Processing Emergent Features in Metaphor Comprehension Asuka Terai Robert L. Goldstone Introduc?on This research was based on a couple of experiments which aimed to study the process of recognizing emergent features when one is presented with a metaphor of the form 'Target' IS 'Vehicle'. For such metaphors, several types of features were presented, namely target features (belonging to the target only), vehicle features (belonging only to the vehicle), common features, and emergent features (not belonging to either the target or the vehicle, but arising as a result of the interac?on between earlier features). An example of such a metaphor is 'Stars are Diamonds'. Here, 'white' is a target (Start) feature, 'rare' is a vehicle (Diamonds) feature, 'beau?ful' is a common feature, and 'amazing' is a feature which results as a result of the interac?on of rare, white and beau?ful, and hence is an emergent feature. Another example is, "The streets were a furnace, the sun an execu?oner." (Cynthia Ozick, "Rosa"). Here, streets may be associated with 'walking', or 'path'. A furnace or the sun may be associated with 'heat'. An execu?oner may be associated with 'ruthlessness'. Emergent features, however, would be 'irrita?ng', or 'unbearable'. Hypothesis: It was hypothesized that since emergent features were a result of the interac?on of non-emergent features, they should take more ?me to be recognized as a valid interpreta?on of the metaphor, as compared to the non-emergent features. Experiment 1 In this experiment, 50 metaphors were used. The features related to those metaphors were categorized as target, vehicle, common or emergent. The experiment used a cue-to-respond method. To ensure that the par?cipants do not simply answer 'yes' to every feature presented, for every metaphor, 2 features which were totally irrelevant were also presented. The experiment was conducted by presen?ng the par?cipant with a fixa?on point to help them concentrate. Next, a metaphor was shown on the screen for 1 second, and the par?cipants were asked to interpret it. Next, a feature was presented and a?er a fixed ?me interval, the par?cipant was buzzed (cue) to give a response indica?ng whether the feature was an interpreta?on of the metaphor or not. There were two parts to the experiment based on the ?me interval as well. In one part, the ?me interval was 0.5 seconds (short), while the other was 5 seconds (long). Results: This experiment did not lead to any interes?ng conclusions as there was no significant effect of the interac?on between the type of feature and the ?me dura?on (short/long). On the contrary, a very weak interac?on suggested that non-emergent features are slightly harder to recognize than emergent features. The authors feel that these results may be explained by saying that perhaps even 5 seconds (the long ?me dura?on) was not enough to judge the aptness of an emergent feature. It is also possible that processing a metaphor is a different process than judging the aptness of a feature. Experiment 2 Since the first experiment did not provide any proof of the hypothesis, and instead le? one with the possibility that the process of processing the meaning of a metaphor is different than the process of judging whether a feature is an apt interpreta?on of the metaphor or not, a second experiment was conducted. In this experiment, the dura?on of presen?ng the metaphor was varied (short and long), in contrast to varying the dura?on of presen?ng the feature. The metaphor was presented for 3 seconds in the short ?me dura?on and for 12 seconds in the long ?me dura?on. A?er that, the par?cipants were asked to respond as quick as possible but in under 6 seconds. Results: A result observed was that par?cipants took longer (~ 2.2 seconds) to respond in the long dura?on case than in the short dura?on case (~2 seconds). This can be explained by arguing that, the shorter dura?on may serve as a nudge to the par?cipants to make a decision quickly. A more interes?ng result was that in the short ?me dura?on, the percentage of posi?ve responses for non-emergent features was greater than that for emergent features. However, for the long ?me dura?on, the earlier observa?on was reversed! This clearly shows that it takes a longer ?me for emergent features to be recognized as an interpreta?on for the metaphor. Another important result was obtained by dividing the trials (on a single par?cipant) into 3 categories - beginning, middle and end (It should be noted that a par?cipant was asked to respond for all the features for all the metaphors.) The observa?on was that, the percentage for emergent features increased in the short dura?on and to the middle in the longer dura?on (but then fell). The percentage for non-emergent features decreased invariably. This may be explained by considering that as ?me proceeds, the non-emergent features get inhibited. The emergent features may also get inhibited as the ?me dura?on gets too long (~18 seconds). It can be seen that these results support the hypothesis. References Processing Emergent Features in Metaphor Comprehension Asuka Terai, Robert L. Goldstone CogSci 2011. Metaphors - grammar.about.com
