Rang clauses in modern mandarin Chinese: a Cardiff Grammar approach

3.1. Basic syntactic categories and relationships in the Cardiff Grammar

Three basic syntactic categories and four relationships are recognized in the CG (Fawcett 2000: 187–260, 2008, 73–75). As represented in Figure 1, the three basic categories are unit, element, and item, and the four basic relationships are componence, filling, exponence, and conflation.⁸ In Figure 1, at the top of the tree diagram is the unit of the clause, which in this example is composed of five elements, i.e. Subject, Operator/Auxiliary, Main Verb, Complement, and Adjunct. The Subject here is filled by a nominal group, which is composed of the element of head only and expounded by the item Ivy. The element of Operator is conflated with Auxiliary and together expounded by the item will. The Main Verb is expounded by the item wash. The clause element of Complement is filled by a nominal group, which is composed of another two elements, i.e., a deictic determiner (expounded by the item her) and a head (expounded by the item hair). The last clause element is an Adjunct, which is filled by a nominal group too, which again is composed of another two elements, a deictic determiner (expounded by the items this) and a head (expounded by the item evening).

Figure 1.

The basic categories and relationships of syntax in the Cardiff Grammar.

As we can see from Figure 1, the CG has abolished the Hallidayan “multiple structures,” and adopted a single, integrated structure in its syntactic representation. The multifunctional nature of language is displayed in the representation of the meanings. Fawcett claims that it is the role of syntax to show the integration of these intermittent “strands of meaning” in a single structure (Fawcett 2000, 147). Thus, for example, the experiential meaning of “Ivy will wash her hair” is realized through the three elements of Subject, Main Verb, and Complement that express the choices in TRANSITIVITY, and the interpersonal meaning of “information giver” is realized through the two elements of Subject and Operator that express the choice of MOOD.

3.2. Basic principles of syntactic description in the CG

Generally, two principles are maintained in the CG’s syntactic description: (1) the principle of function-orientation; and (2) the principle of meaning-centeredness. The principle of function-orientation means that the perspective of syntactic description is functional rather than formal. This is determined by the nature of SFL, which is functional in orientation rather than formal (Halliday 2006, 443–444). Thus, when functional and formal criteria are in conflict, functional criteria are placed above formal criteria as long as they do not lose contact with the level of form (Fawcett 2008, 111). Let us compare example (5) and (6).

(5) Ivy gave Ike money.

(6) Ivy gave money to Ike.

According to Quirk et al. (1985, 59), example (5) has the structure of “S + V+ O_i+O_d,”⁹ that is, Ike is an Indirect Object and money is a Direct Object; while example (6) has the structure of “S + V + O + A,” that is, money is still an Object, but to Ike is an Adjunct. In the CG, both (5) and (6) will have the same functional structure, i.e., “S + M + C + C.”¹⁰ This is because the CG maintains the rule that a Participant Role (PR) is a role that is “expected” by the Process (Fawcett 2008, 137). Since the Process of “give” in both examples is the same (i.e., “someone gives something to someone”), it will predict the same PRs. Thus, the PR played by Ike in both examples will be the same, no matter if it appears in the form of Ike or to Ike. This does not mean that the two examples have the same meaning. The difference between them can be explicated informationally. In normal context, the information focus in (5) falls on money, which may answer the question of “What did Ivy give Ike?”; while in (6), it falls on Ike, which may answer the question of “Who did Ivy give money to?” Therefore, from the CG’s perspective, we can claim that the traditional grammar’s way of analyzing to Ike into an Adjunct not an Object is depending on form rather than function, since in functional description of a language, the fact that a PR that is predicted by the Process is expressed as a prepositional group (in this case, to Ike) is not strong enough to turn a PR into a Circumstantial Role (CR).¹¹

The principle of meaning-centeredness means in syntactic description the starting point is semantics or meaning rather than syntax itself. This principle is also determined by the nature of SFL, which is semantic rather than syntactic in orientation (Halliday 2006, 443–444), since SFL is a theory of language as choice between meanings. However, this does not mean that the syntax of a language is not to be accounted for in a functional linguistic theory, since the form level of syntax is most directly observable (Fawcett 2008, 15). To illustrate this principle, let us look at the CG’s syntactic analysis of example (7) and (8).

(7) He raised the problem again.

(8) He brought up the problem again.

According to traditional grammar, up in example (8) will be analyzed as Adjunct. This analysis is to a great degree based on form not meaning. We know that brought up in (8) means virtually the same thing as raise in (7), except that they may be used in different registers. From the CG’s perspective, brought up in (8) realizes the same Process of “raising” in (7), because semantically they expect almost the same PRs. The difference between them is that in example (8) a single Process (at the level of meaning) is realized in more than one word (at the level of form). In order to solve this “phrasal verb” problems in traditional grammars, Fawcett (2008, 184) proposed another clause element, i.e., Main Verb Extension (MEx), which functions as an “extension” of the Main Verb (M). This clause element together with the Main Verb jointly expresses a Process. Thus, in example (7), the Process is typically expressed in the M raised; while in (8), the same Process is jointly expressed in the M brought and MEx up. Thus, from the CG’s perspective, up is not an Adjunct but a MEx realizing part of the Process.

It is not always easy to draw a distinct line between the two principles mentioned above. What the two principles remind us is that more attention of meaning has been maintained in the CG’s syntactic analysis, because meaning is realized by form and syntactic analysis serves for meaning analysis. In the ensuing section, the syntax of rang clauses will be investigated from the CG’s perspective, and we will see whether typical insights into such clauses that can be derived from an analysis by drawing on the CG.

4.1. The syntax of rang clauses expressing “concession”

This “Concession” usage of rang is the most common one in the corpora, accounting for 46% of the samples. It is often used in situations like competition, argumentation, and entertainment, etc., which means giving advantages to others such as tui rang (退让, concede) and qian rang (谦让, yield), or avoiding something such as bi rang (避让, avoid), etc. Syntactically, this usage of rang is usually followed by a nominal Complement such as rang lu (让路, give the way to), rang zuo (让座, give up one’s seat to), and sometimes followed by two Complements such as rang ta yi ci ji hui (让他一次机会, give him an opportunity). As reviewed in section 2, a consensus has been reached regarding the grammatical function of this rang, i.e., it is a verb, which serves the function of the Predicator. Since this kind of usage of rang seldom causes disputes in the literature, no more discussion will be elaborated. And we just analyze the syntax of this kind of rang clause from the CG’s perspective in order to show how the structure will be represented functionally. Since the rang clause of example (1) mentioned in the introduction section is a typical one expressing “concession” meaning, we will not seek new examples for our illustration (see Figure 2).¹²

Figure 2.

The analysis of rang clauses expressing “concession”.

As we can see from Figure 2, we analyze rang in (1) as a Predicator, which expresses Process meaning. From the perspective of transitivity,¹³ the Predicator rang in the clause realizes an “action process,” which expects two PRs. Thus, the Subject, which is filled by a nominal group and expounded by the item ni, is conflated with the PR of Agent; the Complement, which is filled by a nominal group and expounded by the item ta, is conflated with the PR of Affected. In this clause, zhe is an Auxiliary, marking present tense; and dian er is a quantity group, which fills the Adjunct of the clause.

4.2. The syntax of rang clauses expressing “causation”

This kind of rang usage is also used frequently, taking 34% of the data collected. Here “causation” is used in a broad sense, including meanings like “allowance,” “tolerance,” and “permission” etc. As reviewed in section 2, the causative meaning and its related syntax of rang have been the subject of much controversy in the literature. Some have treated rang as a transitive verb (Li 2007, 27–29), a common verb (Ding et al. 1999, 100), a bifunctional verb (Lv 1999, 461–462), a light verb (Shen, He, and Yang 2001), and there are still others who analyze it as a preposition (Zhu 1999:182; Zhang 1980; Fan 2000). Because these scholars assign different grammatical functions to this rang, various syntactic structures have been offered for different purposes. Taking example (2) to illustrate, let us comment the analyses conducted in the literature to see whether they have consistently matched semantic features to syntactic elements.

First, we deny the view that rang, in this case, is a preposition, since we can still find lexical meanings of “causation” in it. If it is indeed a preposition, it will be categorized as a function word (see Wang et al. 2006, 281). But from our analysis of the corpus data, the causative meaning is till obvious to see. Thus, although the meaning of causative rang is grammaticalized to some degree, it is still too radical to treat it as a preposition. We thus argue that rang in this sense is still a verb as mostly agreed in the literature. Let us first comment Li’s structural analysis. In Li (2007, 27–29), example (2) has the structure of “S + P + O+ O_C.” Semantically, this analysis means that the nominal group ren is the object that is caused or affected by the process of “rang,” and the event of “chong man nuan yi” is used to supply more information about the resulting state of ren. From what is expressed in example (2), we know that what the agent zhe yan se caused is not just a particular target ren, but a whole event. We can only ask questions like “what is caused by this color?” but not “what is caused to the people by this color?” In other words, this example means “someone or something causes an event,” not “someone or something causes someone.” Thus, Li’s analysis to some degree does not match semantic features to syntactic elements. Let us see Lv’s analysis of treating rang as a bifunctional verb. His analysis means that the nominal group ren is of two roles, i.e., it is both the Object of rang and the Subject of the following verb chong man. However, from what just mentioned above, ren chong man nuan yi forms an independent event, which means that ren should not be the Object of rang. We can say “ren xin zhong chong man nuan yi (people feel warm in the heart),” but cannot say “zhe yan se rang ren (this color makes people).” Therefore, Lv’s analysis of this example is also not based on meaning but to some degree on form. Compared to the two analyses just discussed above, the light verb analysis proposed by Shen, He, and Yang (2001) and others (He and Wang 2002; He 2004) is to some degree acceptable, since it treads the VP as a whole event. However, as we know from functional perspective rang does not behave like a light verb, since it still carries content meaning of “causation.”

From the CG’s perspective, the so-called “causative construction” is a type of “influential process” in transitivity in the sense that “someone or something causes some event.” According to Fawcett (2011) and He (2014), an “influential process” is a Process that influences another Process, i.e., it is a “process-influencing” Process. It is “influential” in that it seriously affects our interpretation of the process in the dependent event, i.e., the process in the clause that fills the phenomenon. From the CG’s perspective, a causative verb often has the Process meaning of “someone causes something,” that is, a causative verb normally expects two PRs. One is the causer, and the other is the event that is caused by the causer. The CG assigns these two PRs as Agent and Phenomenon, respectively, where the Phenomenon is often expressed by an embedded clause. Thus, taking example (2) to illustrate, the syntax of rang clauses expressing “causation” can normally be represented in Figure 3 below.¹⁴

Figure 3.

The analysis of rang clauses expressing “causation”.

As Figure 3 shows, we analyze the causative rang as a Predicator. The Predicator expects two PRs, where the PR of Agent is conflated with the Subject, and the PR of Phenomenon is conflated with the Complement. In this clause, the Subject is filled by a nominal group, which is composed of two elements: one is the deictic determiner expounded by the item zhe, and the other is the head expounded by the item yan se. The Complement is filled by another clause. This embedded clause is composed of four elements, where the Predicator expresses a relational Process, which expects two PRs: the Carrier and the Possessed. The Carrier is conflated with the Subject ren, and the Possessed is conflated with the Complement nuan yi.

Our analysis of the causative rang and its related syntactic structure is based on the meaning of rang. From our analysis, this kind of rang clause often consists of three syntactic elements, i.e., Subject, Predicator, and Complement. The nominal group following rang should not be regarded as the Complement of the matrix clause, but as the Subject of the embedded clause. This kind of analysis is thus more function-oriented and meaning-centered.

4.3. The syntax of rang clauses expressing “bei”

This passive “bei” usage of rang accounts for 12% in the samples chosen. As described in section 2, there are mainly three views proposed in the literature regarding the grammatical function of this passive rang: a preposition (Lv 1999: 462; Li 2001, 219), a sub-verb (Ding et al. 1999, 95) and an auxiliary verb (Wang 2002, 80). And because of the different views concerning it, different structures are assigned to this kind of rang clause. We agree with the first view that the passive rang is a preposition. We know that in Chinese language most prepositions are derived from verbs in the process of semantic change or grammaticalization. Since semantically this “bei” usage of rang is completely different from its original meaning of “concession” and “causation,” and its only function in the clause is to introduce the Agent of the action, this kind of usage should be categorized differently. However, we do not agree with the corresponding syntactic analysis of analyzing rang together with its following nominal group as an Adjunct, since this kind of analysis is completely based on form rather than function.

Let us take example (3) to illustrate. In the clause, semantically it is generally agreed that da feng plays the role of Agent. If we agree that it is the Agent, it means that it is a PR instead of a CR. Since a PR is typically conflated with Subject or Complement, we should not analyze rang da feng as an Adjunct. Thus, treating rang da feng as an Adjunct of the clause is a kind of form-centered analysis. In other words, although the PR of Agent predicted by the Process of “chui” is expressed as a prepositional group, it is not sufficiently strong to turn a PR into a CR, because rang has no meaning in its own right. It is merely the passive marker that is introduced automatically whenever a PR that typically occurs as the second (or third) PR in a clause is “promoted” to be the Subject Theme. Thus, rang is not there as the result of a choice in a system to have a meaning realized in rang, but merely as a by-product of the decision to make “chuang hu” the Subject Theme. According to the CG, the functional syntax of example (3) can be represented in Figure 4 below.¹⁵

Figure 4.

The analysis of rang clauses expressing “bei”(1).

As we can see from Figure 4, we analyze chui as the Predicator. Semantically, this Predicator expresses a relational process, which expects three PRs. The Subject, which is filled by the nominal group yi shan chuan hu, is conflated with the PR of Affected-Carrier.¹⁶ The first Complement, which is filled by a prepositional group rang da feng, is conflated with the PR of Agent. More specifically, this Agent is conflated with the completive of the prepositional group. The second Complement, which is filled by a quality group huai, is conflated with the PR of Attribute. In this clause, le is an Auxiliary, marking past tense.

Our analysis of the prepositional group of rang da feng in this case as a Complement rather than an Adjunct have maintained the basic principles of “functional orientation” and “meaning centeredness” in functional description of the syntax. This is because the PR played by “da feng” in the event of “chui” is inherently unchanged, although it happens to be appeared in the form of a prepositional group. This tells us that not all prepositional groups are to be analyzed as Adjuncts. According to the CG’s criteria, only 32% of prepositional groups fill Adjunct, whereas 52% of them fill Complements (Fawcett 2008, 111).

One more point to mention, when discussing the connections and differences between rang and bei, Lv (1999, 463) and Li (2001, 219–223) argue that when bei is used before a solid word (content noun), it is a preposition, when it is used before a verb, it is a passive auxiliary verb. That is to say, bei in bei da (被打, be hit) and bei jie shou (被接受, be accepted) should be treated as auxiliary verbs. Lv claims that rang does not have this kind of usage. However, through careful examination of the three main Chinese corpora, we find that this usage of rang is indeed very infrequent, but there are some examples (see example (9)).

According to Lv’s (1999, 463) and Li’s (2001, 219–223) idea, rang in example (9) may have similar qualities and should be treated as passive auxiliary verbs. This kind of analysis is clearly influenced by the analysis of English passive constructions. In an English passive construction, it typically has the form of “be+V-ed,” where “be” is a passive Auxiliary. But in Chinese language, this passive “be” does not exist. Thus, from functional perspective, we argue that this kind of analysis is putting form before function. From example (3), we know that the function of rang in passive usage is to introduce the Agent of the action and this rang function has not been changed in example (9). Although the Agents of the actions in example (9) are not shown at the level of form, they are still recoverable from the context, that is, still semantically existed. This kind of linguistic phenomenon is well-known in passive constructions, where sometimes the doer of the action is omitted for different discourse purposes. From the CG’s perspective, this kind of omitted Agent is a kind of covert PR. Therefore, functionally rang in example (9) is no different from example (3), which should also be analyzed as a preposition. The only difference between them is that the Agent introduced by rang in example (9) is covert. Thus, taking example (9b) to illustrate, the syntax of rang clauses of this kind can be represented in Figure 5 below.

Figure 5.

The analysis of rang clauses expressing “bei” (2).

As illustrated in Figure 5, in this clause, the Predicator expounded by the item sha expresses an Action Process, which expects two PRs, i.e., “someone kills someone or something.” The Subject, which is filled by a nominal group chai yi, is conflated with the role of Affected. The Complement is filled by a prepositional group, where the covert PR of Agent is conflated with a covert completive.¹⁷ In this clause, there are some other syntactic elements. The item dou expounds the Adjunct, and le expounds the Auxiliary marking past tense. Our analysis of rang in this case still as a preposition not an Auxiliary has maintained the principles established in the CG’s functional syntactic description, i.e., it is based on meaning and function, not on form.

4.4. The syntax of rang clauses expressing “wishes”

Here the meaning of “wishes” is also used in a broad sense, including meanings like “appealing,” “calling for” and “proposing.” Examples like Rang wo men tuan jie qi lai! (让我们团结起来！Let’s get united!), Rang wo zai shuo yi bian. (让我再说一遍, Let me repeat again!), and Rang ta qu ba! (让它去吧, Let it go.) are all included in this category. In our analysis of the corpus data, this kind of rang usage accounts for 8% of the corpus data. As example (4) shows, rang with this usage does not have the meaning of “concession” and “causation,” and does not have the meaning of “bei” either. As reviewed in section 2, this kind of rang usage has not attracted the attention from scholars in the Chinese linguistic field. Very little literature can be referred to. We only find that Lv (1999, 462) treats it as a verb and He et al. (2015, 58) regard it as a Let Element, but no explanations are given. Semantically, rang in this kind of clause has been deprived of lexical meaning, and is also different from the prepositional usage. Syntactically speaking, no syntactic elements can be inserted before rang, i.e., we cannot insert a Subject for rang in this kind of clause (*ni rang wo men yong yuan zai yi qi is ungrammatical). So, functionally, rang in this kind of clause does not behave like a verb or a preposition, which is only a Mood marker expressing “wishes.” There are similar issues in English (see example (10)).

(10) a. Let’s go.b. Let Ivy eat it!

In example (10), let is deprived of lexical meaning, and also used as a Mood marker of the clauses. Functionally, Fawcett (2008, 162) proposes that let in example (10) should not be analyzed as Operator or Main Verb; instead it should be regarded as an independent element of the clause, i.e., Let Element (abbreviated as L).¹⁸ We claim that the CG’s recognition of Let Element is theoretically significant, which has enriched the structural elements in realizing MOOD meaning. More importantly, this Let Element will solve many practical problems in functional syntactic description of some linguistic phenomenon as well. With this Let Element, the syntactic analysis for the so-called let-construction will be more compatible with the grammatical properties the construction displays, and the semantic functions to syntactic elements of the construction will be more systematically and consistently matched.

Adopting Fawcett’s (2008, 162) and Xiang and Liu (2018) analyses of “let’s clauses” in English, we claim that rang in this kind of usage should also be promoted to be the direct element of the clause, since it is uniquely used in modern mandarin Chinese and its only function is to be a Mood marker expressing “wishes.” Thus, the functional syntax of this kind of rang clause can be illustrated in Figure 6. below.

Figure 6.

The analysis of rang clauses expressing “wishes”.

As is shown in Figure 6, we analyze rang in this clause as a direct element of the clause, i.e., Let Element. In this clause, the Predicator expounded by the item zai and Predicator Extension expounded by the item yi qi together express an Action Process, which expects one inherent PR, i.e., the Agent. The Subject, which is filled by an nominal group and expounded by the item wo men, is conflated with the PR of Agent. In this clause, yong yuan is a quality group, which fills the Adjunct of the clause. Based on the CG, our categorization of this new syntactic element of rang, i.e., Let Element, has enriched the structural elements in realizing MOOD meaning in Chinese and solved the practical problems of treating it as a verb in the literature. With this Let Element, the syntactic analysis is more compatible with the grammatical properties this usage of rang displays. Thus, the semantic function to syntactic element of this rang clause is more systematically and consistently matched.