An event integration approach to the family of Chinese instrumental constructions

1 Introduction

The instrument has long been a focus in linguistic literature (Fillmore 1968, Schlesinger 1979, Koenig et al. 2008, Rissman and Rawlins 2017, among others). In Case Grammar, the “instrumental” is proposed to conceptually refer to “the case of the inanimate force or object causally involved in the action or state identified by the verb” (Fillmore 1968: 24). According to Dowty (1991), the instrument is described as a participant that has an equal amount of proto-agent and proto-patient properties, which captures the nature that the instrument is both acted upon by a manipulator and acting on another participant itself (Hooste 2018: 2, Næss 2007: 90). Besides the thematic relations, the instrument is also a hotly discussed topic in linguistic typology. Specifically, the instrument is closely associated with other semantic categories, such as comitative, manner, material (Haspelmath 2003; Narrog and Ito 2007; Schlesinger 1979; Stolz 1996, 2001) and thus the instrumental marking is cross-linguistically multifunctional.

When we move from the instrument or intrstrumental itself to a larger linguistic context representations, it is found that instrumental elements are actually situated in some fixed syntactic patterns, which conform with the definition of construction, “form-meaning pairs” (Goldberg 1995) or “learned pairings of form with semantic or discourse function” (Goldberg 2006). Therefore, these fixed syntactic patterns with elements expressing the instrument concept can be collectively referred to as instrumental constructions, which share “important properties but differing in certain specifics, including their degree of productivity” (Goldberg and Jackendoff 2004), and thus actually form the family of instrumental constructions. In English, semantic contributions of with and use to the instrumental constructions have been widely discussed (Lakoff 1968; Rissman and Rawlins 2017). For example, in (1a-b), with and use are semantic operators used to introduce instrumental components, respectively. Whereas, there are also other instrumental constructions where the instrumental components serve as subjects or objects directly without any markers. For example, the instrumental component hammer in (2a-b) serves as the subject and the object, respectively.

Likewise, there also exist such fixed syntactic patterns in Chinese where the instrumental components can be introduced by prepositions 用 yòng, 拿 ná, 以 yǐ, etc., or act as subjects or objects directly, and thus form the well-known marked instrumental constructions (Shi 1995; Wu 1996), instrument-subject constructions (Chen 2003; Wang 1984) and instrument-object constructions (Ding 1961: 37; Lv 1979: 72), respectively, as shown in (3a-c):

Example (3a) demonstrates a situation where the instrument 刀 dāo ‘knife’ is introduced by prepositional marker 用 yòng. Examples (3b) and (3c) show cases where there are no instrumental markers, and the difference between them lies in that the instrument 缝纫机 féngrènjī ‘sewing machine’ in (3b) functions as the subject, while the instrument 大碗 dà wǎn ‘big bowl’ in (3c) serves as the object.

By combing the studies on Chinese instrumental constructions, we can easily come to the three-way classification. However, as pointed out by reviewers, there may exist other instrumental constructions, for example, the topic instrumental construction, see (4a), where 这把刀 zhè bǎ dāo ‘this knife’ is moved to the initial position and serves as a topic. However, the present study focuses solely on the three widely accepted Chinese instrumental constructions mentioned above, namely those involving instruments as arguments, while excluding other peripheral variants, say the topic instrumental construction, see (4a), instruments as verbal classifiers, see (4b) and instruments as noun quantifiers, see (4c). Such variants deserve attention in future studies.

Another issue that needs to be clarified is the definition of instrumental constructions. Previous studies have taken the three-way classification of instrumental constructions for granted without explaining why. There is no dispute that the marked instrumental constructions are members of the family of Chinese instrumental construction because they semantically include the instrumental markers. More explanation is needed for the other two types of instrumental constructions. Despite the fact that both of the instrument-subject construction and the instrument-object construction belong to SVO constructions, they should not simply be treated as regular SVO constructions. For one thing, in addition to the agent-patient relation in SVO constructions, both of them semantically display instrumental features and thus should be categorized into the family of Chinese instrumental constructions. For another, not all sentences with the instrument NP in the subject or object slots can be categorized as instrumental constructions, see (5a-b). 石头 shítou ‘stone’ only serves as a subject in (5a), whereas it functions as a subject in (5b). But both of them can not be intepreted as the instrumental components and thus examples (5a-b) cannot be categorized into instrumental constructions. This manifests that the instrument-subject construction and the instrument-object construction are different from common SVO constructions and thus should be treated as independent constructions, i.e., the instrumental constructions.

By examining examples in previous literature further, we propose that all instrumental constructions must contain instrumental components which explicitly or implicitly involve the whole “Agent-Instrument-Patient” event sequence, see Figure 1 (Langacker 1991: 217) and thus imply the existence of the instrument-manipulation event and the act-on event.^[2] The instrument-manipulation event refers to the manipulation of an agent on an instrument, while the act-on event refers to a case where an action is performed by an agent or a secondary agent, i.e. instrument, on or in a patient. “Explicitly” means that the event in question is denoted by predicate elements. Take (3a) for example, 我 wǒ ‘I’ denotes the agent, 刀 dāo ‘knife’ the instrument, and 肉 ròu ‘meat’ the patient. The instrument-manipulation event describes a scenario where “an instrument is manipulated by an agent”, and it is exemplified by 我用这把刀 wǒ yòng zhè bǎ dāo ‘I used this knife’, while the act-on event denotes another scenario where an instrument (also referred to as a secondary agent) or agent acts on another entity, and it is instantiated by 这把刀切肉 zhè bǎ dāo qiē ròu ‘this knife cut meat’. “Implicitly” means that the event is implied rather than directly designated by predicate elements. Take (3b) for example, 缝纫机 féngrènjī ‘sewing machine’ is realized as the instrument, while 衣服 yīfú ‘clothes’ as the patient. There is no explicit agent in (3b). However, based on the encyclopedia knowledge, the sewing machine is operated by an agent, because it cannot begin to work spontaneously. In view of the missing of agent in (3b), only the act-on event is explicitly representted, and the instrument-manipulation event is implicitly contained. Generally, these instrumental constructions in (3) are syntactically represented by single clauses, but they are semantically composed of two events. All of these indicates that the semantic concept of Chinese instrumental events basically accords with the theoretical assumption of “event integration”. According to Talmy (2000), event integration refers to a linguistic phenomenon that two events tend to be integrated together and represented by one clause, for example:

Figure 1:

The instrumental action chain (cf. Langacker 1991: 217).

The word down in (6) indicates that the police fulfilled their intention of capturing the fugitive, which was the purpose of their engaging in the hunting activity. Therefore, (6) includes a hunting event and an event of realization, thus embodying the concept of event integration.

Thus far, we have roughly illustrated our research objects, redefined why or how they can be considered as instrumental constructions, and briefly introduced event integration instrumental constructions involved. Actually, the three-way classification is only based on syntax without considering any semantic factors, and thus cannot reflect the differences within a particular type of instrumental constructions or between different instrumental constructions. In addition, previous studies mainly focus on a certain type of instrumental constructions, namely instrument-subject constructions (Niu 2008; Xu 2003; Zhang 2020), instrument-object constructions (Xie and Qiao 2009; Zhu 1982: 110), or marked instrumental constructions (Chen 2001, 2007). To the best of our knowledge, studies concerning the whole family of Chinese instrumental constructions to explore how different Chinese instrumental constructions are structured and what are the differences and similarities among them, have not been conducted so far.

Based on the analyses above, situated in the framework of event integration, this study aims to devote to the family of Chinese instrumental constructions. Specifically speaking, the similarities and differences different Chinese instrumental constructions display during the process of event integration will be intensively explored. Meanwhile, we also try to explore whether this study can theoretically contribute to event integration. In what follows, Section 2 is devoted to introducing the classification of the research objects so as to ascertain the research objects in a fine-grained way. Section 3 introduces the event integration, and further puts forward some new principles suitable for the analysis of Chinese instrumental constructions. Meanwhile, the similarities between Chinese instrumental constructions are generalized as well. Section 4 explores the differences between Chinese instrumental constructions in the degree of event integration. Section 5 discusses implications of Chinese instrumental constructions on event integration. Section 6 concludes the study and points out the future directions.

2 Classification of Chinese instrumental constructions

As the names suggest, the widely accepted three-way classification of Chinese instrumental constructions including marked instrumental constructions, instrument-subject constructions and instrument-object constructions, is classified based on syntax, that is, whether the instrumental component is introduced by a marker or not, and if not, whether it serves as a subject or an object. Such classification embodies its inherent limitations. Specifically, the classification does not take the semantic parameter into consideration and thus is challenged in some cases, as shown in (7):

Although examples (7a-b) belong to 用 yòng instrumental constructions, (7a) can be shifted to an instrument-subject construction, i.e. (7a′), but there is no corresponding (7b′) for (7b). Taking both syntactic and semantic factors into account, Zhang and Li (2022b) innovatively propose an eight-way classification of Chinese instrumental constructions.

Essentially, the instrumental component plays a different semantic role in (7a) and (7b), respectively. Koenig et al. (2008) address the relation between instrumental component and causation: when the instrumental component bears a causative meaning, the instrument is intermediary; meanwhile, the instrumental component may not display the causative meaning, but only express a helping or facilitating meaning. Thus, instruments can be divided into intermediary type and facilitating^[5] type. In (7a), the instrument 刀 dāo ‘knife’ has a direct causative effect on the object 肉 ròu ‘meat’. However, the instrument 筷子 kuàizi ‘chopsticks’ in (7b) has no similar function to 饭 fàn ‘meal’, but only facilitates the occurrence of the eating event. In view of this, this study introduces the causative parameter, and divides the instruments in the instrumental constructions into the intermediary type and the facilitating type. Therefore, the three-way classification can be extended to 6 types correspondingly. However, based on the data from the previous literature and the BCC corpus, Zhang and Li (2022b) also find that besides the 6 types of instrumental constructions mentioned above, there are another two types of instrumental constructions, as shown in (8):

According to the previous syntactic classification, there is a preposition 用 yòng in (8a), so it belongs to “marked instrumental construction”. However, unlike other marked instrumental constructions, the subject in (8a) is the patient rather than the agent, so this study calls it “patient-oriented marked instrumental construction”. The instrumental component 刺刀 cìdāo ‘bayonet’ in (8b) is the subject, and thus this construction falls into the instrument-subject construction. However, unlike other instrument-subject constructions, there is no patient object after the predicate in (8b). The reason is that the construction has already had a patient, namely the instrumental subject, thus referred to as “the patient-oriented instrument-subject construction”.^[6] All of the eight types of Chinese instrumental constructions and corresponding examples are detailed in Table 1.

Figure 2:

The instrumental action chain and its different profile schemas (cf. Langacker 1991: 217).

The eight-way classification of Chinese instrumental constructions is not only reasonable but also holds significant typological importance compared with the instrumental action chain proposed by Langacker (1991: 217). According to Langacker, the agent, instrument and patient depicted in Figure 1 form a sequenced instrumental action chain with Instrument as the intemediary role in terms of the energy transfer, repeated as (a) in Figure 2. Schemas (b-d) profile different portions of the instrumental action chain and they can be instantiated by example (9a-d), respectively.

Example (9c) actually does not imply the existence or involvement of an instrument. More importantly, Figure 1 indicates that the patient is supposed to experience the change of state. However, Table 1 innovatively reveals that the change of state may occur in the instrument rather than the patient, as indicated by the Patient-oriented Instrument-Subject Construction in Chinese. Therefore, this study based on the eight-way classification has a typological significance as well. Therefore, this study takes all these eight types of instrumental constructions proposed by Zhang and Li (2022b) as the Chinese family of instrumental constructions, namely the research objects of this study.

3 Event integration

4 Degree of event integration of Chinese instrumental constructions

The similarities between Chinese instrumental constructions lies in that all of them embody the features of macro-events and can be called the instrumental macro-events, based on which, this section tends to explore the differences between Chinese instrumental constructions from the degree of event integration.

Instrumental constructions are overt linguistic representations of Agent-Instrument-Patient event sequence (cf. e.g. Langacker 1991: 217). Conceptually, the “Agent-Instrument-Patient” event sequence could offer 3 arguments, namely the total number of arguments, but the number of arguments appearing in concrete instrumental constructions may be 3 or less than 3. The marked instrumental constructions involve 3 arguments and profile the whole sequence, while other unmarked instrumental constructions involve 1 or 2 arguments and only designate part of it. However, in terms of the events integrated in all instrumental constructions, in addition to the instrument-manipulation event and the act-on event, some instrumental constructions contain a resultative event as well. This phenomenon that an action may cause a result, conforms with concrete scenarios in the physical world, and Langacker (1991: 217) also introduces the potential result caused by an action when describing the “Agent-Instrument-Patient” event sequence. Therefore, it is a natural and reasonable fact that some instrumental constructions can include a resultative event. Table 2 exemplifies the eight types of Chinese instrumental constructions as well as shows the distributions of the explicit event, the total event and E/TER within them. Meanwhile, in the column of examples, we also illustrate whether a predicate component denoting a resultative event can be added:

In what follows we will elaborate on Table 2 by showing how the explicit event and the implicit event are integrated into different instrumental constructions, how the total event and E/TER are calculated, and why some instrumental constructions can potentially contain a resultative event while others cannot.

1. The intermediary marked instrumental construction includes two certain explicit events, that is, an instrument-manipulation event denoted by 用 yòng (我用这把刀 wǒ yòng zhè bǎ dāo ‘I use this knife’) and act-on event denoted by 切 qiē ‘cut’ (这把刀切肉 zhè bǎ dāo qiē ròu ‘this knife cuts meat’), and potentially one explicit event that represents the possible result of the patient as 碎 suì ‘broken’ in 切碎 qiē suì ‘cut-broken’. Specifically, when the predicate contains the result complements, there is an explicit resultative event denoted by 碎 suì ‘minced’ (肉碎了 ròu suì le ‘meat is minced’). In terms of whether the resultative event is syntactically expressed or not, the number of total events in the intermediary marked instrumental construction is 2 or 3. Therefore, the “E/TER” is 100 %.

2. The facilitating marked instrumental construction also includes two certain explicit events, namely, instrument-manipulation event designated by 用 yòng (我用筷子 wǒ yòng kuàizi ‘I use chopsticks’) and act-on event designated by 吃 chī ‘eat’ (我吃饭 wǒ chī fàn ‘I eat meals’), and potentially one explicit event that represents the possible outcome beared by the patient, i.e. 完 wán ‘finish’ (饭吃完了 fàn chī wán le ‘meals are finished’). The number of total events in facilitating marked instrumental construction is either 2 or 3, and all involved events will be explicitly expressed. Therefore, the “E/TER” is 100 %.

3. The patient-oriented marked instrumental construction includes two explicit events, namely 油墨印课文yóumò yìn kèwén ‘printing ink is used to print texts’ and 课文被印在很粗糙的纸上 kèwén bèi yìn zài hěn cūcāo de zhǐ shàng ‘texts are printed on coarse paper’, where the former is an action-on event denoted by 印 yìn ‘print’ and expressing the relation between instrument and patient, while the latter is a resultative event denoted by 在 zài ‘at’ and undertaken by the patient. Besides the 2 events mentioned above, another is an implied instrument-manipulation event (某人/机器操控油墨 mǒurén/jīqì cāokòng yóumò ‘someone/some machine manipulates printing ink’) because 油墨 yóumò ‘ink’ cannot perform the printing action itself without the involvement of any other forces. Therefore, the number of total events in this construction is 3 and the “E/TER” of patient-oriented marked instrumental construction is 66.7 %.

4. The intermediary instrument-subject construction includes one explicit event, i.e. the act-on event designated by 做 zuò ‘make’ (那台缝纫机做了三百套衣服了 nà tái féngrènjī zuò le sān bǎi tào yīfú le ‘that sewing machine has made three hundred suits of clothes’). When the predicate 做 zuò ‘make’ is followed by a resultative complement 坏‘ruined’, this construction can also contain another explicit event (衣服做坏了 yīfú zuò huài le ‘the clothes are ruined during the process of making’), namely the resultative event. Meanwhile, there also exists an implicit event, namely the instrument-manipulated event indicating that someone manipulates or starts that sewing machine, because 缝纫机 féngrènjī ‘sewing machine’ cannot start itself going. Therefore, the number of total events is 2 or 3 which corresponds to 1 or 2 explicit events repectively, and the “E/TER” of intermediary instrument-subject construction is 50 % or 66.7 % correspondingly.

5. The facilitating instrument-subject construction includes one explicit event and one implicit event. The former is an act-on event denoted by 喝 hē ‘drink’ (喝酒 hē jiǔ ‘drink wine’), while the latter is an implicit instrument-manipulation event implied by the relation between a potential agent and 大碗 dà wǎn ‘big bowl’ (某人用大碗 mǒu rén yòng dà wǎn ‘someone uses the big bowl’) because it is someone rather than a big bowl that performs the action of drinking wine. Therefore, the “E/TER” of facilitating instrument-subject construction is 50 %.

6. The patient-oriented instrument-subject construction includes two explicit events, namely an act-on event designated by 捅 tǒng ‘stab’ (刺刀捅 cìdāo tǒng ‘The bayonet stabbed’) and a resultative event designated by 弯 wān ‘bent’ (刺刀弯了 cìdāo wān le ‘The bayonet bent’), and one implicit event, namely the instrument-manipulated event (某人用刺刀 mǒurén yòng cìdāo ‘someone used the bayonet’). Therefore, the “E/TER” of patient-oriented instrument-subject construction is 66.7 %.

7. The intermediary instrument-object construction includes one explicit act-on event denoted by 捅 tǒng ‘stab’ (刀子捅 dāozi tǒng ‘the knife stabbed something’) and one implicit instrument-manipulated event (耶尔马兹用刀子 yé-ěr-mǎ-zī yòng dāozi ‘Yilmaz used the knife’) implying that 刀子 dāozi ‘knife’ is manipulated by the agent 耶尔马兹 yé-ěr-mǎ-zī. Therefore, the “E/TER” of intermediary instrument-object construction is 50 %.

8. The facilitating instrument-object construction involves one explicit act-on event denoted by 吃chī ‘eat’ (我吃某物 wǒ chī mǒuwù ‘I eat something’) and one implicit instrument-manipulated event implied by the relation between 我 wǒ ‘I’ and 大碗 dà wǎn ‘big bowl’(我用大碗 wǒ yòng dà wǎn ‘I used a big bowl’). Therefore, the “E/TER” of facilitating instrument-object constructions is 50 %.

Note that the intermediary marked instrumental construction, the facilitating marked instrumental construction, and the intermediary instrument-subject construction can include a resultative component because all three of them have an agent or secondary agent that can act on a patient and thus cause a resultative to the patient. Whereas, the facilitating instrument-subject construction, and the facilitating instrument-subject construction and the intermediary instrument-subject construction are usually idioms. Take 大碗喝酒 dà wǎn hē jiǔ ‘drink wine with a big bowl’ for example, it describes a recurring scenario and thus is conventionalized as an idiom to express the bold and forthright character or action. It is true that it may be grammatical in some concrete senario, but it is an accidental example and strictly limited to the context. A search in the Modern Chinese corpus of CCL corpus with “大碗喝$2酒” as a query retrived 417 items, all of which are directly followed by nouns rather than resultative complements.^[8] Therefore, such cases are not within our scope of consideration. By analogy with the “Unique Path (UP) Constraint” (Goldberg 1995: 82), a physical argument X cannot be predicated to move to two different locations at the same time, we conclude that it is impossible to have two results on the same patient in one clause. Both patient-oriented instrument-subject construction and patient-oriented marked instrumental construction have already had a result, they cannot include another resultative complement.

On the basis of two principles proposed in Section 3, the distinction in the degree of event integration of Chinese instrumental constructions is arrived at. Based on principle 1 that the more events if a construction includes, the higher the degree of event integration is, we arrived a continuum of three gradients, as shown in Table 3. It shows that among eight types of instrumental constructions, some integrate two events, some three events, and some two or three events. Furthermore, when two instrumental constructions include the same number of total events, we turn to “E/TER”. That is, the smaller the value of “E/TER”, the higher the degree of event integration. Therefore, according to the values of “E/TER” in Table 2, we can sequence instrumental constructions in every gradient, as demonstrated in Table 3.

Essentially, the degree of event integration means the degree of synthesis. Chinese instrumental construction reflects the process that two or three events with semantic relations between them can merge into a complex event and be syntactically represented by a clause. We have discussed what kinds of events can be integrated together in Chinese instrumental constructions, and now we move on to the semantic relations between events. Specifically, the semantic relation between events can be divided into three types: (1) When the instrument is intermediary, it can form both the “instrument-manipulation event” with the agent and the “act-on event” with the patient, where the causative relation between the two events is the semantic foundation of event integration. (2) When the instrument is facilitating, it can form the “instrument-manipulation event” with the agent, and the agent and the patient form the “act-on event”. In such a case, the semantic relation between events is embodied in the facilitating of the former to the latter. (3) Chinese instrumental constructions may also concern the “resultative event”, like the patient-oriented instrument-subject construction. Therefore, the third relation between events is causation, which exists between the “act-on event” and the “resultative event” and it is the instrument or patient that undertakes the result. To summarize, the “instrument-manipulation event” can cause or facilitate the “act-on event” and the “act-on event” can cause a “resultative event”.

A closer look at Table 3 suggests that marked instrumental constructions are distributed over the first two gradients with patient-oriented marked instrumental construction in the first gradient and intermediary marked instrumental construction and facilitating marked instrumental construction in the second gradient. Two instrument-object constructions only appear in the third gradient, and instrument-subject constructions are scattered over all three gradients with patient-oriented instrument-subject construction in the first gradient, intermediary instrument-subject construction in the second gradient and facilitating instrument-subject construction in the third gradient. By corresponding the distributions of instrumental constructions in gradients and the continuum of the degree of event integration of three gradients, as shown in Figure 3, we proposed that the event integration compactness of eight types of Chinese instrumental constructions can be generally reduced to Figure 4.

Figure 3:

The continuum of degree of event integration of three gradients.

Figure 4 shows that the marked instrumental construction, the instrument-subject construction and the instrument-object construction generally form a continuum of degree of event integration with an increasing trend from left to right. The previous research on these three types of instrumental constructions displays that the decrease in productivity from left to right: the most productive one is the marked instrumental constructions, which can be used for most instrumental expressions, and many unmarked instrumental constructions can be converted into marked instrumental constructions. The productivity of the instrument-subject constructions ranks second, among which the intermediary instrument-subject constructions highlight the agentivity of instruments, while the patient-oriented instrument-subject constructions highlights the patientivity of instruments. And the facilitating instrumental constructions profiling the facilitating function of instruments to the event is quite idiomatic, so its productivity is weak. Instrument-object constructions are mostly idiomatic expressions, for example, 捅刀子 tǒng dāozi ‘stabbing with a knife’ is acceptable, but 捅棍子 tǒng gùnzi ‘stabbing with a stick’ is not. Essentially, the productivity of constructions is based on language use, and its high productivity is a manifestation of its high frequency of use.

Figure 4:

The general continuum of degree of event integration.

Furthermore, this study proposes that the degree of event integration is proportional to the productivity of instrumental constructions, which is in line with the view that “there is a fundamental and recurrent category of complex event that is prone to conceptual integration and representation by a single clause, a type here termed a macro-event” (Talmy 2000: 216). By corresponding the high productivity and frequency of uses of marked instrumental constructions to the high degree of event integration of them, we conclude that Chinese tends to use more compact instrumental construction. This finding accords with Li (2020)’s view that Chinese belong to the macro-event language. That is, despite the fact that all eight types of instrumental constructions fall into macro-events, the more compactly integrated ones are preferred.

5 Implications of Chinese instrumental constructions on event integration

Conceptually, event integration is a cognitive process. It is under the operation of event integration that instrument-manipulation event and act-on event are integrated together so as to form Chinese instrumental constructions. Meanwhile, this process also is accompanied by some syntactic and semantic conditions because it is impossible to integrate any events together. It is therefore reasonable to say that event integration can be interpreted at the conceptual, syntactic and semantic levels. This study starts with the event integration and explores the degree of event integration of Chinese instrumental constructions macroscopically. This section moves on to the microscopic parts of event integration, such as the argument integration and predicates and so on. It’s found that this study displays implications on event integration from the following 4 aspects.

6 Conclusions

Under the framework of event integration, this study devotes to the family of Chinese instrumental construction and examines the similarities and differences among different instrumental constructions. Conversely, the study on the family of Chinese instrumental construction also contributes theoretically to the event integration theory.

Concretely, all constructions in the family of Chinese instrumental construction share similarities but also exhibit differences. In terms of similarities, all of them conceptually includes two events at least, i.e. the instrument-manipulation event and the act-on event with a semantic relation (facilitating or causative) between them, but syntactically are represented by one clause. Based on these similarities, we propose that all Chinese instrumental constructions fall into macro-events and can be explained by event integration. In terms of the differences, it is found that all Chinese instrumental constructions form a continuum of degree of event integration (see Figure 3 and Table 3). By corresponding the productivity and frequency of uses of instrumental constructions to the degree of event integration of them, we conclude that despite the fact that all eight types of instrumental constructions fall into macro-events, the more compactly integrated ones are preferred.

Theoretically, this study makes significant advances in event integration. Firstly, we propose a series of new concepts, such as the instrument-manipulation event versus the act-on event, the explicit event versus the implicit event, the “E/TER”, so as to give a unified description of Chinese instrumental constructions. Secondly, it is found that all Chinese instrumental constructions meet three conditions of macro-events, thus suggesting that the research scope of macro-events can be extended to the instrumental constructions. Thirdly, in view of the pecularity of instrumental constructions, we innovatively propose two principles to examine the degree of event integration, which can offer references for studies on instrumental constructions in other languages. Fourthly, the study on the family of Chinese of instrumental constructions also contributes to the process of event integration in four ways: 1) Event integration can be divided into the single-chain integration and the double-chain integration. 2) Besides explicit argument integration, the implicit argument integration is also a hub of event integration. 3) Event integration involves not only explicit events but also implicit events. 4) The mismatch between arguments and predicate is essentially one of the linguistic representations of event integration.

In conclusion, the definition of instrumental constructions and the concept of instrumental macro-events are clearly put forward in this study, which makes advances in the research of Chinese instrumental constructions. It is also the first attempt to clarify the similarities and differences among constructions in the family of Chinese instrumental constructions from the perspective of event integration. More importantly, this study theoretically supplements the integration types and parameters of event integration and extends the research scope and explanatory power of event integration. However, it is only the first attempt to combine Chinese instrumental construction with event integration and thus more detailed studies on this combination can be carried out in the future. For one thing, this study is only a qualitative analysis, so future studies can be corpus-based with more data so as to test and supplement findings in the present study. For another, this study only zooms in on eight types of instrumental constructions, but more peripherial instrumental constructions or the constructions related to instrumental constructions, for example, the topic instrumental construction mentioned in Section 1, can be studied so as to sketch the family of instrumental constructions and its surroundings.