A Tentative Research On Metaphor Comprehension Mechanism

With the rising of cognitive linguistics, the role of metaphor has been reexaminedand acquired a renewed recognition. In the traditional rhetoric, metaphor, as a figureof speech, is primarily decorative and ornamental in nature; while in the newly risingcognitive linguistics, metaphor is considered as an important tool of our cognitiveapparatus. The importance and ubiquity of metaphors stimulate a strong interest in anin-depth investigation into how metaphors are processed in human mind. For example,how do people understand metaphors such as The lecture was a feast for the mind?Two prevalent psycholinguistic models attempt to offer an explanation - theConceptual Metaphor (CM) Model (Lakoff and Johnson, 1980) and the AttributiveCategorization (AC) Model (Glucksberg and Keysar, 1990). According to the CMmodel, figurative expressions are understood as instantiations of deep conceptualmetaphors, such as IDEAS ARE FOOD. In contrast, the AC model has proposed thatmetaphors are understood as assertions of the topics (e. g., lecture) membership in anattributive category exemplified by the vehicle (e. g., feast). McGlone (1996)conducted a series of experiments to examine whether people used conceptualmappings or attributive category knowledge to understand nominal metaphors.Although his findings favored the AC view over the CM view, his study suffered froma methodological problem - the paraphrasing, similarity rating and cued recall tasksthat he did are not on-line measures. How metaphors are processed in on-lineimmediate comprehension still remains a question which requires furtherexamination.The purpose of the present study was to elucidate which knowledge sources, i. e.CM or AC model, people rely upon when comprehending Chinese metaphors. In thisinvestigation, readers' comprehension of metaphors was examined in both theoff-line/reflective and on-line/immediate senses. Three experiments were conductedto investigate the scope of conceptual metaphors and attributive categories inexplaining people's metaphor comprehension.Experiment One was conducted to investigate the similarities people perceive between different pairs of metaphors under off-line condition. The differences insimilarity ratings are taken to reflect the relative difficulty subjects had incomprehending the metaphorical expressions. According to the CM model, metaphorsthat belong to the same conceptual metaphor were considered easier to understand.The similarity rating, on this account, reflected the use of conceptual metaphor incomprehension. In the experiment, subjects were presented with a list of theCM-consistent metaphors (metaphors that are not only similar in meaning but alsoderived from the same conceptual metaphor), AC-consistent (metaphors that areattributively similar but are not derived from the same conceptual metaphor) andunrelated metaphor pairs. They were then asked to rate the extent to which themetaphors in each pair were similar in meaning to each other on a seven-point scale.It was found that metaphors derived from the same conceptual metaphor wereperceived as more similar than metaphors that are only attributively similar.Conceptual metaphors seem to play a facilitative role when people process metaphorsin a relatively deliberate and conscious way, with minimal time constraints.Experiment Two was also a similarity rating task which followed the same logicas Experiment One but employed a more sensitive, on-line priming measure in whichsubjects were required to give immediate response. The program of Experiment Twowas written through the powerful psychological experiment software - E-prime andpresented to subjects by computer. Therefore, not only the rating scores but also thereaction times for different types of metaphor pairs can be taken into account. Theresults showed that metaphors derived from the same conceptual metaphor were notnecessarily perceived as more similar than metaphors that were only attributivelysimilar in immediate comprehension. The effect of conceptual metaphors seemed tobe weakened or, to put it in other words, not so prominent in the immediate,subconscious understanding of metaphors. The AC model seems to be a morereasonable account for immediate metaphor comprehension.In Experiment Three, a cued recognition paradigm was adopted to investigate thetype of representations people generate and encode when they immediatelycomprehend the metaphors. The experiment had two phases: the acquisition phase and the recognition phase. In the acquisition phase, participants were required to read themetaphors on the computer screen one by one. In the recognition phase, they wereasked to do the recognition task according to three different priming conditions:primed by CM-cues (cues that referred to a metaphor's underlying conceptualmapping), primed by AC-cue (cues that referred to a metaphor's attributive category),and no priming. The results showed that although both CM cues and AC cues provedto facilitate subsequent metaphor recognition, AC cues were more effective. Thefindings also lend support for the claim that the AC model seems to be a morereasonable account for immediate metaphor comprehension, but the influence ofconceptual metaphors should not be overlooked as well.The above findings suggest that the processing mechanisms of metaphors arevery complex. People seem to employ different strategies to understand metaphorsunder different conditions. In the off-line condition, CM model is a better model toaccount for the comprehension of metaphor. In the on-line condition, however, thepower of conceptual metaphor seems to be weakened and the AC model proved to bea more reasonable account for the comprehension of metaphor.