An important component of comprehending any language is the assessment and integration of information about a wordıs meaning. This process can be complicated by the fact that many words have more than one meaning (Gee, 1997). Ambiguous words are words with more than one meaning. For example, consider tthe word BAT. BAT can have the meanings of the flying mouse, the wooden bat, or the action of batting. Ambiguity resolution is the process of selecting a context-appropriate meaning of an ambiguous word. One theory of ambiguity resolution is called a Context-Plus-Frequency theory. In this theory, the meaning of the ambiguous word is selected by means of the present context and the relative frequency of the meaning. If a context is not presented, all of the meanings of the word are activated and then the most frequently occurring meaning (dominant) is selected. The issue of ambiguity resolution is going to be addressed by using a cued recall task.
A cued recall task has two components, the to-be-remembered items (targets) and the words that are used to prompt recall (cues) of the to-be-remembered items. Subjects are typically shown a list of targets and are later given a list of cue words that are meaningfully related to the targets. The purpose of the cues is to help the subjects remember the targets. In this experiment, the cues were either ambiguous or non-ambiguous words, and the targets were also either ambiguous or non-ambiguous words. Previous research has shown that targets that are non-ambiguous are more readily recalled than that are ambiguous (Gee, 1997). Based on this finding it stands to reason that recall will be higher when the cue is non-ambiguous relative to when it is ambiguous. However, what we were particularly interested in was the relationship of cue and target ambiguity and which type of ambiguity played a more important role in cued recall performance.
The third variable manipulated in this experiment was that of task, whether the participants were tested in the intralist or extralist cued recall. Participants in the extralist task were shown the targets in isolation at study, then they were given cues that were meaningfully related to the targets at recall. Intralist participants were shown the cue and the target together at study, and the cues were presented again at test to prompt recall. In essence the intralist condition has better recall rates because explicitly the participant temporarily strengthens the relationship between the cue and target (Nelson, Schreiber, McEvoy, 1992). Gee (1997) claims that since we presented the participants with the cue in the study phase, we pre-imposed the context-appropriate-meaning of the ambiguous word which essentially makes it a non-ambiguous item for the duration of the experiment. In other words, in the intralist task ambiguous words should perform at the same level as the non-ambiguous words.
The participants were 48 undergraduate students at S.U.N.Y, College at Fredonia, 18 or older, taking introductory psychology. The experimental design formed a 2x2x2 mixed-model design. Two lists of 24 words were constructed, such that half of the cues were ambiguous and the other half were non-ambiguous. Additionally, half of both types of cues were linked to ambiguous targets and the other half were linked to non-ambiguous targets. Cue and target ambiguity were the within subjects variables, and type of task was manipulated between subjects.
The main effects of target ambiguity, F(1,46) = 4.69 p<.05, cue ambiguity, F(1,46) = 4.95 p<.05, and type of task, F(1,46) = 45.8 p<.05, were all significant and consistent with our expectations.
When the cue or target word is non-ambiguous, performance is facilitated relative to when it is ambiguous. Similar results have been noted in previous research concerning ambiguity (Gee, 1997). Additionally, it was shown that recall levels were higher in the intralist condition than in the extralist condition, which is also consistent with previous research (Gee, 1997; Nelson, Schreiber, & McEvoy, 1992). According to the Context-Plus-Frequency theories of ambiguity resolution performance should have been the highest in the intralist case because the presence of the cue at study acted to select the appropriate meaning of the ambiguous word, effectively making it a non-ambiguous word for the duration of the experiment. In the extralist case, the context cue was not present at study and the result is that the participants may have selected the wrong meaning of the ambiguous cue or target at study which would result in lower performance at recall.
The most important finding of this experiment is the significant interaction between cue and target ambiguity, F (1,46) = 4.15, p<.05. As you can see from Figure 1 it appears that when the target and the cue are both ambiguous, recall is at its lowest, however recall performance did not vary significantly in the other conditions. In other words it seems that two sources of ambiguity are too difficult to overcome, but the system seems to be able to recover from only one source of ambiguity. This interactive relationship between cue and target ambiguity is very interesting and has not been documented before. Because this interaction appears to have implications for many of the theories of ambiguity resolution it must be replicated. Further research is necessary to determine if this interaction is real and to examine itsı implications.
References
Nelson, D.L., Schreiber, T.A., & McEvoy, C.L. (1992). Processing implicit and explicit representations. Psychological Review, 99, 322-348.
Gee, N.R. (1997). Implicit memory and word ambiguity. Journal of Memory and Language, 36, 253-275.