ISSN : 1226-9654
Two experiments were performed to replicate the results that topic makers had effects on sentence processing(Kim, 1996). In experiment 1, center-embedded relative clause sentences that have topic makers('nun/un') at the first noun were compared to the same structures that have subject case makers('ka/i'). In experiment 2, 'ka/i'-'ka/i' sequence of case markings in the relative structure were compared to 'nun/un'- 'ka/i' sequence in the same structures. The moving-window presentation method were used and self-paced reading times of each words were measured. The results of the two experiments clearly showed that the topic makers significantly changed the syntactic processes. Theoretical models of syntactic ambiguity resolution processes of Korean relative clause sentences were discussed.
We developed guidelines to help users to search information on the internet, and tested their efficacy experimentally. Guidelines for novices on information search were derived from Park and colleagues' earlier study (1998). Search engines were provided either with the above guidelines or without them. Two groups of four novices were assigned to the guideline-provided condition and to the guideline-unprovided condition. They searched on 15 problems of various categories. The results showed that participants provided with guidelines used less number but more various kinds of keywords and solved more problems than participants without guidelines. This indicates that these guidelines could have positive effect on information search of novice users.
People can learn knowledge without realizing what is happening, It is called implicit learning; learning with no the intervention of consciousness. The mechanism of the implicit learning has been explained by three parts: structure-abstraction, instance-encoding, episodic-processing. In this study, using Korean as stimulus, we could verify the generality of the implicit learning. The episodic-processing explanation was tested whether it is the processing mechanism of the implicit learning. In this study, we also compared implicit learning transfer to language learning by investigating whether the rule acquired in implicit learning could be transferred to other language. In experiment 1, we replicated Whittlesea & Wright (1997)'s experiment 3. This experiment consisted of 3 conditions: standard rule condition, training stimuli repeated condition, test stimuli repeated condition. In consequence, we could confirm that a general phenomenon for human being, because the accuracy of well-formedness task in Korean was above chance. As Whittlesea et al. (1997) proposed, among three conditions, training stimuli repeated condition revealed the lowest accuracy and standard rule condition showed the highest accuracy. This result confirmed that implicit learning is not to be acquired automatically without the intervention of consciousness, rather it is acquired with subjects' intention. That is, it is replicated the preceded study that processing mechanism of implicit learning is followed by episodic-processing. Experiment 2 was carried out in two parts; Korean-to-Korean transfer condition and Korean-to-English transfer condition, which have different surface structure of learning phase and test phase. This experiment proved that if learning stimuli and test stimuli have the same deep-structure, implicit learning transfer occur regardless of surface-structure. It demonstrated that even if the learning language and test language are different, implicit learning transfer could be generated cross-linguistically.
Four experiments examined the effects of word fragment type on perceptual and conceptual processing in Korean word completion test. Target words were two-letter words without a final consonant. Word fragment types were manipulated in seven ways by deleting a vowel from second letter, a consonant from second letter, a vowel from first letter, a consonant from first letter, vowels from both letters, consonants from both letters, or a second letter (e.g., '가ㅁ', '가ㅜ', 'ㄱ시', 'ㅗ부', 'ㄴㅂ', 'ㅔㅐ', '계'). When study-test modality shifts were manipulated, significant modality effect was obtained in some word fragment types but not obtained in other types (Experiment 1). When semantic and perceptual study conditions were manipulated between Ss (Experiment 2A) or within Ss in a blocked list (Experiment 2B), significant levels of processing (LoP) effects were obtained but interaction between LoP and word fragment type was not significant. But when semantic and perceptual conditions were manipulated within Ss in a mixed list (Experiment 2C), significant interaction between LoP and word fragment type were obtained. These findings suggest that the relative contribution of perceptual and conceptual processing in word completion priming is dependent on word fragment type, and that perceptual and conceptual word completion tests can be distinguished by word fragment types.