1.2 Research Significance
The traditional divide between “metaphor” and “mixed metaphor” may ultimately beuntenable given the various flexible ways that people employ metaphor in language andelsewhere. Rather than being a deviation from proper metaphor use, mixed metaphors maybe ideal reflections of people’s typical metaphorical experiences in language, thought, andcommunication. In this manner, the st
For this paper, the significance of studying how metaphorical mappings are constructedwhen understanding a mixed metaphor in L1 and L2 can be summarized into three aspects:
Firstly, in view of the special nature of mixed metaphor, the present study will providemore empirical evidence for those assumptions on metaphor, then resulting in strengtheningthe explanatory power and enriching the connotation of certain classical theories of metaphor.If a theory manages to account for the special, complex cases, this would make certain for itsadequacy in handling the simpler examples. Secondly, this study will offer some newperspectives on mixed metaphor comprehension in different language proficiency, which inturn, shed more light on second language teaching and learning. Thirdly, this paper will beconducive to the case that the existence of mixed metaphor is not a reflection of cognitiveconfusion or disorganized linguistic forms, and challenge the traditional views of mixedmetaphor as impermissible forms or cognitive errors.
3 Theoretical Framework
The current study intends to use an experimental approach to investigate the mappingconstruction during mixed metaphor comprehending, and to compare the differences on L1and L2 mixed metaphors comprehension for Chinese (L1) - English (L2) bilinguals. Thetheoretical framework applied in this study is the assumptions on mixed metaphors accordingto Conceptual Metaphor Theory and Non-Conceptual Metaphor Theory.A basic insight of research into conceptual metaphor is that abstract or complex domainscan be understood metaphorically in terms of simpler domains that are amenable to directbodily experience. A conceptual metaphor consists of a bundle of mappings from a sourcedomain to a target domain. But it is worth noting that not all aspects of the source domain aremapped onto the target domain. Just as Grady pointed out, the famous metaphor THEORIESARE BUILDINGS allows speakers to discuss the foundations of a theory, whereas it does notprovide a conventionalized mapping for “French windows”. Metaphorical mappingsbetween two domains are thus rarely fully productive, but remain incomplete. Opponents ofCMT further point out that some conceptual metaphors have emergent properties that cannotbe explained through a simple mapping from source to target domain, for example theConceptual Blending Theory. Both the incompleteness of metaphorical mappings and theexistence of emergent meanings are relevant to the question how mixed metaphors areunderstood.
So, when it comes to the mixed metaphor, from the point of view of the ConceptualMetaphor Theory, since mappings are obtained between a source domain and a target domain,a CMT analysis would predict that mappings are not only constructed between each sourcedomain and the target domain, but also between the source domains themselves. In short, theCMT analysis would claim that people conceptualize a mixed metaphor by maximizingcorrespondences between the target domain and the two or more source domains. Any imageschematic structures that are sufficiently similar should be mapped between the respectivesource domains.
A 2 (Language: Chinese, English) × 2 (Test sentence: mixed metaphor primed, puremetaphor primed) two-way repeated measures ANOVA was conducted on reaction time (RT)with the two factors as within-subject variables. The statistical analysis was only for correctresponses. Reaction time and accuracy rate were calculated after eliminating the data ofreaction time more than 5000 ms. The experiment set the ‘false’ responses in reaction to bothmixed metaphor primed and pure metaphor primed test sentences as right answers, and ‘true’responses in reaction to literally true test sentences, which were fillers, as right answers. Onlythe data with correct responses were reserved as valid. Responses for the literally true testsentences served as fillers were not analyzed. In each condition of interest, the average RTswere statistically analyzed by paired-sample t-test.
The experiment received a total of 1680 responses (35 subjects x 48 stories), of which,116 responses were inaccurate, that is literally false statements were classified as true and viceversa. These data points were excluded, leaving 1564 measurements. The responses of threesubjects were excluded entirely because of their error rates in excess of 30%. The meanreaction time of the remaining 851 data points was 1869 ms with a standard deviation of 703ms. Reaction times of more than 3275 ms were thus counted as outliers, which led to theremoval of 88 data points, reducing the final data to 1308 measurements.
5.2The Mapping Construction during Mixed Metaphor Comprehension
The first aim of the present study was to examine the existence ofbetween-source-domain mapping construction by observing the differences between mixedmetaphor comprehension and pure metaphor comprehension. The data analysis yields anon-significant result both in Chinese and English condition within-subject and within-items,showing that mixed and pure metaphors display no differences with respect to reaction time.In other words, there is no positive evidence that participants constructbetween-source-domain mappings when comprehending a mixed metaphor, namely, mixedmetaphors are not understood by means of maximizing correspondences between sourcedomains. Because if such mappings were constructed by the participants of the present study,they failed to produce the delay in response time that was observed by Glucksberg et al.While it is of course problematic to extrapolate from a relative absence of an effect, themeasurements are consistent with the interpretation that people can process mixed metaphorswithout more efforts because of not running into any incompatibilities. Considering the mixedmetaphors we adopted in the present study are in a special type of metaphor interaction —conceptual complementation, mixed metaphors here are thus more likely to be comprehendedby selectively projection. In other word, in the process of understanding mixed metaphors,people do not mechanically construct the mappings between source domains as many aspossible, but take use of their social and cultural background knowledge and then combine theknowledge with the current context to partly activate the source domains information into acoherent whole, and eventually construct the overall meaning of mixed metaphors.
By conducting a behavioral experiment, the present study aims to investigate one smallaspect of metaphor processing model that bilinguals employ when they process mixedmetaphors. By comparing two conflicting hypotheses of mixed metaphor comprehension forChinese (L1) - English (L2) bilinguals, this research explored the role ofbetween-source-domain mappings in the comprehension of mixed metaphors forChinese-English bilingual. And we have drawn the conclusions as follows:
Firstly, when judging the literal truth of a “Some Xs are Ys” statement, subjects spent nomore time to make a response for the test sentence with a mixed metaphor prime compared toa pure metaphor prime. The results were both observed in Chinese group and English group,which reveals that the participants comprehended mixed metaphor without conductingbetween-source-domain mappings and probably do not run into any incompatibilities as aresult of shallow processing. Secondly, English test sentences consumed more time and wereresponded more correctly under mixed metaphors priming compared to Chinese ones, whichindicates that semantic processing requires more cognitive efforts in L2 than L1 due todifferent language proficiency, and the advantage in L1 for bilinguals may also lead to theirrich associations to attain the reasonability of test sentences resulting in a higher error rate.
Overall, the study has testified the mapping construction in the cognitive mechanism ofnative Chinese who study English as the L2 when they understand mixed metaphors. Aconclusion was drawn that in both L1 and L2, there was no mappings between sourcedomains constructed during mixed metaphors comprehending. In addition, some differenceswas also found between two types of language processing, that is, there was more responsetime as well as higher accuracy rate in the judgement task of the L2 rather than in the L1, asthe advantage in L1 facilitates the comprehension and mental imagery during semanticprocessing.