Grammatical functions and transitivity
Book: Cognitive Linguistics an Introduction
Author: Vyvyan Evans and Melanie Green
Page: C17-601
Recall the prototypical action chain model that we introduced in Chapter 15. According to this model, the prototypical TR is the ‘energy source’ and the prototypical LM is the ‘energy sink’. Langacker (2002: 208) describes the prototypical action in terms of what he calls the ‘billiard-ball model’. This idea relates to the fact that we experience motion, that motion is driven by energy, and that while some entities have an inherent capacity for energy, other entities only receive energy from external entities. The ‘billiard-ball’ metaphor expresses the idea of energy transfer from one entity to another.
According to this model, energy is transferred from AGENT to PATIENT, and results in a change of state for the PATIENT. Figure 15.18 is repeated here as Figure 17.4. According to Langacker, the unmarked active transitive clause with third person participants represents the prototypical action from a canonical view point or perspective. This means that a sentence like George tickled Lily represents the prototypical action.
Grammatical functions and the transitive clause
This characterisation of the transitive clause has implications for how the grammatical functions subject and object are viewed in Cognitive Grammar. Although we have taken the terms ‘subject’ and ‘object’ largely for granted in the discussion so far, grammatical functions are a rather controversial issue in the debate between cognitive and formal approaches. Briefly, in rather the same way that formal approaches reject a semantic characterisation of word classes, they also reject a semantic characterisation of subject and object. Instead, these core grammatical functions are described in terms of distribution (for example, in English, the subject precedes the main verb and the object follows it) and morphology (for example, the subject pronoun is marked for nominative case while the object pronoun is marked for accusative case). We introduced the structural criteria for English subjects and objects in Chapter 14 (see Tables 14.2 and 14.3).
In contrast, as we mentioned earlier in the chapter, the grammatical functions subject and object (like the word classes) receive a schematic semantic characterisation in Cognitive Grammar. In the prototypical action chain, the subject, which elaborates the schematic TR of the verb, is characterised as the volitional energy source. The object, which elaborates the schematic LM of the verb, is the passive energy sink. The transfer of energy between the participants in a scene is described in terms of the action chain model. As we saw in Chapter 15, different participants in this action chain can be profiled, which has consequences for how the clause is structured. Consider the examples in (18), which reflect different construals of the same scene.
The act of igniting love letters involves an AGENT (George), a PATIENT (the love letters) and an INSTRUMENT (a match). In example (18a), each component of this action chain is profiled. The energy is transferred from the AGENT, George, via the INSTRUMENT, a match, to the PATIENT, the love letters. In (18b), on the other hand, only the INSTRUMENT, a match, and the PATIENT, the love letters, are profiled. Despite this, the AGENT is understood as part of the base (or scope of predication) of (18b) because we know that matches generally lack the inherent energy required for independent action. In (18c), only the PATIENT is profiled, but nevertheless the AGENT and the INSTRUMENT are understood as part of the base. The action chain that underlies all these clauses can be represented as in Figure 17.5, where the circles represent each of the participants and the arrows represent the transfer of energy.
As we have seen, the difference between the clauses in (18) concerns which elements of the action chain are profiled. Langacker represents profiling in these diagrams with bold type. For example, because all the participants in the action chain are profiled in (18a), all parts of the diagram are in bold (Figure 17.6). Examples (18b) and (18c) are represented by Figures 17.7 and 17.8 respectively.
As these diagrams show, the subject of the clause in each case is the participant that is closest to the energy source out of the participants profiled. In other words, when the active clause profiles both AGENT and INSTRUMENT as well as PATIENT, the AGENT (as energy source) will be subject. When the clause profiles only INSTRUMENT and PATIENT, the INSTRUMENT will be subject and so on. A number of researchers, including Fillmore (1968), have proposed a thematic hierarchy, which makes predictions concerning the likelihood of a given semantic role occurring as subject of a clause. The hierarchy can be understood as a prototypicality scale, with prototypical subjects on the left and less proto typical subjects on the right. An example of a simple thematic hierarchy is given in (19). This hierarchy predicts that if a language permits any given semantic role in subject position (e.g. INSTRUMENT), it will also allow every semantic role to the left in subject position (PATIENT, BENEFACTIVE and AGENT).
Of course, not all researchers agree about how many semantic roles there are, nor about what labels they should be given. Another point of debate concerns whether semantic roles should be viewed as semantic primitives or as discrete and bounded categories. Dowty (1991) is among those researchers who have attempted to explain semantic roles in terms of a PROTOTYPE model rather than as a set of discrete categories. Dowty (1991: 572) proposes an AGENT proto-role and a PATIENT proto-role. Each proto-role is characterised not by a set of necessary and sufficient conditions, but by a set of properties that a prototypical AGENT or PATIENT will display, whereas less prototypical AGENTS or PATIENTS will display only a subset of these characteristics. The prototypical AGENT is characterised by volition, sentience and movement and by causing a change of state to be effected in relation to another participant. In contrast, the prototypical PATIENT is characterised by being stationary relative to another participant, by being causally affected by another participant and by undergoing a change of state that may be incremental in nature. As the discussion above illustrates, Langacker’s model captures the predictions stated by Fillmore’s (1968) thematic hierarchy in terms of the ‘billiard-ball’ or transfer of energy model.
Of course, the examples we have seen so far encode physical acts and there fore energy transfer. These examples therefore lend themselves well to illustrating the ‘billiard-ball’ model. As Langacker observes, however, not all clauses can be so easily characterised in terms of energy transfer. Consider the examples in (20).
In these clauses, the subject she is not an AGENT but an EXPERIENCER. This semantic role describes a conscious and sentient participant who participates in mental or emotional rather than physical activity. Verbs of perception and cognition therefore have EXPERIENCER subjects. We can describe the object (George; his lovely blue eyes; him) as the STIMULUS. There is no sense in which the subject of these clauses acts with volition or transfers energy in the direction of the object. Despite this, Langacker (2002: 221) suggests that these clauses display the same asymmetry found in clauses describing the prototypical action. While the asymmetry in an action chain arises from the direction of the energy flow, the asymmetry in the EXPERIENCER-STIMULUS relation arises from the fact that the EXPERIENCER is conscious and sentient and is thus responsible for establishing mental ‘contact’ with the STIMULUS by creating a cognitive representation of the experience.
However, some clauses do not encode this asymmetry in terms of ‘directionality’ (either in terms of energy or ‘mental contact’).
These clauses are stative. Furthermore, the participants in the relations that they profile are reversible. Recall that this property is also characteristic of the equa tive copular clauses that we discussed in the previous section. These properties suggest that there is no inherent asymmetry between the participants in these relations. Despite this, the fact that we can reverse the clauses shows that it is possible to construe either participant as TR, which means that the clauses still encode TR-LM asymmetry. In this case, the asymmetry arises from the construal (which participant the speaker chooses to focus attention upon) rather than from the semantics of the verb. As this discussion shows, while the prototypical subject has properties like inherent energy and volition, not all subjects have these properties. The property that all subjects do share in Langacker’s model is that they construe a given participant as TR. In the same way, the prototypical direct object is the PATIENT or ‘energy sink’, but not all objects can be characterised in this way. For example, in the sentence George’s jokes amuse Lily enormously, the object Lily is an EXPERIENCER rather than a PATIENT. The property that all objects share in Cognitive Grammar is that they construe a given participant as ‘second most prominent’(Langacker 2002:225). In other words, the object designates the most prominent aspect of the ground which is the primary landmark.
Intransitive clauses
Given the prototypical action chain, the subject is ‘upstream’ in terms of energy flow and the object is ‘downstream’ from the subject. It follows that a clause can have a subject but no object, but not vice versa. According to Langacker, this is because an object is only meaningful in relation to a subject, while a subject, as TR, is independently meaningful and can thus participate in processes where there is no second participant. Langacker suggests that this explains the properties of intransitive clauses. Here, the subject does not inter act with a second participant in some PROCESS. Instead, the subject interacts with itself, as in (22a), or interacts with the ground by undergoing a change of state that ‘changes the world’ as in (22b). This explains why intransitive verbs, like predicative adjectives, still profile a RELATION.
Ditransitive clauses We turn finally to ditransitive clauses, which are also called double-object constructions. Recall that these contain two objects, a direct object and an indirect object.
The question that arises in relation to double-object constructions concerns how the indirect object (Lily) might be semantically characterised, given the schematic semantic account of subjects and direct objects adopted in Cognitive Grammar. Langacker (1991: 326) argues that a thematic characterisation is most appropriate for indirect objects, since this function shows a greater thematic consistency than the functions subject and object. In other words, Langacker suggests that this grammatical function might be characterised in terms of its semantic role, which means that it is associated with a less schematic (or more specific) meaning than subjects and objects. For example, verbs of transfer like give or send have an indirect object with the semantic role RECIPIENT (24a), while verbs of perception typically have an indirect object with the role EXPERIENCER (24b):
We return to ditransitive clauses in Chapter 20, where we will see that this type of construction has been of particular interest to researchers who take a constructional approach to grammar (particularly Goldberg 1995).
