Joint Detection and Location of English Puns

Proceedings of NAACL-HLT 2019, pages 2117–2123 Minneapolis, Minnesota, June 2 - June 7, 2019. c⃝2019 Association for Computational Linguistics 2117 Joint Detection and Location of English Puns Yanyan Zou and Wei Lu StatNLP Research Group Singapore University of Technology and Design yanyan zou@mymail.sutd.edu.sg, luwei@sutd.edu.sg Abstract A pun is a form of wordplay for an intended humorous or rhetorical effect, where a word suggests two or more meanings by exploiting polysemy (homographic pun) or phonologi- cal similarity to another word (heterographic pun). This paper presents an approach that ad- dresses pun detection and pun location jointly from a sequence labeling perspective. We employ a new tagging scheme such that the model is capable of performing such a joint task, where useful structural information can be properly captured. We show that our pro- posed model is effective in handling both ho- mographic and heterographic puns. Empirical results on the benchmark datasets demonstrate that our approach can achieve new state-of- the-art results. 1 Introduction There exists a class of language construction known as pun in natural language texts and utter- ances, where a certain word or other lexical items are used to exploit two or more separate mean- ings. It has been shown that understanding of puns is an important research question with various real-world applications, such as human-computer interaction (Morkes et al., 1999; Hempelmann, 2008) and machine translation (Schr¨oter, 2005). Recently, many researchers show their interests in studying puns, like detecting pun sentences (Vadehra, 2017), locating puns in the text (Cai et al., 2018), interpreting pun sentences (Sevgili et al., 2017) and generating sentences containing puns (Ritchie, 2005; Hong and Ong, 2009; Yu et al., 2018). A pun is a wordplay in which a certain word suggests two or more meanings by exploiting polysemy, homonymy, or phonologi- cal similarity to another sign, for an intended hu- morous or rhetorical effect. Puns can be gener- ally categorized into two groups, namely hetero- graphic puns (where the pun and its latent target are phonologically similar) and homographic puns (where the two meanings of the pun reflect its two distinct senses) (Miller et al., 2017). Consider the following two examples: (1) When the church bought gas for their annual barbecue, proceeds went from the sacred to the propane. (2) Some diets cause a gut reaction. The first punning joke exploits the sound similar- ity between the word “propane” and the latent tar- get “profane”, which can be categorized into the group of heterographic puns. Another categoriza- tion of English puns is homographic pun, exem- plified by the second instance leveraging distinct senses of the word “gut”. Pun detection is the task of detecting whether there is a pun residing in the given text. The goal of pun location is to find the exact word appearing in the text that implies more than one meanings. Most previous work addresses such two tasks sep- arately and develop separate systems (Pramanick and Das, 2017; Sevgili et al., 2017). Typically, a system for pun detection is built to make …

Content truncated to 500 words