Navigate - Puns

This page summarizes the results of a general project that Kazuko Shinohara (Tokyo University of Agriculture and Technology) and I have been working on. It is about Japanese puns. The project has been inspired by Donca Steriade's theory and aims to test two major theses:

(i) Speakers attempt to minimize the differences between corresponding segments in general, and
(ii) Their measure of similarity has psychoacoustic or perceptual grounds.

To the extent that we can support (i) and/or (ii) in both phonology and in puns, we find interesting parallels between phonological patterns and pun patterns. Although our work has revealed interesting basic patterns, much more needs to be done. However, I feel that we have done enough to pave the way for future work in this domain, and it is my hope that you may take on some of the remaining issues and explore further the intricacy of (Japanese) imperfect puns.

This project is partly funded by a Research Council Grant from Rutgers University to Shigeto Kawahara. #202044, Project title "Verbal art and mental representation of similarity".

Our approach has been corpus-based and experimental, because the traditional, introspection-based study does not seem appropriate for this kind of study--well, it is very difficult to ask ourselves, "is it featural similarity or perceptual similarity that I am making use of when making puns?". Corpus-based and experimental approaches would uncover patterns that may be hidden by other factors. Here is a quote from Shattuck-Hafnagel (1986), with which I cannot agree more (by the way I wish I knew this paper before publishing some of our papers).

A word of caution is in order here about the nature of speech error data. Behavioural data are seldom absolute; if a given factor influences the error corpus more often than chance would predict, then we infer that it plays a role in processing. This does not mean that all errors will be governed by that factor; in fact, a number of counterexamples may arise...This is quite different from the typical line of argumentation in linguistics, where one or a very few unaccountable counterexamples are enough to refute a claim. The difference arises partly from the fact that we are dealing with the results of a complex cognitive process...

(from Shattuck-Hafnagel (1986: 119) The representation of phonological information during speech production planning: Evidence from vowel errors in spontaneous speech. Phonology Yearbook 3: 117-149.)

Although Shattuck-Hafnagel's remark is on speech error data, I share the same philosophy about pun data. This is why we want to get a large corpus of data or do experiments.

If you're interested in this general project, here is a paper that I wrote based on a lecture at Sophia University in July 2008. It gives an overview of this general project. If you prefer a slide with some updated results, here it is.

I have also published a paper that relates this whole project to a larger context of Optimality Theory. It also contains two similarity judgment experiments that support some of the hypotheses made in the papers above. You can download the paper as a pre-publication version.

Our study of consonantal similarity has been published from Journal of Linguistics (the paper). This paper may better contextualize the study in current theoretical debates. The paper builds on my previous study on Japanese rhymes, which appeared in Journal of East Asian Linguistics.

We finished a paper on vocalic similarity, which appears in the Journal of the Phonetic Society of Japan 14 (2010).

We also worked on positional effects--does the position of mismatches matter? The answer seems yes. Speakers do not like to have mismatches in word-initial syllables and in long vowels. Here's the slide and the paper.

We presented our work on syllable intrusion in the Japan Cognitive Linguistics Association and wrote up the proceedings paper.

If you want to get all the papers at once, here is a volume that I put together.

You will find the publication information of the papers in section 6.

So, if you're interested in building on these subprojects, here are some topics that are left for future studies:

Metathesis patterns. We've got some data already. Let me know if you want them (and are willing to collect more examples and analyze them). In particular, it would be interesting to seek for similarities and differences between these data and metathesis patterns in natural languages.

More on positional effects. In one of our experiments we've looked at word-initial consonants, but how about vowels in word-initial syllables (which are not necessarily word-initial)?

A few people have suggested to me that we should look at the role of accents in pun formation. Do accented syllables behave differently from unaccented syllables? (i.e. Do speakers avoid mismatches in accented syllables?) How do accent mismatches affect pun wellformedness? Do they affect more/less than segmental mismatches?

We have looked at short vs. long vowels. How about singletons vs. geminates? i.e. by virtue of being long, do speakers disfavor mismatches in geminates more than in singletons? It seems particularly likely in fricatives and nasals which have internal cues. If we find that this is indeed the case, we may be able to draw interesting parallels with cases of geminate inalterability.

We have used a voicing mismatch to compare initial positions and non-initial positions. The results would be strengthened if we can replicate the same results using other contrasts.

Are positional strengths cumulative? I.e. Would initial long vowels be more prominent than non-initial long vowels or initial short vowels? If so, do positional strengths affect phonological patterns cumulatively as well? In Huariapano (Parker 1998, de Lacy 2001), stressed initial syllables behave differently from other positions, which hints at the cumulativity of strengths.

Positional effects in a containment pattern. In some patterns, one word is properly contained in the other; e.g. Ika-ga ikagawashii 'That squid is suspicious'. In such cases, I have a feeling that the smaller word is contained at the left edge of the larger word, as in the example above. Is this true? If so, why?

Similarity judgment tasks: I think we have a reasonable set of arguments that similarity needs to be defined perceptually. But, nevertheless it is a good idea to obtain a perceptual similarity matrix by some means, and compare that with pun pairing patterns.

More wellformedness judgment tasks. Our studies on consonantal and vocalic similarity are based on corpora. It would be nice to back up our arguments based on more controlled judgment tasks.

Further analyses in relation to other interesting "para-linguistic" patterns, like the slip of the tongue/ear or tongue twister examples. Luckily, there are a tons of websites that report relevant data.

For all of our works, we focused on local similarity rather than global similarity . In other words, we have focused on segmental similarity rather than word-level similarity. Realistically, I think we should represent the similarity by a (linear) equation: Sim_score=a*s1+b*s2+c*s3...x*sx, where (s=similarity of segments, and {a,b,c...x}=weights associated with each position within a word).

In our papers, we have assumed that speakers deliberately choose segments that are similar to make puns. But can it be more passive? Can it be simply that a word actives a set of words that are psychoacoustically similar, and speakers choose a word from the activated set? How do we tease these two options apart? Prof. Otake is pursuing the latter approach these days. See here.

How about other languages? Pun patterns are really understudied (see below). So in my opinion puns are "treasure islands" for linguists...

One reason is that the pun patterns are understudied, so it is of intrinsic academic interest. One practical consequence is that people like it. We usually get positive responses when we submit our paper to a journal or to a conference. Also when I give a general talk on this subject, the audiences--linguists or non-linguists--love it (I can't remember how many times I have given a lecture about Japanese rap rhymes...we get so many invitations!).

Another reason is that it is reasonably easy to study. In fact, one project that we represented at LCC is based on a BA thesis by Yoshida Nobuhiro (Tokyo University of Agriculture and Technology). Nice data are readily available from many websites. Elicitation is easy and fun. If you need to do a more controlled wellformedness judgment experiment, it's not that hard (and you may consider reading this first for an example of how we did it).

I believe in general that pursuing this project can easily result in a presentation at professional conferences or if it's good enough, a journal publication (well, three journals have bought our papers already). I think studying puns would be good for a term paper, a BA or MA thesis, a generals/qualifying paper, and possibly a Ph.D dissertation (well, maybe, if you do it really well and if the results are good). Plus, you will get to practice important research skills like, data collection, data analysis (including stats), and experimental procedures.

So, if you're interested, just email me at kawahara_at_rci.rutgers.edu.

5.1. Some Must-Reads

Kawahara, Shigeto (2009) Probing knowledge of similarity through puns. In T. Shinya (ed.) Proceedings of Sophia University Linguistic Society 23. Tokyo: Sophia University Linguistics Society. pp.111-138. Download Here.

[xxx Well I feel weird to list one of my papers at the top of must-reads, but this paper outlines this project more thoroughly than this website.]

Shinohara, S. (2004). A note on the Japanese pun, dajare: Two sources of phonological similarity. (ms. Laboratoire de Psycheologie Experimentale)

[xxx One of the few phonological analyses of Japanese imperfect puns. Not widely circulated--email me (or Prof. Shigeko Shinohara) if you would like a copy.]

Steriade, D. (2003). Knowledge of similarity and narrow lexical override. In P. M. Nowak, C. Yoquelet, & D. Mortensen (Eds.), Proceedings of Berkeley Linguistics Society 29 (p. 583-598). Berkeley: BLS.

[xxx This project was inspired by this paper.]

Zwicky, A. (1976). This rock-and-roll has got to stop: Junior's head is hard as a rock. In S. Mufwene, C. Walker, & S. Steever (Eds.), Proceedings of Chicago Linguistic Society 12 (p. 676-697). Chicago: CLS.

[xxx This is a classic paper on English rhyming.]

Zwicky, A., & Zwicky, E. (1986). Imperfect puns, markedness, and phonological similarity: With fronds like these, who needs anemones? Folia Linguistica, 20, 493-503.

[xxx One of a few phonological analyses of English puns.]



5.2. And More Stuff To Read...

Kawahara, Shigeto (2010) Papers on Japanese imperfect puns. A collection of papers on imperfect puns.

Fleischhacker, H. (2005). Similarity in phonology: Evidence from reduplication and loan adaptation. Unpublished doctoral dissertation, UCLA.

Holtman, A. (1996). A generative theory of rhyme: An optimality approach. Unpublished doctoral dissertation, Utrecht Institute of Linguistics.

Kawahara, S. (2007) Half rhymes in Japanese rap lyrics and knowledge of similarity. Journal of East Asian Linguistics (JEAL) 16.2: 113-144.

Kawahara, S. and K. Shinohara (2009) The role of psychoacoustic similarity in Japanese puns: A corpus study. Journal of Linguistics. 45.1: 111-138.

Steriade, D. (2001). The phonology of perceptibility effect: The p-map and its consequences for constraint organization. (ms. MIT)

Zuraw, K. (2007). The role of phonetic knowledge in phonological patterning: Corpus and survey evidence from tagalog reduplication. Language, 83 (2): 277-316.



5.3. Scripts & Templates

Some useful resources:

This perl script takes a list of sounds and create a co-occurrence table. Let me know if you want to use it.

This excel sheet contains functions that calculate O/E ratios for you.

6.1. Papers

Get a bibtex file which contains all the references (as of Dec, 2009)

Get a volume that contains all the papers below

Kawahara, Shigeto and Kazuko Shinohara (2011) Phonetic and psycholinguistic prominences in pun formation. In M. den Dikken and W. McClure (eds.) Japanese/Korean Linguistics 18. CSLI. pp.177-188.

Kawahara, Shigeto and Kazuko Shinohara (2010) Calculating vocalic similarity through puns. Journal of Phonetic Society of Japan. An earlier version appears in A. Braver et al. (eds.) RuLing Papers 3.

Shinohara, Kazuko and Shigeto Kawahara (2010) Syllable intrusion in Japanese puns, dajare. Proceeding of the 10th meeting of Japan Cognitive Linguistics Association.

Kawahara, Shigeto (2009) Faithfulness, correspondence, perceptual similarity: Hypotheses and experiments. Journal of the Phonetic Society of Japan. 13.2: 52-61. (A special volume "Development and elaboration of OT in various domains"). Download it as a pre-publication version.

Kawahara, Shigeto (2009) Probing knowledge of similarity through puns. In T. Shinya (ed.) Proceedings of Sophia University Linguistic Society 23. Tokyo: Sophia University Linguistics Society. pp.111-138.

Kawahara, Shigeto and Kazuko Shinohara (2009) The role of psychoacoustic similarity in Japanese puns: A corpus study. Journal of Linguistics. 45.1: 111-138.


Kawahara, Shigeto (2007) Half rhymes in Japanese rap lyrics and knowledge of similarity. Journal of East Asian Linguistics (JEAL) 16.2: 113-144. Eratta.

6.2. Talks

Kawahara, Shigeto (2009) Coda devoicing and perceptual similarity. Colloquium talk. November 20th, 2009. Stony Brook University.

Kawahara, Shigeto (2009) Testing the P-map hypothesis I: Coda devoicing. A workshop on phonological similarity. November 15th 2009. MIT. Slide.

Shinohara, Kazuko and Shigeto Kawahara (2009) Syllable intrusion in a Japanese language game, dajare. The 10th meeting of Japan Cognitive Linguistics Association. Kyoto University, September 27th, 2009.

Kawahara, Shigeto (2009) Testing P-map. Colloquium talk. June 1st 2009. University of Chicago.

Kawahara, Shigeto (2009) Probing knowledge of similarity through puns. MIT phonology group talk. April 18th 2009. MIT. Slide.

Kawahara, Shigeto (2009) Probing knowledge of similarity through puns. A colloquium talk. March 20th 2009. University of Delaware. Slide.

Kawahara, Shigeto and Kazuko Shinohara (2008) Phonetic and psycholinguistic prominences in pun formation: Experimental evidence for positional faithfulness. Japanese/Korean Linguistics 18. City University of New York, November 13th. Slide.

Kawahara, Shigeto (2008) Probing knowledge of similarity through puns. An invited speaker for the workshop "Phonology from the Perspective of Speech Perception" accompanying the 23rd meeting of Sophia Linguistic Society. Sophia University, July 19th, 2008.

Kawahara, Shigeto, Kazuko Shinohara, Nobuhiro Yoshida (2008) Positional effects in Japanese imperfect puns. Language, Communication, and Cognition. Brighton University, August 4th. v

Two Rutgers undergraduate students, Ayanna Beattie and Allan Schwade, have worked on Korean tongue twisters, and presented at the Harvard undergraduate linguistics conference (April 18-19th, 2009).

Vandana Bajaj, a graduate student at Rutgers, has worked on Hindi-English bilingual jokes.

Jonah Katz (former graduate student at MIT) is approaching similar issues using English hip hop rhymes.

Sverre Sausland Johnsen at Harvard has worked on some relationship between perceptual similarity and rhymability.

Prof. Takashi Otake (E-Listening Laboratory) has been studying Japanese dajare from a slightly different perspective.

Kazuko Shinohara, my primary co-investigator on this project, wrote a paper that discusses linguistic restrictions on word games in general, and how speakers can and cannot deviate from such restrictions. In the paper she introduces a novel language game called 'hiragana kookan'.

Kazuko, Shinohara (2009) Designing language games. A special issue of Japan society for the science of design. 16-2. no 62. 45-50.