在议论文中,比喻不仅是一种修辞技巧,也是一种重要的论证方式。该文提出结合比喻识别和论辩挖掘技术自动分析议论文中的比喻及其论证作用。该文构建了一个数据集,标注了约1 200篇学生议论文中的比喻句、论辩角色及论辩质量等级,分析了比喻与论点、论据、阐释和其他论辩角色的作用方式以及比喻运用与篇章质量的关系。该文发现作为常见的修辞手段,比喻句的数量与论辩质量的相关性较弱,但比喻句作为论点时与论辩质量的相关性要强于作为其他论辩角色。此外,该文进一步标注了比喻论点类型以描述比喻的论证作用,包括事实、价值和策略,发现比喻论点的作用主要是传递价值与提出策略。通过比较两类比喻论点类型识别方法,发现基于精调预训练语言模型的方法优于基于提示学习的方法。最后,该文构建了一个集成比喻识别、论辩角色识别与论点类型分类的流水线系统,实验结果显示,该任务具有一定的实用性和挑战性。该研究对于作文自动评分与风格化的论点生成具有很好的应用前景和潜力。
Abstract
In argumentative essays, metaphor is not only a rhetorical skill, but also an important argumentation strategy. We propose an automatic analysis approach to identify metaphors and their argumentative functions in argumentative essays by combining metaphor recognition and argument mining techniques. We construct a dataset with annotations for metaphorical sentences, argumentative roles, and argumentative quality levels in 1,200 argumentative essays by students. We analyze how metaphors function in relation to argumentative elements such as claims, evidence, reasoning, and other argumentative roles, as well as how the use of metaphors relates to discourse quality. Our findings suggest that while the quantity of metaphorical sentences has a weak correlation with argumentative quality, the correlation becomes stronger when the metaphorical sentences serve as claims. Furthermore, we annotate the types of metaphorical claims according to their argumentative functions, including facts, values, and strategies. The results show that the main function of metaphorical claims is to convey values and propose strategies. We compare two approaches to identify the types of metaphorical claims and find that the approach based on fine-tuning pre-trained language models performs better than the prompt-based approach. Finally, we construct a pipeline system that integrates metaphor recognition, argumentative role recognition, and argumentative claim type classification. The experimental results show that this task is challenging. This research has promising prospects and potential in the realms of automated essay scoring and the generation of stylized arguments.
关键词
比喻论证 /
论辩挖掘 /
大语言模型
{{custom_keyword}} /
Key words
figurative argument /
argument mining /
large language model
{{custom_keyword}} /
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 石岳峰,王熠, 张岳. 深度学习在论辩挖掘任务中的应用[J]. 中文信息学报, 2022, 36(07): 1-12.
[2] ERVAS F, OJHA A. Metaphor in argument production vs understanding[C]//Proceedings of the 9th Conference of the International Society for the Study of Argumentation. Sciential International Centre for Scholarship in Argumentation Theory, 2019: 330-341.
[3] PRABHAKARAN V, REI M, SHUTOVA E. How metaphors impact political discourse: A large-scale topic-agnostic study using neural metaphor detection[C]//Proceedings of the 15th International AAAI Conference on Web and SocialMedia, 2021, 15: 503-512.
[4] MOENS M F, BOIY E, PALAU R M, et al. Automatic detection of arguments in legal texts[C]//Proceedings of the 11th International Conference on Artificial Intelligence and Law, 2007: 225-230.
[5] CABRIO E, VILLATA S. Five years of argument mining: A data driven analysis[C]//Proceedings of the 27th IJCAI, 2018, 18: 5427-5433.
[6] PARK J, CARDIE C. A corpus of Rulemaking user comments for measuring evaluability of arguments[C]//Proceedings of the LREC, 2018: 1623-1628.
[7] STAB C, GUREVYCH I. Identifying argumentative discourse structures in persuasive essays[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing, 2014: 46-56.
[8] STAB C, GUREVYCH I. Parsing argumentation structures in persuasive essays[J]. Computational Linguistics, 2017, 43(3): 619-659.
[9] WAGEMANS J. Analyzing metaphor in argumentative discourse[J]. Rivista Italiana di Filosofia del Linguaggio, 2016, 10(2): 79-94.
[10] SONG W, SONG Z, FU R, et al. Discourse self-attention for discourse element identification in argumentative student essays[C]//Proceedings of EMNLP, 2020: 2820-2830.
[11] LAUSCHER A, GLAVA G, DONIETTO S P, et al. Investigating the role of argumentation in the rhetorical analysis of scientific publications with neural multi-task learning models[C]//Proceedings of the EMNLP, 2018: 3326-3338.
[12] 黄华新,祝文昇.科学论证中的隐喻研究: 以合成生物学隐喻为例[J].浙江大学学报(人文社会科学版), 2022, 52(09): 104-113.
[13] LAKOFF G, JOHNSON M. Metaphors we live by[M]. University of Chicago Press, 2008.
[14] WILKS Y. A preferential, pattern-seeking,semantics for natural language inference[J]. Artificial Intelligence, 1975, 6(1): 53-74.
[15] GROUP P. MIP: A Method for identifying metaphorically used words in discourse[J]. Metaphor and Symbol, 2007, 22(1): 1-39.
[16] KRISHNAKUMARAN S, ZHU X. Hunting elusive metaphors using lexical resources[C]//Proceedings of the Workshop on Computational Approaches to Figurative Language. Association for Computational Linguistics, 2007: 13-20.
[17] TSVETKOV Y, BOYTSOV L, GERSHMAN A, et al. Metaphor detection with cross-lingual model transfer[C]//Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics, 2014: 248-258.
[18] STEEN G J, DORST A G, HERRMANN J B, et al. Vu amsterdam metaphor corpus[DP/OL]. http://www.vismet.org/metcor/search[2022-07-19].
[19] MOHAMMAD S, SHUTOVA E, TURNEY P. Metaphor as a medium for emotion: An empirical study[C]//Proceedings of the 5th Joint Conference on Lexical and Computational Semantics, 2016: 23-33.
[20] WU C, WU F, CHEN Y, et al. Neural metaphor detecting with CNN-LSTM model[C]//Proceedings of the Workshop on Figurative Language Processing, 2018: 110-114.
[21] PENNINGTON J, SOCHER R, MANNING C D. Glove: Global vectors for word representation[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing, 2014: 1532-1543.
[22] PETERS M E, NEUMANN M, IYYER M, et al. Deep contextualized word representations[C]//Proceedings of the Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, 2018: 2227-2237.
[23] GAO E, CHOI E, CHOI Y, et al. Neural metaphor detection in context[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing, 2018: 607-613.
[24] DEVLIN J, CHANG M W, LEE K, et al. BERT: Pre-training of deep bidirectional transformers for language understanding[C]//Proceedings of the Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, 2019: 4171-4186.
[25] LIU Y, OTT M GOY AL N, et al. RoBERTa: A robustly optimized BERT pretraining approach[J]. CoRR, 2019: abs/1907. 11692.
[26] ZHANG S, LIU Y. Metaphor detection via linguistics enhanced Siamese network[C]//Proceedings of the 29th International Conference on Computational Linguistics, 2022: 4149-4159.
[27] SONG W, ZHOU S, FU R, et al. Verb metaphor detection via contextual relation learning[C]//Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, 2021: 4240-4251.
[28] LIU L, HU X, SONG W, et al. Neural multitask learning for simile recognition[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing, 2018: 1543-1553.
[29] ZHU D, ZHAN Q, ZHOU Z, et al. ConFiguRe: Exploring discourse-level Chinese figures of speech[C]//Proceedings of COLING, 2022: 3374-3385.
[30] WEI J, WANG X, SCHUURMANS D, et al. Chain of thought prompting elicits reasoning in large language models[J]. arXiv preprint arXiv: 2201.11903, 2022.
[31] WEST P, BHAGAVATULA C, HESSEL J, et al. Symbolic knowledge distillation: From general language models to commonsense models[C]//Proceedings of the Conference of the North American chapter of the Association for Computational Linguistics: Human Language Technologies, 2022: 4602-4625.
[32] BROWN T, MANN B, RYDER N, et al. Language models are few-shot learners[C]//Proceedings of the 33rd International Conference on Neural Information Processing Systems, 2020: 1877-1901.
[33] BURSTEIN J, MARCU D, KNIGHT K. Finding the WRITE stuff: Automatic identification of discourse structure in student essays[J]. IEEE Intelligent Systems, 2003, 18(1): 32-39.
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}
基金
国家重点研究与发展计划专项课题(2022YFC3303504);国家自然科学基金(61876113);北京市教育委员会科技计划项目(KM202010028004)
{{custom_fund}}
PDF(1168 KB)
