corresponding covarying collexemes in the VP slot of the construction. This will be enunciated in the following section.

Meaning patterns of the NP in the NP de VP construction

After manually processing those overlapping lexical items in the NP slot of the NP de VP construction, we finally confirmed 62 types of nouns. Considering their covarying collexemes in the VP slot of the construction, a 62 (types of nouns in the NP slot) × 99 (types of covarying collexemes in the VP slot) contingency table is subsequently formatted. Running the function hclust in R language yields a cluster dendrogram presented in Fig. 2. Figure 2 demonstrates that, except for the difficulty of patterning some lexical items that are significantly attracted to the NP slot of the construction (e.g., shijian ‘time’, wenti ‘issue’, kecheng ‘subject’, etc.), six types of meaning patterns could be abstracted from those items in the NP slot; specifically, lexical items in the NP slot denote senses of “internal traits”, medical names”, regulations”, “results”, “systems”, and “business”, which are to be further expounded in the following paragraphs.

Fig. 2
figure 2

Dendrogram of meaning patterns of NP in the NP de VP construction.

Full size image

Meaning pattern of “internal traits”

The first typical meaning pattern that lexical items in the NP slot clustered pertains to the description of people’s internal traits. Those lexical items generally include nengli ‘ability’, rencai ‘talent’, suzhi ‘quality’, yishi ‘awareness’, zeren ‘responsibility’, zhenxin ‘sincerity’, jineng ‘skill”. In this cluster, there seem to be some other lexical items (e.g., lidu ‘force’, xingwei ‘behavior’, and nianling ‘age’) that are not directly related to the meaning pattern of “internal traits”. Those lexical items prompt the suggestion that hierarchical cluster analysis by means of referring to the covarying collexemes needs to be further improved, which is precluded from further discussion because of the purpose of this study. Concerning the covarying collexemes in the VP slot, corpus data reveal that these verbs by and large incorporate tigao ‘improve’, peiyang ‘cultivate’, and zengqiang ‘enhance’. This could be illustrated by the covarying collexemes zengqiang ‘enhance’ in (9a), peiyang ‘cultivate’ and tigao ‘improve’ in (9b), which significantly cooccur with nengli ‘ability’ in (9a), rencai ‘talent’ and sushi ‘quality’ in (9b), respectively in the NP slot of the NP de VP construction.

(9) a. 感悟水平的提高, 意味着审美体验能力的增强。

ganwu shuiping de tigao yiweizhe shenmei tiyan nengli de

inspiration level de increase mean esthetics experience ability de

zengqiang

enhance

‘The increment of inspiration levels means the enhancement of esthetic ability.’

b. 人才的培养一定要注重全面素质的提高。

rencai de peiyang yiding yao zhuzhong quanmian sushi de

talent de cultivate must be highlight all quality de

tigao

improve

‘The cultivation of talents must highlight the improvement of all kinds of qualities.’

Meaning pattern of “medical names”

Another topic that lexical items in the NP slot of the construction considered are about items that are used in the medical field, including such terms as jibing ‘disease’, xibao ‘cell’, jiyin ‘gene’, and bingfazheng ‘syndrome’. Some others such as shijian ‘event’ (denoting a negative sense) and shigu ‘accident’ are clustered with the meaning pattern of “medical names” because of their denoted negative senses. With respect to their covarying collexemes in the VP slot, verbs such as fasheng ‘occur’, chuxian ‘appear’, and cunzai ‘exist’ are significantly preferred by lexical items that realize the meaning pattern of “medical names”. Examples in (10a) and (10b) suffice to illustrate this point.

(10) a. 并发症的出现与创伤的严重程度有关。

bingfazheng de chuanxian yu chuangshan de yanzhong chengdu youguan

syndrome de appear and trauma de seriousness degree related

“The appearance of the syndrome is related to the seriousness of traumas.’

b. 基因的存在已得到了许多研究的证实。

jiyin de cunzai yi dedao le xuduo yanjiu de zhengshi

gene de exist already get tense part many study de testify

‘The existence of gene has already been testified by numerous studies.’

Meaning pattern of “regulations” and “systems”

The third typical meaning pattern that the NP in the NP de VP construction denotes pertains to zhidu ‘regulations’ and tixi ‘systems’ (except for some irrelevant ones in either of the two meaning patterns). The former further incorporates biaozhun ‘standard’, zhengce ‘policy’, zhidu ‘regulation’, quanli ‘rights’, gongneng ‘function’, and zuoyong ‘function’; the latter incorporates such members as jigou ‘organization’, tixi ‘systme’, tizhi ‘regulation’, and jizhi ‘mechanism’. Investigation of their covarying collexemes in the VP slot demonstrates that they cooccur significantly with verbs that denote senses of “implementation” and “establishment” such as guanche ‘enact’, zhiding ‘implement’, xingshi ‘perform’, and jianli ‘establish’, etc. This could be exemplified by the significant cooccurrences of zhidu ‘regulation’ and guanche ‘enact’ in (11a), and tizhi ‘regulation’ and jianli ‘establish’ in (11b) which is rewritten from example (3).

(11) a. [我们]要认真检查各种制度的贯彻落实。

[women] yao rezhen jiancha gezhong zhidu de guanche luoshi

[we] must carefully check various regulation de enact complete

‘[We] must carefully check the enactment of various regulations.’

b. 这必然阻碍市场体制的建立。

zhe biran zhu’ai shichang tizhi de jianli

this must hinder market regulation de establish

‘This will definitely hinder the establishment of the market regulations.’

Meaning pattern of “results”

It might be a little farfetched to term this meaning pattern as “results”. Another term such as “achievement” might serve this purpose very well. However, all these lexical items denote a sense of moving towards an ending, and thus “results” is employed. Members in this meaning pattern include bufa ‘step’, jiezhou ‘rhythm, chengji ‘result’, chengjiu ‘achievement’, jiazhi ‘value’, and jincheng ‘progress’. Covarying collexemes in the VP slot that are significantly preferred by these lexical items in the NP slot generally incorporate qude ‘achieve’ and shixian ‘realize’. Consider examples in (12), in which covarying collexemes qude ‘achieve” in (12a) and shixian ‘realize’ in (12b) in the VP slot cooccur significantly with chengji ‘result’ and jiazhi ‘value’ in the NP slot respectively.

(12) a. 上述成绩的取得不是偶然的。

shangshu chengji de qude bu shi ourande

abovementioned result de achieve not copular opportunistic

‘The abovementioned achievement of results is not opportunistic.’

b. 言论自由有利于自我价值的实现。

yanlun ziyou youliyu ziwo jiazhi de shixian

speech freedom helpful self-value de realize

‘Freedom of speech is helpful for the realization of self-values.’

Meaning pattern of “business”

The last typical meaning pattern that lexical items in the NP slot could be abstracted is “business” in that these items are concerned with various aspects of the business. For example, the manner that decisions are made in business (i.e., dahui ‘meeting’ and huiyi ‘meeting’), the means that agreements are reached (i.e., hetong ‘contract’ and xieyi ‘agreement’), etc. Concerning the covarying collexemes in the VP slot, these lexical items in the NP slot frequently cooccur with verbs such as zhaokai ‘convene’, qianshu ‘sign’, and lvxing ‘perform’. Lexical items denoting the meaning pattern of “business” and their covarying collexemes in the VP slot are exemplified by (13a) and (13b).

(13) a. 这次会议的召开为今后两国经贸关系的发展打下了基础。

zheci huiyi de zhaokai wei jinhou liang guo jing mao

this meeting de convene for afterwards two nations business trade

guanxi de fazhan daxia le jichu

relation de develop found tense part basis

‘The convening of this meeting laid a foundation for the future development of bilateral economic and trade relations.’

b. 和平协议的签署为两国关系的发展开辟了新的前景。

heping xieyi de qianshu wei liangguo guanxi de fazhan kaipi

peace agreement de sign for two nation relation de develop open

le xinde qianjing

tense part new prospect

‘The signature of the peace agreement opens a new era for the development of the two nations.’

Discussion

Firstly, meaning patterns of the VP in the NP de VP construction identified in this study and Zhan’s (1998) argument that only the VP that denotes weak “action” and strong “event” could enter the VP slot of the construction, together with Qi et al.’s (2004) argument that verbs of the conjoining type could enter the VP slot of the construction, will be discussed in this section; in addition, meaning patterns of the NP in the NP de VP construction observed in this study and Shen and Wang’s (2000) argument that lexical items in the NP slot expressing the prominence that combines informativity and accessibility will also be compared and expounded. Secondly, the principle of linguistic meaning conservation is employed to explain the findings uncovered in this researchFootnote 7. Finally, relevant theories in Construction grammar are further elaborated by means of drawing on features from the NP de VP construction.

In relation to previous studies

Findings in this research, with respect to meaning patterns that lexical items in the VP slot of the NP de VP construction most probably denote, are partially in accordance with those uncovered by Zhan (1998). According to Zhan (1998, p. 25), verbs in terms of their ideational meanings are subcategorized into three types. Verbs of the first type are of a strong sense of “action” but a weak sense of “event”; those of the second type are of a weak sense of “action” but a strong sense of “event”; and those of the third type are both weak in terms of “action” and “event”. Of the three types of verbs, only the ones that denote a sense of weak “action” and strong “event” are capable of entering the VP slot of the NP de VP construction; this argument is also corroborated by Wang (2002) who draws on the numbers of participants in the transitivity system of the verbal processes; that is, the more participants involved in the verbal processes, the stronger the senses of “action” denoted by verbal processes will be. However, concerning the verbs that significantly attracted to the VP slot of the construction, we identified verbs with a sense of strong “action” and weak “event” (e.g., tunshi ‘swallow’, ruqin ‘invade’, qianshu ‘sign”, jianli ‘establish’, etc.) and those with a sense of weak “action” and strong “event” as well (e.g., juede ‘think’, zhaokai ‘convene’, tichu ‘propose’, etc.), although the latter takes a larger proportion than the former (i.e., 82.5% for the latter and 17.5% for the former in terms of verbs that reach the Infinite level of significance). Reasons for the disagreement between the findings of this study and those of previous ones are possibly three-fold. First, different sources of the instances that we counted upon prompt the findings that are not fully in agreement with each other. Zhan (1998) considered verbs from a Dictionary for Usages of Verbs (Meng, 1987) and exemplified his argument with invented examples, while this research considers examples from real data of BCC. It does not indicate that invented grammatical examples are not sufficient to uncover certain features of particular constructions; however, real data from corpus speak more expressly in terms of the intrinsic nature of language phenomena (e.g., meaning patterns of constructions) than those that are particularly made up by researchers. Second, instances that are considered by the two studies differ significantly. Zhan (1998) considered instances of the NP de VP construction in general without taking their probabilities of entering the construction into consideration, while we, by drawing on covarying collexeme analysis considered those instances that are extremely significantly attracted to this construction; by so doing, the typical meanings that this construction could denote will be uncovered. It is inappropriate to regard instantiated verbs with very low occurrences in the construction, if they are considered as a whole, as the typical meaning of this construction; this inappropriateness will be sufficiently avoided by means of covarying collexeme analysis and is regarded as repelled cases to the NP de VP construction. It should be noted that I am not meant to downplay Zhan’s research without considering probabilities of instances with the construction at issue. What he primarily highlights are instances that probably fill the NP and the VP slots of the NP de VP construction, while this research aimed at uncovering the most typical meaning patterns of the construction which will definitely turn to statistical measures such as probabilities of lexical items with the considered construction. Third, the purposes of the two studies are different. Zhan (1998) intended to identify certain types of verbs that could enter the VP slot of the construction, while this research highlights the typical meanings that verbs in the VP slot of the construction could denote. Differences in the purposes also render the different findings between this study and that conducted by Qi et al. (2004). In addition, we approached the same phenomenon with more delicacy than their study. That is to say, although all instantiated verbs in the VP slot identified by this study could be subsumed to those identified by Qi, et al. (2004), we further identified meaning patterns that verbs in the VP slot could denote by means of the hierarchical cluster analysis. Accordingly, with respect to verbs that could enter the VP slot of the construction, this study is significant in that it investigated the typical meanings that instantiated verbs in the VP slot could denote and how these typical meanings could be patterned into different clusters, which is perspicuously absent from these aforementioned studies.

In relation to lexical items in the NP slot of the construction, Shen and Wang (2000) argued that NPs that denote a sense of prominence (i.e., high informativity and/or high accessibility), could enter the NP slot of the construction. Their argument is further corroborated by the findings of this study which has identified such meaning patterns as “internal traits”, “medical names”, “regulations”, “results”, “systems”, and “business”. According to Shen and Wang (2002, p. 30), specific nouns are more accessible than abstract nouns. Therefore, examining the meaning patterns of the NP in the construction identified in this study, we found that these meaning patterns, except for “internal traits”, are actually of some degree of high accessibility. Although lexical items denoting “internal traits” are not of high accessibility (because their meanings are comparatively more abstract than those of other meaning patterns), their meanings are by and large of high informativity. Admittedly, the high informativity of the meaning pattern of “internal traits” is also determined by the context. Consider the example in (14). nengli ‘ability’ is modified by lijie ‘understanding’ and yuedu ‘reading’ in the linguistic context which provide extra information to make the lexical item nengli ‘ability’ in the NP slot of the construction possess a high degree of informativity; in so doing, this lexical item also becomes more specific than the one without such modifiers as lijie ‘understanding’ and yuedu ‘reading’, and thus is also of high accessibility.

(14) 现在对语法不那么重视了, 关键是理解能力、阅读能力的培养。

xianzai dui yufa bu name zhongshi le guanjian shi

now to grammar not so highlight tense part important copular

liji nengli yuedu nengli de peiyang

understanding ability reading ability de cultivate

‘Now grammar is no so important; what matters is the cultivation of understanding and reading ability.’

This paper proceeds further than Shen and Wang’s (2002) study in at least two respects. On the one hand, we considered not only lexical items that could enter the NP slot of the NP de VP construction but also the ones that are representative of the typical meanings of the NPs. By so doing, it will further facilitate our understanding of the typical meaning that the NP de VP construction could denote. On the other hand, we clustered the meaning patterns that lexical items could enter the NP slot by means of the hierarchical cluster analysis, which will objectivize the final meaning groups.

In relation to the principle of linguistic meaning conservation

The range of prototypical meanings for the NPs and VPs of the NP de VP construction, along with the non-prototypical meanings, uncovered in this research appears to be a good example of informational conservation by way of collocational variation. The more variation one finds in lexical items and their types, the more ways there are in conserving information. This is the crux of the Principle of Linguistic Meaning Conservation (Mondal, 2019). The upshot of this is that the variation in patterns of meaning is a function of the variation in lexical items and their types. This is compellingly clear in the wide range of meanings of NPs for the NP de VP construction: “internal traits”, “medical names”, “regulations”, “results”, “systems”, and “business”, and meanings of VPs for the NP de VP construction: “cognition”, “augmentation”, “implementation”, “achievement”, “establishment”, and “report”. Given the greater types of NPs and VPs considered in this research, this does not seem to be a surprise.

In relation to construction theories

This subsection mainly discusses the meaning patterns of the NP de VP construction in terms of construction theories: compositionality, prototypicality, and conventionality.

Meaning patterns of the NP de VP construction uncovered in this research reflect the compositionality in terms of the construction grammar. Compositionality highlights the degree of transparency of association between forms of constructs (i.e., specific NPs and VPs in the construction) and meanings (Traugott and Trousdale, 2013). A construct with compositionality in terms of semantics means that the meaning of a construct could be understood if the meanings of each unit are known. The opposite case is true if the combination of meanings of each unit could not unpack the meaning of the construction. In relation to the NP de VP construction at issue, this construction is of a high degree of compositionality. This viewpoint is evidenced by the three pairings of meaning patterns of the NP de VP construction, i.e., the pairing of “regulations” in the NP slot and “implementation” in the VP slot, the pairing of “systems” in the NP slot and “establishment” in the VP slot, and the pairing of “results” in the NP slot and “achievement” in the VP slot. For instance, if the NP slot is confined to lexical items denoting “regulations”, then the meanings of lexical items in the VP slot are predictable and so is the typical meaning of the NP de VP construction.

Meaning patterns of the construction under investigation also accord with the theory of prototypicality. Prototypicality originally means the degree of category membership (Goldberg, 1995), and in this research, it means the identified meaning patterns are the most typical ones or the most representative of the meaning denoted by the NP de VP construction. Meaning patterns of the NP slot and the VP slot are confirmed by referring to the semantic features of these NPs and VPs. Those lexical items with statistically significant association strengths represent core members of the NP de VP construction. The meanings of these core members in turn represent the prototypical meaning of this construction. Core members of NPs in this construction briefly include zhidu “regulation”, tixi “system”, yewu “business”, etc., and thus the typical meaning patterns are “regulations”, “systems” and “business” (cf. Fig. 2). In a similar vein, meanings of core members of VPs are summarized as such meaning patterns as “implementation”, “achievement”, and “establishment” (cf. Fig. 1).

These meaning patterns reflect the theory of conventionality as well. Both NPs and VPs with significant attraction to the construction are the conventionality of frequently using these lexical items. The higher the observed frequencies of these lexical items or the association strengths between these lexical items and the NP de VP construction (because of the significant correlation between observed frequencies and association strengths (cf. Zhou, 2021)), the higher the degree of the conventionality of these lexical items which in turn leads to the higher degree of the conventionality of the meaning patterns of the construction at issue.

Significance

This research is significant in at least two respects. For one thing, lexical items uncovered in this research are representative of those that could fill the VP and the NP slots of the construction. The covarying collexeme analysis highlights lexical items that are significantly attracted to specific constructions, which indicates that these lexical items are the potential to be typical members of the constructions. More importantly, this research considered a fuller scope of the NP de VP construction than previous ones which investigated lexical items that could enter either the NP slot or the VP slot, in that it considered both cases concurrently; apart from that, lexical items in the NP slot and the VP slot respectively are mutually dependent (due to the nature of this research method employed). For another, the meaning patterns of both the NP slot and the VP slot are scientifically and objectively clustered, which is further attributed to the meaning patterns of the NP de VP construction. This research employed cluster analysis which aims at identifying the meaning patterns of the construction. Previous research unveiled possible lexical items that could be the components of the construction, but they did not inform us what the meaning patterns of this construction are because constructions themselves are meaningful (Goldberg, 1995, 2006).

This research is also of pedagogical significance. Studies have shown the first collexemes in each construction learned by students tend to be those distinctively associated with that construction in the input (Ellis and Ferreira-junior, 2009). In this sense, core lexical items that constitute the typical meaning patterns of the NP de VP construction are firstly learned ones in the input. The findings of this study could feed into the design of pedagogical materials not only for native speakers of Chinese but also for others who study Chinese as a foreign language.

Conclusion

Drawing on the frequency of cooccurrences between lexical items in the NP slot and those in the VP slot in the NP de VP construction in modern Chinese, we identified typical instances that are extremely significantly attracted to the NP de VP construction by implementing a covarying collexeme analysis. Underlying those significantly preferred instances, we further investigated the meaning patterns of those lexical items in the VP slot and confirmed that the most typical meaning patterns are “cognition”, “augmentation”, “implementation”, “achievement”, “establishment”, and “report” by means of the hierarchical cluster analysis. By employing the same cluster analysis, we further confirmed that the most typical meaning patterns of lexical items that are capable of entering the NP slot of the construction incorporate “internal traits”, medical names”, regulations”, “results”, “systems”, and “business”. This study is significant in that it scientifically identified typical instances of the NP de VP construction in modern Chinese, and objectively and precisely uncovered meaning patterns that lexical items in both the VP slot and the NP slot of the NP de VP construction could denote.

Findings uncovered in this study are based on the corpus data in general without referring to genres or text types, such as fiction, magazines, newspapers, and academic discourses, which might display differences in significantly attracting instances to the NP de VP construction and/or meaning patterns that lexical items in both NP and VP slots of the construction could denote. Accordingly, future studies are suggested to highlight the denoted differences of significantly attracted instances and meaning patterns of NP de VP construction in modern Chinese.