Analysis and simulation
ECG claims that when a hearer hears an utterance, he or she has two distinct tasks to perform. The first is analysis (parsing), which involves the hearer mapping the stimulus (the utterance) onto the structured inventory of constructions in his or her grammar and recognising which constructions are instantiated by the utterance. The second task is simulation, which involves the activation of conceptual representations that underlie the interpretation of the utterance and the ‘re-enactment’ of these conceptual representations (recall our discussion of Barsalou’s research on perceptual symbol systems in Chapter 7). It is this process of simulation, together with contextual factors, that gives rise to the hearer’s response. According to ECG, the conceptual rep resentations that are accessed and simulated during language understanding are embodied schemas like the SOURCE-PATH-GOAL schema that we saw in Chapter 7. In other words, it is embodied experience that gives rise to these conceptual representations, and during language processing constructions are specified to prompt for these conceptual representations that arise from embodied experience. This explains why the approach is called ‘Embodied Construction Grammar’. To take a concrete example, consider how a hearer might process the following utterance.
In terms of the analysis stage, each of the phonetic forms maps onto a construction (form-meaning pairing) in the hearer’s inventory of constructions, at morpheme, word, phrase and construction level. The hearer recognises the ditransitive construction, which brings with it the semantics of TRANSFER, as we saw in our discussion of Goldberg’s theory of Construction Grammar. The mapping of participant roles onto argument roles in the construction contributes to the interpretation of the utterance, and the context of the utterance enables the referent of the expression me to be identified (as the speaker). Recall from Chapter 11 that Mental Spaces Theory provides a cognitive account of how this process of reference assignment takes place.
At the simulation stage, Bergen and Chang argue that the interpretation of a ditransitive utterance like this activates three embodied schemas: FORCE APPLICATION, CAUSE-EFFECT and RECEIVE. Each of these is associated with schematic events and schematic roles such as ENERGY SOURCE and ENERGY SINK (Langacker 1987), and it is the mapping of constructions onto these schematic events and roles that gives rise to the simulation process. For example, in (42) the construction instantiated by Lily is ENERGY SOURCE, and the construction instantiated by me is ENERGY SINK. This simulation process gives rise to an ordered set of inferences, some of which are represented in (43), where SMALL CAPS indicate participants and event schemas (based on Bergen and Chang (2005):
Although these inferences seem rather obvious in terms of deconstructing the meaning of the utterance, it is nevertheless important for a model of language processing to explain how such inferences arise in utterance comprehension. According to the ECG model, it is the hearer’s own embodied experience, which results in conceptual representations of that experience in terms of embodied schemas, that gives rise to these inferences via a simulation process. In this way, the hearer mentally re-enacts the event designated by the utterance.
Although we do not go into further detail on the ECG approach here, this brief overview provides a sense of how a constructional approach can be extended to account not only for knowledge of language but also for the dynamic processing of language, while taking seriously the role of embodied knowledge and the notion of mental simulations as the outcome of language comprehension.
