意義與詞義：漢語詞義的計算心理語言學研究─意義能力：新義的心理與神經處理(2/3)

行政院國家科學委員會專題研究計畫 期中進度報告

意義能力：新義的心理與神經處理(2/3)

中 華 民 國 93 年 5 月 26 日

行政院國家科學委員會補助專題研究計畫期中進度報告

意義能力：新義的心理與神經處理 (II)

(Sense Ability: Psycho- and Neuro- Linguistic Processing II)

計畫類別： □ 個別型計畫 □ 整合型計畫

計畫編號：

NSC92-2411-H-002-076-ME

執行期間：92 年 8 月 1 日至 93 年 7 月 31 日

計畫主持人：安可思教授

共同主持人：李佳穎博士、畢永峨教授、陳修元博士、劉齡鶴博士

研究助理： 龔書萍、鍾曉芳、蔡佩舒

成果報告類型：□精簡報告 □完整報告

本成果報告包括以下應繳交之附件：

□赴國外出差或研習心得報告一份

□赴大陸地區出差或研習心得報告一份

□出席國際學術會議心得報告及發表之論文各一份

□國際合作研究計畫國外研究報告書一份

執行單位：國立台灣大學語言學研究所

中 華 民 國 九十二 年 五 月 三十 日

國科會期中報告 NSC NSC92-2411-H-002-076-ME

2

行政院國家科學委員會專題研究計畫期中進度報告

意義能力：新義的心理與神經處理 (II)

Sense Ability: Psycho- and Neuro- Linguistic Processing (II)

計畫編號(II)： NSC92-2411-H-002-076-ME 執行期限： 92 年 8 月 1 日至 93 年 7 月 31 日 主持人： 安可思 國立台灣大學語言學研究所 共同主持人： 畢永峨 國立師範大學英語系 劉鶴齡 私立長庚大學 李佳穎 中央研究院語言學研究所籌備處 陳修元 台北市立師範師院 一、中文摘要 在今年計畫，我們主要探討兩個研究 主題: (1) 創新的隱喻辭彙和約定俗成的隱 喻詞彙是否在不同大腦皮質區處理? (2) 是否能有系統的找出概念隱喻的 對映規則? 首先，我們完成核磁造影實驗。我們 測試了四種中文句子: 約定俗成的隱喻句 子、字面意義的句子、違反隱喻規則的句 子及十字的注視。實驗結果發現，右腦皮 質區的激發表示當閱讀違反隱喻規則的句 子時大腦是在連結二個關聯低的概念，所 以不同的隱喻類型將會影響大腦皮質激 發。 另外，我們也利用 SUMO 和 WordNet 兩個線上詞彙資料庫系統化地找出概念隱 喻的對映規則。 總之，藉著神經影像實驗及語料庫的 研究，我們可以更了解概念隱喻在處理機 制及解釋大腦如何表徵並處理詞彙意義。 關鍵字：核磁造影實驗，概念隱喻，新創 隱喻，對映規則, SUMO, WordNet Abstract

This past year, we investigated two questions:

(1) Are conventional metaphors and novel metaphors processed in different hemispheres?

(2) Can we systematically extract conceptual metaphors from large scale corpora to determine their source domain and generate their mapping principles?

Neuro-imaging studies can now look at whether conventional and anomalous metaphors are processed in different locations in the brain. We created a block-randomized design in Mandarin Chinese containing three conditions: conventional metaphor sentences, literal sentences, and anomalous metaphor sentences. Right hemisphere activation was found and suggests that remote associations are being formed while reading the anomalous metaphor sentences, and that metaphors will vary in recruitment of cortical function depending on their type.

Second, we systematically extract the conceptual metaphors and mapping principles via SUMO and WordNet from both English and Chinese corpora.

To sum up, we can better understand the mechanism of processing conceptual metaphors through conducting the neuro-imaging and corpus-based studies. Our project will aid in understanding how the meanings of lexical items are stored and represented in the mental lexicon.

Keywords: fMRI experiment, conceptual

metaphor, novel metaphor, mapping principles, SUMO, WordNet

二、緣由與研究動機

Where are metaphors processed in the brain? Bottini et al. (1994) ran a positron emission tomography activation study to

NSC NSC92-2411-H-002-076-ME

explore the role of the right hemisphere in metaphor processing. They asked six male subjects to evaluate the plausibility of both metaphorical and literal sentences by raising their left index finger to indicate plausible sentences and to do not do anything for implausible sentences. Examples of the four sentence types are given below.

Metaphorical sentences

The close friends were elastic bands.(Implausible)

Tim had been poured into his clothes and forgotten to say when (Plausible)

Literal sentences

The teacher used her glasses as a blackboard. (implausible)

The policeman taught that arteries can be routes. (plausible)

They found significant activation in the right hemisphere in the prefrontal region (BA46), middle temporal gyrus (BA21) and precuneus (BA31), anterior and posterior cingulate (BA 32/31). They suggested that their findings argued for a role of the right hemisphere in performing complex language operations, such as interpreting connotative meaning.

Bottini’s study, while important for its contribution to understanding the right hemisphere’s role in language processing, has methodological issues that need to be addressed. One main issue is that the level of plausibility for the metaphorical sentences is overall lower than that of the literal sentences (i.e. 79% versus 95%). That is to say, it was harder for subjects to decide if the metaphorical sentences were plausible or not. This could mean that the right hemisphere activation was not a result of processing metaphors, it was instead a result of the difference in difficulty to making a plausibility decision. Newman et al. (2002) and Stowe et al. (1998), for example, argue that the right frontal gyrus is recruited for problem-solving.

Thus, in this project, we will look again at the location of metaphor processing in the brain, but we will only ask subjects to read the sentences to themselves and to press a button when they are done reading. The study will have three conditions: a literal sentence condition, a conventional conceptual metaphor condition, and an anomalous metaphor condition.

三、方法、結果與討論

In the following, we, first, will report the fMRI study. Second, we will give a brief summary of how we extract conceptual metaphors systematically via SUMO and WordNet.

1. fMRI Study Method

Participants

Participants were 8 right-handed, neurologically normal males students from Chang Keng University (mean age = 21). They were all native speakers of Mandarin Chinese and right-handed by self-report. All gave informed consent using a form approved by Chang Keng University Hospital and were paid for their participation.

Materials and Procedures

Three types of sentential stimuli were created: conventional metaphor sentences, anomalous metaphor sentences and literal sentences. Each condition had 36 sentences, for a total of 108 sentences (please see Ahrens 2004 for further information on the methods and procedures.)

Results

Behavioral Performance

The average reaction time to the literal condition was 1483.75 ms, to the conventional metaphor condition was 1539.76 ms and to the anomalous metaphor condition was 1657.32 ms. Reaction times were tested by one-way ANOVA. The statistical results shows that there is a significant difference between each group ( F(2, 841) = 8.016 , p < .01). The Bonferroni post-hoc test also shows that the anomalous metaphors are significantly responded to slowly than the conventional metaphors and literal sentences. However, there is no significant difference in reaction times between the conventional metaphors and literal sentences.

Brain Activations

Conventional metaphors – literal sentences condition. There was greater

activation in the left inferior temporal gyrus (BA21) and right inferior temporal gyrus (BA 20) when the conventional metaphor condition was compared with the literal sentences condition.

Anomalous metaphors – literal sentences condition. There was greater

國科會期中報告 NSC NSC92-2411-H-002-076-ME

4

activation throughout the right frontal gyrus (BA8, 9, 47) and right middle temporal gyrus (BA46) as well as activation in the left middle frontal and temporal gyrus (BA46, BA21), precuneus (BA 19) and fusiform gyrus (BA18).

Anomalous metaphors – conventional metaphor condition. There was greater

activation throughout the right frontal gyrus (BA 8, 9, 46) and precuneus (BA 19), as well as activation in the left middle frontal gyrus (BA 46), and left fusiform gyrus (BA 37).

Discussion

The results presented here demonstrate that conventional metaphors are recruiting some additional left and right hemisphere areas in the inferior temporal gyrus as compared with literal sentences. This finding is unexpected given the conceptual metaphor theory (Lakoff, 1993), which postulates no additional effort for conventional conceptual metaphors.

However, the results presented here could be a result of the block design that we used. In a block design, subjects see sentences from the same condition (in our case, six sentences of the same type). It is possible that subjects are adopting different cognitive strategies when reading these sentences one after another than they would if they were not aware (or less aware) that the sentences they were reading involved conceptual metaphors. It is the nature of the block design to emphasize the sameness of the stimuli under study, and is an implicit confound of block-design experiments. In order to test see if this confound is leading to unwarranted activation effects, we are planning to carry out an event-related fMRI study with the same stimuli in which we will randomly present the stimuli in order to see if these effects disappear.

In contrast to the conventional versus literal condition, when the anomalous condition is compared with the literal condition, increased activation is extensive, occurring in the left and right middle temporal gyrus, as well as throughout the right frontal gyrus (superior, middle and inferior), and the left middle frontal gyrus and precuneus. Although it has been suggested that the right frontal lobe is involved in active responses or judgments (Bookheimer 2002) and problem-solving in relation to working memory constraints (Newman et al., 2002; Stowe et al., 1998),

the task here was simply a reading task. No problem solving or judgments were involved.

Our findings thus point to the idea that the right frontal lobe is involved in semantic processing, especially involving unusual semantic relationships (Seger et al. 2000) as well as in the processing of distant meaning associations (Beeman, 1998; Beeman et al., 1994). These studies by Beeman and colleagues have suggested that the right hemisphere treats meaning associates equally, while the left hemisphere involves semantic processing of close meaning associates. This hypothesis concerning the right hemisphere being recruited to process unusual semantic relationships is further supported by the comparison between the anomalous and conventional metaphor conditions, where the anomalous condition shows comparative activation throughout the right frontal gyrus (superior, middle and inferior) and precuneus. The right hemisphere activation here suggests that remote associations are being formed while reading the anomalous sentences.

This difference in activation volume and area for anomalous metaphors as compared with conventional metaphors is further evidence for theories of metaphor processing that allow for and can explain differences among metaphor types, such as the

conceptual mapping model (Ahrens 2002c)

and the conceptual blending theory (Fauconnier & Turner 1998), and evidence against psycholinguistic theories such as a the property attribution model that defines metaphors as class inclusion statements (i.e. Glucksberg, 1998; Glucksberg & Keysar 1990). When metaphors are defined as class inclusion statements, there is no room for different types – all metaphors are created equal, which is the opposite of what we are arguing here, that there are different types of metaphors, and different gradations of novelty that can be measured both behaviorally and in terms of brain activation once the linguistic criteria for differentiating them can be explicated, as discussed in Ahrens (2002c). However, the fact that there was frontal activation found for anomalous metaphors (as compared with the conventional condition) indicates that this is not, in fact, the case.

2. Corpus Study

The corpora-based analysis of conceptual metaphors started with the work

NSC NSC92-2411-H-002-076-ME

of Chung, Ahrens, & Sung (2003). In this work, they calculated the frequency of conceptual metaphors in the Chinese and English corpora based on the Conceptual mapping Model (Ahrens 2002a, 2002c). The frequency account of conceptual metaphors was then assisted with the Suggested Upper Merged Ontology (SUMO) in Ahrens, Chung & Huang (2003) and Chung, Huang & Ahrens (2003), Chung, Ahrens, & Huang (2003), Chung, Ahrens, Sung (2003) and Chung, Ahrens & Huang (2004a, 2004b). In these studies, the Mapping Principles (MPs) between the source and target domains were verified through observing the related SUMO nodes of the most prototypical (i.e., the most frequent) mappings. The incorporation of SUMO was the first step to make the generation of MPs more automatized. This helped reduce the manual work that was carried out based on intuition.

In 2004, Chung, Ahrens, & Huang (2004a and 2004b) further proposed that the manual selection of the source domains can be made computational through observing their definitions in WordNet and SUMO. This is a crucial step as it proposed that the metaphorical items in a single source domain are inter-connected with one another. Ahrens, Chung & Huang (2004) suggested that the recurrence of the lexical items in the WordNet and SUMO definitions can be used to verify the MPs generated previously. These studies will lead us to systematizing the extraction of conceptual metaphors in corpora. The ultimate goals of these studies will contribute to field such as lexical semantic programming, lexicography and on-line lexical databases.

四、計畫成果自評

This past year, we followed our project schedule and achieved satisfying results. First, we completed the fMRI experiment. In addition, we have published the results at the well-known conference "Mind, Language and Metaphor: EURECO Conference on the processing of Metaphor and metonymy – From computers to neurospychology" at Granada, Spain, on May (Ahrens 2004). Finally, we have published what we extract from SUMO and WordNet at a number of conferences in this past year (Ahrens 2003c, Ahrens, Chung & Huang 2003, 2004, Chung, Huang & Ahrens 2003, Chung, Ahrens, & Huang 2003, Chung, Ahrens & Sung 2003, Chung, Ahrens & Huang 2004a and 2004b).

In the next year, we are going to

continue two tasks. First, since we used the block design in our fMRI study, we are going to use the same stimuli to run another fMRI experiment with event-related design. Furthermore, we will continue to extract the conceptual metaphors via comparing with the lexical and domain information in SUMO and WordNet.

五、參考文獻

Ahrens, Kathleen. (2004). Functional MRI of Conventional and Anomalous Metaphors in Mandarin Chinese. To be presented at Mind, Language and Metaphor

(EuroConference on the Processing of Metaphor and Metonymy – From Computers to Neuropsychology), April 24-29, Granada, Spain.

Ahrens, Kathleen (2003). Verbal Integration: The Interaction of Participant Roles and Sentential Argument Structure. Journal of Psycholinguistic Research, 32.5.

Ahrens, Kathleen (2002a). Metaphor Mapping and Knowledge Representation: Principles and Processing- 3/3. (NSC 90-2411-H-002-060-MC). National Science Council Project Report.

Ahrens, Kathleen (2002b). Using the Conceptual Mapping Model to Explain Metaphors in Foreign Language Classes. Presented at Sixth Foreign Language Teaching Conference at Tamkang University. October, 4-5.

Ahrens, Kathleen (2002c). When Love is not Digested: Underlying Reasons for Source to Target Domain Pairing in the Contemporary Theory of Metaphor. In Yu-Chau E. Hsiao (ed.) Proceeding of the First Cognitive Linguistics Conference, pp 273-302. Taipei: Cheng-Chi University.

Ahrens, Kathleen (2001a). Mapping Principles in Conceptual Metaphors. Presented at the International Conference on Cognitive Linguistics, University of California, Santa Barbara, July 22-27.

Ahrens, Kathleen (2001b). Metaphor Mapping and Knowledge Representation: Principles and Processing. Proceeding of the Symposium on Selected NSC Projects in General Linguistics from 1998-2000, pp, 281-310. Taipei: National Taiwan University.

Ahrens, Kathleen (2001c). On-line Sentence Comprehension of Ambiguous Verbs in Mandarin. Journal of East Asian Linguistics, 10, 337-358.

國科會期中報告 NSC NSC92-2411-H-002-076-ME

6

Lexical Ambiguity Resolution. In J. Nakayama (Ed.) Sentence Processing in East Asian Languages, pp.1-26. Stanford: CSLI.

Ahrens, Kathleen & Chu-Ren Huang (2000) Time is Motion in Mandarin Chinese. Parameterizing Conceptual Metaphors. In H. Y Tai and Chang Y. L. (eds.) Proceedings of ISCLL-7, pp, 27-47. Chai-yi: National Chung-Cheng University.

Ahrens, Kathleen & Dora Lu (2002). Literal Truthfulness and Conceptual Metaphors. 2002. To be presented at 10th ICCPCORAL (International Conference on Cognitive Processing of Chinese and Other Related Asian Languages) December 9-11, 2002. Taipei, Taiwan. Ahrens, Kathleen, Siaw-Fong Chung &

Chu-Ren Huang (2004). “From Lexical Semantics to Conceptual Metaphors: Mapping Principle Verification with WordNet and SUMO.” To appear in the

Proceedings of the 5th Chinese Lexical Semantics Workshop (CLSW-5),

Singapore. June 14-15.

Ahrens, Kathleen, Siaw-Fong Chung & Chu-Ren Huang (2003). “Conceptual

Metaphors: Ontology-based Representation and Corpora Driven

Mapping Principles.” In the

Proceedings of the ACL Workshop on the Lexicon and Figurative Language.

pp. 35-41.

Beeman, M., Friedman, R. B., Grafman, J., Perez, E., Diamond, S., & Lindsay, M. B. (1994). Summation priming and coarse semantic coding in the right hemisphere. Journal of Cognitive

Neuroscience, 6, 26-45.

Bookheimer, S. (2002). Functional MRI of language: new approaches to understanding the cortical organization of semantic processing. Annual Review

of Neuroscience, 25, 151-158.

Bottini G, Corcoran R, Sterzi R., Paulesu E, Schenone P, Scarpa P, Frackowiak RS and CD Frith. 1994. The role of the right hemisphere in the interpretation of figurative aspects of language. A positron emission tomography activation study. Brain, 117 (6): 1241-1253.

Chung, Siaw Fong, Chu-Ren Huang & Kathleen Ahrens (2003). “ECONOMY IS A TRANSPORTATION_DEVICE: Contrasting Representation of Source Domain Knowledge in English and Chinese.” In the Proceedings of the

Special Session for the International Conference on Natural Language Processing and Knowledge Engineering (NLP-KE): Upper Ontology and Natural Language Processing (UONLP).

Beijing. pp. 790-796.

Chung, Siaw Fong, Kathleen Ahrens & Chu-Ren Huang (2003). “ECONOMY IS A PERSON: A Chinese-English Corpora and Ontological-based Comparison Using the Conceptual Mapping Model.” In the Proceedings of

the 15th ROCLING Conference for the Association for Computational Linguistics and Chinese Language Processing, National Tsing-Hwa

University, Taiwan. pp.87-110.

Chung, Siaw Fong, Kathleen Ahrens & Yahui Sung (2003). “STOCK MARKETS AS OCEAN WATER: A Corpus-based, Comparative Study in Mandarin Chinese, English and Spanish.” In the Proceedings of the 17th

Pacific Asia Conference on Language, Information and Computation (PACLIC), Singapore. pp. 124-133.

Chung, Siaw-Fong, Kathleen Ahrens & Chu-Ren Huang. (2004a). “RECESSION IS A DISEASE: Defining Source Domains through WordNet and SUMO.” To appear in the

Proceedings of the 2004 Linguistic Society of Korea (LSK) International Summer Conference, Yonsei University,

Seoul, Korea. July 28-30.

Chung, Siaw-Fong, Kathleen Ahrens & Chu-Ren Huang (2004b). “Using WordNet and SUMO to Determine Source Domains of Conceptual Metaphors.” To appear in the

Proceedings of the 5th Chinese Lexical Semantics Workshop (CLSW-5),

Singapore. June 14-15.

Fauconnier & Turner (1998). Conceptual integration network. Cognitive Science, 22, 1333-187.

Glucksberg, S. (1998). Language and Thought. In R. Sternberg & E. E. Smith (Eds.). The Psychology of Human

Thought. Cambridge: Cambridge

University Press.

Glucksberg, S., & Keysar, B. (1990). Understanding metaphorical comparisons: Beyond similarity. Psychological Review, 97, 3-18.

Lai Tzuyin, V (2002). The Psycholinguistic Processing of Time Perspectives in Mandarin Chinese. MA thesis of

NSC NSC92-2411-H-002-076-ME

Graduate Institute of Linguistics, National Taiwan University, Taipei.

Lai, Tzuyin V,. Kathleen Ahrens, & Chu-Ren Huang (2001). Source Domains for Marriage in Mandarin Chinese. Presented at IACL-10 & NACCL-13, University of California, Irvine, June 22-24.

Lai, Tzuyin, V., & Kathleen Ahrens (2001). Mappings from the Source Domain of Plant in Mandarin Chinese. Proceedings of the 15th Pacific Asia Conference on Language Information and Computation. City University of Hong Kong. February 1-3.

Lai, Tzuyin. V, & Kathleen Ahrens (2002). The Processing of Temporal Perspectives in Mandarin Chinese. To be presented at 10th ICCPCORAL (International Conference on Cognitive Processing of Chinese and Other Related Asian Languages) December 9-11, 2002. Taipei, Taiwan.

Lakoff, G (1993). The contemporary theory of metaphor. In A. Ortony (Ed.), Metaphor and thought (2nd ed.), pp.202-251. Cambridge: Cambridge University Press.

Lu, Dora & Kathleen Ahrens. (2001). A Cross-linguistic Study on Love Metaphors. Presented at the International Conference on Cognitive Linguistics, University of California, Santa Barbara, July 22-27.

Lu, Dora, Kathleen Ahrens, & Sung Ya-hui, and Biq Yong-O (2001). Food Metaphor in Mandarin Chinese. Presented at IACL-10 & NACCL-13, University of California, Irvine, June 22-24.

Lu, Hsin-yi. D (2002). Processing of Conceptual Metaphors in Mandarin Chinese—A conceptual mapping model based study. MA thesis of Graduate Institute of Linguistics, National Taiwan University, Taipei.

Newman, S. D., Just, M. A., & Carpenter, P. A. (2002). The synchronization of the human cortical working memory network. Neuroimage, 15(4), 810-822. Seger, C. A.; Desmond, J. E.; Glover, G. H.,

and J. D. E. Garbrieli (2000). Functional magnetic resonance imaging evidence for right-hemisphere involvement in processing unusual

semantic relationships.

Neuropsychology 14. 361-69.

Stowe, L. A., Broere, C. A., Paans, A. M., Wijers, A. A., Mulder, G., Vaalburg, W., & Zwarts, F. (1998). Localizing

components of a complex task: sentence processing and working memory. Neuroreport, 9(13), 2995-2999.

Tompkins, C. A., Baumgaertner, A., Lehman, M. T., & Fossett, T. R. D. (1997). Suppresion and discourse comprehension in right brain-damaged adults: A preliminary report.

國科會期中報告 NSC NSC92-2411-H-002-076-ME