Script-Based Semantics: Foundations and Applications. Essays in Honor of Victor Raskin 9781501511707, 9781501517433

Citation preview

Script-Based Semantics

Script-Based Semantics Foundations and Applications. Essays in Honor of Victor Raskin Edited by Salvatore Attardo

ISBN 978-1-5015-1743-3 e-ISBN (PDF) 978-1-5015-1170-7 e-ISBN (EPUB) 978-1-5015-1149-3 Library of Congress Control Number: 2019956075 Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.dnb.de. © 2020 Walter de Gruyter Inc, Boston/Berlin Cover image: saemilee / DigitalVision Vectors / Getty Images Typesetting: Integra Software Services Pvt. Ltd. Printing and binding: CPI books GmbH, Leck www.degruyter.com

Preface Years ago, I had suggested to Victor Raskin that we publish something to honor his outstanding contribution to humor studies, in HUMOR, the journal I was editing at the time. With his usual dry wit, Victor suggested that we only ask people who disagreed with him, in a sense producing an anti-Festschrift . I was not particularly keen on the idea of only including critics of Raskin’s semantic-script theory, as were many scholars who would have been quite happy to participate in a Festschrift, but did not feel up to criticize the theory, but I hoped it would make for a strong issue of the journal and so it came to see the light in 2004, as volume 17(4) of HUMOR: International Journal of Humor Research. Yet, the knowledge that the humor of an anti-Festschrift would be inevitably be lost on someone1 kept nagging at me and so I recently approached Victor again and suggested that I edit another Festschrift, a regular one this time, and to boot, since it would be a book, that we do not limit it to his work on humor. Victor agreed and we spent some time discussing who the contributors might be. This was difficult because of course there were so many people whose work and careers had been impacted by his work. Eventually, we settled on a mix of criteria, with on the one hand academic excellence and on the other a certain representativeness of the breadth of his work. This meant we could not ask all the candidates, because a book has its all-too-real limits in size and in the capacity of the editor and publisher to handle the mass of publications we would have received if we had asked everyone who was influenced by Victor’s work to contribute, in all the fields where his work resonates. Once we made the selection of the authors, I proceeded to contact them, not without a certain trepidation because some of them I had never met in person or even interacted with them. All those I contacted agreed immediately and with enthusiasm. I can say that as a result of this project my perception of Victor’s work has changed: I was already a fan, of course, but the sheer breadth of his influence is staggering. I am sure that many of those who are familiar with one side of his work will be surprised and perhaps shocked to find out that not only there’s more, but that in other fields Victor’s influence is just has profound. On top of it, this work has led me to interact with a group of really fine scholars and this has been a pleasure and a blessing in itself.

1 It was. I have had to explain why we did this to several colleagues who approached me in confidence at various conferences. https://doi.org/10.1515/9781501511707-202

VI

Preface

I would like to thank Victor Raskin for his help in the realization of this volume, and of course for all his help and guidance during the many years we have known each other. I would also like to thank Marina Raskin not just for feeding me each time she had a chance, but in general for putting up with my presence for the years I spent in Lafayette and further. I would also like to thank Paul Baltes, Lucy Pickering, Marnie Jo Petray-Covey, Cinzia Di Giulio, and Julia Rayz, who helped me in various ways during this task. Of course, I would like to thank the referees who read the papers and provided valuable feedback to the authors. They must remain anonymous, but my gratitude to them is in no way diminished. Needless to say, I would like to thank the authors for submitting their papers (mostly) on time and working with me on the changes, corrections, and edits with good humor and expeditiousness (all the time). Last but not least, I would like to thank Lara Wysong and Kirstin Börgen, and the editorial staff at Mouton De Gruyter for supporting the project and accommodating my frantic schedule. The Middle of Nowhere Ranch, near Lone Oak, TX 2020

Contents Preface

V

Salvatore Attardo Introduction 1

Part 1: Script-based semantics Salvatore Attardo Scripts, frames, and other semantic objects

11

Irina Kobozeva Script-based approach towards taxis connectors

43

Jonathan Dunn Ontological and grammatical constraints on metaphor productivity Igor Boguslavsky Meaning amalgamation, phrasal stress, and earning money

55

77

Part 2: Humor Giselinde Kuipers Knowledge about humor

93

Willibald Ruch Domains of humor: Challenges from psychology

115

Villy Tsakona Victor Raskin’s overlooked analysis of political jokes Władysław Chłopicki Joke construction and joke structure

139

167

Béatrice Priego-Valverde ‘Stop kidding, I’m serious’: Failed humor in French conversations

191

VIII

Contents

Part 3: Ontological semantics Julia Taylor Rayz Scripts in the Ontological Semantic Theory of Humor

229

Vladik Kreinovich and Olga Kosheleva Which fuzzy logic operations are most appropriate for ontological semantics: Theoretical explanation of empirical observations 257 M.L. Gavrilova Decoding intricacies of human nature from social network communications 269

Part 4: Other applications Dallin D. Oaks A creative approach for linguistic funny business: Using linguistic paradigms and taxonomies 281 John D. Battenburg Tourism after the Arab Spring in Tunisia: An analysis of advertising campaigns 305 Names Index Subject Index

315 321

Salvatore Attardo

Introduction The work of Victor Raskin spans over five decades, several fields of linguistics, and a remarkable variety of topics. It would be impossible to summarize it all in one introduction, however long, not to mention that the task would overwhelm my (finite) skills. The readers will gather nonetheless, form the range of authors and topics included in this collection, how broadly his work has reached. Much more modestly, the task I am setting for this introduction is twofold: (1) briefly provide some historical orientation to the reader unacquainted with either scriptbased semantics or its application to humor studies, and (2) to briefly summarize the papers in the book. Raskin’s work is largely focused on semantics. Raskin’s interest in semantics goes back to his work in Russia and Israel. For example, semantic recursion, which will figure prominently in script-based semantics, is the object of a 1968 paper (Raskin 1968) published in Russian. Simplifying drastically, semantic recursion consists of “filling in” all the information not directly provided by the lexical items in an utterance/sentence. It is completely context-aware/dependent (and in this sense, pragmatic, in the current sense of the word). For example, in a highway sign advising drivers to “watch your speed” “your,” setting aside the deictic external reference provided by the literal meaning of the word, is addressed to drivers and not to pedestrians or joggers. Note how the context affects the determination of who “you” is in the sign. Besides the profoundly original (and still troubling to many readers, to this day) radical pragmatic view on meaning determination, Raskin moved on to consider the lexical part of meaning in as broad a view as that of recursion. Influenced by the Generative Semantics movement, computational linguistics, and supported by a broad approach to semantics, it is perhaps not surprising that he was drawn to the work on meaning representation in Artificial Intelligence, from which the idea of script-based semantics originates (see a review in Attardo’s chapter). Raskin’s first published work explicitly on script-based semantics is Raskin (1981), but obviously his work on script-based semantics had begun much earlier, since Raskin (1979) is an application of script-based semantics. In 1985, Raskin edited two special issues of Quaderni di Semantica, which included a “Round Table Discussion on Text/Discourse” in which a variety of articles by prominent semanticist were published, including three articles by Raskin himself (Raskin 1985c/d/e). Incidentally, Fillmore’s oft-cited (1985) article on frame semantics is part of the special issue. The chapter on script-based semantics from Raskin (1985a) was further republished in an updated version in Raskin (1986). His work https://doi.org/10.1515/9781501511707-001

2

Salvatore Attardo

on script-based semantics continue and eventually evolved and contributed to creating the field of ontological semantics. The evolution of Raskin’s thought into ontological semantics can be fully appreciated in Nirenburg and Raskin (2004). Raskin’s application of script-based semantics to humor was first presented in Raskin (1979) and then in fuller form in Raskin (1985a)1 and subsequently in an article (1985b) for the general public. Needless to say, Raskin has published numerous other works on humor, but the locus classicus of the semantic-script theory of humor is Raskin (1985a). The evolution toward ontological semantics has also affected Raskin’s work on humor, see Raskin et al. (2009). The reader will thus find in this collection articles influenced by Raskin’s view of semantics, by his work on script-based semantics, by his work on ontology, by his work on the application of script-based semantics to humor, and by his long standing interest in the application of linguistics (e.g., Bjarkman & Raskin 1986).

1 Part 1: Script-based semantics Turning now to the summary of the articles, Attardo discusses the origins of object-oriented semantics, a neologism intending to express neutrality among the terminological abundance of scripts, frames, schemata, etc. Attardo is trying, in a sense, to determine the zeitgeist from which Raskin’s script-based semantics emerged. He also highlights aspects that differentiate script-based semantics from other versions of object-oriented semantics and considers some interesting applications of object-oriented semantics, primarily to cognitive linguistics. Kobozeva addresses the temporal organization of scripts in subscripts and the semantic/pragmatic effects that can be achieved by selecting default or marked fillers for these slots. In the case of such “immediate anteriority” connectors as “as soon as” or “once,” the two elements whose temporal distribution is in question can be expressed either via coordination or subordination. Kobozeva shows that if the script does not set the two events in a position of immediate precedence (such as in squeeze the trigger-fire the gun) then “as soon as” or “once” can be employed to set the two events in temporal adjacency: “they married as soon as they had gone out of a few dates” (normally the script for marriage defaults to the two agents knowing each other rather well). Another situation is if the content of the first event is part of the common ground or known information: “as soon as the

1 The colophon lists both 1984 and 1985 as publication dates, which has led to some confusion. Common usage is to quote it as 1985, but the book appeared in 1984.

Introduction

3

party started, Bob got drunk” where serving alcohol at parties is part of the script and thus can be assumed to be known information. Dunn considers the semantics of metaphors. His starting point is that metaphorical mappings are conceptual, as assumed within the now-standard cognitive linguistic analysis, but are nonetheless influenced by the lexical semantics of the words used to implement the mapping. Dunn then proceeds to show that ontological constrains restrict which concepts can be mapped onto another; generally speaking, a less abstract entity is more likely to be metaphorically productive if mapped onto a more abstract entity than the other way around. Finally, he shows that grammatical constraints, more specifically valencies of verbs affect the likelihood of metaphorization. Dunn concludes that “script-based semantics is necessary to describe the linguistic productivity of metaphoric mappings.” Boguslavsky works with the idea of valency, a concept that has been used in the Moscov semantics school and in the meaning/sense-text approach, e.g., Mel’čuk (1981). This idea is one of the foundations of script-based semantics, as seen more clearly in the use of variables in ontological semantics to capture precisely valencies. These are complex structures that postulate underlying levels that will bring to mind the semantic deep structure of the seventies, again unsuprisingly given the origins of frame semantics/script-based semantics in the generative semantics movement.

2 Part 2: Humor Kuipers argues for the addition to the General Theory of Verbal Humor (GTVH; Attardo and Raskin, 1991) and namely of a knowledge resource concerning metahumorous knowledge, “knowledge about humor” embodying the socio-cultural knowledge of “1. how, 2. when, 3. by whom and with whom humor is used.” i.e., “rules and conventions related to the use of humor, as they are held by specific sociocultural groups. “ This interesting approach is sure to attract significant attention in the future. Ruch’s very personal and yet profoundly theoretical article considers the disciplinary differences between the psychological and the linguistic approaches. Essentially, he sees them as differing at the level of granularity, insofar as linguistics distinguishes between jokes that would fall under the same factor in the psychological classification, or as Ruch puts it “there is some overlap between the two approaches but clearly the tools provided by linguistics exceed the diversity that the layperson is able to reflect in their responses” as they are reflected in psychological factors. This methodological differences results in a lack of symmetry

4

Salvatore Attardo

between the approaches, with classifications that are orthogonal vis-à-vis each other: the general theory’s parameters do not describe well the factors in the 3WD and likewise, factor analysis does not fit well in the GTVH’s schema. A possible exception to this is the NON vs. INC-RES factors in the 3WD, which can be distinguished by the presence or absence of a LM (NON has a zero LM). Ruch then goes on to consider Raskin’s 1998 theory of the sense of humor (which is separate and distinct from the SSTH and has received far less attention) and designs a test that could in principle validate it. Priego-Valverde applies the switch to the non-bona-fide mode of communication to conversational failures in which the recipient of a bona-fide-story reacts with humor, unexpected and unappreciated by the original bona-fide speaker. Priego-Valverde incorporates creatively Raskin’s competence-level discussion in a discourse analytic model, in a great example of transdisciplinary crosspollination as rare in scholarly research as it is fruitful. Her discussion focuses on the dynamics of the negotiation between the speakers for who gets to frame the interaction as serious (bona-fide) or humorous (non-bona-fide). To be fair the “negotiation” at time turns into an almost open fight for dominance and indeed Priego-Valverde focuses on the failures of humor that originate from (failed) switches between modes. Tsakona analyzes a little-explored aspect of Raskin’s 1985 classical presentation of the semantic script theory of humor, the analysis of political humor. Raskin is widely seen as the advocate of a purely theoretical analysis of humor, whereas Tsakona provides a useful reversal of this view, highlighting one of the “applied” aspects of Raskin’s work. Other such aspects, are ethnic humor and, in later work, computational humor. Tsakona concludes that Raskin’s model of political jokes transfers fairly well to her Greek data and that therefore further investigation of other political humor data in different contexts is advisable. Chłopicki’s contribution is dedicated to joke construction, in a sense the complementary perspective to the most usual view of Raskin’s theory as a theory of joke analysis. As Chłopicki notes, Raskin does address joke construction in Raskin (1985a), but does not focus on it. This article aims at correcting the imbalance. Chłopicki’s discussion is obviously preliminary as he makes no attempt at a systematic quantitative study, but the anecdotal data nonetheless reveal that joke construction and performance go hand in hand and that speakers may vary significantly in the amount of divergence from the model they were given in their performance of the text. In these choices, which I have called (Attardo 2017) the repertoire of the joke tellers, and in the personality factors that cause the speakers to make the choices are probably to be found cues as to what makes a better or worse performance of a joke.

Introduction

5

3 Part 3: Ontological semantics Rayz re-analyzes some of the classic examples in the STTH and GTVH using the OSTH, thus showing both the continuity between the three theories and the strength and power of the Ontology-based version of the theory. Her article goes a long way toward the goal of providing a fully explicit analysis of some of the examples of the SSTH: due to space and time limitations,2 most of the analyses in the literature are extremely sketchy and often merely mention the opposed scripts, without giving any specific indications of how these are actually found. The article by Kreinovich and Kosheleva tackles a very interesting feature of ontological semantics, i.e., the “fuzzy” nature of many of the restrictions of an ontology. Fuzzy set theory was introduced by Lotfi Zadeh (1965) and, simplifying significantly, can be described as the idea that set membership is not a yes-no question, but a matter of degree. As it turns out, much information in an ontology is fuzzy. For example, consider the script for the action of PAINTING. The agent must be human, but a monkey who, having observed a human paint, does so as well, would still be described as painting. Donkeys have been reported to have painted canvases,3 and recently an AI has also been reported to have “painted” a work.4 So, some of the information in an ontology needs to be duly fuzzy. Kreinovich and Kosheleva’s concern is how to find the right degree of generality. Taylor & Raskin (2010) have provided an answer which Kreinovich and Kosheleva confirm and prove mathematically. The chapter by Marina Gavrilova delves into online identification, a topic seemingly very different from others in this book. However, when one consider the premise that social behavior can be used for biometric applications, such as user identification, and that humor recognition is part of these social behaviors that can be tracked, the connection becomes at once apparent. Gavrilova, in a visionary article, outlines the potential for uses of humor for applications in the biometric space. Unlike previous applications using humor preferences to infer personality traits, Gavrilova outlines how features of the humor produced or

2 To give an idea, in Attardo (2001: 9–15) I analyze the first sentence of a short story by Oscar Wilde. The sentence is 17 words long but would generate over 40 million possible semantic combinations, and those are merely the lexical combinations. Obviously, pragmatic constraints reduce the plausible interpretations, but the point remains. 3 Patty, the painting donkey, can be seen at work in a YouTube video (https://www.youtube. com/watch?v=kzEZNHC3_Y8). The donkey made it on Britain got talent, a TV talent show. 4 The Guardian reports that the painting sold at auction for $432,000. (https://www.theguar dian.com/artanddesign/shortcuts/2018/oct/26/call-that-art-can-a-computer-be-a-painter)

6

Salvatore Attardo

liked by the user could be used to build a “profile” that could then be used to identify the user.

4 Part 4: Other applications Oaks broadens his successful approach consisting of considering linguistic structures as enabling ambiguity, rather than traditionally of linguistic processes reducing it, (see Oaks 2010) by applying it to semantic and pragmatic processes, such as implicatures, presuppositions, and deictics. The underlying idea being of course that ambiguity is often at the root of humor and that therefore by enabling ambiguity one (may) enable humor. Oaks approach is subtly different from other studies considering humor in that the point is not to analyze humor but to enable its production, in this more akin to Raskin’s work with Weiser on writing (Raskin & Weiser 1987) where the point is not to analyze linguistically the writing of students but to help them produce better writing. Battenburg chapter touches on two topics not commonly associated with Raskin’s work, but that instead fit in perfectly: one the one had is the application of the standard semantic-script opposition analysis of humor to advertising discourse in a tourism study. Leisure studies is of course a fast growing new research area developed in the new millennium. On the other hand is the attention to professional discourse. Raskin’s work on applications of linguistics to the professions, mentioned above (Raskin & Weiser 1987) was concerned with the discipline of composition and rhetoric, for example, but see Bjarkman & Raskin (1986) for a broader stance. Battemburg’s work inscribes itself comfortably between these two strands of research.

References Attardo, Salvatore. 2017. The GTVH and humorous discourse. In: Władysław Chłopicki & Dorota Brzozowska (eds.), 2017 Humorous discourse. 93–105. Boston, Berlin: Mouton De Gruyter. Attardo, Salvatore and Raskin, Victor. 1991). Script theory revis(it)ed: joke similarity and joke representation model. HUMOR: International Journal of Humor Research 4(3–4). 293–347. Bjarkman, Peter C., & Raskin, V. (Eds.). 1986. The real–world linguist: Linguistic applications in the 1980s. Ablex Publishing Corporation. Fillmore, Charles. 1985. Frames and the semantics of understanding. Quaderni di Semantica 6 (2). 222–254. Mel’čuk, Igor A., 1981. Meaning–text models: A recent trend in Soviet linguistics. Annual review of Anthropology, 10(1),27–62.

Introduction

7

Nirenburg, S., & Raskin, Victor. 2004. Ontological semantics. Cambridge, MA:MIT Press. Oaks, Dallin D. 2010a. Structural ambiguity in English: An applied grammatical inventory. 2 vols. London & New York: Continuum. Raskin, Victor. 1968. O semanticheskoy rekursii /On semantic recursion/ in: Vladimir A. Zvegintzev (ed.), Semanticheskie i fonologicheskie problemy prikladnoy lingvistiki. Vol.3, Moscow: Moscow State University Press. 268–283. Raskin, Victor. 1971. K teorii yazykovykh podsistem. [Towards a Theory of Linguistic Subsystems] Moscow: Moscow University Press, 1971 Raskin, Victor. 1979. Semantic mechanisms of humor. In C. Chiarello et al. (eds.) Proceedings of the Fifth Annual Meeting of the Berkeley Linguistic Society. 325–335. Raskin, Victor. 1981. Script–based lexicon. Quaderni di Semantica 2:1. 25–34. Raskin, Victor. 1985a. Semantic Mechanisms of Humor. Dordrecht–Boston–Lancaster: D. Reidel. Raskin, Victor. 1985b. Jokes. Psychology Today. 19. 24–39. Raskin, Victor. 1985c. Linguistic and encyclopedic knowledge in text processing. Quaderni di Semantica 6(1). 92–102. Raskin, Victor. 1985d. Script–based semantics: a brief outline. Quaderni di semantica 6(2). 306–313. Raskin, Victor. 1985e. Once again on linguistic and encyclopedic knowledge. Quaderni di Semantica 6(2). 377–383. Raskin, Victor. 1986. Script–Based Semantics. In Donald G. Ellis & William A. Donohue (eds.). Contemporary Issues in Language and Discourse processes. Hillsdale, NJ: Lawrence Earlbaum. 23–61. Raskin, Victor & Irving Weiser. (1987). Language and Writing: Applications of Linguistics to Rhetorics and Composition. Norwood, NJ: Ablex. Raskin, Victor, Christian F. Hempelmann, & Julia M. Taylor. 2009. How to understand and assess a theory: The evolution of the SSTH into the GTVH and now into the OSTH. Journal of Literary Theory, 3(2),285–311. Taylor [Rayz], Julia & Victor Raskin. 2010. Fuzzy ontology for natural language, Proceedings of the Annual Conference of the North American FuzzyInformation Processing Society NAFIPS’2010, Toronto, Canada, July 12.14, 2010. Zadeh, Lotfi A. 1965. Fuzzy sets.” Information and Control 8(3). 338–353.

Part 1: Script-based semantics

Salvatore Attardo

Scripts, frames, and other semantic objects to Victor Raskin, guida e maestro

Abstract: In this article I review the history of various approaches to semantics that share the premise that meaning is (to a greater or smaller extent) encoded in semantic “objects” (scripts, frames, schemata, MOPs, etc.) and their relations (semantic graph). I compare the various approaches, highlighting in particular the common, shared elements of Object-Oriented Semantics (OOS). I will then focus on some of the defining features of OOS, including, primelessness (i.e., the lack of basic, undefined or fundamental privileged objects), fractality (the fact that there exists the same level of complexity at all levels of the network), dynamicity (the fact that semantic objects are built by trial-and-error and are always subject to revision) and falsifiability. Keywords: scripts, frames, schemata, conceptual dependencies, object-oriented semantics

Few would disagree that one of the great unifying themes of the research in Artificial Intelligence, Linguistics, Psychology, Cognitive Science, and related fields is the analysis of meaning and its storage in memory, in terms of constructs that organize and “package” the meanings, memories, or visual scenes. These constructs are defined largely, but not necessarily exclusively, by their associations to other units or other constructs. Another point that is not likely to be challenged is that these constructs were named with a dizzying broad set of terms, which include schemata, frames, scripts, conceptual dependencies, Memory Organization Packets, planes, Idealized Cognitive Models, and much more. In order to avoid prejudging the issue by picking a term to be the “basic” or simplex idea, as does the term “schema theory” (e.g., Sadoski et al. 1991), and defining the others in those terms, even if implicitly, I resort to a neologism,1 objectoriented semantics (OOS): “semantic object” refers to any construct, regardless of 1 Others have remarked on the similarities between script/frame semantics and objectoriented programming, e.g., Lassila & McGuinness (2001). Salvatore Attardo, Texas A&M University-Commerce https://doi.org/10.1515/9781501511707-002

12

Salvatore Attardo

size, internal organization, formalization, and ontological status, that describes an element of meaning. So, when I refer to Frame semantics, the reader should take it that I mean the type of OOS associated with Fillmore and his followers, when I refer to script(-based) semantics, that I refer to the type of OOS associated with Raskin and his followers, and so on. In what follows, we will start with a review of the complex and interleaved history of semantic objects (section 1) and then will consider what features uniquely characterize and define Raskin’s script-based semantics (section 2). Needless to say, what follows is not a comprehensive treatment of semantics and/or the psycholinguistics of semantic memory. I am only trying to delineate some of the influences and intellectual ancestry of Raskin’s script-based semantics.

1 A brief history of objective semantics The work of Bartlett (1932) is widely acknowledged as the starting point of OOS. However, it is fair to say that to consider him just as the inventor of the conceptual schema is to significantly restrict his influence and the nature of his work, in particular the role of discourse and genre in his work (Edwards and Middleton 1987). Bartlett’s work on schemata is inscribed within his research on memory. The central idea is that remembering is constructive and not reproductive (204–205) (i.e., more like a recipe than a photograph). “Human remembering is (. . .) really a construction, serving to justify whatever impression [on the reader] may have been left by the original” (1932: 175–176). “It is an imaginative reconstruction (. . .) built out” of schemata. (213) For Bartlett, schemata are based on experience (1932: 201), or more specifically they are an “organised mass of experience” (210), “active” (213), organized chronologically (203), “interconnected” with each other, i.e., they “overlap” with each other (212). Schemata are dynamic: “Any given schema is continually being transformed in the process of remembering as past and present experiences interact with the enduring interests and attitudes of the individual.” (Johnston 2001: 344) Bartlett has a surprisingly modern view of meaning, which he defines as an “effort after meaning,” i.e., “the attempt to connect something that is given with something other than itself” (p. 227), or in more contemporary terms, we would define this as the semiotic function: a sign is something that stands for something else. Another aspect that would be worth investigating are the similarities between Bartlett’s views and Gestalt theory (see Anderson and Person 1984). Overall, Bartlett’s influence on OOS is striking: virtually every major

Scripts, frames, and other semantic objects

13

scholar in OOS quotes Bartlett, and the general idea of an organized, interconnected, dynamic mental structure has proven very fruitful.

1.1 Artificial Intelligence In the 1970s, Artificial Intelligence, or as it has since been re-labeled “Good Old-Fashioned Artificial Intelligence” (Haugeland 1985) was in its heyday, showing enormous promise. Much of the promise went unfulfilled in the 1970–1980 period and symbolic AI was largely (but not completely) supplanted by connectionism first and big-data fueled machine learning later. Among the early practitioners of AI, the notion of OOS found significant favor, albeit under a variety of names and with significantly different interpretations: e.g., Minsky, (1974); Schank, (1975); Schank & Abelson, (1977). In what follows I will present a selective set of examples, for which no claim of historical representativeness is made. Rather, the studies reviewed below are those that either are central to the development of OOS or that I perceive to be closest to script-based semantics and that may have influenced it. In a sense, what I am attempting to do is to reconstruct the zeitgeist of the origins of script-based semantics.

1.1.1 Conceptual dependencies One of the sources of OOS is the idea of conceptual dependencies. The term “dependencies” comes from Tesnière’s (1959) work. Klein & Simmons (1963), Klein (1965), and Hays (1961, 1964) used it, in its original syntactic sense, in the context of machine translation, as an alternative to phrase-structure and immediate constituents. This work influenced directly Schank (Schank & Tesler 1969; Schank 1972), and Quillian (see below). Schank applies the idea of dependencies to semantics, and hence “conceptual” dependencies. It is unclear how much Schank was aware, at the time, of Tesnière’s work on “valence,” a concept very close to Fillmore’s case grammar. Conceptual dependencies share with OOS the assumption that meaning is an independent linguistic layer, decoupled in principle from the linguistic form (syntax, morphology, phonology) but also from the actual lexical choices. Expressions or sentences with the same meaning should be represented in the same way in conceptual dependencies, regardless of their form/lexical choices. In the context of machine translation, this assumption is reflected in the idea that it is preferable to translate from L1 into a semantic metalanguage and then from the semantic metalanguage into L2 (Nirenburg & Raskin 2004). To put it

14

Salvatore Attardo

differently, the semantic base, or the meaning, of a text is independent from the language in which it happens to be worded. Another influence on conceptual dependencies is that idea that relationships among words are labeled. Ceccato (1962, 1964), again, in the context of machine translation, arrives at an inventory of “correlators” (semantic relationships) between entities. While the number of correlators is well above 150, Ceccato points out that contextual pressure disambiguates very quickly, even just for two words in sequence, thus making the approach computationally tractable (72–73). All the correlations are part of a “correlational net” (74). Ceccato’s model also includes surprisingly contemporary “constellations” which are more semantic, such as the connection between “part” and “whole” or “activity-instrument” and “paintbrush” in the context of painting. (85–86) Conceptual dependencies were based on a set of 10–12 predicates, such as PTRANS (i.e., the transfer of location of an object), PROPEL (i.e., the application of a physical force to an object), MTRANS (i.e., the transfer of mental information between agents). Each predicate had slots, that indicated which conceptual dependencies were associated with it. So, for example, PTRANS had the following slots: – ACTOR – OBJECT – FROM – TO which stood for, the agent of the transfer of location, the object moved, the location where the object was originally, and the location where it ended up, respectively. That the object was moved from the FROM location to the TO location, was an inference, not explicitly stated in the text, but that must be calculated by the system and was thus encoded in the conceptual dependency. Inferences were triggered by a variety of factors, but a particularly interesting one is a missing slot. Suppose I say that John took his car to the shop. We have filled the ACTOR (john), OBJECT (the car), and TO (the shop) slots, but there is no mention of the FROM slot. A conceptual dependency-based inference would be that John’s car had to be somewhere before it was moved to the shop. A more sophisticated inference, not tackled by the original model, would be that it had to be somewhere which was not the shop itself! Scripts were introduced in the conceptual dependencies models as “precompiled” (Lytinen 1992: 52) or “prepackaged sequences of causal chains” (Schank 1980: 253) which were essentially stereotypical events encoding and enabling inferences. The RESTAURANT script (Schank and Abelson 1977) is probably the best known example. As Lytinen (1992) notes, scripts were both meant as inference processing shortcuts but also as structures in memory, and

Scripts, frames, and other semantic objects

15

evidence of this has been found (e.g., Abbot et al. 1985). Memory Organization Packets (MOPs) were further introduced (Schank 1980; 1982) to store knowledge common across scripts. So, for example, the script for DOCTOR VISIT and LAWYER VISIT share the PROFESSIONAL VISIT MOP. Scripts inherit much of their information from MOPs: thus for example, that there may be a waiting room, would come from the PROFESSIONAL VISIT mop rather than from DOCTOR. Summing up, in the Conceptual dependency model there are conceptual structures at different levels: the most concrete are conceptual dependencies, these are organized and stored in scripts, which in turn are selected by plans, which are themselves guided by goals. Finally, goals are organized in themes (Schank 1980: 256; 1999: 137–154) but we will not concern ourselves with this level in the present discussion. In Schank (1980), we find a dynamic view of scripts, which are not seen as stored in memory, but rather dynamically built: “an active memory structure that changes in response to new input” (1999: 11). Scripts are not activated all at once. On the contrary, Schank sees scripts as activated ad hoc, on a asneeded basis in the process of parsing a text: The [. . .] script itself does not actually exist in memory in one precompiled chunk. Rather, it, or more likely its needed subparts, can be constructed as needed. “ [. . .] Why bring an entire script in while processing if it will not be used? Since scripts are being constructed rather than being pulled in whole from memory, only the parts that there is reason to believe will be used (based upon the input text) need to be brought in. (262)

This is a significant point, that has not been adequately stressed in OOS: the instantiation of the abstract script, stored in memory, is not wholesale. Only the relevant parts of the script are activated. In Schank’s view, scripts in memory are aggregates of episodic memories, specific instances of events in which one experience a situation or event (1980: 269–270). Or as he puts it in a later publication summarizing his work, “one’s concept of an event is simply the collocation of one’s repeated encounters with that event” (Schank 1999: 110). However, this ignores the role played by the abstract representations of words in lexical memory. Consider the fact that for many concepts, such as unicorn, average, freedom, and algebra, it is unlikely that one ever experiences them in any other form than the instantiation of a lexical item. It stands to reason to assume that lexical items play a significant role in the formation of memories. Indeed this is the position of contemporary research in concept formation: “abstract lexicalized representations are critically important for the ability to reason and form arguments that could be all but impossible to form by strictly perceptual means” (Sloutski 2010: 1249). Thus it is likely that

16

Salvatore Attardo

when hearing or reading the word “restaurant” one activates a generic, abstract script for restaurant, which is then actualized with the further details provided by the text, if any. This is obviously very different from the experience of walking into a restaurant. It seems impossible that the presence in memory of an abstract representation of the meaning of restaurant would not be connected with the script for restaurant. In conclusion, Schank’s conceptual dependencies, MPs, and scripts, while unquestionably central in the history of OOS, are rather different objects than scripts, as the concept is used in script-based semantics. The two biggest differences are the assumption that conceptual dependencies are based on universal primes and the neglect of the role of lexical semantics in the conceptualization of scripts.

1.1.2 Minsky’s frames According to Minsky (1968: 9), around 1962, the field of AI became more concerned “with the problem of representation of knowledge” and, relatedly with “the representation and modification of plans” (9) moving away from brute-force heuristics. Minsky’s and his students’ work played a very significant part in this process. Indeed, Minsky (1968) is an edited collection of summaries of several PhD dissertations directed by Minsky, which include crucially Quillian’s work, which had a massive influence on OOS. Minsky (1974), widely considered “one of the more influential contributions to Knowledge Representation” (Brachman and Levesque 1985: 245), was originally published as an MIT technical report and then reprinted in several other versions. It introduces the concept of “frame” as a “data-structure for representing a stereotyped situation” (1974: 1; 1985: 245) as well as the idea of seeing it as a “network of nodes and relations” (1974: 1, 1985: 245), the idea of a slot-filler format (slots are called confusingly “terminals”; 1974: 1, 1985: 245), and the idea that frames may contain sub-frames. Frames also encode “expectations” by having some of their slots pre-filled with “default assignments” which corresponds to the stereotypical expectations of the speakers. Thus, for example, a frame for BAR will have a slot for the agent preparing and serving the drinks, typically filled by a BARTENDER sub-frame. These default assignments are “attached loosely” to their slots “so that they can be easily displaced by new items that fit better the current situation” (1974: 2). Indeed, upon walking in a bar in which the person behind the counter is wearing a clown suit, one would need very little effort to re-align to the new situation (a clown-bartender or a bartender-clown). Minsky sees frames as representing both word meaning and larger nonlinguistic concepts. He presents a “story” frame, containing slots for “setting,

Scripts, frames, and other semantic objects

17

protagonists, main events, moral” (30) that gets progressively filled with information: “As the story proceeds, information is transferred to superframes whenever possible, instantiating or elaborating the scenario” (30). A striking feature of Minsky’s proposal is how significant the debt is to Bartlett’s work, which gets quoted repeatedly (e.g., 1974: 3, 6, 24, etc.), along with Charniak’s (1972) thesis directed by Minsky.2 Another remarkably prescient aspect of Minsky’s proposals is how broad the potential applications of frame are, encompassing sentence processing and understanding, visual processing, reasoning about a story, semantic cases (which we will see fully developed in Fillmore, see below; Minsky discusses Fillmore’s cases explicitly (1974: 29)) learning, memory, clustering, and analogical reasoning. Finally, Minsky addresses the claim that frames are just a notational variant of first order logic, which has been advanced repeatedly (Hayes 1979; Lassila and McGuinness 2001: 2; see also the discussion in Woods (1991: 48); Woods himself holds a different position: Woods (1975)). This claim is directly contradicted by Minsky’s pellucid discussion of the non-monotonic nature of knowledge (Minsky 1974; 1985) and of several other problems, which lead him to the conclusion that “‘Logical’ reasoning is not flexible enough to serve as a basis for thinking.” (262) From a different perspective, Sowa (1984; 1992; 2008) conceptual graphs combine Tesnière’s valencies, akin as we will show below to Fillmore’s case grammar (mediated via Quillian’s work), Tesnière’s dependency-based syntax (mediated via Hays’ work), and predicate calculus, thus showing that perhaps the two approaches are not incompatible. The first formulation of Sowa’s conceptual graphs go back to 1968, in a course with Minsky. Only in Sowa (1984) was the connections with Peirce’s graphical treatment of first degree logic added. Sowa’s interest is primarily logical and has had little impact on objective semantics.

1.1.3 Charniak Charniak (1972) introduced the term “deep semantic processing” to describe the processing that results in the semantic representation of the meaning of sentences. His model is rich with a panoply of curiously named mechanisms (demons,

2 Another author quoted repeatedly is Piaget. We cannot pursue this observation in this context, but Piaget’s constructivism and his schemes, distinct from schemata, deserve more consideration within the context of object-oriented semantics.

18

Salvatore Attardo

for example, which are subroutines that look for a given bit of information). Some of these are however familiar: “base routines express what we know about a given concept” (1972: 39) regardless of context. Charniak (1972) is focused mostly in the demons’ reasoning on the basis of information they glean from the text or that is somehow given to them by the author of the program in which they run. While this may be useful in the processing of text by a computer, it is of limited interest from a linguistic standpoint.

1.2 Fillmore’s cases and frames Fillmore’s contribution to linguistics was truly remarkable. In what follows we will very briefly sketch two of his contribution that most directly affected the development of OOS: Case grammar ad Frame semantics.

1.2.1 Case grammar Case grammar is probably one of the most successful ideas in modern linguistics having been adopted both in purely formal/syntactic models, such as “theta roles”, and in semantic approaches, such as frame-semantics. Originally proposed in Fillmore (1968), it was then updated in Fillmore (1977). One of the sources of the idea of semantic cases is in Tesnière’s (1959) ideas of actants and circumstants and more generally of semantic valencies of verbs (Fillmore 1982: 114). Fillmore (1968: 38) follows Tesnière in rejecting the need for a VP, or in other words, posits the centrality of the verb,3 rather than the predicate (Fillmore 1982: 114). Essentially, Fillmore sees a sentence as a “flat” structure in which the verb is associated with each NP in the sentence with a particular semantic relationship (case), irrespective of its morphological or syntactic realization (e.g., through affixes, word order, etc.), and with the crucial constraint that each case may occur only once per sentence (Fillmore 1968: 42). Each verb then comes with a list of cases that it takes and with semantic specifications, such as [+/- animate], so that for example, KILL should have a [+animate] objective case.

3 An interesting observation, in Minsky (1974: 29), is that while clearly sentences are centered around the verb, its centrality diminishes as the transphrastic structures grow in size. This is consonant with Fillmore’s description of discoursal frames.

Scripts, frames, and other semantic objects

19

In the 1968 version of case grammar, Fillmore posits six cases: agentive, instrumental, dative, factitive, locative and objective (1968: 46); however, he leaves open the possibility of adding to the list, and indeed in the 1971 version the number of cases grows to eight: agent, instrument, experiencer, object, location, source, goal, time and benefactive (Fillmore 1971). The value of the cases is fairly standard, except perhaps for the factitive, which describes the result of an action or state, as in John baked a pie, and the objective, which is not the accusative or direct object, although it will correspond in many cases to those, as in John smoked a salmon. It is obvious how deep the influence of case grammar was on OOS: the characterization of verbs as having a set of valencies, to use Tesnière’s term, which can be described as slots and fillers (the “case frame” of the verb; Fillmore 1968: 48) and the limited inventory of semantic relations between the verb and its valencies, are part and parcel of the theoretical apparatus of OOS. A comprehensive discussion of case theory, including discussion of other models, presented by Wallace Chafe, Jeffrey Gruber, and Kenneth Pike, among others, can be found in Cook (1989).

1.2.2 Frame semantics Fillmore (1977, 1982, 1985) discusses relativizing cases to “scenes”: in the COMMERCIAL EVENT scene; there are several prototypical components: the buyer, the seller, the goods, and the money. Another example are the “verbs of judging” such as accuse or criticize, which assume a scene in which there is a “Judge,” a “Defendant,” and a presupposition that the situation is “blameworthy.” Fillmore sums up his views as follows: we have here not just a group of individual words, but a ‘domain’ of vocabulary whose elements somehow presuppose a schematization of human judgment and behavior involving notions of worth, responsibility, judgment, etc., such that one would want to say that nobody can really understand the meaning of the words in that domain who does not understand the social institutions or the structures of experience they presuppose. (1982: 116)

This is a common thread in Fillmore’s thought. Already in Fillmore (1976: 28) we find the same argument for “alimony”: “understanding this word requires knowing the whole scenario” (28) of divorce, the court system, etc. Note, in passing, how this comes close to Trier’s “word field” (1931). However, Fillmore stresses the methodological differences between frame semantics and wordfield theory (Fillmore & Atkins 1992: 76).

20

Salvatore Attardo

These scenes have another level of significance. Fillmore notes that “any single clause that we construct in talking about such an event requires us to choose one particular perspective on the event” (1977: 5). This idea shows Fillmore’s evolution toward the idea of frames, to which we turn presently, but not before noting that the idea of perspectives in a scene will be very fruitful in cognitive linguistics (see below). In 1975, Fillmore defines a frame as “any system of linguistic choices [. . .] that can get associated with prototypical instances of scenes.” (124) Scenes are defined as “any kind of coherent segment of human beliefs, actions, experiences or imaginings” (124) including but not limited to “visual scenes [. . .], familiar interpersonal transactions, standard scenarios defined by the culture, institutional structures” (124). Clearly, Fillmore was very much aware of the research that was taking place in that period on the subject. Right from the onset, Fillmore’s (1975, 1976) definition of frame (or “module” as in modular-furniture) is eclectic,4 encompassing Bartlett’s work, Minsky’s and Quillian’s work in AI, field-semantics (Trier 1931; Geckeler 1971). Fillmore (1982: 130) also quotes Schank and Abelson (1977) on scripts. Fillmore espouses a broad interpretation of frame semantics as applying to and providing a “uniform representation for word meanings, sentence meanings, text interpretations and world models” (1976: 28) which we saw already in Minsky (1974). Fillmore stresses the networked aspect of frame semantics: speakers have a “model of the world” which consists of a network of interlinked relationships representing bits of knowledge and the ways in which these bits of knowledge are integrated into a more or less coherent model or image of the world. (Fillmore 1976: 26)

and its prototypical organization, in contrast to the so-called check-list theories of meaning (Fillmore 1975). Prototypicality will be addressed in more detail below, in the section on cognitive linguistics. It is interesting to note that Fillmore’s frames share with script-based semantics the encyclopedic semantic network approach, but with one difference, namely that Fillmore, seems to be less interested in the hierarchical (ontological) aspect of the network, and later Framenet will display this bias as well (Ruppenhofer et al. 2016: 9). Another similarity between script and frame semantics is that Fillmore, like Minsky, anticipates the work on narrative scripts: “the development on the part of the interpreter of an image or scene or picture of the world, as that gets built up and filled out between the beginning and the

4 Although the term “frame” was initially used by Fillmore in reference to syntactic frames (1982: 374).

Scripts, frames, and other semantic objects

21

end of the text-interpretation experience. (Fillmore 1975: 125) An example of this approach, within script-based semantics can be found in Attardo (2001). Fillmore’s frame semantics veered toward lexicography (e.g., Fillmore & Atkins 1992) and was eventually implemented into a computational linguistic lexicon called Framenet. The progression in that direction is well documented in Petruck 1996, which is also a good discussion of Fillmore’s frame semantics, overall. The literature on Framenet is large. See Fillmore et al. (2003); Fontenelle (2003); Boas (2001, 2005). Recent work on Framenet can be found in Fried & Nikiforidou (2013); Ruppenhofer et al. (2016); Boas & Dux (2017); see also Andor (2010). The influence of Fillmore’s work on OOS was very broad. We mentioned the prototypical nature of frames. This idea, based on the work on prototypicality by Rosch and her colleagues (e.g., Rosch & Mervis 1975), will become central in some versions of OOS and in cognitive linguistics (see below). Semantic cases are explicitly mentioned in Minsky (1974: 29, 40–41; 1985: 253) and in Schank’s (1972) limited set of conceptual cases part of his conceptual dependencies.

1.3 The semantic graph Quillian’s (1966; 1967) work on semantic networks is certainly one of the most significant influences on OOS. Explicitly designed to have psychological reality (i.e., be a model of human memory and language processing), Quillian’s semantic network5 is defined as “a mass of nodes” which are “interconnected by different kinds of associative links” (1966: 13). The meaning of each node is given by links to “a configuration of other nodes that represents the meaning” of the given node. Note the completely intensional, non-referential definition of meaning. The associative links are labeled: “several different kinds of associative links, rather than the simple undifferentiated associations assumed in most classical psychological studies of word association” (23). Overall, then “the memory [=semantic network] is a complex network of attribute-value nodes and labeled associations between them.” (32). Incidentally, word meanings, as linguistic units, are encoded in Quillian’s memory constructs. This is stated explicitly: e.g., in discussion of the disambiguation of “words whose meanings are encoded in

5 The term “semantic net” is used in Masterman (1961: 442) but to describe a Boolean lattice, in which “interlingual aspect indicators” i.e., “recurrent aspects of basic situations which occur in real life” (449).

22

Salvatore Attardo

the memory model” (Quillian 1966: 35). Quillian quotes an early version of Fillmore (1968) on semantic cases, as part of his discussion on disambiguation. Nodes can be types or tokens (instantiations). Tokens of nodes can be arranged in configurations, which yield planes: the configuration of interlinked token nodes [. . .] represents a single concept as comprising one plane in the memory. Each and every token node in the entire memory lies in some such plane, and has both its special associative link pointing “out of the plane” to its type node and other associative links pointing on within the plane to other token nodes comprising the configuration. (1966: 14)

Quillian’s model is explicitly not based on primitives (1966: 21) and does not have an ISA hierarchy (21), so in this sense it is a flat network. Also, nodes are not lexical items, but rather concepts or “properties” (26) since concepts and mental images “may be thought of as some bundle of properties (attribute values) and associations among them.” (26). A “full concept” for Quillian is the sum of all the connections of a given concept, the “patriarch,” it include[s] all the type and token nodes one can get to by starting at the initial type node, or patriarch, and moving first within its immediate definition plane to all the token nodes found there, then on “through” to the type nodes named by each of these nodes, then on to all the token nodes in each of their immediate definition planes, and so on until every token and type node that can be reached by this process has been traced through at least once. (18)

This definition presents a problem, in that the algorithm described has no stopping point, and would eventually encompass the entirety of all the concepts in the network. However, this is easily fixed by a spreading-activation like mechanism (see below) or by a hardwired or pragmatic limit to how many recursive calls to other concepts one should perform. However, a far more interesting point can be made on the subject, to which I turn next.

1.3.1 Unlimited semiosis This aspect of OOS will require the introduction of a little-studied aspect of Charles Sanders Peirce’s semiotics and namely unlimited semiosis. Peirce’s thought is complex and evolved throughout his life, so that it is often difficult to pinpoint definitions. However, his conception of the sign as a tripartite entity, remains constant: “a sign is something, A, which brings something, B, its interpretant sign determined or created by it, into the same sort of correspondence with something, C, its object, as that in which itself stand to C.” (1902a) A sign

Scripts, frames, and other semantic objects

23

creates an interpretant which is itself a sign and therefore will be interpreted by another interpretant, and so on. Peirce is quite explicit about the infinite progression, so to speak, of the semiotic process: a sign, is, in Peirce’s definition, Anything which determines something else (its interpretant) to refer to an object to which itself refers (its object) in the same way, the interpretant becoming in turn a sign, and so on ad infinitum. (1902b[1931 vol. 2: 203])

Eco (1975: 176) notes that since, in Peirce’s view, any concept is defined by its connections with other concepts (interpretants, in Peirce’s terminology) each concept may be interpreted (defined) by all the other concepts, as we saw above: starting from any concept one can reach any other concept in the network. This corresponds, Eco argues, to the Peircean notion of “unlimited semiosis” (176). Hayes (1979) in the process of rebutting the claimed difference between frames and other semantic representations, notes that frames contains named ‘slots’, which can be filled with other expressions – fillers – which may themselves be frames, or presumably simple names or identifiers (which may themselves be somehow associated with other frames, but not by a slot-filler relationship: otherwise the trees formed by filling slots with frames recursively, would always be infinitely deep) (Hayes 1979: 452)

Thus, he is coming tantalizingly close to the correct intuition that the trees are in fact infinitely deep as they are connected in a vast semantic network. Eco brings up an interesting problem: the two-dimensional representation used by Quillian works only for a very limited number of nodes and links. Visualizing or even conceptualizing the complete graph would require a “polydimensional” graph, with “changing topological properties” i.e., in which the path from one node to another may change in response to its neighboring nodes, dynamically. Eco uses the metaphor of a box full of spheres, which when shaken, rearrange themselves in different relations of proximity/distance. I believe that a non-planar graph, i.e., one that allows edges crossing, could handle the complete semantic graph. I see no particular reason to require that the semantic graph be planar. So, if we combine Quillian’s and Peirce’s ideas about infinite graphs and unlimited semiosis, we arrive at a conception of a purely intensional semantics, in which the meaning of a given entity is given by its connections to other entities. Obvioulsy the complete absence of referential function needs to be accounted for, but this can be handled by what we may call “referential anchors” i.e., semantic entities within the graph that have a clear and socially accepted referent and are probably lexicalized for convenience (say /cat/). So, if one needs to use language to refer to something, one can do so by either using a referential anchor

24

Salvatore Attardo

directly (e.g., “the cat is on the mat”) or by using an intensional descriptor built from the referential anchor: for example, in the science fiction series Red Dwarf the character “Cat” a humanoid cat, or more precisely it “is a descendant of Dave Lister’s [another character in the series] pregnant pet house cat Frankenstein, whose descendants evolved into a humanoid form over three million years while Lister was in stasis (suspended animation).” (Wikipedia, accessed Jan 1, 2019) Needless to say this entity is fictional, but the point is that we can understand and even visualize this entity, despite its lack of referent, due to its connection to a referential anchor.

1.3.2 Spreading activation Quillian’s model represented “the gradual activation of each concept outward through the vast proliferation of associations originating from each patriarch, by moving out along these links, tagging each node encountered” (38) Collins and Loftus (1975) named this process “spreading activation” (the term is not in Quillian 1966) and in the process modified it somewhat. Spreading activation is a psychological process and as such merits discussion here only because it is used in Attardo (2001) as a general process to build textual scripts from the activation of lexical and other scripts in the text. In a slightly different sense, Raskin’s combinatorial rules of meanings (1985a: 86; Quillian 1966: 111–132) can be seen as an abstract form of spreading activation, with the combinations rejected by the combinatorial rules being assigned zero activation.

1.4 General considerations We can now attempt to sum up a general consensus of features pertaining to OOS. Let us start as a “zero degree” case, with Brachman (1979) overly pessimistic conclusion that the only thing that all the above proposals (and a few similar others we have not considered for limitations of space) share is “their connectivity – the fact that they all claim to be made of links and nodes.” (203) The previous discussion should be enough to belie this statement. However, Brachman does have a point, in that a network or graph is merely a representation or an implementation of a semantic theory. There isn’t any special semantic theory embodied in a graph of nodes and links. In fact, according to Sowa (1984), in its implementation of Peirce’s existential graphs, it is equivalent to first-order predicate calculus.

Scripts, frames, and other semantic objects

25

In fact, OOS’ graphs have far more specificity than Brachman claims. First, they are intensional descriptions of the meaning of the nodes. They do not rely on the connection with a referent or on truth-values. In other words, the meaning is a function of the (sub-)graph. Second, the links are labeled and correspond to semantic relations (e.g., Mel’čuk 1981). Third, in most versions of OOS, the graphs are primeless or the primes are for convenience. Fourth, OOS is based on the idea that meaning is encyclopedic and there are no theoretically tenable boundaries between lexical and encyclopedic semantics nor between semantics and pragmatics. In this sense, a good slogan for OOS would be “meaning is meaning, wherever it may be found.” Fifth, semantic objects are recursive, nesting entities. Sixth, under most versions, semantic objects are prototypical and/or stereotypical, i.e., they describe typical situations or entities with relaxable criteria. I think that the above list should satisfy the skeptics that OOS is making substantive testable claims about its object. I will now turn to delineating more specifically script-based semantics.

2 Fundamentals of script-based semantics In this section I will examine, briefly and non-technically, some defining features of script-based semantics, trying to determine its place relative to the other forms of OOS, we have discussed. These features are not exhaustive, i.e., I am not claiming that script-based semantics can be reduced to these features. My claim is that these features are significant to distinguish script-based semantics from other forms of OOS. The features, in no particular order, are the following: 2.1 primelessness 2.2 fractality 2.3 dynamicity 2.4 falsifiability 2.5 broad definition of knowledge

2.1 Primelessness Primelessness is one of the crucial stances of script-based semantics, which is rather notable considering that some of its historical sources, such as conceptual dependencies were not primeless (Schank 1972); likewise, cases were assumed to

26

Salvatore Attardo

be very limited in number (Fillmore 1968). Indeed, many structural semanticists tried to propose lists of “simpler” or “more fundamental” semantic markers (Katz and Fodor’s 1963 markers and distinguishers, for example) only to have to acknowledge that the purported distinctions were untenable. Ultimately, this mistaken avenue contributed to the fall of structural semantics. However, Lytinen’s claim that primitives in semantics have “disappeared” (p. 51) is in error: primitives are alive and well, only in non-mainstream approaches such as Wierzbicka’s universal metalanguage. This is not the place for a discussion of Wierzicka’s ideas, but let us note that she and her followers assume that there is a smallish (20+) inventory of semantic entities that are a) universal, i.e., they appear in all languages, and b) they are self-explanatory and cannot be further decomposed. (E.g., Wierzbicka 2011: 393). Among the supposedly indivisible and self-evident semantic entities are “live,” “die,” “move,” and “moment” (Wierzbicka 2011: 382). I will not address the claim of universality, but the claim of indivisibility and self evidence seem prima facie obviously false. Not long ago, linguists decomposed “die” as “become not alive” and “moment” seems quite a complex concept to be self-evident. It is patent that “moment” can be further analyzed as “a short period of time.” Script-based semantics sides firmly on the side of the lack of primes, quite simply because it can do without them and thus does not need the carry the (meta)theoretical baggage of postulating universal semantic structures or implausible units of meaning that cannot be further analyzed. In fact, the lack of primes of script-based semantics follows from its definition of scripts as graphs, as we will see below.

2.2 Fractality It’s turtles all the way down (Ross 1967: 4)

I will now show, briefly and without any mathematical proof, that the fractality and the unlimited semiosis aspects of the semantic network are corollaries of, i.e., follow from, the definition of scripts as subgraphs of a connected labeled graph. The technical definition of a script is based on graph theory (Raskin 1985a/b; Bales & Johnson 2005).6 Each node is a vertex and the edges are the

6 Bales & Johnson (2005) provide a clear introduction to graph-theoretic semantic representation. See also Attardo et al. (2002) for an application of graph-theoretical modeling of scripts to humor.

Scripts, frames, and other semantic objects

27

links between them. Since we have different links (such as isa, agent, hasparts, etc.) the edges of the graph are labeled and directed (i.e., the direction in which they are traversed matters: BANANA isa FRUIT is allowed but *FRUIT isa BANANA is not). A subgraph is a graph in which all vertices and edges are a subset of another graph. So, if we consider the overall semantic graph as the original graph, each script is a subset of the overall graph. Since there are no restrictions on cyclicity or more generally on what the fillers of the slots may contain, it is intuitive that any filler may be described by a combination of slot-filler structures (and for that matter, any slot can be so described as well, probably to the consternation of the purists wanting to keep the metalanguage distinct). Since any filler may contain a slot-filler structure, this makes scripts recursive. This recursion is potentially infinite. A different way of expressing this is that we will find at each level of the semantic network the same level of complexity (i.e., scripts), which is the nature of some fractals (Mandelbrot 1977: 18; “self-similarity”). A different terminology is used by Bales & Johnson, who refer to “scale-free topology, which tends to have a similar appearance when examined at varying scales.” (2005: 451). Needless to say, the fractal nature of the semantic network is in sharp contrast with the idea of semantic “primitives.” These primes would need to somehow exist “outside” of the regular semantic space (either because they are genetically encoded in the human DNA, with all the complications that this entails, or because they are Kantian a-priori forms of the intellect which human reason is bound to rediscover). The idea of the fractal nature of the semantics script network correspond to Kobozeva’s observation (this volume) that sub-events within a script are also scripts “at a different level of granularity.” Generally speaking, granularity is a significant problem for semantic analysis, since different situations may need to instantiate progressively more specific (finer granularity) versions of the script (or other semantic object). For example, consider the script for INSIDER TRADING. A basic definition, gleaned from Wikipedia, is the following: Insider trading is the trading of a public company’s stock or other securities (such as bonds or stock options) by individuals with access to nonpublic information about the company. (Wikipedia. Accessed December 24, 2018)

A script for INSIDER TRADING might include that it is a kind of trading, done by an agent who has some kind of information which other agents (the public) do not have. This is correct, but as we find out, still on Wikipedia, a few lines below, an “insider” is a specific class of people: “a company’s officers, directors and any beneficial owners of more than 10% of a class of the company’s equity securities.” There is nothing that cannot be handled by a script, all we need is

28

Salvatore Attardo

to add this information in the slot for the “agent” of the trading. However, a further few lines below we find out that insider, in the US, “can include any individual who trades shares based on material non-public information in violation of some duty of trust.” What that means, I have no exact idea, and I assume all my non-legally trained readers share this ignorance, but a lawyer at the U.S. Securities and Exchange Commission, the legal body governing these violations, presumably does. In fact, such an individual would probably add even more specificity to the definition. All these different “levels” of specificity exemplify granularity: at a broad level, for general communication, we don’t need too much detail, but when the discourse becomes specialized, more specificity (smaller granularity) can be added. It is possible that there is a “bottom” to granularity, i.e., for example, actual instances of a phenomenon: so, for example, a specific legal case in which actual individuals did specific things might be as deep as granularity needs to go, but of course there may be further necessity of granularity, for example to distinguish committing insider trading on different days that results significant gains or in actual losses. On levels of abstraction see also Reed (2016).

2.3 Dynamicity We have seen that the structures built in OOS are dynamic, i.e., they are updated as needed: already Bartlett saw schemata as continuously changing; Schank conceived of scripts as changing in response to new information; Minsky’s frames are built “loosely” so that new information can be added or changed as needed. Raskin’s scripts are built and upgraded through a process that follows closely Popperian falsifiability: essentially, one can consider one’s current version of a script for a given concept as a hypothesis on the content of said concept. I will explore this in more detail below. An interesting issue, which I cannot explore in this context, is what happens when the system is confronted with the equivalent of cognitive dissonance, i.e., two bits of information that are not compatible and cannot be assimilated in one away or another (for example, as we did above with the humanoid cat assimilated as fictional). In some cases, this triggers humor (a.k.a., script opposition; Raskin 1985), but in some cases the other constituents of humor may be missing, for example, signals of the intention to be funny or of a non-bona-fide mode of communication (see Priego-Valverde, this volume), overlap of the scripts, etc. In these cases of unredeemed contradiction other mechanisms will need to handle the situation.

Scripts, frames, and other semantic objects

29

2.4 Falsifiability A strong objection has been leveled against the entire field of “frame-based” accounts of reading/memory (Brown 1979; Sadoski et al. 1991). Generally speaking, the criticism that the field is fairly murky and that the definitions are vague and therefore essentially impressionistic is a fair one. What is striking in Fillmore’s definition of frame (above) is how eclectic it is. Brown (1979: 221) suggests polemically a test of relevance: removing the word “schema” from any sentence that contains it and, according to Brown, the sentence would still function. The idea being that “schema” provides little or no semantic meaning but merely a positive connotation. From this perspective, Raskin’s script-based semantics (1985a/b)introduces a welcome corrective by formalizing the process for determining the content of a script: by noting that scripts are essentially falsifiable, i.e., they can be tested against the native speaker’s intuition or any new information available to a system, the vagueness and openness of the definitions is replaced by a Popperian falsificationist process. Essentially, a script can be considered a hypothesis (put forward by a speaker or a system) relative to the known information. For example, consider the script for CRYPTOCURRENCY, and assume a minimal amount of information known, namely a) it is a currency, and b) it is based on some cryptographic algorithm. An hypothetical speaker would consider this the meaning of cryptocurrency. When faced with novel information, such as the fact that any transaction in cryptocurrency can be traced, he/she would add this to his/ her knowledge. Essentially, in Popperian terms, this is equivalent to falsifying the hypothesis and revising it in light of new evidence. From this it follows that scripts are never static (proven true) but are always subject to revision (not yet proven wrong). While elegant theoretically, this also provides a very strong empirical method to build up scripts: start with a general meaning and refine the hypothesis by considering an increasingly large number of sentences/contexts, until the definition stabilizes, i.e., it no longer needs to be revised frequently. This heuristic is probably the least acknowledged aspect of script-based semantics and yet it is methodologically possibly the most important one, as it addresses the criticism mentioned above, that object based semantics is just a fancy notational variant of plain lexical semantics.

2.5 Broad definition of knowledge Finally, I think that a significant difference between script-based semantics and other OOS models is the breadth of its descriptions.

30

Salvatore Attardo

2.5.1 Encyclopedic knowledge Last but not least, scripts/frames describe not just lexical knowledge but also encyclopedic knowledge. Leaving aside the fact that no easy theoretical boundary can be found to distinguish the two, let us assume that empirically the kind of information likely to be found in a dictionary is lexical knowledge, whereas the kind of information one is likely to find in an encyclopedia, which exceeds lexical knowledge, is encyclopedic knowledge. This has been widely assumed and discussed (Raskin 1985a/b), so an example will suffice. Consider the following: (1)

I went to have a donut, but the coffee was terrible.

The presence of “but” and of “the” are indicators of the depth of semantic information associated with the lexical handle “donut” (at least, in American culture). “But” is a logical connector which contrasts two entities that are somehow connected; consider (2)

John is smart but impractical

(3)

Mary is tall but does not play basketball

being smart is expected to correlate with being practical, being tall is likewise expected to correlate with playing basketball, for obvious reasons. This can be shown even more forcefully by considering example in which there is no correlation: (4)

*John is smart but does not play basketball.

(5)

*Mary is tall but impractical.

From this it follows that the script DONUT must contain information about the fact that donuts are often eaten while drinking coffee and that therefore, coffee is sold in most places where donuts are sold. Needless to say, no lexical description of “donut” might accommodate that sort of information.

2.5.2 Ontological and lexical knowledge We have seen that Fillmore and Quillian have relatively “flat” models of the semantic network. Indeed, as we will see in the discussion on cognitive linguistics

Scripts, frames, and other semantic objects

31

uses of OOS, this has already been remarked upon. Conversely, script-based semantics incorporates as a central feature an ISA hierarchy which contributed significantly to its evolution into ontological semantics (e.g., Nirenburg and Raskin 2004). There is much that could be said about the significance of the hierarchical component of the semantic network, including for example, the “depth” of the ontology, a common benchmark in the field, or the problems tied to inheritance of features from superordinate concepts. However, this would require delving into ontological semantics, which is beyond the scope of this paper. I will however note that we know from psychological research (e.g., Brown 1958), that concepts come with a “basic level (e.g., dog). The superordinate (or hyperonym) canine, and the hyponyms (great pyrenees, poodle, collie) are less accessible, less salient, and are acquired later by children and thus presumably will require greater “effort” to be activated. One way of seeing this, in a topological metaphor, is that the distance between dog and cat is smaller than the distance between cat and great pyrenees. To the best of my knowledge this insight has not been incorporated in OOS. Ontological knowledge is not language specific. However, script-based semantics and to some extent its ontological counterpart, have remained “close” to lexical semantics, seeing its ultimate task as analyzing and understanding actual text, in a way that has prevented the loss of specificity of projects that are not directly related to linguistic analysis. I think a good example of this is the fact that in its original formulation (Raskin 1985a/b) the only way for scripts to be activated was by occurring in a text. Later versions of script-theory allow for other forms of activation (see Rayz this volume)

3 Applications, criticisms, and empirical evidence In this final section I will, very briefly review some applications of OOS, for example to cognitive linguistics and reading theory, as well as some criticisms of OOS. Obviously, one of the most significant applications of OOS to artificial intelligence (or computational linguistics or to the “semantic net”) is the theory of ontology. However, we will not pursue this discussion because essentially Raskin’s script-based semantics morphed into ontological semantics and so an entire part of this volume is already dedicated to the subject. Thus, keeping in mind these limits, we can start from a prima facie surprising source of empirical evidence in favor of OOS, i.e., syntactic theory.

32

Salvatore Attardo

3.1 Syntactic arguments in favor of OOS Lambecht (1984) provides evidence for the psychological reality of frames from the capacity to occur as bare binomials (i.e., NPs without an article) outside of frozen expressions, as in (6)

It was cold in the apartment. She took bucket and shovel to get coals from the cellar.

(6) contrasts with (7) (7)

#She was furious. She took bucket and shovel and threw them out of the window. (Lambrecht 1984: 786; the # symbol indicates pragmatic anomaly)

“Bucket and shovel” is not a frozen expression, along the lines of “salt and pepper,” and so it can occur as a bare binomial only in the context of the frame COAL-HEATING, whereas there is no frame connecting a bucket and a shovel to throwing objects, and so the objects have to be determined by the article. The coordinate structure constraint (Ross 1967) states, roughly speaking, that nothing can be extracted out of a coordinate structure, so if we have a coordinate structure such as (8)

Mary ate cake and John ate pizza.

extraction from one of the elements is ungrammatical: (9)

*What cake did Mary eat and John ate pizza?

There is one known exception to this rule, the so-called across-the-board extraction, i.e., if the same constituent is extracted from both coordinates. Lakoff (1986) argues that a purely syntactic analysis is untenable and that the constraints on Coordinate structure extraction are governed by frame-like information. Consider example (10) (10) Sam is not the sort of guy you can just sit there, listen to, and stay calm. (1986: 153) in which “Sam” is “extracted” from the second conjunct (“listen to [Sam]”) but nothing is extracted from the other coordinates. What accounts for this and

Scripts, frames, and other semantic objects

33

other similar examples? According to Lakoff, extraction is allowed from any coordinate if the coordinates are part of an established frame. Thus (11) What did he go to store, buy, load, in his car, drive home and unload? is acceptable because it follows the frame for SHOPPING, at least in the American suburban environment.

3.2 Cognitive linguistics It is fair to say that cognitive linguistics has adopted the idea of OOS. In fact, Lakoff’s (1987) Idealized Cognitive Models (ICM) and Langacker’s (1987) domains are central aspects of the development of cognitive linguistics. A quick survey of introductions to cognitive linguistics shows that they all have some significant discussion of OOS, usually under guise of Fillmore’s frame semantics. Croft and Cruse (2004), for example, actually start their treatment of the field with a discussion of frame semantics (7–28), capped by discussion of ICMs and domains. Ungerer & Schmid (1996) also have a discussion of frame semantics (205–222) as do Evans & Green (2006: 222–230). Cognitive linguistics emphasis is inspired primarily by Fillmore’s frame semantics, so it is not surprising to see that they emphasize more the different perspectives within a frame, such as for example the COMMERCIAL EVENT frame, which can be seen as buying, from the perspective of the purchaser, or selling, from the perspective of the seller, or spending, from the perspective of the cost, etc. Furthermore, cognitive linguistics is interested in distinguishing “aspects” within the frames themselves. One such differentiation is between profile and base, so for example, DIAMETER or RADIUS makes sense on within the base of CIRCLE (Evans & Green 2006: 237; Croft & Cruse 2004: 15), BAILIFF, in the US sense, within the base of COURT and ALIMONY within the base of DIVORCE. Another differentiation are “active zones” which correspond to the parts of a frame that are activated or foregrounded: Evans & Green (2006: 238) use the example of a soccer player (“footballer”) kicking the ball or waving at the crowd and specify that the foot is active in the first example, whereas the arms and hands are in the second. Likewise, our FOOTBALLER has a liver and a spleen, barring extremely unlucky circumstances, which are inherited from HUMAN, but these are backgrounded.

34

Salvatore Attardo

3.2.1 Lakoff’s ICM Idealized Cognitive Models are based on Fillmore’s frame semantics, with an emphasis on prototypicality and the construction of categories: Lakoff uses the example of “Tuesday”, which can be defined only with reference to the concept of week, which is idealized as a series of seven days, labeled in a given order in which Tuesday comes third (Sunday, Monday, Tuesday). Note the idealization whereby we are not referring to any specific week, but to an ideal sequence of seven days. Lakoff considers then the classical example “bachelor” and notes that there are unmarried males that are better or worse instances of bachelor: the Pope and the Dalai Lama are not prototypical examples of “bachelor,” for example. Cognitive models may be augmented metonymically or metaphorically, others are scalar or radial, i.e., there is a core meaning and derived meanings organized around it; “mother” is such a case. There is a core meaning of mother which is a female who gives birth to a child, supplies half of the genes, takes care of the child, is married to the father of the child, etc. There exist many “satellite” cases in which some or most of these conditions are not met: for example, a step mother is married to the father of the child but did not give birth to it; a biological mother gave birth to the child but is not raising it, for some reason. At this point, it becomes clear that ICMs are rather broader than other constructs in OOS, but in fact ICMs are even broader still: according to Lakoff they structure mental spaces, such as hypotheticals, counterfactuals, etc. While scripts, frames, schemata, etc. are relatively small, linguistic structures (fittingly “objects,” in my metaphor) ICMs are more cognitive, non-necessarily linguistic, broader categories of thought, that structure and underlie our cognitive processes. While they are very interesting and, for example, the work on prototypicality by Rosch and her colleagues (e.g., Rosch & Mervis 1975), which underlies most ICM discussion, is truly an epochal finding, to pursue this train of thought any further would take us too far away from OOS.

3.2.2 Langacker’s domains Langacker (1987: 147) introduces the idea of domains by noting that “most concepts presuppose other concepts and cannot be adequately defined except by reference to them, be it implicit or explicit. The concept [KNUCKLE], for example, presupposes the conception of a finger.” A finger assumes the concept of hand, which assumes the concept of “arm,’ etc. in a “hierarchy of complexity” which ends in a “basic domain” which does not assume any other domain. It is

Scripts, frames, and other semantic objects

35

unclear if Langacker sees these as primitives (149). This line of reasoning is obviously the same idea underlying the semantic network and the hierarchy of domains is none other than a partonymy, in this case. Langacker himself (150) notes that “an abstract domain is essentially equivalent to [. . .] [an] ICM and what others have variously called a frame, scene, schema or even script (at least in some uses)” [emphasis in the original, SA]. However, there are significant differences between Langacker’s domains and semantic objects such as frames or scripts: domains are explicitly on a scale of abstraction and ultimately the basic domains are founded on embodied cognition (Evans and Green 2006: 231), a concept which is foreign to, but not incompatible with, OOS. The hierarchical organization, which Evans & Green (2006: 231) note as a significant difference from Fillmore’s frame semantics, is not a difference in other versions of OOS, if they incorporate a ISA hierarchy, which most do. In conclusion, it is fair to say that within cognitive linguistics there has been a widespread acceptance of OOS. In some cases, the acceptance is a tad too optimistic, e.g., Croft (2009, p. 7) “A semantic frame is essentially identical to what other cognitive linguists have called a domain” since we have seen that domains and ICMs are not “essentially identical” to frames or scripts. However, and Croft is right about that, they are the starting point for some of the most characteristic aspects of cognitive linguistics. Using a metaphor, we might say that OOS and cognitive linguistics share 95% of their DNA.

3.3 Schema theory in reading and its critics One of the greatest success stories of OOS is the field of reading. There is widespread consensus, even among its critics, that schema theory, as it has become known in the field of reading, has been extremely influential, perhaps even more influential than script theory has been in humor studies. In reading theory the standard term has become “schema,” rather than frame or script, under the influence of Bartlett (1932) and Rumelhart (1980).

3.3.1 Schema theory and reading It is likely that the success of schema theory in reading is due, in part, to the fact that “among educators, something like schema theory has driven conceptions about reading” (Anderson & Pearson 1984: 258). Anderson et al. (1977) describes an experiment in which a text which was deliberately ambiguous

36

Salvatore Attardo

between two situations (a wrestling match and a prison break) shows that “high-level schemata provide the interpretative framework for comprehending discourse” (367). Rumelhart’s work with schemata (1980) was applied primarily to story understanding. Anderson & Pearson (1984) review some of the studies that support schema theory in reading. For example, recall of lists of words or sentences are improved if they activate a schema. Likewise, vague stories are clarified by the presence of a title that activates and overall schema (Bransford & Johnson 1972). The fortunes of schema theory in the field of reading have declined, in the past 20 years, as the cognitive approach has been replaced by more sociocultural approaches and a broader interest in literacy. However, even as interest in schema theory has flagged, within the reading scholars, it still remains a “required reading” for reading scholars (cf. McVee et al. 2005: 534; Alvermann et al. 2013) and, according to McVee et al. (2005: 241) the research simply migrated to cognitive science journals and stopped influencing the field of reading.7

3.3.2 Criticisms of schema theory in reading It goes without saying that there have been criticisms of objective semantics. Some of them are obviously warranted, such as the critique that, within reading theory, “schema theory has ignored the roles of imagery and emotional response in reading” (Sadoski et al. 1991: 465). Others, such as the claim we have seen, that OOS is “nebulous” (Brown 1979) are easily dismissed. Others still are misguided: for example, Krasny et al. (2007) claim that schema theory is “incommensurable” with embodied cognition, which is of course questionable, as seen by the above discussion of cognitive linguistics’ embodied versions of OOS. Sadoski et al. (1991: 466) sum up a general malaise about schemata The multitude of terms and metaphors suggests confusion about the basic nature of schemata [. . .]. There is also confusion regarding the nature of the knowledge that constitutes a schema and regarding the degree of abstraction that must occur before a schema can be said to exist. Therefore, it is not clear in schema theory whether we have an overall schema for color that is made up of particular colors (e.g., red, yellow) or schemata for particular colors that are made up of particular examples of these colors (e.g., fire engine

7 A similar phenomenon can be observed in the relationship between theoretical (Chomskian) linguistics and psycholinguistics: a period of intense interest (1960–70s) followed by an almost complete abandon.

Scripts, frames, and other semantic objects

37

red, lemon yellow). This problem is sometimes solved by postulating embedded schemata, or schemata within schemata. However, this approach invites infinite regress and theoretical circularity, problems that trouble many cognitive theories.

These criticisms do not apply to the definition of semantic objects as subgraphs, which morph regress and circularity into essential features of the semiotic property, as we have seen. Further discussion of reading theory and in particular of dual coding theory (Paivio 1986) would be interesting but is outside the scope of this article.

4 Conclusion In this paper I hope to have shown on the one side, a number of influences on script-based semantics, which should help shed light on the roots of the theory, and on the other hand, I hope to have highlighted several aspects on which script-based semantics set itself apart from other OOS theories. I then explored some applications and criticisms of OOS theories, to highlight the productivity of this model, even besides its current incarnation in ontological semantics. While I can only hope to have been successful in my attempts, I am reasonably sure that the significance, profound originality, and richness of applications of OOS and more specifically script-based semantics have emerged from this treatment.

References Abbott, Valerie, John B. Black, and Edward E. Smith. 1985. The representation of scripts in memory. Journal of memory and language 24(2) 179–199. Alvermann, Donna E., Norman Unrau, & Robert B. Ruddell. 2013. Theoretical Models and Processes of Reading. 6th ed. Newark, DE: International Reading Assoc., 2013. Anderson, Richard C. & P. David Pearson. 1984. A schema–theoretic view of basic processes in reading comprehension. In. P. David Pearson (ed.) Handbook of reading research Vol. 1. New York: Routledge. 255–291. Anderson, Richard C., Ralph E. Reynolds, Diane L. Schallert, & Ernest T. Goetz. 1977. Frameworks for comprehending discourse. American educational research journal 14(4). 367–381. Andor, József. 2010. Discussing frame semantics: The state of the art: An interview with Charles J. Fillmore. Review of Cognitive Linguistics. 8(1). 157–176. Attardo, Salvatore. 2001. Humorous Texts. Berlin: Mouton De Gruyter. Attardo, Salvatore, Hempelmann, Christian F, & DiMaio, Sara. 2002. Script oppositions and logical mechanisms: Modelling incongruities and their resolutions. Humor: International Journal of Humor Research 15(1),3–46

38

Salvatore Attardo

Bales, Michael E., and Stephen B. Johnson. 2006. Graph theoretic modeling of large–scale semantic networks. Journal of Biomedical Informatics 39(4). 451–64. Bartlett, Frederic Charles. 1932. Remembering: A study in experimental and social psychology. Cambridge: Cambridge University Press. Boas, Hans C. 2001. Frame Semantics as a framework for describing polysemy and syntactic structures of English and German motion verbs in contrastive computational lexicography. In Proceedings of corpus linguistics, 113, 64–73. Boas, Hans C. 2005. From theory to practice: Frame Semantics and the design of FrameNet. Semantik im lexikon. 479. 129–159. Boas, Hans C., and Ryan Dux. 2017. From the past into the present: From case frames to semantic frames. Linguistics Vanguard 3 (1). Brachman, Ronald J. 1979. On the epistemological status of semantic networks. In. Findler, Nicholas V. (ed.) Associative Networks: Representation and Use of Knowledge by Computers. New York: Academic press 3–50. Reprinted in Brachman, Ronald J. and Hector J. Levesque (Eds.) 1985. Readings in Knowledge Representation. San Mateo, CA: Morgan Kaufmann. 191–215. Brachman, Ronald J. and Hector J. Levesque (Eds.) 1985. Readings in Knowledge Representation. San Mateo, CA: Morgan Kaufmann. Bransford, John D., and Marcia K. Johnson. 1972. Contextual prerequisites for understanding: Some investigations of comprehension and recall. Journal of verbal learning and verbal behavior 11(6). 717–726. Brown, Ann L. 1979. Theories of memory and the problems of development: Activity, growth, and knowledge. In Cermak, Laird S., & Craik, Fergus I. M. (Eds.). Levels of processing in human memory. Oxford, England: Lawrence Erlbaum. 225–258. Brown, Roger. 1958. How shall a thing be called?. Psychological review 65(1) 14–21. Ceccato, Silvio. 1962. Automatic translation of languages. Presented at the NATO Advanced Study Institute 55. Ceccato, Silvio. 1964. Automatic translation of languages. Information Storage and Retrieval 2(3). 105–158. Charniak, Eugene. 1972. Toward a model of children’s story comprehension.” PhD dissertation. Massachusetts Institute of Technology. Collins, Allan M., & Quillian, M. Ross. 1972) How to make a language user. In E. Tulving & Wayne Donaldson (eds.). Organization of memory. Oxford, England: Academic Press. Cook, Walter A. 1989. Case Grammar Theory. Washington, DC: Georgetown University Press. Croft, William .2009. Connecting frames and constructions: a case study of ‘eat’ and ‘feed’. Constructions and Frames, 1(1).7–28. Croft, William and D. Alan Cruse 2004. Cognitive Linguistics. Cambridge: Cambridge University Press. Edwards, Derek, and David Middleton. 1987. Conversation and remembering: Bartlett revisited. Applied cognitive psychology 1.2 77–92. Fillmore, Charles J. 1968. The case for case. In Emmon Bach & Robert T. Harms (eds.) Universals in Linguistic Theory. New York: Holt, Rinehart & Winston. 1–90. Fillmore, Charles J. 1971. Some problems for case grammar. Georgetown University Round Table on Language and Linguistics 1971. Washington, DC: Georgetown University Press. 35–56. Fillmore, Charles J. 1975. An alternative to checklist theories of meaning. In Annual Meeting of the Berkeley Linguistics Society, 1, 123–131.

Scripts, frames, and other semantic objects

39

Fillmore, Charles J. 1976. Frame semantics and the nature of language. Annals of the New York Academy of Sciences 280(1). 20–32. Fillmore, Charles J. 1977. The case for case reopened. In Peter Cole & Jerrold M. Sadock (eds.) Grammatical Relations. Syntax and semantics, 8, New York: Academic Press. 59–82. Fillmore, Charles J. 1982. Frame semantics. in Linguistic Society of Korea (ed.), Linguistics in the Morning Calm, Seoul: Hanshin Publishing Company. 111–137. Reprinted in Geeraerts, Dirk (ed.) Cognitive Linguistics: Basic Readings. Berlin: Mouton De Gruyter. 2006. 373–400. Fillmore, Charles. 1985. Frames and the semantics of understanding. Quaderni di Semantica 6(2). 222–254. Fillmore, Charles J., and Beryl T. Atkins. 1992. Toward a frame–based lexicon: The semantics of RISK and its neighbors. In Lehrer, Adrienne, Eva Feder Kittay, and Richard Lehrer. (Eds.) Frames, fields, and contrasts: New essays in semantic and lexical organization. New York: Routledge. 75–102. Fillmore, Charles J., Christopher R. Johnson & Miriam R. L. Petruck. 2003. Background to FrameNet. International Journal of Lexicography 16 (3). 235–250. Fontenelle, Thierry. 2003. Framenet and frame semantics. International Journal of Lexicography 16(3),231–231. Fried, Mirjam and Nikiforidou, Kiki, (eds.). 2013. Advances in Frame Semantics. Amsterdam: John Benjamins Geckeler, Horts. 1971. Strukturelle semantik und wortfeldtheorie. München: Fink. Haugeland, John. 1985. Artificial Intelligence: The Very Idea, Cambridge, Mass: MIT Press. Hays, David G. 1961. Grouping and Dependency Theories. In Harold P. Edmundson (Ed.), Proceedings of the National Symposium on Machine Translation, Held at the University of California, Los Angeles, February 2–5, 1960. Englewood Cliffs, N. J.: Prentice–Hall, 1961. Hays, David G. 1964. Dependency theory: a formalism and some observations. Language 40 (4). 511–525. Johnston, Elizabeth B. 2001. The repeated reproduction of Bartlett’s Remembering. History of psychology 4(4) 341–366. Klein, Sheldon. 1965. Automatic paraphrasing in essay format. Mechanical Translation & Computational Linguistics 8(3–4). 68–83. Klein, Sheldon, and Robert F. Simmons. 1963. Syntactic dependence and the computer generation of coherent discourse. Mechanical Translation 7(2). 50–61. Krasny, Karen A., Mark Sadoski, and Allan Paivio. 2007. Unwarranted Return: A Response to McVee, Dunsmore, and Gavelek’s (2005) ‘Schema Theory Revisited.’ Review of Educational Research 77(2). 239–44. Lakoff, George. 1986. Frame semantic control of the coordinate structure constraint. Chicago Linguistics Society. 22. 152–167. Lakoff, George. 1987. Women, Fire, and Dangerous Things. What categories reveal about the mind. Chicago: University of Chicago Press. Lambrecht, Knud. 1984. Formulaicity, frame semantics, and pragmatics in German binomial expressions. Language 60(4).753–796. Lassila, Ora & Deborah McGuinness. 2001. The role of frame–based representation on the semantic web. Linköping Electronic Articles in Computer and Information Science 6(5). Lehrer, Adrienne. 1993. Semantic fields and frames: Are they alternatives?. Lutzeier, PR (Hg.): Studien zur Wortfeldtheorie. Tübingen. S, pp.149–162. Lytinen, Steven L. 1992. Conceptual dependency and its descendants. Computers and Mathematics with Applications 23(2–5). 51–73.

40

Salvatore Attardo

Mandelbrot, Benoit B. 1977. The Fractal Geometry of Nature. New York: Freeman. McVee, Mary B., Kailonnie Dunsmore, & James R. Gavelek. 2005. Schema theory revisited. Review of educational research 75(4). 531–566. Mel’čuk, Igor A., 1981. Meaning–text models: A recent trend in Soviet linguistics. Annual review of Anthropology, 10(1),27–62. Minsky, Marvin (ed.). 1968. Semantic Information Processing. Cambridge, MAL MIT Press. Minsky, Marvin. (1974). A framework for representing knowledge. MIT technicalreport. http://hdl.handle.net/1721.1/6089. Minsky, Marvin. (1985) A framework for representing knowledge. In Brachman, Ronal J. and Hector J. Levesque (Eds.) (1985). Readings in Knowledge Representation. San Mateo, CA: Morgan Kaufmann. 246–262. [this is the reprint of an abridged version of the text published in John Haugeland (ed.) (1981) Mind Design. Cambridge, MA: MIT Press.] Minsky, Marvin. 1985. The Society of Mind. New York: Simon & Schuster. Nirenburg, S., & Raskin, V. 2004. Ontological semantics. Cambridge, MA:MIT Press. Paivio, Allan. 1986. Mental representations: A dual coding approach. Oxford: Oxford University Press. Peirce, Charles Sanders. (1902a[1976]) The New Elements of Mathematics. Carolyn Eisele (ed.) 4 vols. The Hague: Mouton, 1976. Peirce, Charles Sanders. (1902b) Sign. In James Mark Baldwin (ed.). Dictionary of Philosophy and Psychology. New York: McMillan. Vol. 2, 527. Reprinted in James Hooper (ed.) Peirce on Signs. Chape. Hill, NC: University of North Carolina Press. 239–240. Petruck, Miriam R. L. 1996. Frame semantics. In Jef Verschueren, Jan-Ola Ostman, Jan Blommaert, & Chris Bulcaen (eds.). Handbook of Pragmatics. Philadelphia: JohnBenjamins. 1–13. Petruck, Miriam R. L. 2011. Advances in frame semantics. Constructions and Frames 3(1),1–8. Quillian, M. Ross. Semantic memory. 1966. No. Scientific-2. Bolt Beranek and Newman Inc Cambridge MA, 1966. Quillian, M. Ross. “Word concepts: A theory and simulation of some basic semantic capabilities.” Behavioral science 12, no. 5 (1967): 410–430. Raskin, Victor. 1985a. Semantic Mechanisms of Humor. Dordrecht: D. Reidel. Raskin, Victor. 1985b. Script–based semantics: a brief outline. Quaderni di semantica 6(2). 306–313. Raskin, Victor & Irwin Weiser. 1987. Language and Writing: Applications of Linguistics to Rhetoric and Composition. Norwood, NJ: Ablex. Reed, Stephen K. 2016. A Taxonomic Analysis of Abstraction. Perspectives on Psychological Science 11(6). 817–837. Rosch, Eleanor & Carolyn B. Mervis. 1975. Family resemblances: Studies in the internal structure of categories. Cognitive psychology 7(4). 573–605. Ross, Haj [John Robert]. 1967. Constraints on variables in syntax. PhD diss., Massachusetts Inst. of Technology Cambridge. Rumelhart, David E. 1980. Schemata: the building blocks of cognition. In Rand J. Spiro, Bertram C. Bruce, and William E. Brewer (Eds.) Theoretical issues in reading comprehension, Hillsdale, New Jersey: Lawrence Erlbaum. 33–58. Ruppenhofer, Josef, Michael Ellsworth, Miriam RL Petruck, Christopher R. Johnson, & Jan Scheffczyk. 2016. FrameNet II: Extended theory and practice. Institut für Deutsche Sprache, Bibliothek. http://www.sun-rise.com.vn/java/framenet.pdf

Scripts, frames, and other semantic objects

41

Sadoski, Mark, Allan Paivio, & Ernest T. Goetz. 1991. Commentary: A critique of schema theory in reading and a dual coding alternative. Reading research quarterly 26(4). 463–484. Schank, Roger C.1972. Conceptual dependency: A theory of natural language understanding, Cognitive Psychology, 3. 552–631. Schank, Roger C.1975. Conceptual Information Processing. Amsterdam: North Holland. Schank, Roger C. 1980. Language and memory. Cognitive science 4(3) 1980. 243–284. Schank, Roger C. 1982. Dynamic memory. Cambridge: Cambridge University Press. Schank, Roger C., 1999. Dynamic memory revisited. Cambridge: Cambridge University Press. Schank, Roger C. & Larry Tesler. 1969. A conceptual dependency parser for natural language. In Proceedings of the 1969 conference on Computational linguistics, pp. 1–3. Association for Computational Linguistics. Schank, Roger C. & Robert Abelson. 1977. Scripts, Plans, Goals and UnderstandingNew York: Wiley. Sloutsky, Vladimir M. 2010 From perceptual categories to concepts: What develops?. Cognitive science 34(7). 1244–1286. Tesnière, Lucien. 1959. Eléments de syntaxe structurale. Paris: Klincksieck. Trier, Jost. 1931. Der deutsche Wortschatz im Sinnbezirk des Verstandes. Ph.D. diss. Bonn. Ungerer, Friedrich & Hans-Jörg Schmid. 1996. An Introduction to Cognitive Linguistics. London/New York: Longman. Wierzbicka, Anna. 2011. Common language of all people: The innate language of thought. Problems of information transmission 47(4). 378–397. Woods, William A. 1975.What’s in a link: Foundations for semantic networks. In Bobrow, Daniel G., and Allan Collins (Eds.) Representation and understanding. Cambridge, MA: Academic Press. 35–82. Woods, William A. 1991. Understanding subsumption and taxonomy: A framework for progress. In John F. Sowa (ed.) Principles of Semantic Networks: Explorations in the representation of knowledge. San Mateo, CA: Morgan Kaufmann. 45–94.

Irina Kobozeva

Script-based approach towards taxis connectors Abstract: The paper shows the role of scripts in explaining the use of taxis connectors of immediate anteriority such as as soon as, once in English. Taxis connectors (conjunctions and syntactic phrasemes) such as after, as soon as, before have been studied in linguistic typology as the means of expressing taxis relation between situations, i.e. anteriority, simultaneity and posteriority with their variations, e.g. immediate (contact) vs. distant anteriority. The same sequence of events may be presented either as an asymmetrical structure with subordination or as a symmetrical coordinate structure. In both cases speakers have multiple choices between different patterns of expressing one and the same taxis relation. We argue that one of the influential factors of this choice is the factor of scripts. In the vein of V. Raskin’s Script-Based Semantic Theory, we show that the natural usage of immediate anteriority connectors presupposes: 1) that the time interval between component events, denoted by connected clauses, is not fixed as ‘zero’ in the corresponding script; 2) that the subordinate clause of immediate anteriority has the informational status of “given”, and that one of the main sources of such a status is a script activated earlier in discourse, the script that contains as its part the event denoted by the clause. We also show how the deviation from the standard usage may be used to achieve special semantic effects. Keywords : script, connector, taxis relation of immediate precedence, complex sentence, discourse analysis

1 Introduction In the paper, we discuss the crucial role of scripts in the use of connectors and adverbials expressing the taxis relation of immediate anteriority. The grammatical category of taxis was introduced by R. Jakobson (Jakobson 1957). This category encodes temporal relations between two situations P1 and Р2 in terms of simultaneity / anteriority / posteriority of P1 to P2, which is localized in relation to time of the utterance. More recently, taxis has been treated as a functionalsemantic category that can be expressed by grammatical and / or lexical means (Bondarko 1991). Taxis connectors (conjunctions and syntactic phrasemes) Irina Kobozeva, Moscow State University https://doi.org/10.1515/9781501511707-003

44

Irina Kobozeva

such as after, as soon as, before belong to grammatical means on a par with relative tenses (e.g. passé anterieur in French), while adverbs (e.g. then or its Russian equivalent potom ‘then’) and adverbial phrasemes (e.g. at once, right away) belong to lexical means. Taxis connectors are treated in traditional grammars as temporal subordinating conjunctions. In linguistic typology, their semantics has been represented by features of hierarchically organized parameters. The main parameter is the temporal order of events with the features anteriority, simultaneity and posteriority; the supplementary parameter is the temporal closeness of events, or the interval between them with binary features contact vs. distant. In this paper, we confine our attention to the use of connectors (and to some extent adverbials) that express contact anteriority (CA) of P1 to P2, i.e. the idea that the event P1 precedes P2 and the time interval between P1 and P2 is practically zero.1 In English those are as soon as and once. In Russian we have a whole row of synonymous connectors: subordinating conjuction kak tol’ko (the most frequent of them), four morphologically simple subordinating particle-conjunctions (tol’ko, lish, jedva, chut’), that can form several two- and occasionally three-word combinations (e.g. jedva tol’ko, jedva lish tol’ko),2 not to mention a correlative construction kak . . ., tak in its temporal use. In both languages, connectors of CA may be redundantly accompanied in the main clause with one of the synonymous adverbials of immediate posteriority of P2 to P1, e.g. at once, immediately, right away in English, srazu, momental’no, totсhas zhe in Russian (Kobozeva 2016). From the discourse analysis perspective connectors are considered as expressing semantic relations between discourse units. In terms of Rhetorical Structure Theory (RST) (Mann and Thomson 1988) connectors of CA as well as all the other temporal connectors would be treated as markers of asymmetric rhetorical relation of Circumstance between the satellite discourse unit and the nuclear discourse unit, as in (1a) below. The same sequence of events may be presented in discourse by two nuclear discourse units connected by symmetric rhetorical relation of Sequence, expressed by a corresponding taxis adverb in a coordinate construction, as in (1b), or in a paratactic construction, as in (1с):

1 When we say that P1 precedes P2 (or P2 follows P1) with “zero interval” we mean, that the time interval between P1 and P2 was so short (on the scale of the corresponding type of events) that its value can be neglected, cf. (Boguslavsky 1996, 52). The synonymous expression used in traditional descriptive grammars is “immediate”, e.g. in (Russkaja grammatika 1980) the meaning of CA-connectors is characterized as “immediate succession.” Note that the latter mode contradicts the fact that these connectors mark the anterior event of the subordinate clause and not the succeeding event of the main clause. 2 The constraints on combining those particles are formulated in (Kobozeva 2016).

Script-based approach towards taxis connectors

(1)

45

a. As soon as the door had closed, he grabbed the phone. (ParRNC)3 b. The door closed and at once he grabbed the phone. c. The door closed. Нe immediately grabbed the phone.

The only difference between (1a), on the one hand, and (1b) and (1c) on the other is the discourse status of the first clause. In (1b, c) its informational value is equal to that of the second clause while in (1a) it has the lower value of supplementary information that can be omitted if we want to reduce the text to its summary. At this point we can ask: what is it that makes us choose a subordination strategy, i.e. the use of CA connector, as opposed to the coordination strategy of expressing the same temporal relation of events? At the same time, we know that the taxis of CA may be conveyed implicitly, without any overt markers. Consider the following examples: (2)

a. The phone rang again. Roseanne picked up the receiver. b. They’ll tell his mother that he is dead and she’ll burst into tears. c. The egg fell down and broke.

In all such cases the order of events is iconically reflected in the order of corresponding clauses, as is generally the case. “The temporal order of speech events tends to mirror the order of narrated events in time . . . ” (Jakobson 1965: 27). This tendency later was explained by G. Leech in his “Principle of pragmatics” as one of the special cases of following the Maxim of Iconicity: “All other things being equal make the text imitate aspects of the message;” the Maxim is subsumed under the Principle of Expressivity (Leech 1983: 68). At the same time, such a manner of narration respects Grice’s Cooperative Principle (Grice 1975) by satisfying one of the conversational Maxims of Manner supporting this principle, namely “Be orderly”. The information about the order of events in (2), i.e. that the event denoted by the first clause precedes that of the second, is an implicature: if the speaker did not have in mind such order of events he would have flouted the maxim. Вut how do we know that the interval between the two events was “zero” when this information is not linguistically encoded in (2)? Thus, from the above observations on the use of CA-connectors, two questions arise: 1) what are the factors that make us prefer subordination against coordination for expressing the meaning ‘P1 & P2; P1 precedes P2; the time interval between P1 & P2 is zero’, i.e. use a CA-connector; 2) what are the factors that

3 Examples marked as ParRNC are taken from the Parallel English-Russian subcorpus of the Russian National Corpus.

46

Irina Kobozeva

favor explicit mentioning of the zero interval between P1 & P2 i.e. use a CAconnector and/or CP-adverbial while often it can be conveyed implicitly. Both questions are relevant for modeling text generation according to a given meaning representation. In what follows, I will try to show that these questions can be answered on the basis of the principles of the somewhat extended Script-Based Semantic theory (Raskin 1986). We begin with the second question.

2 Scripts and explication vs. implication of the time interval between the narrated events The notion of script is widely used in cognitive semantics and its applications for modeling the process of text understanding. Scripts are conceptualized as “routine”, “typical” sequences of events (often called complex events) that happen in the world with a well-specified set of object-like entities that appear in different roles throughout that sequence. (. . .) The component events of a script are often optional; alternatively, some component events stand in disjunctive relation with some others (. . .), and their relative temporal ordering may be fuzzy” (Raskin et al. 2003: 2) Thus the script of ORAL EXAM (at least in Russia) is a sequence of sequential component events, or sub-events: the examinee takes an examination paper, (s)he prepares the answers, (s)he delivers them to the examiner, the examiner asks additional questions, the examinee answers them, the examiner gives the grade to the examinee. The sub-event of asking additional questions is optional; the sub-event of preparation stands in disjunctive relation to the sub-event of the examinee’s statement of either 1) not being able to answer the questions of the exam paper, or 2 being able to answer without preparation. The relative temporal ordering of the examinee’s answering the exam paper questions and of the examiner asking additional questions is fuzzy (the examiner may start asking additional questions before the examinee has finished answering to questions of the exam paper). Every sub-event of a script may be considered a script in itself, at a different level of granularity. E.g. the sub-event of GRADING is a script in itself consisting of a sub-event of the examiner’s announcement of the grade to the examinee and a sub-event of the examiner’s recording of the grade into the corresponding official document. At the highest level of conceptual granularity, we find such “routine”, “typical” sequences of events that are not lexicalized, that is, there is no word in a language that refers to this sequence. Thus, the sequence of two

Script-based approach towards taxis connectors

47

events: ‘Х stumbles’ and ‘Х falls’ (disjunctively related to ‘Х regains balance’) has no corresponding verb neither in English nor in Russian. Still such sequences are also scripts, although they do not have a corresponding name in the lexicon and a corresponding concept in the ontology of a given language. In Ontological semantics (Nirenburg & Raskin 2003), such micro-scripts are captured in the ontology by the relation EFFECT. Scripts include information about temporal ordering of sub-events in terms of temporal relations between them, such as BEFORE or, conversely, AFTER. Thus, the sub-event of ANSWER-EXAM-PAPER stands in a BEFORE relation to the sub-event GRADE in the ORAL-EXAM script. These relations, as we can see, are just the taxis relations of anteriority and posteriority. A sub-event Q that stands in an EFFECT relation to a sub-event P, always stands in an AFTER relation to P (the effect always takes place after its cause). Now let us return to examples in (2) and the problem of how we know that the interval between the two events mentioned (between the telephone ringing and Roseanne picking up the receiver, between the announcement of one’s son death and the parent’s bursting into tears, between the falling of an egg and its breaking) was “zero”, although this information is not linguistically encoded. Now it is common knowledge in linguistics that implicit information is conveyed in communication mainly because linguistic structures and the situational context activate knowledge representation structures (frames and scripts) in the minds of interlocutors from where this information is extracted and merged with what was said. That means that we know about the zero interval between particular events described in (2) because such a time interval is fixed as zero by default in the scripts activated by linguistic units. So we need to supply scripts with an additional parametric property of pairs of sequential events TIME-INTERVAL that may have either absolute values, such as “9 months” or “10 years”, or relative values [0–10]. From the point of view of text generation, we do not explicitly mark time intervals between events if we expect that this information can be extracted from the script we activated. This is just one manifestation of Economy (or Rationality) Principle, governing speech activity as well as human activity in general. It is easy to see that the insertion of overt markers of zero interval into the utterances in (2), makes them unnatural as is shown in (3), because they violate this principle4:

4 Redundant structures with repetitive expression of this or that meaning component are fully justified and natural when they serve such purposes as cohesion or expressivity. But here we deal with the case of unnecessary verbalization of what is clear from common knowledge with no surplus.

48

(3)

Irina Kobozeva

a. The phone rang again. ?At once Roseanne picked up the receiver. b. ?They’ll tell his mother that he is dead, she’ll burst into tears right away. c. ?As soon as the egg fell down, it broke.

Such utterances may become natural in special contexts that inform us that we have to do with circumstances under which the typical script cannot be applied. In (3a), it can be the situation where everyone knows that one should wait for a given number of rings in order to answer the call. Thus the default value (zero) of time interval between telephone ringing and someone picking up the receiver is cancelled in favor of some other value. In such a context mentioning of zero interval becomes fully informative. (3b) could become acceptable in a context where the mother is a sort of “iron lady”, that never cries, at least in public. Thus, if such a person Х plays the role of ADRESSEE in an ANNOUNCE event and at the same time the role of EXPERIENCER in the EMOTIONAL-REACTION event, and the THEME of ANNOUNCE is DEATН event with its EXPERIENCER instantiated as a CНILD of Х, then mentioning the zero interval between the two events in (3d) is justified by contrast to default [> 0] value for this particular mother. (3c) is acceptable if we knew that actually the egg is an artifact that can break or just crack or stay intact, after falling. Thus, the conditions, that favor the implicit transmission of a CA taxis relation between events P1 and P2, denoted by the first and second clause of an utterance respectively, are as follows: 1) the first clause activates a script and refers to the instantiation P1 of a sub-event En of this script; 2) the second clause refers to the instantiation P2 of the next sub-event En+1 of the same script; 3) the default value of the time interval between En and En+1 in this script is zero. Accordingly, overt marking of CA taxis (either by CA-connector and / or CAadverbial) is motivated by the intention of the speaker to attract attention to the non-default or sometimes unusual zero value of the time-interval parameter between events P1 and P2 instantiating successive sub-events of some script, as illustrated in (4): (4)

I just mean that once you get into bed, I bet you’ll fall right asleep. . . [Lemony Snicket. The Ersatz Elevator (2001)] (ParNRC)

In (4) the script GOING-TO-SLEEP is activated, and in this script the time interval between successive sub-events of going to bed and falling asleep is not fixed (it may be zero, may be not). That is why the use of CA taxis markers once and right in (4) is quite natural. It makes the utterance more informative and expressive. When the zero time intervаl between successive sub-events P1 and P2 is not fixed in a script, the fact that in a given case it was or will be zero is normally

Script-based approach towards taxis connectors

49

marked by narrators as an unexpected feature of the succession in question, as was done in (4). The violation of this pragmatic rule, i. e. the omission of the corresponding taxis marker may be used as a figure of speech with a special semantic effect. It is this very effect that explains why Rosalind in Shakespeare’s “As you like it” can justly call the Caesar’s famous motto “I came, saw and overcame” “Caesar’s thrasonical brag”. Let’s have a closer look at this famous phrase in its original form: (5)

Veni, vidi, vici.

In (5), the order of events triggers an implicature. The interval I1 between coming (Veni) and seeing (Vidi) is fixed as zero in the corresponding typical microscript: coming somewhere has an immediate effect of seeing what’s going on there. The omission of explicit markers of CA taxis is a norm for such cases. The interval I2 between seeing and conquering (Vici) the enemy is not fixed in a typical script of a military clash: the time needed to overcome the enemy in a battle may be different, from very short to quite long depending on many factors. The norm in such cases is to somehow mention the interval (or the duration of the battle). Omission of such an information, further supported by the syntactic parallelism generates the implicature that Caesar has fixed a zero interval between seeing and conquering the enemy.

3 Scripts and subordination vs. coordination strategy in an utterance describing successive events We now turn to the other question posed in section 1: what are the factors that favor the choice of subordination against coordination for expressing the meaning ‘P1 & P2; P1 precedes P2; the time interval between P1 & P2 is zero’, i.e. the use of CA-connector as opposed to CA-adverbial. Let’s have a closer look at the examples in (1) repeated below as (6) for convenience: (6)

a. As soon as the door had closed, he grabbed the phone (ParRNC) b. The door closed and at once he grabbed the phone. c. The door closed. Нe immediately grabbed the phone.

50

Irina Kobozeva

The example (6a) is taken from the following text fragment: (7)

The guard watched as the two newcomers made their way to the elevators, inserted their key, boarded the lift, and disappeared. As soon as the door had closed, he grabbed the phone. [Dan Brown. The Da Vinci Code (2003)] (RNC)

In (7) the script of using an elevator is represented. Some sub-events of this script (those that can be seen by the observer from outside) are verbalized: coming to the elevator, inserting the key (one of the options in this script), boarding the elevator, closing оf the elevator’s door. Closing of the door is a sub-event of the script. As such it is predictable. Its predictability determines its low information value. That is why it can be construed as a satellite discourse unit (DU) in terms of Rhetorical structure theory. Its explicit mentioning in the text has only one function: this sub-event serves as a reference point, or a landmark, for the temporal localization of the nuclear DU he grabbed the phone. Even if we did not know the context (7) of (6а), we would presume that the closing of the door was something expected. By contrast, constructed examples (6b) and (6c) with the same event presented as nuclear DU do not require of this event to be a sub-event of the previously introduced script. The closing of the door in this case may be a totally unexpected event as when someone is trapped in a room deliberately or automatically and grabs the phone to call for help. Consider some more of the numerous eхamples of the use of CA-connectors in contexts revealing the status of subordinate clauses as sub-events of a previously introduced script. (8)

In the author’s own experience, the withdrawal of a single land plot from joint shared ownership requires up to one year of constant occupation. For example, as soon as the joint share owners have reached an agreement regarding the withdrawal of a specific plot, the plot should be marked on the ground and registered in the cadastre. [Natalya Shagaida, Zwi Lerman. The Land Market Living with Constraints (2004)] (ParRNC)

(9)

Harry followed Fred and George out of the locker room and, hoping his knees weren’t going to give way, walked onto the field to loud cheers. Madam Hooch was refereeing. She stood in the middle of the field waiting for the two teams, her broom in her hand. “Now, I want a nice fair game, all of you, she said, once they were all gathered around her”. [Joanne Kathleen Rowling. Harry Potter and the Sorcerer’s Stone (1997)] (ParRNC)

Script-based approach towards taxis connectors

51

(10) Mark had thoughtfully put me between Geoffrey Alconbury and the gay vicar. Actually, though, once we all got a few drinks down us conversation was by no means stilted. [Helen Fielding. Bridget Jones’s Diary (1996)] (ParRNC) (11) “You insert your key in that slot there. . .” The man pointed to a large electronic podium facing the conveyor belt. The podium had a familiar triangular hole. Once the computer confirms the markings on your key, you enter your account number, and your safe-deposit box will be retrieved robotically from the vault below for your inspection. When you are finished with your box, you place it back on the conveyor belt, insert your key again, and the process is reversed. [Dan Brown. The Da Vinci Code (2003)] (ParRNC) In (8) the clause, headed by CA-connector as soon as refers to a sub-event of the script “withdrawal of a single land plot from joint shared ownership” mentioned in the previous sentence. In (9) although the paragraph given in the corpus does not contain the standard name of the script, we easily recognize with the help of overt lexical representatives (locker room, field, cheers, refereeing, two teams, fair game) that we have to do with a description of a match of some invasion game. The clause headed by CA-connector once corresponds to a specific sub-event of the script of such a match, namely the teams gathering around the referee at the beginning of the game. The context (10) is even shorter, but it is absolutely certain that somewhere before in the text the script of a party was activated and the two sentences of (10) instantiate its sub-events, one of which, namely having drinks, is presented by the once-clause. At last, we give example (11) just to show that in order to use a CA-connector it is enough that the script be present in the mind of a speaker. The addressee can be ignorant of it. In such cases the presence of CA-connector can be taken as an overt marker for the addressee of a “subevent of a script” status of the headed clause. In all examples the event in subordinate clause provides a landmark for temporal localization of the main clause event and in case of once the former event is always the cause of the latter. In many cases the instantiated sub-event referred to by a clause headed by a CA-connector belongs to the script that needs no verbal activation in the previous text because it is, so to say, omnipresent, in the sense that such scripts are always in the process of unfolding in the communicative situation or in the world of discourse, as e.g. our common everyday routine (waking up in the morning, brushing one’s teeth, having breakfast, going to work and so on and so forth and going to sleep at the end of the day) or “natural scripts” of the succession of seasons, or of sun rising and then setting and so on. Нere are just two examples of CA-connectors heading instantiated sub-events of such scripts:

52

Irina Kobozeva

(12) Too feverish to rest, I rose as soon as day dawned. [Charlotte Brontë. Jane Eyre (1847)] (ParRNC) (13) As it so happens, anyway, as soon as I woke up I began to have more sensible ideas, it’s just like me to see only the dark side last thing at night and to wake up different. [John Fowles. The Collector (1963)] (ParRNC) The third kind of subordinate clauses headed by CA-connectors is represented by clauses that are related to some foregoing sentence either by (partial) coreference or by “action – result” relation as in examples below: (14) She and Cornflower went out of the room. As soon as they were gone Dunno caught sight of a white smock and cap hanging on a hook. [Nikolay Nosov. The Adventures of Dunno and his Friends (Margaret Wettlin, 1980)] (Par RNC) (15) He slammed on the brakes and avoided entering the intersection just as a line of four Alfa Romeos appeared out of nowhere and tore by in a blur. Once past, the cars skidded, decelerating, and cut sharply left one block ahead, taking the exact route Glick had intended to take. [Dan Brown. Angels and Demons (2000)] All the three semantic kinds of clauses have one feature in common. They contain information that is “given” (cf. Chafe 1976) because their meaning has either been expressed earlier in the text or activated by some script, introduced earlier or presupposed as belonging to constantly unfolding cyclic scripts of social, individual or natural life. It is not surprising that CA-connectors select such clauses, because the events they refer to should provide landmarks for the temporal localization of other events, and such landmarks should be familiar in order to perform their function. Note that not all taxis conjunctions require of their complement clauses that they should denote “given” events. See (16) with before introducing new information. (16) Brasher didn’t notice Bosch on the bench. She (Brasher – I. K.) was almost to the station door before he spoke. [Michael Connelly. City Of Bones (2002)] The same is true of conjunction when and its translational equivalent kogda ‘when’ in Russian. Subordinate clauses like before-clause in (16) are atypical because they look like sattelite DUs (they are syntactically dependent and they contain conjunction as a marker of a rhetorical relation) but semantically they

Script-based approach towards taxis connectors

53

are nuclear DUs, because the information they express is more significant to the discourse than that of the main clause.

4 Conclusion In the paper we tried to explain the observed regularities in the natural usage (i.e. not only grammatical but also pragmatic well-formedness) of taxis connectors of contact (immediate) anteriority on the basis of the data from the parallel English – Russian sub-corpus of the Russian National Corpus. The data comprised contexts with CA-connectors from original English texts and from English translations of Russian texts made by professional translators whose native language is English. Firstly, we tried to find out the principle underlying overt marking of CA taxis by CA-connector and / or CA-adverbial as opposed to its more economical implicit rendition. It turned out (not surprisingly) that implicit conveyance of the CA-relation is preferred when events P1 and P2 described in a sentence correspond to successive sub-events of some script and the interval between the corresponding sub-events in this script is fixed as “zero”. Accordingly taxis connectors are fully appropriate when events P1 and P2 described in a sentence either do not belong to one and same script, or in case they belong to the same script, the interval between the corresponding sub-events is not fixed as “zero” in this script. When there is a script with successive sub-events P1 and P2 in which the interval between them is generally “more than zero” than the absence of any overt marker of the time interval between P1 and P2, as in P1, (and) P2, signals that in this case a special sub-type of the same script is activated – the one with “zero” interval (the sub-type may differ in this respect from the general type due to some special properties of the event’s participants). Such a “coercion” of a particularized script over the general one may be used to tacitly achieve special semantic effects. Secondly, we tried to explain the choice of subordination strategy with the use of CA-connector (e.g. As soon as P1, P2), as opposed to coordination strategy with a CA-adverbial (e.g. P1 and immediately P2). Our corpus data show that subordination strategy with the use of CA-connector can be chosen only when the preceding event P1 has the informational status of “given”, and one of the main sources of such a status of P1 is that it corresponds to a sub-event of the script that was introduced earlier in the text. The regularities that have been observed in this paper must be taken into consideration in computational modeling of text generation and text understanding.

54

Irina Kobozeva

The crucial role of scripts in formulation of such regularities requires the formalization of this notion. The Script-based Semantic Theory proposed by Victor Raskin in (Raskin 1986) and elaborated by him and his colleagues in subsequent papers provides the needed formalism.

References Boguslavsky, Igor M. 1996. Sfera dejstvija leksicheskih edinic. [The scope of lexical items]. Moscow: Shkola “Jazyki russkoj kultury”. Bondarko, Alexander V. 1991. Functional grammar: A field approach. Amsterdam & Philadelphia: John Benjamins Publishing (Linguistic and Literary Studies in Eastern Europe. Vol. 35). Chafe, Wallace L. 1976. Givenness, contrastiveness, definiteness, subjects, topics, and point of view. In Charles N. Li (ed.) Subject and topic, 25–55. New York: Academic Press. Grice, H. Paul. 1975. Logic and conversation. In Peter Cole and Jerry L. Morgan (eds.) Syntax and Semantics. Vol. 3. Speech Acts, 41–58. New York: Academic Press. Jakobson, Roman. 1957: Shifters, verbal categories and the Russian verb. Cambridge Mass.: Harvard University Press. Jakobson, Roman. 1965. Quest for the essence of language. Diogenes 13. 21–37. Kobozeva Irina M. 2016. Kognitivno–semanticheskii podhod k opisaniju sredstv sv’azi predlozhenij (na primere konnektorov neposredstvennogo sledovanija) [Cognitive–semantic approach towards the description of connectors (Connectors of immediate precedence)]. Trudy instituta russkogo jazyka im. V.V. Vinogradova [Proceedings of the V.V. Vinogradov Russian language institute] 10. 120–133. Leech, Geoffrey N. 1983. Principles of pragmatics. Lоndon & New York: Longman. Mann, William C. & Sandra A. Thomson. 1988. Rhetorical structure theory: Toward a functional theory of text organization. Text 8. 243–281. Nirenburg, Sergei and Victor Raskin. 2003. Ontological Semantics. Cambridge, MA: MIT Press. Raskin, Victor. 1986. Script–based semantic theory. In D. G. Ellis and W. A. Donohue (eds.) Issues in language and discourse processes. Hillsdale, NJ: Erlbaum. 23–61. Raskin Victor, Sergei Nirenburg, Cristian F. Hempelmann, Inna Nirenburg & Katrina E. Triezenberg. 2003. The Genesis of a script for bankruptcy in ontological semantics. In Proceedings of the HLT–NAACL. Workshop on Text Meaning. http://aclweb.org/anthology/ W03-0905. Russkaja grammatika. 1980. Moscow: Nauka.

Jonathan Dunn

Ontological and grammatical constraints on metaphor productivity Abstract: Traditional approaches view metaphor as a semantic/pragmatic phenomenon that occurs at a conceptual level as mappings between independent concepts. These conceptual mappings are then lexicalized into observed metaphoric expressions. In this view, the lexical and grammatical structure of a metaphoric expression is not relevant to the underlying metaphor’s level of productivity. This paper argues that lexical constraints, ontological constraints, and grammatical constraints are all required to explain the productivity of metaphors. The productivity of metaphor lexicalizations is used to argue for the usefulness of a systematic script-based and ontology-based approach to meaning. Keywords: metaphor, conceptual mapping, lexicon, ontology, script

1 Productivity and conceptual metaphors The now-traditional approach to metaphor views it as a conceptual mapping between two independent concepts (Lakoff & Johnson 1980, Lakoff & Johnson 1999). These independent concepts are discrete entities defined by (i) their place in a hierarchy of other concepts and (ii) their relationships to other concepts (i.e., as modeled by an ontology: Nirenburg & Raskin 2004). For example, we could posit a concept CAT that is a type of ANIMAL and, more generally, a type of ANIMATE_BEING. This concept is defined, further, by its relations to other concepts: DOESN’T_LIKE -> DOGS or DOES_LIKE -> MILK. The traditional approach views a metaphoric mapping as an additional, perhaps temporary, relationship between concepts that largely mimics other types of mappings between concepts. For example, the metaphoric utterance in (1a) could be described using the conceptual metaphor in (1b), notated in all capitals by convention. In this case, there is a concept CAT that has a property such as DOES_LIKE -> SCRATCHING. And there is a concept BOSS with a property such as IN_CHARGE_OF -> EMPLOYEES. The simplest conceptual view of metaphor is that the metaphoric mapping overlaps or merges these concepts CAT and BOSS, but only temporarily in order to support the interpretation of this utterance. Jonathan Dunn, University of Canterbury https://doi.org/10.1515/9781501511707-004

56

Jonathan Dunn

(1a) My boss is a tom cat and I’m his scratching post. (1b)

BOSS IS A CAT

These metaphoric mappings are seen as temporary additions to an ontology (i.e., to a formal representation of the conceptual system). However, conceptual metaphors are most often cited within the usage-based Cognitive Linguistics paradigm. One of the important tenets of this approach to linguistics is that frequent perception becomes entrenched to such a degree that it influences future production. This, in turn, influences future perception and creates a production-perception loop until particular patterns become grammaticalized. In the case of the BOSS IS A CAT metaphor, this process has not taken place: this is still a novel metaphor. In the case of (2a), however, this conceptual metaphor has been produced and perceived so many times that it is posited to have grammaticalized a conceptual or metaphoric link between these two concepts. Informally, grammaticalized here means that a new linguistic structure has been created by repeated and consistent usage. The linguistic evidence for this grammaticalized conceptual link is the productivity of unmarked utterances such as (2b), (2c), and (2d). Over time, the metaphoric mapping itself becomes like a concept and influences both language use and reasoning. (2a)

ARGUMENT IS WAR

(2b) The boss beat back our advances for better treatment. (2c) John defended his position. (2d) Mary rebelled against the established paradigm. The problem with a usage-based approach to the relationship between production and perception is that it becomes a just-so story. In other words, simply positing that metaphoric mappings become more entrenched the more they are used is not sufficient. What we need is a model of metaphor productivity: Which metaphoric mappings are able to produce new metaphoric utterances (novelty)? Which metaphoric mappings have left behind fossilized metaphoric utterances but cannot produce new tokens? What are the ontological or conceptual constraints on which metaphoric mappings are possible and likely to become productive? What are the grammatical constraints on how conceptual mappings can be lexicalized into

Ontological and grammatical constraints on metaphor productivity

57

actual observed utterances (c.f., Steen 2007)? This sort of model has been neglected by conceptual approaches to metaphor. This paper synthesizes work on three types of constraints on the productivity of conceptual metaphors: First, lexicalized properties can influence the meaning of conceptual metaphors even when the same concepts are involved in the underlying mapping; Second, ontological relationships between concepts constrain which metaphoric mappings are possible and how metaphoric the resulting utterances are; Third, grammatical constraints on the realization of metaphoric material in various case roles influences how metaphoric a particular utterance becomes and whether that utterance is ambiguous in context.

2 Lexical constraints on metaphor productivity Although conceptual approaches to metaphor view this phenomenon as an emerging mapping between concepts, there are few attempts to determine what the conceptual system is like outside of metaphor. What are the independent concepts that are available for metaphoric mapping? How do we know that metaphor operates on concepts and not just lexical items? How do we know what the baseline relationships between concepts are? Ontological semantics (Nirenburg & Raskin 2004) provides a systematic approach to these questions by attempting to develop a formal model of the conceptual system. Here we draw on previous work (Dunn 2013a) and test how concepts behave as part of metaphoric mappings using the technique of forced metaphorization, which is essentially a semantic substitution test. Starting with a non-metaphoric utterance, we replace one of the concepts with a new lexical item. This changes the meaning of the utterance in one of three ways: first, the two utterances can have equivalent meanings; second, the two utterances can have different meanings; third, the utterances can have no meaning (i.e., interpretation is not possible). The purpose of applying a semantic substitution test like this is (i) to map out independent concepts that are then posited to be involved in metaphor and (ii) to understand the influence of non-conceptual lexical properties on metaphoric meaning. A syntactic substitution test relies on intuitions of grammaticality; a semantic substitution test relies on intuitions about implications. First, when two utterances have the same essential or propositional meaning, the utterances should also then share the same strong implications. This is not strictly speaking an entailment. The test for this kind of equivalence is whether there is a strong implication that differs between the utterances. A strong implication here is not necessarily an entailment but, more simply, an implication that is

58

Jonathan Dunn

not easily canceled (non-defeasible). The point is that, if substitution changes one of these strong implications, then the propositional or essential meaning of the utterance has changed as well. For example, (3a) has the strong implication in (3b); this strong implication, and thus the same essential meaning, is shared by the examples in (3c) and (3d). The implications have metaphoric sources; the point here is that the metaphoric implications are consistent. (3a) Market confidence tumbled last Friday. (3b) Strong Implication: Market confidence is lower now than two weeks ago. (3c) Market confidence dived last Friday. (3d) Market confidence plummeted last Friday. Not all semantic substitutions will produce utterances that have the same essential meaning. The example in (4a) has the strong implication in (4b). This example does not indicate substantial change but instead implies that market confidence has remained at the same general level. The substitution in (4c) shares this strong implication but the substitution in (4d) does not. (4a) Market confidence ambled last Friday. (4b) Strong Implication: Market confidence may or may not be lower now than two weeks ago. (4c) Market confidence drifted last Friday. (4d) Market confidence sky-rocketed last Friday. Using this method, the utterance in (5a) has the strong implication in (5b.a) and the weak implication in (5b.b). The negation test is being used to distinguish between strong and weak implications: (5b.a′) is unacceptable, but (5b.b′) is acceptable. This use of dive in requires that parliament at least attempted to take care of this problem (on the assumption that the participants intended to dive; the intention is important for the implication). However, it is possible given this scenario that parliament is unable to avoid the problem any longer so that this is a grudging move. For example, (5c) may be more or less grudging given the intentions of parliament. The substitutions in (5c) through (5e) hold the same strong and weak implications: thus, they have the same essential and non-essential meanings.

Ontological and grammatical constraints on metaphor productivity

(5a)

59

Parliament dived into a difficult problem this week.

(5b.a) Strong Implication: Parliament debated and tried to fix a difficult problem. (5b.a′) *Parliament did not debate or try to fix the problem, however. (5b.b) Weak Implication: Parliament is active or purposeful in doing this. (5b.b′) Parliament is grudgingly facing a problem it can no longer avoid. (5c)

Parliament galloped into a difficult problem this week.

(5d)

Parliament jumped into a difficult problem this week.

(5e)

Parliament strutted into a difficult problem this week.

This methodology of semantic substitution is important for showing that lexicalizations of metaphor do, in fact, operate on independent concepts that can be modeled in an ontology. The claim is that lexical items either (i) point to concepts or (ii) lexicalize semantic properties of individual concepts or the utterance as a whole (i.e., aspect). Combining a script-based approach to semantics with metaphor as mappings between concepts, the first prediction is that semantic substitution of words that point to the same concept will not change metaphoric meaning. This term, metaphoric meaning, refers to the weak and strong implications discussed above. There may be additional emotional or affective meanings as well. The second prediction is that semantic substitution of the same concept with different lexicalized properties will only change the nonessential meaning of the utterance. The third prediction is that lexical items pointing to different concepts will create utterances with different essential meanings. This is important as a baseline for understanding the productivity of metaphor lexicalization: how do metaphoric mappings interact with the mapping from lexical items to concepts? The original utterance in (6a) shows a non-metaphoric use of ignore. This lexical item points to a concept that could be labeled COMMUNICATION or KNOWLEDGE (the label is not important). The utterances in (6b–d) contain semantic substitution with the verbs flee, evade, and abandon. These lexical items point instead to a concept that could be labeled MOTION. In other words, the substitutions in (6b-d) all point to the same concept. As predicted, these three utterances share the same strong implication in (6e) (i.e., they share the same essential meaning). This is

60

Jonathan Dunn

evidence that the metaphoric mapping can be modeled at the conceptual level in a way that provides generalizations across many lexical items. (6a) Non-metaphoric: Most politicians ignore scientific findings. (6b) Metaphorized 1: Most politicians flee scientific findings. (6c) Metaphorized 2: Most politicians evade scientific findings. (6d) Metaphorized 3: Most politicians abandon scientific findings. (6e) Strong Implication: Most politicians do not make conclusions based on scientific findings. Starting with the original utterance in (7a), the verb made points to the concept Metaphoric utterances are produced in (7b-d) by substituting went along with, entered, and reached. As before, these substitutions point to a single concept (MOTION) and also lexicalize similar properties: there is intended movement towards a goal. The strong implication in (7e) is shared by each of these new utterances, again matching the prediction that the metaphoric mapping is between concepts and that all lexical items which point to the same concept will experience the same metaphoric interpretation. Some may consider (7a) metaphoric on its own because the arms treaty is reified as an object; that, however, does not change this analysis.

CREATE.

(7a) Non-metaphoric: Russia has made an arms treaty with the United States. (7b) Metaphorized 1: Russia has went along with an arms treaty with the United States. (7c) Metaphorized 2: Russia has entered an arms treaty with the United States. (7d) Metaphorized 3: Russia has reached an arms treaty with the United States. (7e) Strong Implication: Russia and the United States now have an arms treaty. Even though metaphoric mappings are conceptual, not all lexical items that point to the same concept share all the same semantic properties. If this is the case, then metaphor productivity will be influenced by lexicalized properties

Ontological and grammatical constraints on metaphor productivity

61

in addition to concepts: these utterances will share the same essential meanings but have different non-essential meanings. The utterances in (8b–d) have substituted meandered, sauntered, and struggled. While these all point to the concept MOTION, they also lexicalize different semantic properties. Meandered implies that there is no specific destination and that the movement is slow. Sauntered implies a slow and self-assured manner of movement and leaves the destination vague (there is neither a specific destination nor the lack of a specific destination). Tripped implies that the motion is difficult, prevented by something. These utterances share the strong implication in (8e), evidence that they also share the same essential meaning. But they do not share the same weak implications. Meandered (8b) carries the weak implication that there is no fixed purpose for the communication. Sauntered (8c) carries the weak implication that the communication is slow and self-assured. Tripped (8d) carries the weak implication that the communication is difficult to understand. The fact that these different substitutions carry different implication is evidence that they differ in their non-essential meaning. This is an important point for the productivity of metaphors because the underlying conceptual mapping is filtered or influenced by lexicalized meaning. (8a) Non-metaphoric: The recorded message sounded through the poor cell connection. (8b) Metaphorized 1: The recorded message meandered through the poor cell connection. (8c) Metaphorized 2: The recorded message sauntered through the poor cell connection. (8d) Metaphorized 3: The recorded message tripped through the poor cell connection. (8e) Strong Implication: The recorded message could be heard in spite of the poor connection. A final example in (9b–d) substitutes the motion verbs stampeded, wandered, and crawled; these words point to the same concept but provide different properties on top of that concept. After metaphoric mapping, these utterances share the same essential meaning, as shown by the shared implication in (9e). But again the additional lexicalized meaning alters the metaphoric mapping. The manner of motion is specified by stampeded. The weak implication as a metaphor is that

62

Jonathan Dunn

these foreign nationals were not particularly welcome but came anyway. Again, the lack of a specific destination is lexicalized in wandered. As a metaphor, this produces the weak implication that the foreign nationals do not have a clear plan and did not all invest in the same way. A slow manner of movement is lexicalized in crawled. As a metaphor, the implication is the foreign nationals considered entering the New Zealand market very carefully. (9a) Non-metaphoric: Foreign nationals bought into New Zealand when its currency was weak. (9b) Metaphorized 1: Foreign nationals stampeded into New Zealand when its currency was weak. (9c) Metaphorized 2: Foreign nationals wandered into New Zealand when its currency was weak. (9d) Metaphorized 3: Foreign nationals crawled into New Zealand when its currency was weak. (9e) Strong Implication: Foreign nationals are now invested in New Zealand. The point here is that these lexical items point to the same concept and so produce metaphors with the same essential meaning. But the additional lexicalized properties of these verbs change the metaphor. The implication is that metaphoric mappings are conceptual but are still influenced and constrained by lexical properties. Thus, the normal processes of lexical semantics apply to metaphoric mappings and the final utterance meaning is not simply a result of conceptual processes. The idea that lexical semantics contributes to metaphor is not in itself a new proposal. This discussion is important, however, because a focus on conceptual mappings has left a gap in the literature that overlooks the influence of linguistic factors.

3 Ontological constraints on metaphor productivity So far we have assumed that a metaphoric mapping involves two discrete and independent concepts. It is clear, however, that not all concepts can be mapped together. This section discusses constraints at the ontological level (i.e., at the

Ontological and grammatical constraints on metaphor productivity

63

level of relations and properties of concepts) that constrain metaphor productivity. These ontological constraints are used to describe which concepts cannot be mapped onto one another. The main constraints used in the literature are (1) semantic similarity, (2) abstractness/concreteness, and (3) a sufficient number shared attributes. We can synthesize these constraints using Searle’s social ontology (1995, 2010). The first idea is that most concepts contain some meaning that can be experienced by the senses together with some meaning that cannot be. A concept like NURSE or DOCTOR, for instance, combines physical information about the expected characteristics of these social roles together with non-physical social information about the roles of nurses and doctors within the script of health-care. The concepts that are potentially available for metaphoric mappings can be incapable of being experienced by the senses because (i) they are inherently created by humans or (ii) they are physical objects that have been assigned nonphysical functions by humans. This is an important point because it allows us to be more precise about the oft-cited generalization that metaphoric mappings have less abstract source concepts and more abstract target concepts. This generalization is, in many cases, not true. If we make more precise predictions about source and target concepts, we can maintain accuracy. The second idea is that there is not a clear distinction between concepts about external reality (the physical world) and those about internal reality (the mental or social world). This two-way distinction is impoverished. Searle says, “The world of Supreme Court decisions and of the collapse of communism is the same world as the world of the formation of planets and of the collapse of the wave function in quantum mechanics” (1995: 120). Searle’s social ontology can be used to collapse abstractness, semantic distance, and shared attributes into a single continuum based on two notions: fact-status and function-status. The first, fact-status, is a property that transitions from the purely physical to the purely social. Some concepts reference items that are present in the external world. These brute-physical-facts are independent of human consciousness and account for prototypical concrete concepts like BIRD, TREE, ROCK, and WATER. These concepts are not problematic for the traditional definition of abstractness, even if they are not universal “reflections” of some external reality. Other concepts, called mental-facts, are products of human consciousness: concepts created by humans. The simplest of these mental-facts are non-intentional sensations: PAIN, EXHAUSTION, HUNGER. These concepts are mental states experienced passively. Searle distinguishes these from active intentional-mental-facts: DESIRE, IDEA, BELIEF. Searle defines intentionality as “that capacity of the mind by which it is directed at, or about, objects and states of affairs in the world, typically independent of itself” (2010: 25). Such intentional-mental-facts “are always about, or

64

Jonathan Dunn

refer to, something” (25) while non-intentional-mental-facts are sensations or perceptions that do not refer to something outside the mind. Non-intentionalmental-facts exist in solitary individuals, even individuals who have only lived in isolation. Intentional-mental-facts, however, can exist either in individuals or in groups of individuals. Going a step further, many concepts are cultural. Their existence depends upon being shared among many individuals. These collective-intentional-mental-facts exist only if many individuals believe or accept that they exist: GOVERNMENT, UNIVERSITY, FOOTBALL, CONFERENCE. Searle in this way creates a distinction between physical-facts and mentalfacts, between non-intentional and intentional mental-facts, and between singular and collective intentional-mental-facts. This fact-status property arranges concepts on a continuum of how much or how little they depend on human beings. In terms of metaphoric mappings, the more a concept depends on humans the more available it is for metaphoric reinterpretation by humans. The second part of Searle’s social ontology involves the function information that concepts contain. As Searle argues, function information is inherently ontologically subjective even if that function information is supplied for purely physical objects: “This consequence follows from the observer-relative character of all functional attributions” (1995: 123). The function-status of a concept is the purpose of a concept as defined by humans: “[A] function is a cause that serves a purpose. And the purposes have to come from somewhere; in this case, they come from human beings” (2010: 59). Function information depends upon humans for its existence: even concepts referring to physical-facts can contain human-dependent function information, and this information is more available for metaphoric reinterpretation by humans. Many concepts contain no function information: ROCK, TREE, BIRD. Other concepts contain embedded function information that is independent of human interpretation: non-agentive functions (also called natural functions) like HEART (purpose: circulating blood), DIGESTION (purpose: creating energy from food), BREATHE (purpose: intaking oxygen). Functions like this are achieved independently of human intentionality. In Searle’s ontology these concepts are less concrete than functionless concepts, but only slightly. Non-agentive functions add information about the concept’s purpose in addition to causal information. This allows these functions to be evaluated against their ability to achieve their purpose (Searle 2010: 59). Going further, a large number of concepts contain functions that depend entirely upon human consciousness. Agentive functions contain human-centered information that situates these concepts according to their use for or by humans. Causal-agentive-functions have a use that is performed by the essential properties of the concept itself: HAMMER, CAR, CLOTHING. This category of concepts are used for

Ontological and grammatical constraints on metaphor productivity

65

specific purposes by humans and embed this functional information. Institutionalagentive-functions are performed not by the concept itself but rather by the consent of the humans who have created that concept: MARRIAGE, MONEY, TREATY. This category of functions can only be fulfilled if humans believe or accept that they have been fulfilled. Searle’s social ontology now provides a second hierarchy: concepts with no function information (BLUE, SWEET, MOUNTAIN), followed by non-agentive functions (NOSE, CAT, MOONLIGHT, TIDE), causal-agentive-functions (UNDERWEAR, CIGARETTE, TRUMPET) in which much of the definition of the physical object comes from how humans use that object, and finally institutional-agentive-functions (POET, ELECTION, PROTEST) in which much of the definition of these concepts comes from their use by humans and yet that use is a matter of human convention and acceptance (e.g., a treaty cannot be enacted unless a group of humans believes or accepts that it has been enacted). The important point here is that (i) concepts and (ii) the functions of concepts can depend more or less on human beings under this system. Semantic similarity is measured by how close in this hierarchy two concepts are situated. This is not a complete semantic similarity, in the sense that antonyms and converses actually end up rather close on this hierarchy. Abstractness is also measured by where concepts fall on this hierarchy: concepts that depend more on human beings are more abstract. The essential property of concepts, in this view, is how dependent they are on humans for their existence. This hierarchy is made more tangible by using the decision trees below, reproduced from Dunn (2015b). The purpose of Tables 1 and 2 is to make tangible this hierarchy of concepts. First, the categories in both tables move from least abstract to most abstract. This captures the property of abstractness that is important but undefined in metaphor theory. Second, the closer that concepts come within these hierarchies the more similar they are. This partially captures the property of semantic similarity that is important for restricting which metaphoric mappings are possible. It is important to note that more difficult mappings remain possible but are marked or highly metaphoric. We now have a way to define the relationship between concepts in order to formulate ontological constraints on metaphor productivity. To accomplish this, we will systematically test the acceptability of metaphors by drawing source and target concepts from adjacent parts of the hierarchy, taking examples from Dunn (2015b). How similar (in semantic distance and abstractness) can two concepts be while still supporting a metaphoric mapping? Singular-mental-facts are more abstract than physical-facts. Starting in (10) we take a concept from each category, in this case sharing a function-status: WEED and PASSIONS. Mapping from less abstract to more abstract produces the utterance in (10b),

66

Jonathan Dunn

Table 1: Annotating for Fact-Status (from Dunn 2015b). 

A: Does the concept depend strictly upon definition (e.g., DEDUCTION)? B: Does the concept exist only within a formal system (e.g., INTEGER, INTEGRAL)? C: Does the concept exist in multiple domains (such as PHYSICAL, MENTAL, SOCIAL)? If yes to any: fact-status is ABSTRACT; Else: continue to 



A: Can the concept be experienced by the senses (sight, sound, smell, touch)? B: If yes, is the concept distinguished from similar concepts by characteristics that can be experienced by the senses (e.g., TABLE v. CHAIR as opposed to JUDGE v. POLICEMAN)? If yes to both: fact-status is PHYSICAL; Else: continue to 



A: Can the characteristics that distinguish the concept from similar concepts exist in the mind of a single individual in isolation (e.g., IDEA v. LAW)? If yes: fact-status is SINGULAR-MENTAL; Else: continue to 



A: Do the characteristics that distinguish the concept from similar concepts exist only when common to the minds of a group of individuals (e.g., WIFE and WIDOW)? B: Does the concept exist only if its existence is accepted or acknowledged by a group of individuals (e.g., LAW, TREAT, MONEY)? If yes to either: fact-status is COLLECTIVE-MENTAL; Else: If no: restart decision tree

Table 2: Annotating for Function-Status (from Dunn 2015b). 

A: Does the concept contain or carry by default any information about the purpose, function, use, or role of the concept (e.g., not CLAY or MEADOW)? If no: function-status is NO-FUNCTION; Else: continue to 



A: Is the concept’s purpose, function, use, or role carried out without human intervention (e.g., bees make honey and cows produce milk)? B: Does the concept have a natural cause-and-effect relationship that is interpreted by humans as a purposeful act (e.g., the heart pumps blood)? If yes to either: function-status is NON-AGENTIVE; Else: continue to 



A: Is the purpose, function, use, or role of the concept carried out by means of or according to the intrinsic properties of the concept itself (HAMMER v. STOP-SIGN)? If yes: function-status is CAUSAL-AGENTIVE; Else: continue to 



A: Is the purpose, function, use, or role of the concept carried out only by the consent of a group of individuals (e.g., MARRIAGE and TREATY)? B: Is the purpose, function, or role of the concept performed only when a group of individuals accepts that it has been performed (e.g., CHRISTENING, BAPTISM)? If yes to either: function-status is INSTITUTIONAL; Else: restart decision tree

Ontological and grammatical constraints on metaphor productivity

67

an acceptable utterance with a metaphoric meaning. The reverse mapping, however, produces a non-metaphoric utterance. The only reading of (10c) is one involving literal weeds. This suggests that a metaphoric mapping between these concepts is not possible in this direction. (10a) Source: WEED (physical-fact; no-function) (10a’) Target: PASSIONS (singular-mental-fact; no-function) (10b) The passions grow up like weeds wherever the soul is left untended. (10c) *The weeds in our flower garden need to be brought up in confession. To what degree does function-status influence the possibility of mapping these concepts? We repeat the same test in (11), this time drawing concepts that have a causal-function categorization. In this case, metaphoric utterances are possible with mappings in both directions. The mapping from more abstract to less abstract in (11c) is, however, more metaphoric (i.e., more marked). This suggests that this type of mapping is less productive. (11a) Source: CAMERA (physical-fact; causal-function) (11a’) Target: OBSERVING (singular-mental-fact; causal-function) (11b) She watched his walk and stored the images in her head for later. (11c) The camera was her eyes and the roll of film her only memory. Staying on the fact-status hierarchy, the next two adjacent categories are singular-mental and collective-mental. Searle’s distinction here essentially separates individual and social cognition. In the example in (12) an individual mental state is mapped with a collective mental state. As above, both directions of mapping can produce acceptable metaphoric utterances. As above, however, the reverse mapping from less to more abstract produces a more marked utterance in (12c). For further investigation along these lines, including corpus and psycholinguistic evidence, see Dunn (2015b). (12a) Source: KNOWLEDGE (singular-mental-fact; no-function) (12a’) Target: MARRIAGE (collective-mental-fact; institutional-function)

68

Jonathan Dunn

(12b) A good marriage takes time, study, and teaching to mature. (12c) His knowledge of tax law stayed with him despite his many adulteries with corporate law and only occasionally nagged him about his misdoings. We now switch to differences between concepts as measured using the function-status hierarchy. Again, the question is whether particular mappings are productive and, if so, which direction of mapping produces a more marked metaphoric utterance. The example in (13) maps between two purely physical concepts which differ only in the function that is assigned to them. Is this function information enough to influence the productivity of metaphoric mappings? Mappings in both directions produce acceptable metaphors; but mapping from more to less abstract concepts produces marked metaphors. This case may be unique, however, because physical concepts provide rich source domains for metaphors. (13a) Source: GROWING (physical-fact; non-agentive-function) (13a’) Target: RECYCLING (physical-fact; causal-function) (13b) Recycling allows us to grow new products out of old products. (13c) Every fall the trees turned in their empty leaves to the recycling center and waited to receive their refurbished leaves in the spring. The next adjacent pair on Searle’s hierarchy is between non-agentive and causal functions, but here we draw the concepts from separate fact-status domains as well. This increases the distance between concepts and allows us to probe whether reversed mappings are acceptable given larger ontological differences between concepts. The example in (14) shows that the expected mapping produces an unmarked metaphoric utterance (14b), but the reverse of that mapping produces a very marked and perhaps unacceptable utterance in (14c). This may be a one-off case and the coupling is tried with different concepts from the same categories in (15), with the same effect. We can now begin to generalize that reverse mappings between concepts close in the ontology produce marked but acceptable metaphors while reverse mappings between more distant concepts produce unacceptable metaphors.

Ontological and grammatical constraints on metaphor productivity

69

(14a) Source: DIGEST (physical-fact; non-agentive-function) (14a’) Target: IDEA (singular-mental-fact; causal-function) (14b) Ideas take time to digest newly discovered facts into new theories. (14c) ?My indigestion is an idea that sometimes comes out of nowhere and then leaves as quickly. (15a) Source: SLEEPWALKING (physical-fact; non-agentive-function) (15a’) Target: REHEARSE (collective-mental-fact; causal-function) (15b) We will rehearse until we can do all this walking in our sleep. (15c) ?My younger brother sleepwalked every night, rehearsing for the next day at school. Does this asymmetry extend to institutional-function concepts? The test case in (16) shows that it does. The distance between concepts is greater here and the reverse mapping produces a questionable metaphoric utterance in (16c). Even if it is possible, we can say that this sort of mapping is unproductive given how marked and highly metaphoric it is. (16a) Source: KIDNAPPING (collective-mental-fact; causal-function) (16a’) Target: ELECTING (collective-mental-fact; institutional-function) (16b) They kidnapped the election with campaign contributions and received their ransom later. (16c) ?The kidnappers campaigned hard to pass their rival gang in the primaries. We finish this line of investigation with the paired examples in (17) and (18). Again the ontological distance between concepts comes from function-status, with fact-status held constant. The interesting difference here is that the reverse mapping in (17c) is clearly unacceptable while the reverse mapping in (18c) produces an unmarked metaphoric utterance. The reason seems to be that physical concepts provide rich source domains even when the distance between concepts is large. This points to ontological constraints specific to source domains.

70

Jonathan Dunn

(17a) Source: STAGE-HAND (collective-mental-fact; causal-function) (17a’) Target: LAWMAKER (collective-mental-fact; institutional-function) (17b) Members of congress are just stage-hands doing what the big corporations require. (17c) *The stage-hands passed a bill to change the scene from a forest to a lake. (18a) Source: UNDRESSING (physical-fact; causal-function) (18a’) Target: THEATER (physical-fact; institutional-function) (18b) Acting in front of audiences leaves you exposed and naked up there on the stage. (18c) He finished undressing with a dancer’s flourish and bowed to the audience in the mirror. Ontological constraints on metaphor productivity, in which relationships between two concepts determine whether a metaphoric mapping is possible and how marked or unmarked that mapping is, have been examined here using Searle’s social ontology to situate concepts against one another. This is an important undertaking for the overall program of predicting metaphor productivity, but one that has been neglected in the metaphor literature with its lack of interest in concepts and conceptual systems outside of metaphoric mappings. An important take-away from this short discussion is that ontological constraints are less likely to produce unacceptable utterances than to produce marked and highly metaphoric utterances.

4 Grammatical constraints on metaphor productivity The final and least considered constraint on metaphor productivity is grammar: how is a conceptual mapping lexicalized in the observed linguistic structure of an utterance? This section draws from Dunn (2011, 2015a) and argues that the lexicalization of metaphoric mappings is constrained by the density of metaphoric fillers for case roles. The more metaphoric case roles an utterance

Ontological and grammatical constraints on metaphor productivity

71

contains, the more marked and less likely to occur that utterance becomes. At a certain point, called metaphoric saturation, an utterance on its own becomes ambiguous between a very metaphoric reading and a literal reading. Such highly metaphoric and saturated utterances are less likely to be productive, providing evidence that there are also grammatical constraints on metaphor productivity. Case roles describe the relationships between an event and its participants. In analyzing the lexicalization of metaphor, we draw on the case roles recognized in Nirenburg and Raskin (2004): agent, beneficiary, destination, experiencer, instrument, location, path, purpose, source, theme. In addition, because the verb often also lexicalizes metaphor, event is also considered a case role. Depending on the particular frame or script of an utterance, some of these case roles are necessary (obligatory) or unnecessary (optional). The point of this section is to systematically probe how filling different case roles influences the final density of a metaphoric utterance. Each example provides instances of a single underlying conceptual mapping, ARGUMENT IS WAR. One problem with a conceptual metaphor analysis here is that, given the linguistic expression, there is no reason this could not be analyzed as an instance of ARGUMENT IS BUILDING. The examples in (19) through (22) lexicalize the same underlying metaphor into different case roles. Each of these lexicalizations produces a metaphoric utterance because there is a semantic clash: no literal interpretation is possible (this is an important observation that contrasts with lexicalization into multiple case roles, as we will see soon). On the other hand, there is no clear increase in the markedness of the metaphor based on which case role receives the lexicalized metaphor, except that perhaps that optional case role in (22) is more marked. (19a) Event (19b) Mary [demolished] John’s argument. (20a) Agent (20b) [The enemy] produced a strong argument. (21a) Theme (21b) Mary disproved [John’s strongest weapon]. (22a) Instrument (22b) Mary disproved John’s best argument [using a double-edged sword].

72

Jonathan Dunn

When the metaphor is lexicalized into multiple case roles, the density of the metaphoric expression increases. We have seen above that mappings between certain concepts create marked metaphoric utterances. In this case, increased density also creates marked metaphoric utterances. In each of these examples in (23) through (25), the metaphoricity of the utterance increases with each additional case role. On the other hand, only (24) is a potentially ambiguous utterance: this could be a literal description of a military action. This sort of ambiguity seems to be possible only if there is no case role that is filled by nonmetaphoric material. (23a) Event-Theme (23b) Mary [destroyed] [John’s strongest weapon]. (24a) Agent-Event-Theme (24b) [The enemy] [besieged] [John’s best stronghold]. (25a) Agent-Event-Instrument (25b) [The enemy] [besieged] John’s dissertation argument [using a horde of empirical warriors]. A metaphoric utterance becomes saturated when it contains enough metaphoric case roles (i.e., with sufficient density) to be ambiguous between a metaphoric reading and a literal reading. The example in (24a) is continued in (26); it is ambiguous between the two readings, which here means that it could be followed by the utterances in either (26b) or (26c). Metaphor saturation provides an upper limit for density: an utterance without metaphoric case roles is not at all metaphoric but an utterance with only metaphoric case roles is ambiguous between literal and metaphoric readings. (26a) [The enemy] [besieged] [John’s stronghold]. (26b) He was denied tenure after his theory was disproved. (26c) He was captured and convicted of treason. How do metaphoric utterances become saturated in actual usage? Previous corpus studies found that quite a few saturated metaphors are used (Dunn, 2013a,

Ontological and grammatical constraints on metaphor productivity

73

2015a). In this section we look at causes for saturation together with examples from previous corpus studies. The first two examples are caused by simple ambiguity: in (27), we do not know whether “it” refers to a literal or metaphoric computation and in (28) we do not know whether “case” is a physical object or a legal abstraction. In both examples, saturation is simplistic in the sense that this ambiguity most likely does not exist in the complete communicative context. (27) Referential ambiguity (27a) I didn’t know what it was. It would not compute. It had a kind of clarity about it. (28') Lexical ambiguity (28a) So – and they can’t throw the case out. Unspecified arguments are a more significant reason for saturation. In (29a), we can assume that what is jumping is the number of chapters and conferences. But we do not know for sure. In (29b), the missing argument of butchered also dictates whether the utterance is metaphoric or literal. This means that pragmatic implicatures used to fill in that missing argument determine whether the utterance is metaphoric or not. (29) Unspecified arguments (29a) India, the number of members increased 18% and chapters and conferences also saw significant jumps. (29b) Mom butchered ours. There’s only two pictures left of Dad in all twelve albums. Background knowledge is another significant cause of metaphor saturation insofar as the utterance underspecifies its interpretation. In (30a) the town is not legally owned, because that is not possible. However, without script-based background knowledge to tell us this (forcing a metaphoric reading of the utterance), a literal reading is possible. (30) Background Knowledge (30a) It’s about a town bought and paid for and subverted by a gangster element.

74

Jonathan Dunn

The more case roles are filled with metaphoric material, the denser the lexicalized metaphor becomes. This means that examples like (31a), which have a large number of optional and obligatory case roles, can potentially become very dense metaphors. (31) Many metaphoric constituents (31a) Veterans of many an evening at the routine pursuit of duty in the field, their emotional sacs barnacled with cynicism and their minds programmed for sardonic wit, sprang to their feet. The presence of grammatical metaphors (Lakoff 1987; Sullivan 2013) also creates highly metaphoric utterances. The corpus example in (32a) is more marked than its constructed counterpart in (31b). But the only difference between these utterances is their grammatical structure. Unusual constructions or argument structure can increase metaphoricity. (32) Altered or unexpected argument structure (32a) My normal sensations flooded back, although I had to migrate my eyes to get them pointed forward. (32b) My normal sensations flooded back, as my eyes migrated forward again. The purpose of these examples is to show that grammatical constraints on the lexicalization of metaphoric mappings also influence the productivity of those mappings, both in the form of (i) which metaphoric utterances are acceptable and (ii) how marked or metaphoric an utterance is in actual usage. An alternate analysis is that some of these examples involve mixed metaphors that increases their markedness (Crisp 2005; Goatly 2011). In some cases, the two analyses overlap, but the observations about metaphoricity here apply to many cases in which only one conceptual metaphor is present.

5 Summarizing constraints on metaphor productivity The purpose of this paper has been to synthesize findings from recent work in order to make the argument that script-based semantics is necessary to describe

Ontological and grammatical constraints on metaphor productivity

75

the linguistic productivity of metaphoric mappings. Even though linguistic utterances are the primary evidence for metaphor as a phenomenon, most theories of metaphor have little if anything to say about the two main questions that have been asked here: First, what metaphoric utterances are possible? Second, what metaphoric utterances are marked or highly metaphoric?

Even though metaphor is a mapping between concepts, there are a number of predictable constraints on which concepts can be mapped and on how metaphor is lexicalized into observed linguistic utterances. The first type of constraint is on lexicalized meaning: in a script-based approach to semantics, words can point to concepts in an ontology. However, as shown in section 2 words are not always simple pointers and often encode other non-concept meanings as well. These lexicalized meanings persist after metaphoric mappings and change the meaning of the produced metaphoric utterances. The point here is that lexical semantics provides a constraint on the productivity of metaphoric meanings in a way that goes beyond simple relationships between concepts. Ontological constraints on metaphor productivity influence which concepts can be mapped together and in which direction the mapping can occur (source vs. target). We use Searle’s social ontology to show that the degree to which a concept depends on humans for its definition is the primary factor required to explain mapping constraints. The difficulty here is that, in most cases, metaphoric mappings do not become unacceptable but rather become more marked. A more marked mapping produces a highly metaphoric linguistic utterance (Dunn 2014a, 2014b, 2015a). Grammatical constraints on metaphor productivity influence the lexicalization of conceptual mappings into specific utterances. The increased density of metaphoric case roles increases the markedness of metaphoric utterances in the same way that certain ontological constraints increase metaphoricity. At a definable point, an utterance becomes ambiguous between literal and metaphoric readings. Taken together, these constraints support models that make predictions about metaphor-in-language: what utterances are metaphoric (i.e., Dunn 2013b, 2013c) and how metaphoric are these utterances (i.e., Dunn 2014a, 2014b)? The ability to make predictions in this way is essential for a falsifiable approach to metaphor and it depends heavily on an ontology-based and script-based approach to meaning.

76

Jonathan Dunn

References Crisp, Peter. 2005. Allegory and Symbol – A Fundamental Opposition. Language and Literature, 14(4) 323–238 Dunn, Jonathan. 2011. Gradient Semantic Intuitions of Metaphoric Expressions. Metaphor & Symbol, 26(1): 53–67. Dunn, Jonathan. 2013a. How Linguistic Structure Influences and Helps to Predict Metaphoric Meaning. Cognitive Linguistics, 24(1): 33–66. Dunn, Jonathan. 2013b. Evaluating the Premises and Results of Four Metaphor Identification Systems. In Proceedings of the Conference on Intelligent Text Processing and Computational Linguistics, Vol. 1 (CICLING 2013). Heidelberg: Springer. 471–486. Dunn, Jonathan. 2013c. What Metaphor Identification Systems Can Tell Us About Metaphor–in–Language. In Proceedings of the Conference of the North American Chapter of the Association for Computational Linguistics: First Workshop on Metaphor in NLP (NAACL 2013). Stroudsburg, PA: Association for Computational Linguistics. 1–10. Dunn, Jonathan. 2014a. Measuring Metaphoricity. In Proceedings of the Annual Meeting of the Association for Computational Linguistics (ACL 2014). Stroudsburg, PA: Association for Computational Linguistics. 745–751. Dunn, Jonathan. 2014b. Multi–Dimensional Abstractness in Cross-Domain Mappings. In Proceedings of the Annual Meeting of the Association for Computational Linguistics: Second Workshop on Metaphor in NLP (ACL 2014). Stroudsburg, PA: Association for Computational Linguistics. 27–32. Dunn, Jonathan. 2015a. Three Types of Metaphoric Utterances That Can Synthesize Theories of Metaphor. Metaphor & Symbol, 30(1): 1–23. Dunn, Jonathan. 2015b. Modeling Abstractness and Metaphoricity. Metaphor & Symbol, 30(4): 259–289. Goatly, Andrew. 2011. The Language of Metaphors. Abingdon: Routledge. Lakoff, George. 1987. Women, Fire, and Dangerous Things: What Categories Reveal About the Mind. Chicago: University of Chicago Press. Lakoff, George. & Johnson, Mark. 1980. Metaphors We Live By. Chicago: University of Chicago Press. Lakoff, George. & Johnson, Mark. 1999. Philosophy in the Flesh. NY: Basic Books. Nirenburg, Sergei. & Raskin, Victor. 2004. Ontological Semantics. Cambridge, MA: The MIT Press. Searle, John. 1995. The Construction of Social Reality. New York, NY: The Free Press. Searle, John. 2010. Making the Social World: The structure of human civilization. Oxford, UK: Oxford University Press. Steen, Gerard J. 2007. Finding Metaphor in Grammar and Usage. Amsterdam: John Benjamins. Sullivan, Karen. 2013. Frames and Constructions in Metaphoric Language. Amsterdam: John Benjamins.

Igor Boguslavsky

Meaning amalgamation, phrasal stress, and earning money For Victor, looking back at old times at Manežnaja Square

Abstract: The major mechanism that governs word meaning amalgamation is the semantic valency filling, or, in a different terminology, establishing predicate-argument pairs. However, this process is much more complex than it is customary to believe and by no means boils down to inserting nouns into the valency slots of verbs. One of the complications is that the slot-filling semantic units do not always correspond to the meanings of words. Often, a slot filler is one of the smaller elements which make part of the word meaning. Another difficulty is due to the fact that the prosody may interfere in the process of semantic interpretation. We demonstrate several examples of different types that illustrate these difficulties. In particular, we show a curious example of prosody-induced semantic shift which results in a humorous effect. Keywords: semantic valency, internal scope, phrasal stress, semantic shift, humor

1 Introduction There is a Russian saying: (1)

Vsex deneg ne zarabotaeš.

Literally, it means ‘You cannot earn all the money’. Its more profound message is that one should not work too much, one should be able to stop working at some moment, in the chase of prosperity one can miss life’s simple pleasures. Here is a typical example extracted from the novel “A Starry Ticket” by Vasily Aksenov: Šabašte, rebjata! – skazal stepennyj gruzčik Nikolaev. – Vsë ravno vsex deneg ne zarabotaete ‘Stop working, guys! – said the staid loader Nikolaev. You will not earn all the money anyway’.

Igor Boguslavsky, Institute for Information Transmission Problems, Russian Academy of Sciences / Universidad Politécnica de Madrid https://doi.org/10.1515/9781501511707-005

78

Igor Boguslavsky

Recently, I came across a joke in which this saying received an unexpected continuation: (2)

Vsex deneg ne zarabotaeš, čast’ vsë ravno pridëtsja ukrast’. ‘you cannot earn all the money, there is always some you have to steal’.

As we know from Victor Raskin, each good joke should contain something that is interpreted according to two different scripts at a time. There is no doubt this is the case in (2). It is equally certain that the twofold interpretation of (1) in context (2) has a linguistic basis. This paper will try to bring it to light. We will see that the differences between the interpretations under analysis bear a direct relation to how word meanings are connected to each other. Therefore, before analyzing joke (2), we will dwell on how the meanings of separate words are combined to form the full meaning of the sentence (section 1). It is relevant for the interpretation of (2) that the place of the phrasal stress can make a significant contribution to sentence semantics, not only to its communicative organization. In section 2, we will demonstrate it by the Russian negative particle ne ‘not’ and adjective edinstvennyj ‘sole, unique, single’. Finally, in section 3 we will come back to joke (2) and show how the two interpretations of (1) are formed.

2 Valency filling as semantic glue Suppose we have a sentence and we have to construct its semantic structure. To do that, we need at least two things. First, a dictionary is needed that contains semantic defintions of all meaning-bearing words of our sentence. Second, the syntactic structure of the sentence should be available. Given all this information, one can pose a question about how semantic definitions of words combine to form the semantic structure of the whole sentence. We proceed from the assumption that the main mechanism of meaning amalgamation is (semantic) valency filling. The notion of valency is wellknown, although its understanding by different scholars may vary – cf., for example a survey of different approaches in Somers 1987. We adopt the Moscow Semantic School (MSS) approach, which claims that the set of valencies of each word is determined by the meaning of the word and its behavior. Roughly speaking, a word L has valency P if L denotes a situation which has an obligatory participant P and this participant can be systematically expressed alongside L. A word that has valencies is called predicate, a word that

Meaning amalgamation, phrasal stress, and earning money

79

fills the valency is called actant (or argument). For more on valencies and actants see Apresjan 1974, Melčuk 1974, 2004a, 2004b, Boguslavsky 1985, 1996. Often, valencies are discussed in the context of government properties of verbs and nouns and constitute the theoretical foundation of valency dictionaries. We would like to make a different emphasis: valency filling is the major instrument of joining together word meanings, a kind of semantic glue, which makes the sentence meaning a connected structure. This is one of the similarities that exists between the Moscow Semantic School approach and that of Formal Semantics. Formal Semantics took in this revolutionary idea in the beginning of the 70s directly from R. Montague. To do justice, it should be mentioned that starting with the famous 8th issue of “Machine translation and applied linguistics” (MPPL 1964), which initiated the Meaning – Text approach in the Soviet Union, and numerous subsequent publications on the Explanatory Combinatorial Dictionary, it was explicitly claimed that the semantic definition of many words contains valency slots for the arguments. In the semantic definition, these slots are represented by variables. To construct the semantic structure of the sentence, one has to identify the actants with the help of the Government Pattern (≈ Subcategorization Frame) and substitute them for the variables. Linguistic phenomena rarely exist in their pure form. Most often, there is a core zone, in which the properties of the phenomenon stand out very clearly, and a periphery zone, in which these properties are weaker or undergo certain modifications. In terms of the dependency structure, the actants of the core zone are connected to their predicate directly, and not through some intermediate nodes. Besides, the actants depend on their predicates, and are not subordinated by them. As far as semantics is concerned, we will single out the impermeability of the word meaning to predicate-argument relations. The syntactic properties mentioned above are obvious and seem not to require explanations. We will only comment on the impermeability property. It means that actant A participates in the predicate-argument relation with predicate P “as a whole”. That is, predicate-argument relations cannot hold between word P and some internal semantic component of A. These properties are manifested the most obviously in the verbal constructions, in which the actants are expressed by the subject of the verb and different types of complements. The periphery zone is much looser, and these prototypical properties can be easily violated. As far as the syntactic aspect of valency filling is concerned, the deviation from the prototype is determined first of all by different syntactic potentials of valency-bearing words. There is a wide range of syntactic positions that an actant

80

Igor Boguslavsky

may take with respect to its predicate. From this point of view, three types of valency slot filling could be distinguished: ACTIVE, PASSIVE, and DISCONTINOUS ones (Boguslavsky 2003). If lexeme L subordinates its actant A by means of an immediate dependency relation, we will say that such a valency filling is ACTIVE (the boy [A] runs [L], the search [L] for [A] the solution). This is the most typical (prototypical) case. If a lexeme is subordinated to its actant, we will say that the filling is PASSIVE (green [L] leaves [A], run [A] quickly [L]). This kind of valency filling is characteristic of adverbials, adjectives, particles, conjunctions and prepositions. If there is no direct syntactic link between the lexeme and its actant, we will call such valency filling DISCONTINOUS. This is a relatively infrequent type, typical of adverbials and adjectives. Let us give two examples of discontinuous valency filling. First, there are many subject-oriented adverbials, which are syntactically connected to the verb, and at the same time one of their valencies is filled by the subject of this verb, for example: By habit [L], John [A] got up early. As the second example, we will take a very instructive Russian adjective edinstvennyj that is translated in different contexts as only, only one, unique, sole. The context, in which this adjective expresses its valencies in the most explicit way is the following: (3)

Pëtr – edinstvennyj čelovek, kotoromu Ivan doverjaet. ‘Peter is the only person whom Ivan trusts’

Therefore this construction can be used as an input of the semantic definition: (4)

Q jest’ edinstvennyj R, kotoryj P ‘Q is the only R which P’ = ‘(Q is R which P); among all Rs there is no one except (this) Q, which has property P’

Here, the parentheses enclose the part of the sentence meaning which serves as the context for edinstvennyj. In sentence (3), the subject ‘Peter’ fills valency Q, ‘person’ – valency R, and the clause ‘whom Ivan trusts’ – valency P. Given these variable instantiations, applying the definition (4) to (3) will yield the following meaning: ’Peter is a person whom Ivan trusts; among all the (relevant) people, there is no other (= different from Peter) person whom Ivan trusts’. The valency that is filled in (3) in the discontinuous way is P: the relative clause kotoromu Ivan doverjaet ‘whom Ivan trusts’ is syntactically connected to čelovek ‘person’, but fills a valency of its modifier edinstvennyj. This adjective can also serve to illustrate the semantic deviation from the prototype, i.e. the situation when a valency is filled with some internal component of the lexical meaning of the word, leaving all other components outside its scope.

Meaning amalgamation, phrasal stress, and earning money

81

We say that in such cases the predicate has an “internal scope”. Let us compare two sentences: (5)

U staruški slomalsja edinstvennyj stul. lit. at-old-lady broke the-only chair ‘the only chair of the old lady broke’

(6)

U staruški pogib edinstvennyj syn. lit. at-old-lady died the-only son ‘the only son of the old lady died’

Sentence (5) means: ‘the chair of the old lady broke; she had no other chair’ (= ‘there is no x different from this chair, such that the old lady has x and x is a chair’). We see that valency P of edinstvennyj is filled with stul ‘chair’. Sentence (6) is very similar to (5), but its meaning is slightly different. What this sentence implies under the most probable interpretation is not that the old lady has no other sons, but that she has no other children: ‘the son of the old lady died; she has no other children’. This amounts to saying that valency P is not filled with the whole meaning of son (‘male child’), but only with its top component ‘child’. The phenomenon of internal scope frequently occurs when the predicate is expressed by an adjective or an adverbial. It is typical for these words to affect an internal component of the word they are connected to. It seems reasonable to assume that if word A semantically affects word B in some way, then B should contain a meaning component for A to act upon. In other words, if word A semantically affects word B, in most cases there is a semantic valency behind it. Word B or its component fills a semantic valency of word A in one way or another. After all, all semantic connections between the elements of the semantic structure are connections between predicates and their arguments, and there are no other semantic connections. This principle is very useful when one is trying to find a semantic definition of a word. One has to examine all possible combinations of word A under investigation with other words. If there exist a combination A + B, where B semantically affects A and characterizes semantic element x, this is a good argument in favor of including x in the definition of A. To give one example, the Longman Dictionary of Contemporary English defines accent as ‘the way someone pronounces the words of a language, showing which country or which part of a country they come from’. According to this definition, Southern accent is interpreted as the way somebody pronounces the words of a language, showing that the speaker is from the South. However, this

82

Igor Boguslavsky

definition does not explain the combinability of this word with intensifiers: strong accent. The definition does not contain any quantifiable component that is affected by these adjectives. What do these adjectives intensify? When we say that somebody speaks English with a HEAVY Essex accent we mean that his pronunciation of English words (a) is typical for people from Essex and (b) is VERY different from the standard. This is a good reason for revising the definition of accent and including the component ‘different’ in this definition: (7)

X has an A accent (in B) = ‘the way X pronounces the words of language B is different from the way speakers of B usually pronounce them and typical for speakers of language, group, or locality A’.

In this section we discussed the way meanings of words connect to each other. We saw that the major mechanism of word combination is valency filling and that finding semantic material to fill valencies of predicates might be not an easy task. However, valency filling is not the whole story. There are other linguistic instruments that contribute to the sentence interpretation, sometimes in an unexpected way. One of these is prosody. In the next section we will show some relevant data to illustrate it.

3 Semantic shift under phrasal stress We will examine some facts that show that phrasal stress can modify the meaning of the word and sometimes result in different valency filling.

3.1 Negation meaning ‘less than’ Normally, the phrasal stress shift leads to the modification of the communicative organization of the sentence, but not its denotation. For example, in examples (8a-c) the situation described is the same – the purchase of bread by Ivan – but it is presented in different communicative perspectives: (8a) Ivan ↓kupil xleb Ivan ↓bought bread ‘Ivan did buy bread’

Meaning amalgamation, phrasal stress, and earning money

83

(8b) Ivan kupil ↓xleb. Ivan bought ↓bread ‘It is bread that Ivan bought’ (8c) ↓Ivan kupil bread. ↓Ivan bought bread. ‘It is Ivan who bought bread’. On the other hand, it is well known that the communicative organization of the sentence, that defines the topic-focus articulation, is essential for constructing the negative sentence given the asserted one. Since the negation affects only the focal part of the sentence, the negative particle is placed before the focus usually marked by the phrasal stress (Paducheva 1974). (9a) ‘it is not the case that (8а)’ (9b) Ivan ne kupil xleb. ‘Ivan did not buy bread’ (10a) ‘it is not the case that (8b)’ (10b) Ivan kupil ne xleb. lit. Ivan bought not bread ‘It is not bread that Ivan bought’ (11a) ‘it is not the case that (8c)’ (11b) Xleb kupil ne Ivan. lit. Bread bought not Ivan ‘It is not Ivan who bought bread’. Also, it was long noted that instead of its normal meaning ‘it is not the case that’ the Russian negative particle ne ‘not’ may mean ‘less than’ in some contexts: (12) Eto pal’to ne stoit tysjači dollarov lit. this coat does not cost thousand dollars ‘this coat costs less than a thousand dollars’

84

Igor Boguslavsky

Sentence (12) is not a standard negation of (13): (13) Eto pal’to stoit tysjaču dollarov ‘this coat costs a thousand dollars’ because the negation of (13) should mean ‘the price of the coat is not a thousand dollars’. Should the conclusion be that the negative particle in (12) has a specific meaning? I believe not. The origin of the relationship between (12) and (13) lies elsewhere (Boguslavsky 1985). Let us compare sentence (13) pronounced with the neutral intonation and sentence (14) pronounced with the phrasal stress on the last word: (14) Eto pal’to tysjaču dollarov ↓stoit. lit. this coat thousand dollars ↓costs The two sentences are used in somewhat different communicative contexts and have slightly different meanings. Sentence (13) simply informs us on the price of the coat. Sentence (14) has a different point of interest. It is appropriate in a dialogue as a reply to the interlocutor’s doubt as to the adequacy of such a high price. The speaker of (14) does not claim that that the coat costs just thousand dollars. Rather, he says that the price of the coat is no less than a thousand, and maybe even more. Looking at (12), we see that this sentence, just as (14), is perceived as a reply to an earlier opinion (maybe imaginary) on the price of the coat. We can conclude that the peculiarity of sentence (12) is not related to the specific meaning of particle ne (‘less than’). Its meaning is explained by the fact that it is the negation of the communicatively marked sentence (14). Indeed, if (14) means ‘costs a thousand dollars or more’, its negation should naturally mean ‘costs less than a thousand’. As for the communicatively neutral sentence (13), it is intended to answer the question “How much does the coat cost?” Therefore, its communicative focus and the phrasal stress fall on the final noun phrase. It would be a mistake to expect that the negation of (13) would look like (12). The negative particle should be placed at the focal element of the sentence, that is at the final noun phrase: (15) Pal’to stoit ne tysjaču dollarov.

Meaning amalgamation, phrasal stress, and earning money

85

Our conclusion is as follows: what makes sentence (12) specific is not a particular meaning of the negative particle but the shift of the meaning of the affirmative sentence induced by the prosody.

3.2 Edinstvennyj ‘only, unique, sole’ In this section we will show how the prosody may affect valency filling. We will speak about the valencies of the adjective edinstvennyj ‘only, unique, sole’, which claims that a certain object is – in some aspect – unique (a more precise definition will be given below). Edinstvennyj is used in several syntactic contexts, and in each of them the valencies are filled in a different way. A detailed analysis of valency instantiation of edinstvennyj can be found in (Boguslavsky 1996). Here we will only give some examples. It is the copulative construction that is the most transparent from the point of view of valency instantiation. In this construction all the valencies are filled by clearly distinguishable phrases: (16) Pëtr – edinstvennyj čelovek, kotoromu Ivan doverjaet. ‘Peter is the only person whom Ivan trusts’ Therefore, this construction can be used as an input of the semantic definition: (17) Q jest’ edinstvennyj R, kotoryj P ‘Q is the only R which P’ = ‘(Q is R which P); among all Rs there is no one except (this) Q, which would have property P’ Here, the parentheses enclose the part of the sentence meaning which serves as the context for edinstvennyj. In sentence (16), the subject ‘Peter’ fills valency Q, ‘person’ – valency R, and the clause ‘whom Ivan trusts’ – valency P. Given these variable instantiations, applying definition (17) to (16) will yield the following meaning: ‘Peter is a person whom Ivan trusts; among all the (relevant) people, there is no other (= different from Peter) person whom Ivan trusts’. Valency P is obligatory in all contexts, R and Q are optional. For this paper, the construction “edinstvennyj + NP”, where edinstvennyj serves as a modifier of NP, is especially relevant. This construction may be contrasted to the co-predicative construction, in which the adjective also agrees in number, case and gender with the noun, but is linked syntactically to the verb and not to the noun. In co-predicative and modificative constructions, edinstvennyj fills its valencies quite differently. This becomes obvious if we compare the copredicative phrase (18a) and the modificative one (18b):

86

Igor Boguslavsky

(18a) Angličanin edinstvennyj prišel vovremja. lit. Englishman only came on-time ‘the Englishman was the only one to come on time’ (18b) Edinstvennyj angličanin prišel vovremja. lit. only Englishman came on-time ‘the only Englishman came on time’ In both cases the Englishman is set off to other people according to some property, but these properties are quite different in (18a) and (18b). In (18a) no one else came on time, and in (18b) no one else was an Englishman. Another thing that is important about the modificative construction (Adjective + Noun) is that the modified noun fills two valencies of edinstvennyj at a time – Q and P. If we come back to sentence (18b), we will see that ‘Englishman’ occurs twice in its semantic structure: ‘besides (this) Englishman, there is no one who is an Englishman’. It is to be noted that these two occurrences of ‘Englishman’ differ in their referential status: in the position of Q the status is referential (‘this Englishman’), while in the position of P it is predicative (‘be an Englishman’). Now we are ready to demonstrate the influence of phrasal stress on the valency filling of edinstvennyj in the modificative construction. Let us compare sentence (19a), in which edinstvennyj is pronounced with a neutral intonation, and (19b), where this word is stressed: (19a) My upustili edinstvennuju vozmožnost’ perelomit’ situatsiju. lit. we missed only opportunity to reverse the situation ‘we missed the unique opportunity to reverse the situation’ (19b) Eto neprijatno soznavat’, no, poxože, my vospol’zovalis’ ↓edinstvennym blagom svobody. ‘It is frustrating to realize, but it seems we made use of ↓only-one asset of freedom (there were several assets, but we made use of only one of them)’. In both cases, edinstvennyj is a noun phrase modifier without any restrictive attributes, which makes both sentences similar from the point of view of the valency instantiation rules. However, sentence (19a) is interpreted as predicted: ‘we missed an opportunity to reverse the situation; there was no other opportunity’. This interpretation is obtained with Q = ‘opportunity’, P = ‘be an opportunity’, and R not instantiated. In (19b), on the contrary, the modified noun phrase (‘asset of freedom’) fills valency R, P is instantiated by the

Meaning amalgamation, phrasal stress, and earning money

87

predicate of the sentence (‘make use’), and Q is not instantiated at all: ‘we made use of an asset of freedom; of all the assets of freedom, there is no other one we made use of’. This dramatic shift in valency instantiation has been provoked by the phrasal stress that falls on edinstvennyj. The semantic shift induced by the phrasal stress in (19b), should not be confused with another phenomenon which is also related to the phrasal stress. It is well-known that if the phrasal stress is moved from the top element P of the noun phrase to its internal component, it has a semantic consequence: the noun phrase is considered against the background of a larger class consisting of all Ps. For example, in the sentence I want a ↓chocolate ice-cream what the speaker claims is not that he just wants a chocolate ice-cream, but also that this variety of ice-cream is selected out of a larger class of ice-cream varieties. Here the phrasal stress introduces contrast (chocolate ice-cream and not some other kind of ice-cream). This phenomenon has nothing to do with the valency shift discussed here.

4 Earning all the money Let us come back to joke (2) and examine the two interpretations of (1) to which it refers. (2) You can’t earn all the money, there is always some you have to steal. (1) You can’t earn all the money. First of all, one should note, that the two interpretations differ in what exactly is quantified by all. Under the first interpretation, similar to the English proverb There is always more fish in the sea, money denotes all the existing money: you can’t earn all the money that exists around just as you can’t catch all the fish that exists in the sea. Under the second interpretation, suggested by the continuation there is always some you have to steal, what is meant here is the money which the subject needs. You can’t earn all the money needed, some part of this money will have to be obtained in a different way. However this difference is not the only one and the most important trait of these interpretations. Let us look into the meaning of the verb zarabatyvat’ ‘to earn’. The Longman and Macmillan dictionaries define to earn as ‘to receive a particular amount of money for the work that you do’. Examples:

88

Igor Boguslavsky

(20) He earned nearly £20,000. I was the only person in the house who was earning. She was earning good money at the bank. Chris will pay – he’s earning a fortune. These examples confirm that among the two components of the definition – ‘to receive a particular amount of money’ and ‘for the work that you do’ – the first component is the central one. Sentence He earned £20,000 normally entails that he received £20,000. However, in some contexts this is not true (at least, in Russian). Cf.: (21) V prošlom mesjace ja zarabotal mnogo deneg, no ne poluchil ix, poskol’ku rabotodatel’ otkazalsja platit’. ‘I earned a lot of money last month, but never received it because my employer refused to pay’. In such sentences the fact that somebody earned some money does not imply that he got it. What can be inferred is that he obtained the right to this money as return for work. It hardly makes sense to postulate different lexemes zarabatyvat’ ‘to earn’ represented in (20) and (21). Both variants should be combined disjunctively in the same definition: (22) А1 zarabotal А2 ‘A1 earned A2’ = А1 (a) received money А2 or (b) obtained the right to receive money A2 as return for work’. It is essential that components (a) and (b) in the definition of zarabatyvat’ are not of equal footing. Component (a) is the main one and is realized by default – cf. (20). For component (b) to be realized, special contextual conditions are needed. The first of these is the context which excludes the realization of (a). This is what we see in sentence (21). The second condition which can trigger realization of (b) is the phrasal stress on the verb. Cf. sentences (23) and (24): (23) Ja ↓zarabotal eti den’gi. ‘I ↓earned this money’ ‘I obtained the right to receive this money as return for work’ (24) Ty ne ↓zarabotal eti den’gi. ‘you did not ↓earn this money’ ‘you did not obtain the right to receive this money as return for work’

Meaning amalgamation, phrasal stress, and earning money

89

Let us emphasize that both in (23), and in (24) component (a) of the definition is not realized. It is not presupposed that the subject has the money in question. For example, (23) is possible both in the case the subject received the money (cf. I earned this money and can spend it as I please) and in the case he did not (cf. I earned this money and I demand that it be paid immediately). Here is an example from a journal illustrating component (b): (25) Den’gi ostajutsja, no ostaëtsja i dolgovoe objazatel’stvo, tak kak gosudarstvo ne zarabotalo etix deneg (sobralo v vide nalogov), a vzjalo v dolg. (from Russian National Corpus) ‘The money remains, and also remains the debt, because the state did not earn this money (did not collect it as taxes), but borrowed it’. It is to be noted that if zarabatyvat’ is placed at the end of the sentence and bears the phrasal stress (e.g. Ty ničego ne ↓zarabotal, lit. you nothing not ↓earned, ‘you didn’t earn anything’), then the phrasal stress is no longer semantically distinguishing, because it is a standard prosody of the final group. Therefore, this sentence can be interpreted both in the sense (a) (You have no money), and in the sense (b) (you have no money received as return for work). Now let us come back to joke (2) with which we began this paper and show how phrase (1) obtains both its interpretations. (1)

Vsex deneg ne zarabotaeš. ‘you can’t earn all the money’

The first interpretation is the default one. Component (a) of the zarabatyvat’ definition is realized, and the sentence means: you can’t receive all the money (that exists or is otherwise in sight). The sentence is in the same row as sentences You can’t read all the books (there will be always books you didn’t read), You can’t catch all the fish, etc. The second interpretation is possible due to the phrasal stress falling upon the verb. Component (b) of zarabatyvat’ is realized, which gives the meaning: you can’t obtain the right to receive all the money (you need) as a compensation for work. In other words, you may have the right of receiving only a part of the money you need.

90

Igor Boguslavsky

5 Conclusion The major mechanism that governs word meaning amalgamation is the semantic valency filling, or, in a different terminology, establishing predicateargument pairs. However, this process is much more complex than it is customary to believe and by no means boils down to inserting nouns into the valency slots of verbs. One of the complications is that the slot-filling semantic units do not always correspond to the meanings of words. Often, a slot filler is one of the smaller elements which make part of the word meaning. Another difficulty is due to the fact that the prosody may interfere in the process of semantic interpretation. We demonstrated several examples of different types that illustrated these difficulties. In particular, we showed a curious example of prosodyinduced semantic shift which results in a humorous effect.

References Apresjan Juri D. 1974. Lexical Semantics [Leksicheskaja semantika]. Moscow: Nauka. Boguslavsky I. 1985. Studies in syntactic semantics: scope of logical words [Issledovanija po sintaksicheskoj semantike: Sfery dejstvija logicheskix slov. ] Moscow: Nauka. Boguslavsky Igor M. 1996. Scope of lexical units [Sfera dejstvija leksicheskix edinic]. Moscow: Shkola Jazyki russkoj kul’tury. Boguslavsky Igor M. 2003. On the Passive and Discontinuous Valency Slots. In: Proceedings of the 1st International Conference on Meaning–Text Theory, pp. 129–138. Ecole Normale Supérieure, Paris, June 16–18. Mel’čuk, Igor A. 1974. Esquisse d’une théorie des modèles linguistiques du type “Sens ⇔ Texte”. La sémantique, la syntaxe. [Opyt teorii lingvističeskix modelej “Smysl ⇔ Tekst”. Semantika, Sintaksis.], Moscow: Nauka. Mel’čuk, Igor A. 2004a. Actants in semantics and syntax I: actants in semantics, Linguistics 42(1). 1–66. Mel’čuk, Igor A. 2004b. Actants in semantics and syntax II: actants in syntax, Linguistics 42(2). 247–291. MPPL 1964. – Machine Translation and Applied Linguistics (Mašinnyj perevod i prikladnaja lingvistika). Moscow, 1964. Paducheva, Elena V. 1974. On the semantics of syntax [O semantike sintaksisa]. Moscow: Nauka. Somers, Harold L. 1987. Valency and Case in Computational Linguistics (Edinburgh Information Technology Series 3). Edinburgh: Edinburgh University Press.

Part 2: Humor

Giselinde Kuipers

Knowledge about humor Abstract: Humor is one of the forms of communication that is most likely to fail. Such humor failure is of interest to humor scholars because it highlights the mechanisms in making humor work (or not). Existing work on humor failure is mainly done in linguistics, and focuses on humor failure as a result of a lack of knowledge. This article argues that humor success and failure depend on Knowledge about Humor: widely shared (though not uncontested), group specific or culture-specific cultural rules and conventions regarding the use of humor. Using examples from humor research, my own empirical work, everyday life and current affairs, I show that different groups have different rules and expectations regarding 1. how, 2. when, 3. by whom and with whom humor is used. Such knowledge about humor is what sociologists refer to as practical knowledge: learned behavior that native users experience as natural, but with clear regularities that researchers and clever individuals can “decipher” as rules or rule-like. I argue that knowledge about humor should be included in script-based theories of humor as an additional knowledge resource: KaH. Hierarchically, KaH can be positioned “outside” or “above” previously defined knowledge resources. This is a major contribution sociology can make to the rich tradition of script-based humor theories as developed originally by Victor Raskin in 1985. Keywords: failed humor, knowledge about humor, script-based humor theories, practical knowledge, sociology

1 Introduction: Humor failure as a source of knowledge Humor is one of the forms of communication that is mostly likely to fail. This failure is often experienced as painful and embarrassing, because it is taken to be a sign that someone is an outsider, not sufficiently knowledgeable of group culture or shared knowledge (Coser 1960; Smith 2009; Shively 2013; Bell 2015; Kuipers 2015; Hale 2018). Among comedians, a failed performance is known as “dying”. This idiom highlights how excruciating failed humor can be, both for Giselinde Kuipers, University of Amsterdam https://doi.org/10.1515/9781501511707-006

94

Giselinde Kuipers

the would-be joker and their audience. A successful joke, met with laughter, is the ultimate proof of social attunement. A failed joke signals social failure: we are not on the same wavelength. For academics, failed humor is interesting because it lays bare the workings of humor. Not unlike the way mechanics gain insight into the processes “under the hood” when trying to figure out why the machine stopped working, humor scholars can learn a lot from the moment when a joke falls flat. In a failed joke, like a broken car or computer, the cause of the problem is often not immediately clear. The problem may be in the mechanics of the joke itself, but it may also be the topic, the joker, the audience, the relationship between would-be joker and audiences or something else about the social context. Very often, it is a combination of these things. Thus, failed humor shows us the many elements that should be in place and properly functioning to perform this everyday miracle: a successful joke. As is often the case in humor research, the more sophisticated work on failed humor is done by linguists. This linguistic work on humor has acknowledged this possibility of failure in various ways. First, semantic theories such as the SSTH, GTVH and OSTH (explained below) provide many reasons why a joke itself might be of low quality or even fundamentally flawed, and thus unsuccessful. Second, both purely semantic and more pragmatic approaches may help explain why people may not enjoy a well-constructed and well-delivered joke. From a strictly semantic perspective, the most important reason for joke failure is insufficient knowledge. People who fail to get a joke lack the required “knowledge resources” (Attardo and Raskin 1991; Bell and Attardo 2010) to understand, and therefore appreciate, a joke. This is clearly demonstrated in Nancy Bell’s (2015) analysis of failed humor. The starting point for her research was her work on humor in classrooms where English is taught as a second language. In such settings, participants lack many of the required knowledge resources – obviously language, but also many other forms of group-specific, often implicit knowledge required to “get” a joke. However, linguistic analyses tend to focus on the knowledge implied in the joke itself, such as the linguistic knowledge required understand puns, or more general cultural understandings to see why two scripts are experienced as oppositional. In this contribution, I look at another form of knowledge required for a successful joke: knowledge about humor. By this I mean: group-specific or culturally specific knowledge about social and cultural conventions regarding humor. The success of a joke also depends on people’s ability to abide by – or play with – such conventions. However, this knowledge is usually not implied by the joke itself. Every society has settings where humor is not allowed, or where humor is almost mandatory (think of a wedding speech); semi-ritualized forms of

Knowledge about humor

95

interaction revolving around humor (like a roast in American culture); or social roles that do or do not allow for joking. This is a knowledge resource that falls largely outside of the scope of linguistic approaches to humor. Instead, the analyses of such “knowledge about humor” falls squarely in the domain of sociology, the discipline that studies social relations and institutions. Sociology as a discipline can therefore make a unique contribution to script-based theories of humor by analyzing Knowledge about Humor (or KaH) as a knowledge resource. In this article, I argue that in order for a “single jokecarrying text” (Raskin 1985) to be understood and appreciated as humor, people have to observe shared (though usually implicit) cultural rules stipulating 1. how, 2. when, 3. by whom and with whom humor is used. These rules are usually embedded in societal “humor regimes” (Kuipers 2011) and “joking cultures” (Fine and DeSoucey 2005). Hierarchically, KaH can be positioned “outside” or “above” the existing knowledge resources. It is therefore an extension of, rather than an amendment to this script-based humor theories. However, it is more specific than Villy Tsakona’s (2017) notion of context, also presented as an additional knowledge resource that captures sociocultural knowledge needed to understand humor. I will return to this below. This addition to script-based humor theories has both theoretical and practical implications. Theoretically, it provides a further conceptualization of the knowledge resources needed to produce and share humorous interactions. It broadens the horizon of linguistic semantics by trying to understand not only whether a particular piece of language is a joke, but also under what conditions this joke will be amusing, and to whom. More practically, the realization that KaH is necessary for successful humor is relevant for further computational applications of humor research. While people may be willing to forgive or ignore people’s failure to create humor, or their failure to appreciate their jokes, in computers or other devices it will certainly be met with less consideration. Indeed, failure to use humor successfully – that is, with suitable and suitably funny content for all parties in the interaction, given the situation and the context – will likely be seen as the definitive proof of the computer’s inhumanity.

2 Humor and knowledge: The semantic approach to humor and beyond In 1985, Victor Raskin published Semantic Mechanisms of Humor, the book that has made contemporary humor studies into what it is today: an interdisciplinary field that models itself on the sciences rather than the humanities. This is

96

Giselinde Kuipers

unexpected given the (seeming) frivolousness of the topic, as well as the early origins of humor studies in philosophy, literature, theatre and psychoanalysis. Since the 1980s, the field has become a gathering point for empirically minded scholars from all walks of life, but dominated by linguistics and psychology. Consequently, the field has not been very receptive (one might even say hostile) towards the critical, cultural, reflexive or post-structural approaches that have become prominent in much of the social and cultural sciences. At the core of humor studies, we see a remarkable consensus that humor is based on incongruity. Typically, humor scholars dutifully discuss the “three classical humor theories” before moving on to identify incongruity theory as the most promising in terms of the development of a systematic, overarching, maybe even testable theory of humor. Raskin’s script-based approach, and later iterations of this theory, is by far the most sophisticated version of incongruity theory. It stipulates that a “single joke-carrying text” is compatible with two scripts, and that “the two scripts with which the text is compatible are opposite” (1985: 99). Thus, the rather vague notion of incongruity is conceptualized and operationalized as “script opposition”. Scripts are seen as culturally specific knowledge: ‘‘a cognitive structure internalized by the native speaker and it represents the native speaker’s knowledge of a small part of the world.’’ (Raskin 1985: 81) The central assumption of this approach therefore is that the production and the reception of humor is a cognitive process that is based in knowledge. In subsequent adaptations, this theory became the General Theory of Verbal Humor, or GTVH (Attardo and Raskin 1991; Hempelmann and Ruch 2005; Attardo 2001; Hempelmann & Attardo 2011), and the Ontological Semantic Theory of Humor or OSTH (Raskin, Hempelman and Taylor 2009; Taylor and Raskin 2012; Rayz 2017). Today, the theory is still generally referred to as the GTVH, possibly because the OSTH, with its formulas and strings of code, requires more computational ability and ambitions than most (humor) scholars have. In these subsequent iterations, the theory has become increasingly sophisticated both in its conceptualization of script oppositeness, and in its conceptualization and operationalization of knowledge resources. The GTVH refined the notion of “script oppositeness” by introducing a hierarchical classification of knowledge resources. These can be represented as concentric circles, from the more basic “script oppositions” (SO) and “logical mechanisms” (LM) that are relatively limited in number and possibly shared across cultures, to the more diffuse and definitely group-specific knowledge resources: situation (SI), target (TA), narrative strategy (NS) and language (LA). Over the next three decades, the script oppositions and logical mechanisms have been classified and refined, critiqued and revised (Davies 2011; Oring

Knowledge about humor

97

2011; Hempelmann and Attardo 2011), as part of the ongoing debate in humor studies on what “incongruity” means and how it can be defined. However, the more diffuse or cultural-specific knowledge resources – the situations, targets, narrative strategies and language – received much less attention in these debates (however, see Attardo 2001, Tsakona 2017). Not only do these knowledge resources show much more variability, they also seem less directly related to the “core” of the joke. As all humor scholars know, jokes with the same punchline – that is, with similar script oppositions and logical mechanisms – can be found around the world. Such global jokes are set in different situations, with different targets, in different narrative forms (cartoon, joke, riddle, story), told in different languages, but they still work (Davies 1990; 2002). The knowledge resources in the “outer circles” beyond SO and LM therefore seems less essential to the joke. Consequently, they are easily dismissed as irrelevant to the central question of humor research: “what is humor?” However, from the perspective of humor failure or success, these “outer” knowledge resources are as relevant as the knowledge resources of the inner circle. Here, the question is not “what is humor?” but “under what conditions does a joke work?”. Both the failure to produce a successful joke, and the failure to understand or to appreciate a joke, may be related to all these knowledge resources. In their analysis of failed humor among non-native speakers, Bell and Attardo (2010) show that humor production fails, for instance, because of insufficient language skills or inability to employ fitting narrative strategies. Failure to understand humor, in this article, is related to the entire array of knowledge resources: from failure to process language (LA) to failure to recognize the incongruity (SO). Thus, the matter of joke success versus failure leads us to consider all forms of knowledge involved in humor, not just the knowledge needed to create and decode the right kind of incongruity. In her fascinating book on failed humor, Bell (2015) discusses in detail how script-based theories help us understand failed humor. She starts by noting that this theory “goes much further” (Bell 2015: 26) than superiority and release theories in explaining failed humor, because it allows for various reasons for humor to fail, related to the joke, the joke-teller and to the hearer. Thus, hearers may have fewer scripts available, or have no or less understanding of the oppositeness of the scripts; while jokers may be unsuccessful in triggering the right scripts or oppositions. Interestingly, Bell extends her discussion to a later paper by Raskin on the sense of humor (1998). She translates Raskin’s analysis of four dimensions of a sense of humor into four reasons for humor failure: cognitive, communicative, experiential and volitional. Thus, apart from lack of general knowledge (cognitive) or specific knowledge (experiential), humor may fail for volitional reasons: because people are unwilling to accept a certain kind of

98

Giselinde Kuipers

joke, or refuse to acknowledge the joke even though they are amused by it. Finally, failed humor can be due to communicative reasons: because of a “lack of familiarity with communicative norms or inability (or conscious refusal) to engage with these norms” (Bell 2015: 26) Raskin’s original formulation suggest some impatience with these “volitional” and “communicative” aspects of the sense of humor. One can’t help but think that Raskin believes that a joke, like a plate of food or a drink, should be accepted when offered – out of simple politeness, or maybe because one can never be sure when the next one will be offered, and by that time it may be too late. However, I argue that such volitional and communicative aspects of humor are part and parcel of all processes of humorous communication. First, personal circumstances or situational specifics may preclude people from switching to “non bona fide communication”. Second, people may refuse to acknowledge their amusement because of personal, social or political reasons. This does not mean they are “hypocritical” as Raskin (1998) suggests. Certainly, failing to acknowledge amusement at some forms of humor (even when amused) does not mean people have no sense of humor. On the contrary, common-sense understandings of the sense of humor would include knowing when, where, and how to joke (Martin and Ford 2018.; Ruch 1998). Thus, the volitional and communicative aspects of humor are central to the ability to use humor successfully, as a fully integrated member of one’s language community. This article expands the notion of “knowledge resources”, to include not only things that are inherent in a joke text, as is the case for Attardo and Raskin’s original knowledge resources. It also goes beyond contributions made by sociolinguists and conversation analysts such as Nancy Bell (2015) or Jennifer Hay (2001), who add pragmatics to this semantic approach. I aim to add a sociological perspective that focuses not on the joke text or on joke performance, but instead highlights the wider cultural knowledge related to humor: the (generally implicit) rules and conventions related to the use of humor, as they are employed by specific sociocultural groups. The sociological tradition I am working in typically sees such cultural knowledge as both both practical and cognitive (cf. Lizardo and Strand 2010). People behave according to cultural rules and conventions, but they would be hard-pressed to formulate them explicitly. Yet, people respond with surprise, unease and annoyance when these implicit rules are broken. The regularities in such cultural practices and conventions can be “decoded” by academics (and other clever individuals) as rules. This is the “practical” part of cultural rules: people enact them even when they cannot put them into words. In addition, cultural knowledge is cognitive in the sense that it provides knowledge about the world. People employ culturally specific classifications to draw boundaries

Knowledge about humor

99

in the relative chaos of the world surrounding them. They distinguish, for instance, categories of persons (sociologists vs. linguists; women vs. men); and social settings that require or forbid specific forms of communication (Bourdieu 1991). I assume that this approach feels quite natural to linguists. Language, too, entails people following rules without being aware of the formal version of them. Language, too is both practical and cognitive. Finally, in language too, rules are identified most quickly by looking at moments of failure.

3 Knowledge about humor: The how, when, and who of joking So what does knowledge about humor consist of? Drawing on examples from everyday life, literature on humor, and my own research, I will show that in order for a “single joke-carrying text” to be understood and appreciated as humor, people have to observe shared (though usually implicit) cultural rules stipulating 1. how, 2. when, 3. by whom and with whom humor should be used.

3.1 How to use humor (and laughter) All humorous communication – both face-to-face and mediated humor – relies on humorous “framing”: some sort of indication that we are entering a non-bona fide form of communication. This framing is regulated by cultural conventions that show considerable variations. Different communities may have different rules on how to signal and acknowledge humorous intent in face-to-face communication. Should one laugh, produce a benevolent grin, or keep a straight face when attempting to joke? In accepting a joke, what is the best response: a belly laugh? A smile? A straight face, with maybe just a small twitch of the mouth? Should one try to joke back or not? Some forms of framing have a clear relationship to humor – for instance, a grin. However, many conventions for signaling humor cannot be easily traced back to universal human expressions of playfulness or non-seriousness (Kuipers 2009). For instance, in German and Dutch the beginning of a (canned) joke is signaled by the inversion of subject and verb. In English the same can be done by dropping the article: “Guy walks into a bar”. All regular viewers of American television recognize a sitcom because there is a couch in the middle of a living room. Among both French and Dutch, references to Belgians will activate humorous scripts related to stupidity and French fries. There is nothing inherently

100

Giselinde Kuipers

humorous about dropped articles, couches or Belgians, but anyone aware of these conventions will recognize the cue. Failure to recognize humorous intent leads to confusion (“Are you serious?” “Is this a joke?”) and often to humor failure. And while an inappropriate or unrecognizable response to humor does not mean that a joke has failed, it does hamper a successful switch to humorous interaction. This will discourage further exchange, in particular further attempts at humor. One of the most striking cultural variations in the “how” of humor has to do with a cultural or subcultural preference for more restrained/implicit versus exuberant/explicit styles of humorous communication. In a very enlightening article on “learning how to be funny in Spanish”, Rachel Shively (2013) describes learning not to use deadpan humor as one of the key elements of this process for her subject, a midwestern twenty-something male living in Spain. Clearly, his Spanish and Spanish-speaking interlocutors require a more exuberant and explicit signaling of humorous intent. This was a major cause of humor failure in the early period of his visit, until he learned to adapt to local styles. Stylistic conventions for signaling humor vary across, but also within countries. In my research in the Netherlands (Kuipers 2015), I found that one of the main objections to joke-telling was the overly clear signaling of humorous intent: a canned joke is so obviously intended as funny, with its strong genre characteristics, the lengthy narration, and the clearly signaled punchline, that some people failed to find it humorous anymore. This objection is couched in a wider difference in class cultures. Dutch upper-middle class people distinguish themselves from lower classes with a restrained “civilized” communication style, that requires controlled laughter, subtle body language, and understated, ironic humor that is not signaled too clearly. This general style led them to dismiss joke-telling as a genre. Thus, even while educated Dutch find the joke text understandable and acceptable, if it is delivered in an exuberant, overt way, this “ruins” their appreciation. Consequently, upper-middle class Dutch tend to look down upon people telling jokes: they are “not amused”. Conversely, lower middle class or working class Dutch often respond with confusion, annoyance or boredom to the implicit, ironic joking style of highly educated Dutch. This pattern of humorous miscommunication sustains class divisions in Dutch society. The humor failure on both sides of this class divide is only partly based in joke content or even in successful performance. Instead, it has to do with cultural rules regarding how to express and signal humor. Long before the joke reaches the punchline, people decide that this joke is “not for them”. Many (sub)cultures tend to prefer either exuberant or more restrained expression of humorous intent and “humor support” (Hay 2001). In much of (Northern) Europe, this division is class-coded. See the classic example of the

Knowledge about humor

101

English “stiff upper lip”, which is not so much a national character trait as a symptom of the most class-ridden society of Europe (yes, American readers, more so than France). As one moves east in Europe, the straight face appears to be become less of a signal of class, and more generally a signal of a good sense of humor. In Russia and Poland, over-enthusiastic signaling of humorous intent is believed to kill even the best joke (it is also seen as a sign of stupidity). Moving west across the Atlantic, in contrast, humorous signaling is allowed to be more explicit. This can be explained from wider societal characteristics: because of the relative lack of class divisions in the US and Canada both humorous intent and humorous appreciation can be signaled freely and abundantly without any danger of losing status or dignity. Moreover, in such “low context” (Hall 1976) immigrant societies, implicit communication is likely to be more risky and thus socially sanctioned. However, there are both regional and subcultural exceptions to this general “exuberant” signaling. The example of Shively’s Midwesterner may be a case in point. Various social and cultural groups are known to cultivate a more implicit style, for instance American Jews (Elliott Oring, personal communication). And some – typically controversial – humorous genres revolve around unsignalled humorous intent, such as teasing (Bell and Attardo 2010) and practical jokes (Marsh 2015). This divide between restrained/implicit versus exuberant/explicit humor and laughter is probably one of the most globally recognizable sets of cultural rules regarding the “how” of humor. But scattered throughout the literature on humor, one can find many more examples. For instance, the divide between “humor” and “wit” that was important in early humor theories (Morreall 1987), is primarily a stylistic division between more emotional/sociable versus more intellectual/cognitive forms of humor. While these divides are clearly visible in narrative strategies, cultural rules about the signaling of humor are not the same as the GTVH knowledge resource “narrative strategy”. The same narrative strategy – say, joke-telling – can be signaled both in a deadpan and an exuberant way. Moreover, a cultural of group-specific preference for more exuberant/ explicit versus controlled/implicit humor affects preferences for all knowledge resources. For instance, a more implicit/controlled humor style is associated with less explicit narrative strategies, but probably also with a distaste of obvious or unsophisticated script oppositions. Thus, knowledge regarding the “how” of humor is distinct from other knowledge resources. In line with the original formulation of the GTVH (Attardo & Raskin 1991), KaH adds another layer of “joke similarity”. Humorous utterances that draw on exactly the same knowledge resources, can still differ in terms of the KaH required to create a successful humorous exchange. In the case of knowledge about the “how” of humor, success requires, first, knowledge of

102

Giselinde Kuipers

conventions for signaling and accepting humor. Second, it requires shared humor and communication styles. Skilled users of a language typically employ such stylistic conventions without realizing it. They typically adopt the style most common in their surroundings, and tend to think of this as self-evident, natural, or as a sign of “a good sense of humor”. People with different humor styles are felt to be strange, boorish or boring, long-winded or oddly succinct. In all such cases, their attempts of humor are likely to fail. Knowledge about how to signal and respond to humor then, is a form of practical, culture-specific knowledge about humor, that all fully functioning members of a culture or group have, generally without realizing it.

3.2 When to use humor There are occasions when humor is more or less mandatory. For instance, in all the cultures that I am familiar with, a wedding speech requires some attempts at humor. There are also occasions where humor is typically considered inappropriate, and attempts at humor are likely to be received with grim silence, such as formal rituals, funerals, and the immediate aftermath of disastrous events. However, beyond such broad generalizations, there are striking differences in cultural rules stipulating when humor is considered appropriate or inappropriate, or even mandatory or completely forbidden. Around the world, we find examples of mandatory humor. Many tribal cultures know so-called joking relationships: relations where two people with a specific position, such as uncle and nephew, are expected to joke with each other. The joking in such relationships is often asymmetrical and quite harsh, but the butt of the joke is expected to show amusement (Radcliffe-Brown 1940; Apte 1985). Early anthropologists wrote about such relations with great puzzlement. In contemporary complex societies, comparable “teasing” relationships can be found, especially in work relations. These conventions may be as puzzling to uninformed observers as they were to early 20th century Africanists. For instance, in many classical orchestras the players of violas are often the butt of jokes, and they are expected to respond to this with good humor. There are also many examples where humor is a mandatory element of specific occasions. Around the world and throughout history, people have celebrated carnivals and rituals of reversal which generally involve humor (Apte 1985). Participants in such rituals are expected to behave in humorous ways, and all participants are expected to show appreciation. This is not always easy: in contemporary carnivals, for instance in the Netherlands, one can often see the participants, especially dignitaries such as mayors and priests, trying hard

Knowledge about humor

103

to look amused. In such rituals, humor failure is not an option. But there is leniency for those who try. The important thing is that the humorous frame is not contested. Versions of such “humor rituals” exist across many contemporary societies, but with striking cultural variations. An example that never fails to surprise cultural outsiders is the American tradition of the “roast”: a ritualized mocking, again quite harshly, at the expense of one of the participants, who is expected to show amusement (Oring 1994). The entire event is characterized by strong “humor coercion”: the event in effect forbids humor failure. Even presidents are not exempt from this tradition: American presidents usually undergo this ritual at the annual Correspondents Dinner. However, the current president Donald Trump so far has refused invitations to this Correspondents Dinner.1 In 2018, the performance of comedian Michele Wolf at this event (without the president) led to an unusual outcry.2 Both the non-attendance of the president, and the strong negative responses to the skit are interesting in the light of this discussion of KaH. They show what happens when people refuse to go along with such mandatory humor rituals. All parties concerned are aware that this is a pointed flaunting of cultural rules. This sends a clear message: “I am not part of your community.” Michael Billig (2005, cf. Smith 2009) called this unlaughter: the pointed non-acceptance of humor. Such mandatory humor is perhaps the most dramatic example of cultural rules stipulating when to engage in humor exchange. Because of the clear boundaries of such joking relations and humor rituals, these rules are relatively easy to decipher. In such cases, humor failure may happen, but it is probably not accidental or unintended (as the case of President Trump illustrates). However, most cultural rules regarding when (not) to joke are more diffuse and unspecific. In general, not joking when one should is not particularly problematic. However, joking when one shouldn’t can be remarkably painful and embarrassing. It also happens quite easily, especially in intercultural communication. The danger of “joking at the wrong moment” is a common trope in handbooks and websites provided advice on intercultural management and negotiation. Typically, English speakers are advised something like this: “when making a presentation to a cross-cultural audience, leave the jokes in your briefcase. An eloquent speech or a clever argument are one thing but most audiences, wherever you are in the world, don’t need a stand-up routine. Dignity and

1 https://www.politico.com/story/2018/04/06/trump-plans-to-again-skip-the-white-housecorrespondents-dinner-505507 2 https://www.nytimes.com/2018/04/30/business/media/michelle-wolf-white-housecorrespondents-dinner.html

104

Giselinde Kuipers

professionalism are better. Tell jokes in a presentation and you may be regarded as amateurish, or lightweight.”3 Advice in Dutch management and intercultural communication guides is typically more double edged. Using humor in the interaction with Brits and Americans is recommended. However, there are often severe warning against using humor in formal interactions with our neighbors to the east: “You think you should break the ice by a bit of clowning? Better not. The pitfall for the Dutch is to become informal too quickly. Germans like to keep business and private separate: [in German] ‘Dienst ist Dienst und Schnaps ist Schnaps’ ([Dutch translation in original] work is work and Schnapps [liquor] is Schnapps). So: no shoulder slapping or jolly behavior.”4 Intercultural management and communication handbooks typically relate such cultural codes to national variations in strictness of dividing lines between public and private, formal and informal. When boundaries between formal and informal settings are more strictly guarded, humor is relegated to informal situations. This seems a plausible explanation. However, there is surprisingly little academic research on this. Probably, such cultural conventions are more fine-grained than the boilerplate generalizations of management handbooks. Beside general divides between formal and informal cultures, private and public spheres, rules for when to joke or not have to do elusive aspects of communication like “tact” and “politeness” – vaguely circumscribed categorizations of proper and improper interpersonal behavior that refer to a sort of interpersonal “ease” that cannot be captured in rules. Moreover, while informal rules on when to joke vary across countries, they also vary considerably across micro-cultures (Fine 1979). Organizations, groups of friends, academic communities, or even spontaneously emerging fleeting groups in for instance public transportation quickly develop their own “joking culture” (Fine and DeSoucey 2005) that may diverge quite a bit from general conventions. For instance, while funerals are often cited as places where humor would be improper, I am personally familiar with specific micro-cultures where this is not out of place at all (including a microculture surrounding a well-known humor scholar). Moreover, with intercultural or bicultural communities becoming increasingly common, such rules become even more diffuse. Learning the rules about when to joke or not is a constant process of socialization into fluctuating, local knowledge about humor, and of constant cocreation of local joking cultures. While such local humor knowledge is

3 https://countrynavigator.com/blog/cultural-intelligence/humor-across-cultures/ 4 https://www.sprout.nl/artikel/internationaal/8-dingen-die-je-moet-weten-over-zakendoenduitsland

Knowledge about humor

105

suspended in larger humor regimes, it cannot be reduced to it. It is also open to negotiation and exploration. Therefore, it is also easily contested, and may lead to disagreement. Thus, “knowledge about humor” is practical knowledge that cannot be easily translated into a clear set of rules that people always follow. As I wrote above, such practical knowedge is sometimes much closer to “tact” or “politeness”. Interestingly, these terms are not unlike the notion of “sense of humor”: they are all forms of practical knowledge, a “feel for the game” (Bourdieu 1998) that is culture-specific, has quite a bit or regularity, but is always contested. At the lower end, a certain “feel for the game” is the minimal requirement for group membership (“You gotta have a sense of humor”). At the high end, it is a special talent only available to some unique individuals. Thus, knowledge about humor, like all forms of knowledge, is a matter of learning, competence – and some talent.

3.3 By whom and with whom humor to use humor A final element of KaH I would like to explore here is the who of humor. Who can use humor, with whom and to whom? Who can be joked about? Here, too, we see marked differences between culture and social groups. Moreover, in this case, too, we see that not having such knowledge not only lead to humor failure, but also to social discomfort, indignation, or worse. First, skilled members of a community need to be able to recognize socalled humor specialists, and to respond appropriately to this specific category of people. Across societies, we find special functions and positions that are associated with humor (Apte 1985). This ranges from humorous clowns in Native American culture and joking monks in Buddhism, to medieval court jesters and fools, to contemporary clowns, comedians and satirists. Such people have the license and the duty to be funny. “Normal” people in turn have an obligation to respond in kind: to accept the humor, and to accept more from such humor specialists than from others. Such humorous roles are usually clearly demarcated – for instance by a distinctive outfit or clear genre conventions – which limits the risk of humor failure because of KaH. As in other forms of signaling humor, these conventions may be quite random, and not directly related to conventional expressions of humor or mirth. Therefore, they are not only culture-specific but they can also change quite quickly. An interesting example of this is the traditional clown outfit, with a red nose, red wig, and big shoes. As a result of horror movies such as Poltergeist (1982) and It (1990), clowns are increasingly seen as scary rather than fun. Failure to recognize such a humorous specialist can therefore easily happen,

106

Giselinde Kuipers

for instance in cross-cultural encounters, or in less clearly demarcated settings such as street theatre. Some humorous genres also make use of the fact that humorous specialists may not be recognizable to everyone. Examples are candid camera shows, or the comic oeuvre of comedian/film maker Sacha Baron Cohen (aka Borat, Ali G, Bruno). However, most humor is produced by non-specialists, in everyday interactions. Everyday communication is organized according to a complicated set of roles, expectations and conventions. As ample research, in particular in conversation analysis and sociolinguistics, has shown, the success of any attempt at humor strongly depends on who is attempting to be funny, and who is supposed to be amused (Hay 2001; Kotthoff 2006; Bell 2015). Many of the factors involved in this process are beyond KaH as I define it here. As Bell shows, humor success and failure strongly depend on relations between joker and audience (eg. power, levels of familiarity and intimacy), and on purely situational aspects ranging from group size to general mood. Such dynamics of relations and situations are always delineated by cultural understandings of appropriate or inappropriate humor. These often have to do with the “who” of humor. Thus, in the US and in most European countries, there is widely held normative understanding that humor should “punch up” rather than “punch down”. In other words: it is better to joke at those in power than those who have a weaker social position. A related, commonly shared cultural rule across contemporary societies is that laughing at oneself is better than laughing at others. Both rules define “who” can joke “with whom”. Such understandings about humor are not universal. Even within European societies, there are clear exceptions to this general rule. For instance, canned jokes have a marked tendency to mock the less powerful, as Christie Davies (1990; 2002) has shown. Jokes in contemporary societies only rarely target those in power. Self-mocking jokes, while universally praised, are actually rather rare. In premodern and early European societies, the notion that humor should not go “downward” seems quite absent. In a recent survey of the historical development of Schadenfreude (Kuipers 2014) I encountered some rather dramatic historical examples of humor at the expense of others that was relished and certainly not sanctioned. For instance, joking and mockery was a common element of punishment of criminals. Moreover, the famous Roman gladiator fights, which included the killing of countless humans and animals, were often described by contemporaries as fun events, involved humor and mockery of (wounded and dying) losers. Even today, the cultural convention sanctioning joking down is a complicated and somewhat contradictory norm, with many exceptions, clauses and subclauses (cf. Kuipers 2015, Chapter 7). A particular problem is that it is actually not that easy to decide what is “up” or “down” in

Knowledge about humor

107

society. For instance, in recent Dutch debates about so-called “political correctness” versus “freedom of speech” in the use of humor, both parties describe the opposing party as more powerful or “elitist”, and therefore deserving of jokes at their expense. All members of a group of culture therefore employ normative rules or guidelines that set some limits on who can joke about whom with whom. As the examples above suggest, these rules and conventions are not uncontested. Failure to comply with such moral rules regarding humor may lead to typically moral responses: indignation, offense, social exclusion and conflict. Alternatively, people may choose not to respond to humor they find offensive or transgressive (what Raskin called “volitional” aspects of the sense of humor). Such clashes over the moral boundaries of humor have endless ramicifiations, and can easily explode into full-fledged humor scandals (Kuipers 2011). In the unfolding of such controversies, much depends on further dynamics of power and solidarity, embedded in further conventions: who is able to object, and who will receive support from whom for their objections? What procedures are in place for raising objections to humor? Are there cultural repertoires in place that recognizes “offense” at a joke as legitimate, for instance because it is based in a cultural system of honor? Or is this almost inevitably seen as lacking a sense of humor? Here, knowledge about humor easily blends in with larger moral and even ideological frameworks, to the point that they become almost indistinguishable. As I noted earlier, the failure and success of conversational humor is strongly shaped by power dynamics. Such power dynamics within conversations are shaped by wider societal relations. One can see this, for instance, in gendered patterns in humor use and humor production. Women and men tend to use humor differently, and men tend to be more successful in using humor, partly as the result of power differentials (Kotthoff 2006). Such patterns have of course drawn societal attention, and they have led to a specific form of knowledge about humor: beliefs or stereotypes about the humor of a specific group. Thus, women are often said to have “no sense of humor” or “less of a sense of humor than men”. Similar patterns can be noted in the perceived lack of humor of other minority or underprivileged groups. This may feed into a lowering of expectations regarding these sense of humor not only of these groups as a whole, but also individual members of this group. Such beliefs and stereotypes become a self-fulfilling prophecy. The simple fact that a group is believed to have no sense of humor may dissuade members from trying. The cliché also increases the risk of humor failure because people will not expect humor from specific persons. This sets into motion a self-reinforcing cycle. The mirror image of such “humorless groups” are the groups that are widely expected to have a good sense of humor. The classic example of course,

108

Giselinde Kuipers

are Jews. There are other examples, ranging from travelling salesmen in the Netherlands to Neopolitans in Italy. Here too, there is a wide-spread belief that this particular group has a good sense of humor. Note that these groups typically are also not characterized by high social status, although their status is more ambiguous than the “humorless groups”. However, in these “high humor” groups, group members make little attempt to contest this stereotype. But this particular form of knowledge about (Jewish, salesmen, Neapolitan) humor functions in the same way. It starts as cultural assumption, which affects (and maybe reflects) humorous practices, leading to the belief becoming more and more real. This, stereotyped knowledge about the “who” of humor shapes and reinforces pattern of humor failure and success. At the same time, it reflects and affects wider dynamics of social status and social inclusion and exclusion.

4 Conclusion: Dissecting an everyday miracle (and then teaching it to computers?) Successful humor is an everyday miracle. Many times a day, people manage to perform this stunning feat of sophisticated communication, without any visible effort. Rarely do they realize how much linguistic, cultural and social knowledge is required to pull off this feat. Only when it fails, people may realize how difficult it is. Humor failure, while painful and unpleasant to most, is therefore a pleasure and a treat to humor scholars: it is a pathway to understanding the mechanisms of humor. In this article, I have argued that humor failure and success depend on knowledge about humor: culture-specific understandings of when, how and with whom to joke. This overview was not meant to be exhaustive. I am quite sure other interrogative words would lead to more insights about, for instance, the what, why and where of humor. My aim was to show the relevance of the notion of KaH as additional knowledge resource. While it may not be essential to generating a “single joke-carrying text”, knowledge about humor plays an important role in making such a single joke-carrying text work in actual social situations. Even the most perfect joke can fall flat when the “how, when and who” (etc.) of humor are not conforming to cultural conventions, rules and practices. This knowledge about humor is what sociologists refer to as “practical knowledge”: people are usually not aware of it, and certainly do not think of it as knowledge, because it manifests itself as semi-automatic behavioral patterns that feel

Knowledge about humor

109

“natural”. Yet, it is knowledge: learned, group-specific information that has clear regularities. These regularities are not commonly understood as rules, However, when they are broken or violated people are aware that something is wrong. In such cases, social responses may range from puzzlement and mild embarrassment, to annoyance, lack of communicative ease and consequent avoidance, to strong emotional responses: offense, moral outrage, social exclusion and conflict. I offer “Knowledge about humor” as a sociological expansion to scriptbased approaches to humor. Linguistics– from semantics to pragmatics, from computational approaches to conversation analysis –so far has been most successful in developing theories for understanding how humor succeeds and fails. In many linguistic humor theories, but in particular semantic approaches, humor is a cognitive process and therefore primarily a matter of knowledge. While the six original knowledge resources from the GTVH may be sufficient to establish whether an utterance is a joke, they do not suffice when trying to predict or explain joke failure or success (Bell 2015). Thus, while KaH may not be essential to the production of a “single joke-carrying text”, it is crucial in making this joke work in actual social situations. This suggestion to introduce a knowledge resource intended to capture cultural variations in the use of humor echoes recent work by Villy Tsakona (eg., Tsakona 2017), who proposes to expand script-based humor theories with a context (CO) knowledge resource. Tsakona’s CO-KR encompasses both “sociocultural presuppositions” required for production and interpretation of jokes, and “metapragmatic stereotypes” on humor. KaH is both more and less specific than CO-KR. First, KaH does not include sociocultural presuppositions, a rather large category referring to any knowledge a member of a culture may have. Of course, such general cultural knowledge is essential to the production and understanding of humor (cf. Chiaro 2010). However, I find it helpful to make an analytical distinction between cultural understandings and assumptions specifically related to humor, and wider cultural presuppositions. KaH is close to Tsakona’s second notion of “metapragmatic stereotypes”, but here KaH has a broader scope that KR-CO. Tsakona defines these stereotypes as “ideological assumptions and stances on whether a specific text can be considered humorous or not, why, how, when and to whom”. As I have attempted to show in this short article, KaH is certainly not exclusively ideological. Some of the cultural conventions regarding how, when and to whom are certainly ideological, such as conventions about joking up or joking down. Other humor rules, such as those related to occasions where joking is mandatory, and jokes at one’s expense should be accepted with good humor, are certainly morally charged, as evidenced by the strong responses when they are

110

Giselinde Kuipers

not observed. However, they are not necessarily embedded in a wider ideological framework. Finally, some forms of KaH are neither ideological nor moral – think of the conventions for signaling a joke, or cultural preferences for humorous communication styles. The notion of practical knowledge that I employ here encompasses ideological, moral and conventional knowledge: all three are self-evident, automated forms of learned behavior, with a rule-like (though contested and contestable) character. Thus, they can be described as “normative” only in the very general sense of “establishing norms”. Such norms, however, do no have to more morally charged, and they certainly are not held together by ideological frameworks (although ideologies may be used to legitimate them). This variation in types of norms is also reflected in responses to humor failure related to KaH. Ideological and moral norms lead to emotional responses, social conflict, exclusion, but also to refusal to acknowledge humor. This is an interesting question for further research: are differences in responses to humor failure related to ideology vs morality? On the other hand, failure to comply with mildly moral or non-moral norms is more likely to lead to puzzlement that to emotional reponses such as offense or indignation. This raises the question whether there is a wider framework that connects the various bits of humor knowledge of well-socialized member of a culture. How systematic is “knowledge about humor”? Is it mostly a rather random collection of cultural beliefs, norms and practices, or are they connected through some wider logic? If so, what sort of logic would this be? Earlier discussions of this question – including my own – have suggested that coherence within “joking cultures” (Fine and DeSoucey 2005) is mostly based in morality and group regulation. Thus, Tsakona and Kramer (2011) speak of “humor ideologies” and their relation to “metapragmatics”. In my previous work on humor scandals, I focused on “normative communities of humor”, later renamed “humor regimes” (Kuipers 2011). However, the cognitive approach taken here suggests that KaH may not be rooted in normative, but in cognitive frameworks. These could be conceptualized as “folk theories of humor”: cultural beliefs about what humor is, that lead to culture-specific knowledge about humor. For instance, do people believe humor to an individual characteristic or a form of communication? What class of objects do people believe it belongs to? Is humor a duty? A virtue? A form of subversion? An emotion? A mental capacity? What is it opposed to? To reason? To seriousness? To anger? One could hypothesize that such folk theories are at the heart of general understandings and beliefs regarding humor, which in turn inform everyday rules and conventions. Thus, an understanding of humor as a virtue (which seems to be dominant in the contemporary US) lead to a very different understanding of everyday humorous practices and expression than an

Knowledge about humor

111

understanding of humor as emotion, which has been a dominant understanding of humor in much of Europe from the Middle Ages until Freud (cf. Morreall 1987; Kuipers 2015). The even older conceptualization of humor as bodily fluid, still alive in the contemporary word but otherwise difficult to relate to for 21st century people, obviously led to yet another set of understandings and conventions. This conceptualization fueled the belief that humor and laughter are healthy – a form of “knowledge about humor” now scientifically proven to be quite wrong (cf. Martin and Ford 2018), but still widely believed. Such musings are theoretically interesting, but also have practical applications. Since the introduction of the first script-based theory of humor, both the inventor of the theory, and the field of linguistics have moved towards computational humor. This is mostly a very theoretical field, dedicated to producing ontologies and formal theories. However, the more appealing part of this field, from a perspective of a relative layperson, is the attempt to make computers recognize, understand, and potentially even generate humor. One can easily imagine that computers will at some point become quite adept at mastering knowledge resources such as logical mechanisms, targets or even language. But the real test will eventually be the computer’s skill not in using humor, but in using it right: the right joke, at the right time, to the right person. Without KaH, no automated system will even be able to pass this humorous equivalent of a Turing Test.

References Apte, Mahadev. 1985. Humor and Laughter: An Anthropological Approach. Cornell: Cornell University Press. Attardo, Salvatore. 2001. Humorous Texts: A Semantic and Pragmatic Analysis. Berlin: Walter de Gruyter. Attardo, Salvatore and Victor Raskin. 1991. Script theory revis(it)ed: Joke similarity and joke representation model. HUMOR. International Journal of Humor Research 4(3). 293–348. Bell, Nancy. 2015. We Are Not Amused: Failed Humor in Interaction. Berlin/Boston: Walter de Gruyter. Bell, Nancy and Salvatore Attardo. 2010. Failed humor: Issues in non–native speakers’ appreciation and understanding of humor. Intercultural Pragmatics 7(3). 423–447. Billig, Michael. 2005. Laughter and Ridicule: Towards a Social Critique of Humour. London: Sage. Bourdieu, Pierre. 1991. Language and Symbolic Power. Cambridge: Harvard University Press. Bourdieu, Pierre 1998. Practical Reason: On the Theory of Action. Stanford: Stanford University Press. Chiaro, Delia (ed). 2010. Translation and Humour. London: Bloomsbury Publishing. Coser, Rose Laub. 1960. Laughter among colleagues. Psychiatry 23(1). 81–95.

112

Giselinde Kuipers

Davies, Christie. 1990. Ethnic Humor Around the World: A Comparative Analysis. Bloomington: Indiana University Press. Davies, Christie. 2002. The Mirth of Nations. New Brunswick: Transaction Publishers. Davies, Christie. 2011. Logical mechanisms: A critique. HUMOR: International Journal of Humor Research 24(2). 159–165. Fine, Gary Alan. 1979. Small groups and culture creation: The idioculture of little league baseball teams. American Sociological Review 44(5). 733–745. Fine, Gary Alan, and Michaela DeSoucey. 2005. Joking cultures: Humor themes as social regulation in group life. HUMOR. International Journal of Humor Research 18(1). 1–22. Hale, Adrian. 2018. “I get it but it’s just not funny”. Why humour fails, after all is said and done. European Journal of Humour Research 6(1). 36–61. Hall, Edward T. 1976. Beyond Culture. New York: Anchor. Hay, Jennifer. 2001. The pragmatics of humor support. HUMOR International Journal of Humor Research 14(1). 55–82. Hempelmann, Christian and Salvatore Attardo 2011. Resolutions and their incongruities: Further thoughts on logical mechanisms. HUMOR International Journal of Humor Research 24(2). 125–149. Hempelmann, Christian and Willibald Ruch. 2005. 3 WD meets GTVH: Breaking the ground for interdisciplinary humor research. HUMOR International Journal of Humor Research 18(4). 353–387. Kramer, Elise. 2011. The playful is political: The metapragmatics of internet rape–joke arguments. Language in Society 40(2). 137–168. Kuipers, Giselinde. 2009. Humor styles and symbolic boundaries. Journal of Literary Theory 3(2). 219–239. Kuipers, Giselinde. 2011. The politics of humour in the public sphere: Cartoons, power and modernity in the first transnational humour scandal. European Journal of Cultural Studies 14(1). 63–80. Kuipers, Giselinde. 2014. Schadenfreude and social life: a comparative perspective on the expression and regulation of mirth at the expense of others. In van Dijk, Wilco and Jaap Ouwerkerk (eds). Schadenfreude. Understanding Pleasure at the Misfortune of Others. Cambridge University Press: Cambridge, 259–294. Kuipers, Giselinde. 2015. Good Humor, Bad Taste. A Sociology of the Joke. Berlin/Boston: De Gruyter. Kotthoff, Helga. 2006. Gender and humor: The state of the art. Journal of Pragmatics 38(1). 4–25. Lizardo, Omar and Michael Strand. 2010. Skills, toolkits, contexts and institutions: Clarifying the relationship between different approaches to cognition in cultural sociology. Poetics 38(2). 205–228. Marsh, Moira. 2015. Practically Joking. Boulder: University Press of Colorado. Martin, Rod and Thomas Ford. 2018. The Psychology of Humor: An Integrative Approach. London (etc): Academic Press. Morreall, John. 1987. The Philosophy of Laughter and Humor. New York: SUNY Press. Oring, Elliott. 1994. Humor and the suppression of sentiment. HUMOR: International Journal of Humor Research 7(1). 7–26 Oring, Elliott. 2011. Still further thoughts on logical mechanisms: A response to Christian F. Hempelmann and Salvatore Attardo. HUMOR: International Journal of Humor Research 24(2). 151–158

Knowledge about humor

113

Radcliffe-Brown, Alfred. 1940. On joking relationships. Africa 13(3). 195–210. Raskin, Victor. 1985. Semantic Mechanisms of Humor. Dordrecht: Reidel Publishing Company. Raskin, Victor. 1998. The sense of humor and the truth. In Ruch, Willibald (ed.), The Sense of Humor. Explorations of a Personality Characteristic, pp. 95–108. Berlin: Mouton De Gruyter. Raskin, Victor, Christian Hempelmann and Julia M. Taylor. 2009. How to understand and assess a theory: The evolution of the SSTH into the GTVH and now into the OSTH. Journal of Literary Theory 3(2). 285–311. Rayz, Julia. 2017. Computational humour and Christie Davies’s basis for joke comparison. European Journal of Humour Research 5(4). 169–178. Ruch, Willibald. 1998. The Sense of Humor. Explorations of a Personality Characteristic. Berlin: Mouton De Gruyter. Shively, Rachel. 2013. Learning to be funny in Spanish during study abroad: L2 humor development. The Modern Language Journal 97(4). 930–946. Smith, Moira. [Marsh] 2009. Humor, unlaughter, and boundary maintenance. Journal of American Folklore 144(484). 148–171. Taylor, Julia and Victor Raskin. 2012. On the transdisciplinary field of humor research. Journal of Integrated Design and Process Science 16(3). 133–148. Tsakona, Villy. (2017). Humor research and humor reception: Far away, so close. In Chłopicki, Władysław and Dorota Brzozowska, Humorous Discourse, 179–203. Berlin: De Gruyter.

Willibald Ruch

Domains of humor: Challenges from psychology Abstract: This chapter discusses two topics about which Victor Raskin and I have a history of debate and exchange. Given the nature of our disciplines we do have different assumption to start from and therefore also different outcomes. While some disciplines focus on analyzing the humor artifacts from their discipline and describe an idealized model, the nature of psychology (and in particular personality psychology) requires considering a more complex compound of “humor, researcher, recipient” to be analyzed to also consider and describe different human functioning. We share an interest in the issue of classification of jokes and cartoons and co-organized a symposium on the topic and presented our approaches (General Theory of Verbal Humor, 3WD) and further on cooperated on an empirical verification of the GTVH (Ruch et al. 1993) and whether or not it helps describing and further understanding the 3WD humor factors (Hempelmann and Ruch 2005). Likewise, his theory of the sense of humor (Raskin 1998) provides a conceptual basis for empirically testing hypotheses on the sense of humor. An index representing his criteria may serve as a marker, also when evaluating the relative importance of the newly found six factors of humor. Linguistics and psychology may profit from each other but it is also clear that the realm of the disciplines are different and linguistics may assume an idealized model while psychology needs to take the human factor, in all its diversity, into account. Keywords: humor appreciation, sense of humor, cheerfulness, seriousness, gelotophobia

1 Introduction Good things often come at once. After my slightly disappointing first humor conferences (Tel Aviv in 1984, Cork in 1985) I stopped going to them for a few years until Sheffield 1990 where I attended (more for private reasons) and organized my first symposium (Innovations in psychological humor research) at a ISHS conference. Before this I moaned in a bold letter to a Professor Raskin that

Willibald Ruch, University of Zürich https://doi.org/10.1515/9781501511707-007

116

Willibald Ruch

I did not hear back about my submission to HUMOR for a way too long time. At the Sheffield conference Victor attended my symposium and later took me aside and told me that the article had been reviewed meanwhile and got accepted for publication in HUMOR. He also invited me to the board of editorial consultants of HUMOR and to be a co-editor of the book series “Humor Research.” The symposium plus additional articles were turned into a special issue (published in 1993) and as Don Nilsen invited me to serve in the executive board of ISHS I can assert that this conference certainly constituted a good second start for me and since then I have organized a symposium at almost every conference. Most importantly, discussions with Victor started that were then continued at the following conferences in hotel rooms and during my visits at Purdue University and his home. At this time, plans about future research and publication projects developed and I got to know his best students at that time, among them Salvatore Attardo and Amy T. Carrell. Victor and I were interested in similar topics but approached them with our own and different disciplinary backgrounds. For example, when writing about “humor and personality” (Ruch 1992) for “The world and I” these differences became apparent. I created my 3WD (3 Witz-Dimensionen) humor test of personality to measure funniness and aversiveness of incongruity-resolution (INC-RES), nonsense (NON) and sexual (SEX) humor. In INC-RES jokes and cartoons first an incongruity is detected that is then resolved completely using information found elsewhere in the joke; that is, the recipient has “gotten the point”. It seems that resolving the incongruity is the key element and this led to the idea that reducing uncertainty (i.e., preferring redundancy) is the rewarding element in this form of humor. Nonsense humor also has a surprising or incongruous punch line, but “ . . . the punch line may 1) provide no resolution at all, 2) provide a partial resolution (leaving an essential part of the incongruity unresolved), or 3) actually create new absurdities or incongruities” (McGhee et al. 1990: 124). Thus, in nonsense humor the resolution information gives the appearance of making sense out of incongruities without actually doing so. People who successfully process nonsense humor know that they have “gotten” what there is to get. Enjoyment results from playing with absurd ideas, the contrast of sense and nonsense; it is not that they enjoy what they did not understand. The common element in this form of humor is that people apparently enjoy the higher degree of both incongruity and residual incongruity (compared to INC-RES). This led to the hypothesis that nonsense humor is an indicator of generally liking uncertainty (in the information theory sense) as it can be found in asymmetry, complexity, incongruity, or novelty. Both the incongruity-resolution and the nonsense structure can be the basis for harmless as well as tendentious content (e.g., sexual humor). (For the development and validation of

Domains of humor: Challenges from psychology

117

this two-mode model of humor appreciation see, for example, Ruch 1992; Ruch and Hehl 1998). The different forms of the 3WD use 10 or 15 jokes and cartoons for each of the categories. Averaging across 10 items made the test reliable, which enhanced the chance for better validity coefficients in personality studies (see Platt and Ruch 2014). This is common practice in test construction to ensure higher psychometric properties. The big surprise came when Victor and I organized a symposium on taxonomic issues in humor research for the 1991 ISHS conference at Brock University, St. Catharines, Ontario, Canada and Victor and Salvatore presented their General Theory of Verbal Humor (GTVH) and boldly suggested to classify jokes with an N = 1 in the cells. Of course, each pair of jokes is different and so a linguistic analysis should reveal this. When laypeople process jokes, subtle differences (e.g., slight variations in some of the six knowledge resources) do not matter and lead to comparable responses. For me, jokes needed to fall onto an entirely different factor to be considered different enough. I remember that it took the three of us at least one afternoon at Brock University to understand each other’s position, and to discuss the basic misunderstandings about each other’s disciplines before we could come to constructive work and publish a pioneering first interdisciplinary study (Ruch et al. 1993). To me the collaboration with linguists was inspiring, and it helped me to see the strengths and limitations of the different disciplines involved in humor research including my own. Clearly, I did profit from further exchange with linguistics, be it the GTVH analysis of the 3WD jokes (Hempelmann and Ruch 2005) or a defense of why I considered Gary Larson cartoons as good markers for nonsense humor (Ruch 1999). It became apparent that while a linguistic theory will be very well suited to describe peculiarities of jokes in comparison to other texts (and then by definition no two jokes are identical), the challenge from psychology is that individuals do not process all the differences that there are and they don’t matter to them; that is, the response to two “different” jokes can be highly interchangeable. This led to the suggestion of distinguishing at least three ways of constructing a classification of humor at a more fundamental level (Ruch 1999). In one approach we only consider “humor” (Model A), the second approach acknowledges the existence of both “humor” and “researcher” (Model B), and in the third approach “humor,” “researcher,” and “recipients” are considered (Model C). Model A assumes “ . . . that jokes/cartoons are different and the intrinsic structure only needs to be identified by somebody who takes the effort to do so (plus perhaps has the skills which some academic discipline would provide)” (Ruch 1999:72). Most saliently, the role of the researcher is not an explicit factor in this approach, as it is not considered to determine the outcome and typically

118

Willibald Ruch

there is no need to have second person undertaking the same steps. Limitations emerge when more researchers indeed aim for the same goal and disagree, as this leads to the question of who is right and who is wrong, rather than acknowledging that researchers as persons are an integral part of the research endeavor and that different taxonomies can coexist. Model B takes into account that “ . . . different researchers have different backgrounds, different aims, and apply different tools, and thus also the resulting taxonomy will most likely be different” (Ruch 1999:73). Depending on one’s background a researcher might want to cluster for joke features such as structural properties, repressed needs, historical or regional origin etc. A taxonomy that was generated for one purpose will not necessarily suit for other aims. Therefore one can expect that there will not be a universal taxonomy that fits the needs of all researchers. Thus, a classification of jokes accomplished by psychoanalysts will not get linguists excited and vice versa. Only Model C takes the naive recipient of humor into account. Here the importance of the joke’s core ingredients (i.e., factors that make jokes similar or different) is not primarily determined by the researcher’s direct analysis of the joke but would also incorporate the perception of the respondents. Here we do not analyze jokes, or jokes as examined by the expert, but jokes as perceived by the layperson. We would place two jokes into the same category (or consider them to be similar) if they are perceived to be similar in a direct comparison, or if they independently from each other elicit comparable responses in the receiver. These two jokes might be different on many dimensions identified a priori by experts; if so, then apparently that identified dimension does not affect the naive person’s response very much (Ruch 1998:73).

The work on the 3WD humor test (as well as other factor analytic approaches) is an example for Model C. This approach expects that if variations in a postulated joke parameter are below the threshold for a recipient’s response (i.e., this parameter does not affect the responses), then it is not important when creating a taxonomy. For example, when jokes were pre-rated for the initial study that finally led to the 3WD, “aggression” was a category that was considered. When the results of the first analysis were inspected a cluster around aggressive was expected but did not emerge (unlike the cluster of sexual humor). Instead they were distributed across the two factors later interpreted as incongruity-resolution and nonsense humor. These two factors seem to be characterized by the common joke structure and not a common content. So how could jokes considered to be “aggressive” end up on two different factors? Obviously, the ingredient identified did not produce a lot of variance and hence jokes sharing this conceptually identified content did not correlate well enough to form a separate factor. It should be mentioned that also in model C the researcher is

Domains of humor: Challenges from psychology

119

essential, for example, to determine what tools or participants to chose, etc. These choices to be made will determine the outcome so therefore a replication of the taxonomy (exchanging items and participants), ideally in different countries, is essential before finally arriving at a stable classification. Here the differences in the approaches are most apparent. The work of Victor Raskin and his group (Attardo and Raskin 1991; Raskin 1985; Raskin et al. 2009) allows for a more fine-grained description of jokes that may place each joke as a single node in a complex net. For the GTVH there are six knowledge resources (Script Opposition, Logical Mechanism, Target, Situation, Narrative Strategy, Language) – each with separate parameters – that describe jokes. The jokes and cartoons of the 3WD were selected to be pure markers of the factors. There is no intent to describe the differences among the jokes of a factor, rather the scores are averaged across all jokes. Hence the information obtained is rather basic and reduced to the essential. But what is it? Hempelmann and Ruch (2005) provided a GTVH analysis of the 3WD jokes and cartoons and it turns out that no single parameter distinguishes perfectly among the three humor categories, but several of them do contribute to the characterization of the humor types arrived at by factor analysis. Above all, it is the degree of residual incongruity that distinguishes among the humor types. NON (and NON-based sexual humor) has the highest amount of residual incongruity, followed by INC-RES and SEX. Initial incongruity is a discriminating feature too, with NON exceeding INC-RES and SEX, primarily because there is a higher frequency of possible/ impossible SOs and a lesser number of actual/ non-actual script oppositions in NON” (Hempelmann & Ruch 2005:379–381).

There are fewer targets for NON than for incongruity-resolution jokes. Thus, the knowledge resources were not distinguishing well, as the strongest differences were found for the derived index of degree of incongruity and degree of residual incongruity, one element that came from the 3WD theory and which was backed up by linguistic analysis. Thus, there is some overlap between the two approaches but clearly the tools provided by linguistics exceed the diversity that the layperson is able to reflect in their responses. The three factors of incongruity-resolution, nonsense and sexual humor were replicated in several countries. However, this was always done with the help of the original 3WD items and never supplemented by “local” humor to really test its comprehensiveness. Research also found a convergence between the funniness ratings in the 3WD and films (Ruch and Hehl 1998) and TV ads (Hofmann and Ruch 2017) selected to represent these three categories as well as questionnaire measures of nonsense (Heintz in press). That is, someone who finds nonsense funny in cartoons also finds nonsense funny in films, TV ads,

120

Willibald Ruch

and reports nonsense humor to be a typical humor style in self-reports. This suggests that they measure the same construct although of course method variance exists. There is also evidence for hereditary and environmental influences (Weber et al. 2014) and neurophysiological correlates of INC-RES and NON (Samson et al., 2009). Studies of convergent and discriminant validity are very important to establish construct validity and should be undertaken for any instrument before it is recommended for use (Heintz 2017a). Similar tests were constructed that measure the 3WD or similar constructs (e.g., Carretero-Dios et al. 2010; Sulejmanov et al. 2018). Funniness of the humor factors were repeatedly linked with individual differences in other domains, such as personality (in self and peer reports), performance tasks, or intelligence tests, and the two structures clearly yielded different correlational patterns (e.g., Hehl and Ruch 1985; Sulejmanov 2018). Figures 1a and 1b summarize what was found in different studies. Recent developments. While the GTVH allows for a multimodal classification of jokes (i.e., the six knowledge resources), and most classifications are unimodal (e.g., sexual, aggressive, scatological humor), the 3WD favors a bimodal classification (i.e., considers structure and content) but this is not fully developed as noted by Ruch and Platt (2012). More by coincidence it was discovered that sexual humor had a second loading on either INC-RES or NON, depending on the structure of the joke or cartoon. This allows stipulating that once structure is controlled for maybe more content categories might be identified. Ruch and Platt (2012) argue that more such content categories should be utilized with jokes and cartoon with INC-RES and NON structure. They argue that . . . . . . to verify this model the first step should be a theoretical analysis of thematic and schematic properties of the pool of humor items to be taxonomized. In the second empirical step, jokes and cartoons of the different categories is given to a large sample that rates them for funniness. In the analysis, different structural models should be tested against each other and the one with the best fit should be retained. For example, one model might represent a unimodal taxonomy of jokes according to their content; another unimodal model might represent structural factors only. These and other models might be tested against a bimodal model that simultaneously specifies one content and one structure loading for each joke. It is expected that a bimodal model will yield the best fit. The empirically derived weights then can tell how important the postulated structural and thematic properties are for a given joke or cartoon (or clusters of jokes/cartoons) (Ruch and Platt 2012:24).

Heintz (2019) conducted a first study to empirically investigate the extent to which a bimodal model of humor appreciation yields empirical support by comparing structure-only, content-only, and bimodal structural equation models of a set of humor stimuli that capture two structures (INC-RES and NON) and two

Domains of humor: Challenges from psychology

Incongruity-Resolution Humor (INC-RES)

Low appreciation characterized by

High appreciation characterized by

Conservative Attitudes intolerance of minorities, militarism, religious conservative fundamentalism, education, traditional family ideology, capitalistic attitudes, property/money, law and order attitude, punitiveness, conventional values

liberal/ radical

General Inhibitedness disinhibited

inhibited

superego strength, inhibition of aggression, selfcontrol, rigidity, need for order, antihedonistic, sexually not permissive, low Psychoticism, agreeableness

Uncertainty Avoidance low

intolerance of ambiguity, avoiding new & complex experience, low intensity (AISS), prefers simplicity & symmetry, conventional vocational interests, liking of simple, non fantastic art, likes low grotesque texts, likes simple line drawings (BWAS), likes simple music (Pop, Country)

high

Depressivity depressed

depressiveness

not depressed

Social Desirability frank

social desirability, “lying”, low frankness

acts socially desirable

Age younger

older

Figure 1: Summary of findings on personality correlates of incongruity-resolution (1a) and nonsense humor (1b) clustered by domain (above the line) and the individual variables studied (below the line).

121

122

Willibald Ruch

Nonsense Humor (NON)

Low appreciation characterized by

High appreciation characterized by

Openness to Experience avoids new experience

openness to values, ideas, aesthetics, fantasy, mental experience seeking, seeks new experiences, avoids repetition, novelty (AISS), grotesque texts, interest in plastic arts, sculpture, imaginative

seeks out experience

Complexity prefers simplicity

likes complex fantastic art paintings, likes complexity in line drawings (BWAS), produces complexity in black/white pattern, enhances visual incongruity (”prism glasses”)

prefers complexity

Intelligence low

“fluid” intelligence”, speed of closure

high

Sexual Libido weak

high sexual experience, pleasure, libido, activity, desire

strong

Nonconformism conventional

not obedient, low social desirability, “lying”, frank

non-conform

Age older Figure 1 (continued)

younger

Domains of humor: Challenges from psychology

123

contents (sex and aggression). As principal component and exploratory factor analyses aim at simple structure and thus do not allow one item to have several high loadings, structural equation modeling is needed to adequately test such a model. Specifically, one humor stimulus can be modeled to load on more than one latent factor, in this case one content factor and one structure factor, to consider both sources of variance simultaneously. The results suggest that a bimodal model is superior to a four-factor model (which represented the four content-structure combinations, INCR-RES sex, INC-RES aggression, NON sex, and NON aggression) and to unimodal models, in which only the two contents or the two structures were included. Heintz (2019) also found that funniness was mostly influenced by the structures, while aversiveness was mostly influenced by content. The correlations with the regular 3WD scales supported the convergent and discriminant validity. This study suggests that bimodal models of humor appreciation more adequately represent and assess individual differences in humor appreciation. This progress is promising and will ultimately lead to a robust classification of jokes and cartoons, which allows distinguishing more elements than were identified in the 3WD model created almost 40 years ago. To be able to accumulate research findings a classification is needed as a reference rather than developing new categories that are often only used in one or two studies. Hence taxonomic studies are needed with modern and independent material from all over the world. The parameters of the GTVH (or other linguistic models) might be of interest in this process. Here a different type of work is needed to be useful for psychology. Studies need to determine how much variance the systematic variation of the parameters in each of the six knowledge resources accounts for. What knowledge resources matter to what degree? It should be possible to derive empirical weights for, for example, substituting the target of a joke. It might matter when it is the own reference group (or the opponents group), but probably not matter at all when the targets are from two unrelated reference groups. Even if it is the own group, how much does it matter compared to variations in script opposition or in the logical mechanism? Are there interactions between knowledge resources; for example, the variations in script opposition might depend on the type of target. Of course, the resulting text will always be different, but what would the rated funniness be? Informal observations when talking with my colleagues from linguistics taught me that they consider their own judgment as sufficient. So here we are in a Model A setting. Accounting for differences among researchers implies Model B. Deriving weights for postulated ingredients through empirical research is necessary as it would allow more definite statements about their relative relevance; for example, about the knowledge resources. Experimental philosophy is not an oxymoron anymore as philosophers started to conduct empirical studies. For

124

Willibald Ruch

interdisciplinary humor research an empirical underpinning or extensions of linguistic models seems equally desirable.

2 The sense of humor My first decade of research was mostly about jokes and cartoons and I did not consider questionnaire approaches to humor fruitful, as you cannot measure your IQ with a rating scale. In the 80s a trend started that seemed to favor questionnaire measures over humor tests. Clearly, a lot of humor emerges in everyday life and studying the responses to a joke test will not allow predicting whether someone, for example, will like to entertain others through humor and how well he or she will be. My own studies led me to study emotional responses to humor. We looked at the face and tried to explain individual differences in propensity to smile and laugh through what we called trait cheerfulness (seriousness and bad mood). We tried to avoid “sense of humor” as a concept as it seemed to be more of a folk concept rather than a scientific construct. This started the second domain of exchange with Victor. I organized a symposium at the ISHS conference in Ithaca, NY in 1994 on the sense of humor and this lead to a special issue of the measurement of the sense of humor in 1996 and later I edited my sense of humor book (Ruch 1998) and asked for a contribution from Victor. The State-Trait Cheerfulness Inventory (STCI) proved to be a good predictor of facial responses to humor (e.g., Hofmann et al. 2015), liking of nonsense, for humor creation and a variety of other humor-related behaviors (Ruch and Hofmann 2012; Ruch and Köhler 1998). But it does not cover all aspects of humor. So starting with Ruch (1994) we aimed at looking at the intercorrelations among all available humor scales at that time. Indeed cheerfulness clearly marked the major factor extracted from the humor scales but there was the need for a second factor as well, which was marked by low seriousness. While it was soon clear that the sense of humor is not unidimensional it remained unclear how many dimensions are needed to account for all individual differences in humor, and whether or not all should be covered under the umbrella of “sense of humor” or whether this refers to a subset of humor-related traits. More than 30 years have elapsed since the renewed interest in the sense of humor in the 80s but no comprehensive validated model exists. The analysis of the existing scales is an imperfect way to accomplish this as the selection is biased. Then how to build such a model? Ideally, the model should entail a rich description of the field of humor behavior but also allow for a parsimonious

Domains of humor: Challenges from psychology

125

representation of the variety. This way all the important indicators of humor may be represented but there is also the possibility to build a structural model and describe humor through dimensions aggregating behaviors. The former approach is important to be able to distinguish individuals at the level of humor habits and one can assume that no or few individuals will share the same habits. At this level also the training of humor will be most practicable as the habit to be learned, modified or replaced can be described or displayed most clearly. The most developed approach to this is the list of statements about everyday humorous conduct (Craik et al. 1996). These authors gathered 100 non-redundant statements (e.g., Displays a quick wit and ready repartee; Maintains group morale through humor) derived both from a survey of the theoretical and empirical psychological research literature on humor and from observations of everyday social life. They devised the Humorous Behavior Q-sort Deck (HBQD; Craik et al. 1993), which ranks theses statements into nine categories (very uncharacteristic to very characteristic) according to a pre-specified normal distribution. A second more recent approach compiles a list of 45 daily humor behaviors (e.g., Made someone laugh, Laughed at someone), which were defined as “ . . . a single concrete behavior that involved humor and/or laughter, and that can be performed by the participants” (Heintz 2017b). These behaviors were derived from three instruments including the HBQD and participants indicated the frequency with which they showed each of these humor behaviors in one day on a five-point scale, rating from 1 (Not at all, zero times) to 5 (very often, 11+ times). Both the lists of the 100 statements on everyday humorous conduct and the 45 daily humor behaviors exhibit a peculiarity typically observed in personality research, namely that clusters of correlated items emerge. That is, for each element of a cluster it applies that if someone finds one of the statements to be very characteristic (or the humor behavior is shown very frequently), the other elements of the same cluster also tend to be considered more characteristic (or shown more frequently), and vice versa; if the one item in the cluster yielded a low score (was very uncharacteristic, shown infrequently), the others in the same cluster tended to yield low scores. The covariance among the items are used to define “factors” (latent constructs) that are supposed to contain the essence of the clusters. In research of personality and intelligence hierarchical models have proven to be useful. Such a model would distinguish many narrow lower-order factors, which may be combined to form fewer, broader higher-order factors, which, in turn, might be combined to form one very broad general factor on the top. This postulated g-factor of humor should not be confused with older measures that treat sense of humor as a unitary construct (e.g., the Situational Humor Response QuestionnaireSHRQ; Martin and Lefcourt 1983), as such measures typically do not cover a broad range of behaviors in one concept but only a small slice of behavior. The

126

Willibald Ruch

question then is, what would be an appropriate model for humor behaviors? The 45 daily humor behaviors can be represented in a smaller space; that is, there were seven lower-order dimensions of cheerful, witty, deriding, amused, sarcastic, self-directed, and canned (Heintz 2017b). Also Craik et al. (1996) applied factor analysis to the intercorrelation of the 100 statements and found five lower order bipolar dimension of (1) socially warm v. cold humorous styles (socially constructive v. asocial and socially distant displays of humor), (2) reflective v. boorish humorous styles (recognizing humor v. being competitive and rude in humor) (3) competent v. inept humorous styles (ability v. inability for wit and joke-telling), (4) earthy v. repressed humorous styles (enjoying v. inhibition toward jokes about taboo topics), and (5) benign v. mean-spirited humorous styles (enjoying harmless but mentally stimulating humor v. laughing at and making fun of others). Recently, Ruch and Heintz (2019) investigated this set of statements again but also added a vertical dimension to the model. They found six factors just as Craik et al. (1996) also considered four and six factors before deciding to take five. Four factors where more important (mean-spirited/earthy, entertaining, inept, and reflective/benign), and two explained slightly less variance (laughter and canned) and were kept until their nature and importance will be identified by future studies. The six factors and some marker items (partly adapted from the HBQD) are shown in Figure 2. Earthy/mean-spirited will go along with the aggressive humor style in the Humor Styles Questionnaire (HSQ; Martin et al. 2003), the dark styles (sarcasm, cynicism, satire, irony) and wit in the Comic Styles Markers (CSM; Ruch et al 2018), low seriousness in the STCI and katagelasticism in the disposition to laughter and ridicule (Ruch and Proyer 2009). Young males will have higher scores, and the personality dimensions of conscientiousness and agreeableness will be at lower levels. The high scorer in factor 2 will score higher in affiliative and self-enhancing humor (HSQ) and fun (CSM), and be a gelotophile (i.e., liking to make others laugh at him/herself). In the temperament scale they will be cheerful and lower in bad mood. Their personality will show elevated levels of extraversion and agreeableness and low levels of neuroticism. Factor 3 covers the inept (but also cold and repressed) humor style and this obviously is not inherent in any of the comic styles. The high scorer will be described by gelotophobia (i.e., the fear of being laughed at) and the self-defeating (and low affiliative) humor style of the HSQ. In the humorous temperament as measured by the STCI it will be trait bad mood and low levels of cheerfulness. These will be younger individuals who are primarily neurotic and introverted but also low in agreeableness and conscientiousness. The cognitively orientated benign/reflective humor factor will have correlations with the self-enhancing humor style of the HSQ and humor from the comic style markers. There will be a smaller

5: Laughter/expressive Laughs with whole body. Laughs at the slightest provocation. Laughs before finishing a joke. Has an infectious laugh. Has a bland sense of humor. (-) Can’t control the urge to laugh. Doesn’t remember many jokes. Smiles grudgingly. (-) Uses nudges/gestures when telling a funny anecdote. Uses funny facial expressions.

Six dimensions of humor

2: Entertaining Is a practical joker, plays the clown Maintains group morale through humor. Tells long complex anecdotes well. Displays a quick wit and ready repartee. Frequent contributor of humorous anecdotes. Tells humorous stories in dialect. Has a good sense of timing when telling jokes. Uses funny facial expressions. Varies intonation in speech to be funny. Does comic impersonations.

6: Canned (vs. spontaneous) Is more responsive to spontaneous humor than to jokes. (-) Enjoys exchanging topical jokes. Prefers jokes to recounting comic episodes from real-life. (-) limitates a professional comedian. Likes ethnic jokes. Enjoys routiness of stand-up comedians.

3: Inept/socially cold Is crushed when people don’t like their jokes. Reacts exaggeratedly to mildly humorous comments. Chuckles to flatter others. Misinterprets good-natured kidding. Responds with short-lived smiles. Explains own jokes. Makes jokes when uncomfortable. Seldom laughs at own personal failings. Smiles inappropriately. Is only humorous with close friends.

Figure 2: Marker items for the six dimensions of humor derived from the study of the HBQD statements.

4: Reflective/benign Likes intellectually word play and witticisms Uses humor to express life contradictions. Enjoys puns and punning Enjoys the humor of visual juxtapositions. Enjoys limericks and nonsense rhymes. Detached perspective on self and others. Finds humor in animal behavior. Challenges social expectations through humor. Appreciates humorous potential of persons/situations. Likes bemused reflections on self and others.

1: Mean-spirited/earthy Makes jokes about the macabre and the grotesque. Indulges in vulgar humor. Likes obscence parodies. Occasionally makes cruel jokes. Jokes about other’s imperfections. Pokes fun at the naive or unsophisticated Likes bathroom humor. Is sarcastic. Laughs “at” others. Tells bawdy stories. Needles others jokingly

Domains of humor: Challenges from psychology

127

128

Willibald Ruch

correlation with gelotophilia and none with the STCI. The prime personality correlate will be openness to experience. Laughter propensity will relate to trait cheerfulness and low seriousness. The intensity of laughter factor is expected to most strongly correlate with high trait cheerfulness and low seriousness, and younger females will be most expressive. The canned humor factor will be positively correlated with the funniness scales of the 3WD; that is, there will be an appreciation of humor incongruity-resolution and nonsense humor. Based on the results of this initial factor analysis new items were written (six for each of the six scales) to specifically measure these six factors and the hierarchical nature of humor was assessed again (see Figure 3).

1/1 First unrotated principal component .85

.64

2/1 Sense of Humor .84

2/2 Mockery / Inept Jokes .64 .83

.68

3/1 Laughter Fun .88 .57 4/1 Laughter Fun .96 .43 5/1 Laughter .99 6/1 Laughter

3/2 Fun / Cognitive Mockery .74 .77 4/2 Entertaining .72

5/2 Fun Reflective .97 .31 –.34 6/2 Fun

.88 5/3 Cognitive

.95 –.40 –.31

6/3 Cognitive

3/3 Inept Jokes .56 .90 4/3 Mockery Inept .99

4/4 Inept Jokes –.97

5/4 Mockery Inept –.85 .65

5/5 Inept Jokes –.48 .93

6/4 Mockery

6/5 Inept

6/6 Canned

Figure 3: Hierarchical factor analysis (HFA) of the 36 items measuring the six factors. (In a HFA the analysis of the items starts from the top by extracting the first unrotated principal component, then two components (that are rotated orthogonally), then three, and so forth. The factor scores extracted at each level are correlated with the factor scores of the next level to determine which factors remain stable across the different steps in the hierarchy and which ones change; e.g., by splitting into smaller components).

Domains of humor: Challenges from psychology

129

Figure 3 replicates the findings by Ruch and Heintz (2019) in different ways. First, at the sixth level the same factors appear relatively pure (as indicated by the color-coding of the items). Furthermore, the first unrotated factor has items from most (but not all) scales; in the present figure the liking of canned humor over spontaneous one is not represented at all but only emerges at step two. Third, at higher levels in the hierarchy compounds emerge, which are more complex and resemble prior conceptualizations. Finally, and most importantly, there is evidence that sense of humor seems to emerge at level two. First, in the study by Craik et al. (1996) several indicators of sense of humor correlated mostly with the socially warm and competent humor styles. These statements that were related to sense of humor also emerged as a sense of humor factor at level two in the study by Ruch and Heintz (2019), and then gradually decomposed into elements of reflective/benign humor, competent (low inept), laughter and entertainment/fun, which still correlated with the sense of humor index but to a much lower magnitude. Also in the present study the sense of humor-compound contains the cognitive/reflective element, the competent (coping with adversity) element, the enjoyment element and laughter. A correlation of benevolent humor (and laughing at yourself) and the factors were highest with this factor at level two but less so at lower levels. Thus, the folk concept of sense of humor ranges on top of the positive humor scales and is independent from canned vs. spontaneous humor and from mockery. It contains elements of enjoyment/fun, reflection, coping with adversity and laughter/positive emotion. Prior unidimensional sense of humor scales were lacking at least one of these components (e.g., the SHRQ; Martin and Lefcourt 1983, does not tap into the reflective element but highly emphasized the laughter component). For building a complete hierarchical model several further steps need to be taken. First, the bottom of the hierarchy needs further testing, regarding the number of factors although between four and six is already a fair number. Furthermore, it needs to be tested whether all existing humor scales fit into this structure, if not it might need to be expanded. For example, the recently created measures of fun, benevolent humor, wit, nonsense, corrective humor, irony, sarcasm and cynicism will neatly fit into three of the six factors (Ruch et al. 2018). The conceptualization of sense of humor by McGhee (2010) will be spread across some of the factors, mostly the fun/enjoyment factor (verbal humor, laughing at yourself), but also the reflective/cognitive (verbal humor, humor in everyday life), inept (not laughing at yourself), laughter (laughter) and canned (enjoyment of humor). For a solid measurement each of the factors will need to be composed of facets. The level above will need further study too. Repeatedly the humor scales were shown to fall into clusters of positive valence and

130

Willibald Ruch

negative valence and in the study by Ruch et al. (2018) three such factors were distinguished that were positively intercorrelated. This gave rise to the assumption of a general factor of humor. Further research will see whether this assumption is tenable and what are the best indicators for it. How does Raskin’s account (1985, 1998) of humor, the humor mode of communication, and the sense of humor fit to the above? Where are overlaps and where are the differences or even incompatibilities? An apparent difference is that the focus of theorizing is primarily on the humor item or artifact (e.g., a joke) or on the individual; that is, peculiarities of the text or the person’s stance towards the artifact (e.g., appreciation of production of a joke) are highlighted. Of course linguists refer to how a layperson will perceive or understand a joke. This is an illustration but not necessarily part of the theory, which by definition is about the text. Another difference relates to the conditions facilitating or accompanying the emergence of humor. Raskin (1985) distinguishes between the bona-fide (serious, truth-committed) mode of communication and the non-bonafide (humorous) mode of joke telling and argues that the non-humorous, serious person wants to function exclusively in the bona fide mode of communication. This is very precise and compatible with the view proposed by psychologists that see a dimension of playfulness vs. seriousness (or telic vs. paratelic state) as conducive to humor. However, Raskin (1998) goes further by stretching the continuum to even include the misoghelus’es (i.e., ‘laughter-haters’) and aghelastos’es (i.e., ‘refrainers from laughter’). He proposes: One can see here the making of a sense of humor scale with the aghelastos/misoghelus type on the left, negative pole of the scale and the humor-message supertruth people filling the middle between the poles, closer to either of them depending on the number of joke types in which they are willing to discover the exonerating elements of truth. But who occupies the positive pole of the scale, that is, who else besides this writer? The answer follows easily from the preceding discussions. Those people who do not need nor expect any “serious” message in a joke to enjoy it, who do not expect any remnant of the commitment to the truth on the part of the joke author or teller, who do not believe in the reality of purely humorous stereotypes. Such people are never offended by a Polish or Jewish joke because they do not give any credence to the necessary stereotypes, namely, that the Poles are dumb and the Jews crafty, respectively. These people can deduce a serious message from a joke but they do not have to. They possess the implausible stereotypes necessary for quite a few jokes but do not extend them back into BF (Raskin, 1998:93–94).

For Raskin there seems to be one reason for being humorless (“A humorless person is always serious and refuses to laugh” 93) but I wonder whether at the lower end there is really only one type; i.e., whether ‘laughter-haters’ and ‘refrainers from laughter’ are only one type. Refusing to laugh implies volition, but depressed people just don’t feel like laughing, they cannot even if they

Domains of humor: Challenges from psychology

131

wanted. Likewise, people that fear being laughed at (i.e., the gelotophobes; see Ruch, Hofmann, Platt, & Proyer, 2014) misperceive laughter and joy and rather show contempt (Hofmann, Platt, Ruch, & Proyer, 2015) and enjoyable emotions other than amusement also don’t lead to laughter (Platt, Hofmann, Ruch, & Proyer, 2013). Thus, not wanting to laugh or refraining form laughter and fearing (or hating) laughter will be of different quality. Finally, there are differences regarding the conceptualization of the sense of humor. Linguistic theories tend to ignore the differences among various senses of humor, but rather strive to identify the nature of “the funny” and assume that everybody has the same sense of humor. Thus, they hit the idealized core disregarding the variations and idiosyncrasies. For Raskin (1998) the sense of humor is the ability to perceive, interpret, and enjoy humor and its performance and motivational components. The scope is much larger for psychological research. For example, the list of 100 statements transcends dealing with humor artifacts by far, by referring to everyday behaviors, attitudes, intentions and motivations, etc., and also relates to failing humor and humorlessness. Thus, once humans come into play the list of relevant phenomena quickly expands. Unfortunately, psychology uses the term “sense of humor” inconsistently. Some see “sense of humor” as an umbrella term for all forms of humor, including the darker sides as well. Occasionally, the sense of humor relates to positive aspects only and Figure 3 (and results by Craik et al. 1996; Ruch and Heintz 2019) shows that sense of humor is related to a subset of humor behaviors. Whatever theorists saw as part of the sense of humor, it is typically more than dealing with jokes (or even explicitly not dealing with jokes or humor artifacts). Hence it is not surprising that there is no match in the resulting models. There are overlaps though, and personality trait may amplify the specific elements of perceiving, interpreting, and enjoying humor. The factor of canned humor refers to liking of humor artifacts over spontaneous humor and hence has some overlap with the above definition. The reflective/benign related to the perception of intentional or unintentional incongruities in life (that may or may not be considered funny), laughter amplifies the “enjoying humor” element in Raskin’s definition by adding individual differences in laughter propensity and intensity of expression. Entertaining (i.e., liking to share humor and entertain others) is implicitly present and overlaps with the competence component and it is contingent on the non-bona fide mode of communication. Here the ‘refrainers from laughter’ (i.e., the serious) might be on the lower end. Mean-spirited/earthy refers to a motivational component. Inept (e.g., misunderstanding humorous attempts; not being able to laugh at personal failings) clearly is outside the domain of linguists as it describes how (neurotic) individuals fail at performing appropriately with self-directed humor. Still, the ‘laughterhaters’ might be located on this dimension.

132

Willibald Ruch

Thus, the scope of what psychologists study is broader, as it is the individual that adds deviations from the idealized sense of humor. The general approach is different too as psychology is happy to admit that the sense of humor is merely a construct that needs to be constructed on a more objective basis and may need to show its utility by being able to describe and predict important outcomes. Raskin (1998) describes linguists looking at psychology as follows: It is customary for a linguist to suspect a null-hypothesis psychologist of having a perfectly non-null hypothesis and of being unwilling to admit it. The truth is probably that most psychologists are not accustomed to search for such untested hypotheses inside their own minds. It is quite possible that the existence of the sense of humor is such an untested premise, and then all the measurement techniques yield what is already assumed to characterize the sense of humor. (108)

What is the premise (untested or not) here? It is commonly observed that individual differences in humor performance and appreciation exist and these differences are relatively stable over time and generalized across situations. To account for these differences a personality trait is needed. Surely the psychology of humor could do without a sense of humor construct as individual differences in humor behavior and experience can also be predicted by other, more general constructs (e.g., extraversion, agreeableness and openness). However, the description of the humor profile of an individual will need content-saturated descriptors. The approach taken by Craik et al. (1996) is certainly only one of the possible routes but an ecologically valid one as it depicts the humor use in everyday life. The assumption is that the collection of descriptors and studying their co-occurrence in people may lead to a representative set of dimensions of individual differences in humor. The empirical part allows examining whether the indicators indeed measure the intended dimension. The truth is that psychologists are aware of the pitfalls of relying merely on assumptions or intuition and therefore utilize empirical methods to see how individuals actually tick. Such methodology is also needed to examine, for example, whether playful and serious are indeed opposites in terms of behavior not only conceptually (see Proyer and Rodden, 2013). Raskin (1998) also offers extensions of his theory ready to be tested empirically. If the theoretical perspective on the sense of humor offered by him is valid then a list of personality tendencies should be predictive. Three are counter indicative of the sense of humor, namely a) a commitment to the literal truth of what is said (under any circumstances and in any mode of communication;truthfulness); b) acceptance of stereotypes as factually true (i.e., a refusal to accept purely fictitious stereotypes), and c) an indiscriminate view and

Domains of humor: Challenges from psychology

133

rejection of lying (which is seen as different from truthfulness). These ideas can also be tested by studying individuals with Autism Spectrum Disorders (Kirchner et al. 2016; Samson et al. 2013). Indeed this group was, among others, higher in seriousness and not liking jokes and cartoons but also lower in trait cheerfulness. Two traits are seen as indicative of a sense of humor, namely d) playfulness (for Raskin this means being ready to switch modes of communication at ease) and e) tolerance of fiction (e.g., love for literature, fairy tales, myths, and fantasies, including horror stories and science fiction). Indeed, need for play (Ruch and Hehl 1993) goes along with liking nonsense humor and playfulness is predicting engagement of various components of humor (Ruch and Heintz 2018). The latter aspect was not tested directly but indirectly through openness to experience (Ruch and Hehl 1998), which will entail tolerance of fiction as a component. All in all, Raskin’s theory of the sense of humor is in need of further testing, in particular with respect to its yet untested elements. As the theory is now, it is not specified whether the five traits are assumed to predict all elements of the “ability to perceive, interpret, and enjoy” humor alike; i.e., whether this is a unitary dependent variable or a set of three different (but related) variables which overlap but also have unique parts (that require different predictors). When testing the theory one needs to keep in mind that this theory does represent the essence of humor and not its variants; that is, it will be necessary to use heterogeneous sets of jokes to balance out variation in content. Even when this approach aims at the idealized version of a sense of humor it will be interesting to find out how close the individual conceptualization of the sense of humor comes to this prototype. The six criteria may be turned into a brief instrument (see Appendix) and could be used in studies as an anchor for Raskin’s theory. For example, how close is the high scorer in an instrument (or a subscale of an instrument) to this anchor. For example, I would predict that appreciation of nonsense would be closer to that anchor than appreciation of incongruity-resolution humor, and both would be higher than, for example, the six dimensions described above (of whom the reflective might be closest). So we have the possibility to jointly study both a marker for an idealized sense of humor and diverse humor behaviors to eventually build a comprehensive space in which the prototype and variations have their space. Linguistics will not care about these differences but this is the domain psychology–and in particular personality–is interested in. Victor’s central admonition at every board meeting of the HUMOR journal was that we should encourage interdisciplinary work and also demand from authors to be informed by other disciplines; that is, at least quote articles from outside their own field. I tried to highlight in this manuscript that different disciplines have different aims and a different scope but it is possible to relate to each other. The idealized core of a sense of humor or the linguistic models of

134

Willibald Ruch

jokes may serve as a reference for psychologists, but matters are more complex when humans come into play and these models fall short of describing everything. There are places where disciplines have their own goals to pursue, but cross-fertilization in humor research will continue, in particular among equal disciplines.

References Attardo, Salvatore & Victor Raskin. 1991. Script theory revis(it)ed: Joke similarity and joke representation model. Humor – International Journal of Humor Research, 4, 293–348. Carretero-Dios, Hugo, Cristino Pérez, & Gualberto Buela-Casal. 2010. Assessing the appreciation of the content and structure of humor: Construction of a new scale. Humor – International Journal of Humor Research, 23, 307–325. Craik, Kenneth, Martin D. Lampert, & Arvalea J. Nelson. 1993. Research manual for the Humorous Behavior Q–sort Deck. Berkley, CA: University of California, Institute of Personality and Social Research. Craik, Kenneth, Martin D. Lampert, & Arvalea J. Nelson. 1996. Sense of humor and styles of everyday humorous conduct. Humor – International Journal of Humor Research, 9, 273–302. Heintz, Sonja. 2017a. Putting a spotlight on daily humor behaviors: Dimensionality and relationships with personality, subjective well–being, and humor styles. Personality and Individual Differences, 104, 407–412. Heintz, Sonja. 2017b. Do others judge my humor style as I do? Self–other agreement and construct validity of the Humor Styles Questionnaire. European Journal of Psychological Assessment. Advance online publication Heintz, Sonja. in press. Locating eight comic styles in basic and broad concepts of humor: Findings from self–reports and behavior tests. Current Psychology. Heintz, Sonja. 2019. Separating content and structure in humor appreciation: A bimodal structural equation modeling approach. Manuscript submitted for publication. Hehl, Franz-Josef, & Willibald Ruch. 1985. The location of sense of humor within comprehensive personality spaces: An exploratory study. Personality and Individual Differences, 6, 703–715. Hempelmann, Christian F., & Willibald Ruch. 2005. 3 WD Meets GTVH: Breaking the ground for interdisciplinary humor research. Humor – International Journal of Humor Research, 18, 353–338. Hofmann, Jennifer, Tracey Platt, Willibald Ruch, Radoslaw Niewiadomski, & Jerome Urbain. 2015. The influence of a virtual companion on amusement when watching funny films. Motivation and Emotion, 39, 434–447. Hofmann, Jennifer, & Willibald Ruch. 2017. Humorous TV ads and the 3WD: Evidence for generalizability of humour appreciation across media? The European Journal of Humour Research, 5(4), 194–215. Hofmann, Jennifer, Tracey Platt, Willibald Ruch, & René T. Proyer. 2015. Individual differences in gelotophobia predict responses to joy and contempt. Sage Open, 5. doi:10.1177/ 2158244015581191

Domains of humor: Challenges from psychology

135

Kirchner, Jennifer, Willibald Ruch, & Isabel Dziobek. 2016. Brief report: Character strengths in adults with autism spectrum disorder without intellectual impairment. Journal of Autism and Developmental Disorders, 46(10),3330–3337. Martin, Rod A. & Herbert M. Lefcourt. 1983. Situational Humor Response Questionnaire: Quantitative measure of sense of humor. Journal of Personality and Social Psychology, 47, 145–155. Martin, Rod A., Patricia Puhlik-Doris, Gwen Larsen, Jeanette Gray, & Kelly Weir. 2003. Individual differences in uses of humor and their relation to psychological well–being: Development of the Humor Styles Questionnaire. Journal of Research in Personality, 37(1),48–75. McGhee, Paul E. 2010. Humor as survival training for a stressed–out world: The 7 humor habits program. Bloomington: AuthorHouse. McGhee, Paul E., Willibald Ruch & Franz-Josef Hehl. 1990. A personality–based model of humor development during adulthood. Humor – International Journal of Humor Research, 3, 119–146. Platt, Tracey, & Willibald Ruch. 2014. 3 WD Humor Test. In S. Attardo (Ed.), Encyclopedia of humor studies (pp. 763–765). Thousand Oaks, CA: Sage. Platt, Tracey, Jennifer Hofmann, Willibald Ruch, & René T Proyer. 2013. Duchenne display responses towards sixteen enjoyable emotions: Individual differences between no and high fear of being laughed at. Motivation and Emotion, 37, 776–786. doi:10.1007/ s11031-013-9342-9 Proyer, René T., & Frank A. Rodden. 2013. Is the homo ludens cheerful and serious at the same time? An empirical study of Hugo Rahner’s notion of Ernstheiterkeit. Archive for the Psychology of Religion 35(2). 213–231. Raskin, Victor. 1985. Semantic mechanisms of humor. Dordrecht: D. Reidel. Raskin, Victor. 1998. The sense of humor and the truth. In W. Ruch (Ed.), The sense of humor: Explorations of a personality characteristic, 95–108. Berlin, Germany: Mouton de Gruyter. Raskin, Victor, Christian F. Hempelmann, & Julia M. Taylor. 2009. How to understand and assess a theory: The evolution of the SSTH into the GTVH and now into the OSTH. Journal of Literary Theory, 3(2),285–311. Ruch, Willibald. 1991. An appeal for a cross–cultural approach to a taxonomy of humor. In Willibald Ruch and Victor Raskin (chair), Approaches to the Taxonomy of Humor. Methodological Issues and Preliminary Results. Symposium, Ninth International Conference on Humour and Laughter. St. Catherines, Canada, June 26–30, 1991. Ruch, Willibald. 1992. What is funny to whom and why? Humor and personality in Germany. In: V. Raskin (Ed.) The mission of humor: Part 8. The World and I, 7, pp. 681–691. Ruch, Willibald. 1994. Temperament, Eysenck’s PEN system, and humor–related traits. Humor – International Journal of Humor Research, 7, 209–244. Ruch, Willibald. 1998. The Sense of Humor. Explorations of a Personality Characteristic. Berlin: Mouton De Gryter. Ruch, Willibald. 1999. The sense of nonsense lies in the nonsense of sense. Comment on Paolillo’s (1998) “Gary Larson’s Far Side: Nonsense? Nonsense!” Humor – International Journal of Humor Research, 12, 71–93. Ruch, Willibald. 2001. The perception of humor. In Alfred. W. Kaszniak (Ed.), Emotion, qualia, and consciousness (pp. 410–425). Tokyo, Japan: Word Scientific Publisher. Ruch, Willibald, Salvatore Attardo, & Victor Raskin. 1993. Toward an empirical verification of the general theory of verbal humor. Humor – International Journal of Humor Research, 6(2),123–136.

136

Willibald Ruch

Ruch, Willibald & Franz-Josef Hehl. 1993. Humor appreciation and needs: Evidence from questionnaire, self– and peer–rating data. Personality and Individual Differences, 15, 433–445. Ruch, Willibald & Franz-Josef Hehl. 1998. A two–mode model of humor appreciation: Its relation to aesthetic appreciation and simplicity–complexity of personality. In W. Ruch (Ed.), The sense of humor: Explorations of a personality characteristic, 109–142. Berlin, Germany: Mouton de Gruyter. Ruch, Willibald, Sonja Heintz, Tracey Platt, Lisa Wagner, & René T. Proyer. 2018 Broadening humor: Comic styles differentially tap into temperament, character, and ability. Frontiers in Psychology: Personality and Social Psychology 9(6). Ruch, Willibald & Sonja Heintz. 2018. Psychometric evaluation of the revised sense of humor scale and the construction of a parallel form. Humor – International Journal of Humor Research, 31, 235–257. Ruch, Willibald & Sonja Heintz. 2019. On the dimensionality of humorous conduct and their association with humor traits and behaviors. Humor – International Journal of Humor Research, 32, 643–666. Ruch, Willibald & Hofmann, Jennifer. 2012. A temperament approach to humor. In Paola Gremigni (Ed.), Humor and health promotion (Chap. 5; pp. 79–112). Hauppauge, NY: Nova Science Publishers. Ruch, Willibald, Jennifer Hofmann, Tracey Platt, & René T. Proyer. 2014. The state–of–the art in gelotophobia research: A review and some theoretical extensions. Humor – International Journal of Humor Research, 27, 23–45. Ruch, Willibald & Gabriele Köhler. 1998. A temperament approach to humor. In W. Ruch (Ed.), The sense of humor: Explorations of a personality characteristic, 203–230. Berlin, Germany: Mouton de Gruyter. Ruch, Willibald & Tracey Platt. 2012. Separating content and structure in humor appreciation: The need for a bimodal model and support from research into aesthetics. In Anton Nijhold (Eds.), Proceedings of the International workshop on Computational Humor (TWLT14; pp. 57–70). Enschede, The Netherlands: University of Twente. Ruch, Willibald & René T. Proyer. 2009. Extending the study of gelotophobia: On gelotophiles and katagelasticists. Humor – International Journal of Humor Research, 22(1–2). 183–212. Samson, Andrea C, Christian F. Hempelmann, Oswald Huber, & Stefan Zysset. 2009. Neural substrates of incongruity–resolution and nonsense humor. Neuropsychologia, 47, 1023–1033. Samson, Andrea C., Oswald Huber, & Willibald Ruch. 2013. Seven decades after Hans Asperger’s observations: A comprehensive study of humor in individuals with Autism Spectrum Disorders. Humor – International Journal of Humor Research, 26, 441–460. Sulejmanov, Filip. 2018. The relation of humor structure appreciation with sensation seeking and judgments of complex–abstract art and sophisticated music. Psihologija, 51. 229–241. Sulejmanov, Filip, Ognen Spasovski, & Tracey Platt. 2018 The development of the Humour Structure Appreciation Scale and its relation to Sensation Seeking Inventory and Need for Closure Scale. European Journal of Humour Research, 6, 124–140. Weber, Marco, Willibald Ruch, Rainer Riemann, Frank M. Spinath, & Alois Angleitner. 2014. A twin study on humor appreciation: The importance of separating structure and content. Journal of Individual Differences, 35, 130–136.

Domains of humor: Challenges from psychology

137

Appendix RCSOH In the following we would like to assess your point of view of some statements. Please choose one of the statements, which is closest to what you personally believe or do. 1. How strong is your degree of commitment to the literal truth a) I typically believe everything that is said to be literally true b) I most often believe things said to be literally true, unless the speaker gives away it is not c) Sometimes I do not believe things said to be literally true, for example, when people blink their eye d) Sometimes I do not believe things said to be literally true, for example, when said in between jokes 2. How much truth is in stereotypes? a) I believe that virtually all so-called stereotypes are factually true; no matter how things are said b) I believe that most stereotypes are factually true; that is, they state what is known c) I believe that stereotypes usually contain a kernel of truth d) I can accept that sometimes stereotypes are purely fictitious; that is, they are made up and not true, in particular when this fits well in a joke 3. Playfulness and Seriousness in Conversations a) I enjoy switching often between seriousness and jocularity during a single conversation b) I can easily switch between serious and funny in a conversation when needed c) I’d rather keep serious conversations and joking apart d) I don’t like it when a conversation switches between serious and funny 4. Rejecting or tolerating all sorts of “Lying” a) I believe saying something that is not true to fact is always “lying” and one should not lie under any circumstances b) It is ok to say the opposite of what one believes; saying things that are not true to fact can be humorous in the right situations.

138

Willibald Ruch

c) It is ok to say the opposite of what one believes; especially when it is clear that it is only said in fun. d) It is fun to say the opposite of what I believes; people will find out what I really believe. 5. Facts and Fiction a) I only love to read factual (e.g., scientific) books. I do not like fiction at all b) I mostly read factual books and prefer them over fiction (e.g., fairy tales, myths and fantasies) c) I love fictional literature (e.g., fairy tales, myths, fantasies, horror stories and science fiction) as well as factual books d) I prefer fictional literature (e.g., fairy tales, myths, fantasies, horror stories and science fiction) over factual books

Villy Tsakona

Victor Raskin’s overlooked analysis of political jokes Abstract: In his seminal book on the Semantic Script Theory of Humor, Raskin employs, among other kinds of jokes, a corpus of political jokes to apply the theory and demonstrate its analytical potential. Still, research on political jokes, even within linguistics, appears to have so far overlooked Raskin’s analysis and classification of political jokes. The present study attempts to use this classification to investigate contemporary political jokes referring to the current Greek financial crisis in order to underscore the validity and significance of Raskin’s proposal. By tracing both similarities and differences between Raskin’s data and the Greek ones, it is argued that Raskin’s classification constitutes a useful heuristic tool for the analysis of political jokes, which could be further exploited and enriched to bring to the surface the particularities of political jokes originating in diverse linguocultural and sociopolitical contexts. Keywords: Semantic Script Theory of Humor, political jokes, Greek financial crisis, denigration, exposure

1 Introduction One of the less discussed aspects of Victor Raskin’s work is his analysis of political jokes. In his seminal book (Raskin 1985), and after building and elaborating on the Semantic Script Theory of Humor (henceforth SSTH), Raskin chooses three thematic categories of jokes to demonstrate how the proposed analysis could be performed: ethnic, sexual, and political jokes. To the best of my knowledge, Raskin never returned to the investigation of such jokes to expand or revise it. On the other hand, some studies in political humor may briefly mention Raskin’s SSTH and the concept of script opposition (Laineste 2009; Santa Ana 2009; Tsakona and Popa 2011, 2013), but do not specifically concentrate on his proposal concerning political jokes (with the exception of Tsakona 2017b). The present study sets out to bring to the fore Raskin’s work on political jokes and discuss its potential and significance for the analysis of contemporary jokes.

Villy Tsakona, National and Kapodistrian University of Athens https://doi.org/10.1515/9781501511707-008

140

Villy Tsakona

More specifically, I will analyze political jokes referring to the current financial crisis in Greece using the categories identified by Raskin (1985: 222–246), in order to demonstrate that both the similarities and the differences between Raskin’s analysis and mine render his proposal a useful heuristic tool for the investigation of political jokes. To this end, first I present Raskin’s main observations and claims about political jokes (Section 2) and then I provide a description of my own corpus as well as of the sociopolitical circumstances that gave rise to the jokes examined here (Section 3). The analysis follows in Section (4), where I use the thematic categories of political jokes identified by Raskin to classify the jokes included in the Greek corpus. Moreover, I show that a few jokes cannot be subsumed under any of the above–mentioned categories. Finally, Section (5) summarizes the findings of the present study and accounts for both the similarities and the differences between Raskin’s political jokes and the Greek ones. The potential of Raskin’s classification as a heuristic tool for analyzing political jokes is also discussed.

2 Raskin’s chapter on political jokes In order to apply the SSTH, Raskin (1985: 222–246) decides to collect,1 among other kinds of data, a corpus of canned political jokes. This hardly comes as a surprise if one considers that jokes and, in general, humorous texts referring to political affairs and figures are common and popular in most sociocultural contexts (that we know of). The data collected and analyzed originate in diverse languages and cultures and are translated, wherever necessary, into English by Raskin himself (1985: xvii). So, in the book, we read jokes coming from the USA, Central Europe, Israel, Germany, Russia, the former Soviet Union, etc. dating from the 1890s until the publication of Raskin’s book. The jokes pertain either to prominent political figures (e.g. the Austro–Hungarian Emperor Franz Joseph, Mussolini, Hitler, Lenin, the Greek dictator George Papadopoulos), to groups of people with institutional roles or political power (e.g. Gestapo members, Nazis, Russian populists, Russian liberals, communist regimes), or to ethnic groups (e.g. US Americans, Israelis, Russians). Even though this is a sample of convenience, it cannot be denied that it covers different eras, sociocultural traditions, and political contexts, thus offering an appealing and encompassing image of political joking. Raskin appears to try to

1 In a few cases, it seems that Raskin himself has created one joke or has asked other people to do it for him (Raskin 1985: xv–xvi).

Victor Raskin’s overlooked analysis of political jokes

141

avoid being accused of building a theory based on texts typical of a single sociocultural community and, instead, to offer support for SSTH’s “universal” potential, namely its ability to account for jokes across countries and languages. Still, there seems to be a noticeable preference for Russian/Soviet jokes (to which he also dedicates a separate section; see Raskin 1985: 237–246), which he justifies in the introduction of the book: In many [. . .] cases I had to opt for Russian/Soviet examples, which were obviously clear of any copyright restrictions, having never been printed before. The predominance of Soviet humor in some of the chapters [. . .] was not really intended, therefore, but it is not at all unrepresentative of the world treasury of jokes as a whole because of the enormous Russian productivity in this domain (Raskin 1985: xvi).2

The application of the SSTH to political jokes brings to the surface two overarching script oppositions, that is, “good/bad” and “proper/improper”, since such texts criticize political leaders, regimes, representatives, groups, etc. for behaving in an unreliable, damaging, and eventually inappropriate manner, while the opposite would be expected: political figures and institutions are supposed to care for the people and work for their benefit. In particular, the more specific script oppositions identified allow Raskin to classify the jokes examined into two broad thematic categories: 1. denigration jokes targeting “a person, a group, an idea, or the whole society”; this category includes jokes against politicians and their policies and ideas; 2. exposure jokes targeting “a political regime as a whole and contain[ing] a reference to an event or series of events, which are not widely publicized, and quite often actively suppressed by the regime”. Included here are jokes referring to censored facts, such as arrests, acts of state terror, and food or other shortages (Raskin 1985: 222). The overlapping between these categories is unavoidable, since, as Raskin (1985: 222) himself admits, “some exposure may actually occur in a denigration joke, and the purpose of all exposure is usually denigration”. These categories are further subcategorized as follows: 1. Denigration jokes – Denigration of a political figure – Denigration of a political group or institution – Denigration of a political idea or slogan

2 Russian/Soviet jokes as well as Israeli and American ones could also be considered as related to Raskin’s places of origin and residence.

142

2.

Villy Tsakona

Exposure jokes – Exposure of national traits – Exposure of political repression3 – Exposure of shortages – Exposure of specific political situations

This classification appears to be comprehensive, but has never been tested in other sets of data after Raskin (1985). So, in what follows, I will exploit these categories to classify and analyze the Greek political jokes examined here in order to identify potential similarities and differences, and assess the validity of Raskin’s proposal.

3 The data of the present study and the Greek sociopolitical context Before proceeding with the analysis, a presentation of the Greek corpus as well as of the recent and current Greek sociopolitical context is necessary. The data examined here comes from a large corpus of canned political jokes (596 texts) referring to the Greek financial crisis and collected from January 15, 2010 until December 12, 2013, that is, during the first 4 years of the crisis. All of them were sent to the author’s personal email account by friends and relatives. The author was spontaneously included in the informants’ lists of recipients when forwarding the emails including the humorous material, so the data collected were not actually solicited by her. None of the emails received was excluded from the collection and, at the same time, no other material was added by the author (e.g. downloaded from websites or coming from printed collections). Although the corpus does not claim representativeness, it could be suggested that it was randomly selected and this selection was not biased by the author’s personal preferences (see also Tsakona 2015, 2017a, 2017b, 2018a, 2018b). The Greek jokes under scrutiny could be divided into two broad categories: a) jokes directly targeting and discrediting politicians and their political decisions and actions concerning the Greek debt crisis (322 jokes, 54.02%); and b) jokes referring to Greek people’s everyday lives and problems due to that crisis,

3 Raskin’s (1985: 232) original title is “Exposure of political expression”, but this hardly makes sense and seems to be a typo, especially if we consider the content of the category (see Section 4.5).

Victor Raskin’s overlooked analysis of political jokes

143

thus only indirectly and by implication targeting politicians and their policies (274 jokes, 45.97%). As the analysis in Section (4) will show, such a classification roughly corresponds to Raskin’s (1985: 222) distinction between denigration and exposure jokes (see Section 2). Such humor first surfaced with the onset of the Greek financial crisis at the end of 2009, and later on with the signing of the first Memorandum of Understanding between Greece and the European Commission, the European Central Bank, and the International Monetary Fund in spring 2010.4 Greek people gradually realized that the financial landscape in Greece would change dramatically – and it did. The austerity measures imposed on them were5 too harsh: salaries and pensions were cut off (first in the public sector), the minimum wage was lowered, taxes were significantly increased, unemployment rates started to grow rapidly, and at the same time labour laws were liberalized. In addition, a significant decrease in public expenditure and the privatization of state–owned enterprises led to, among other things, the deterioration of the social services provided by the state (e.g. health, education, transportation). Greek people were soon faced with the hard reality: they would have to pay for the high deficits and public debts accumulated for several decades, while the whole country sunk into recession. The standards of living for the majority of Greeks became (much) worse, the (lower) middle classes were strangled, a significant number of Greeks became homeless, and the number of crimes and suicides increased especially in the big cities. Only within a few months’ time, anxiety, frustration, and disappointment grew higher and higher among Greeks due to the new economic conditions. On the other hand, the political system could not bring relief to the Greek people. Chronic problems of the Greek state such as an inefficient and highly bureaucratic public administration, lack of transparency, weak institutions, corruption, clientelism, and cronyism further decreased Greek people’s limited confidence in the state and its representatives.6 The Greek political parties, especially the two major ones at that time, the centre–left socialist PASOK and the centre–right conservative New Democracy, proved, in the eyes of the citizens, incapable of handling the new conditions and of protecting people from the

4 For a more detailed account of the Greek sociopolitical context, see Hatzidaki and Goutsos (2017) and Tsakona (2017b: 139–141), and references therein. 5 The use of the past tense to describe the Greek context does not mean that things nowadays have changed for the better; it rather indicates that the present situation began a few years ago. 6 On the low levels of political trust among Greeks even before but also after the eruption of the crisis, see among others Koniordos (2011: 52), Ellinas and Lambrianou (2014), Pappas and O’Malley (2014: 1608–1609), Ellinas (2015: 11–12).

144

Villy Tsakona

consequences of the crisis. They were accused, among other things, of inconsistency between pre–electoral promises and post–electoral actions, and of not being able to strike a balance between the measures imposed by the creditors and the EU, on the one hand, and the wishes and needs of the Greek people, on the other. Greek politicians more often than not strived to exonerate themselves by putting the blame for the crisis and/or the austerity measures on previous governments and on the international partners and creditors of the country, thus blurring responsibility for policy outputs. Moreover, it became abundantly clear that they would not reach any kind of consensus concerning state policies; coalition governments had not been at all common in Greece before the crisis and proved difficult (but necessary) to form during the crisis. In this context of political and parliamentary instability, citizens watched politicians voting against their political values, defecting, and neglecting their commitments to their voters. Greek politicians were accused, among other things, of hypocrisy, cynicism, and opportunism (see also Koutsoulelou 2017). Opposition and mistrust towards politicians also extended beyond the Greek borders, as Greek people resisted the interventions of foreign politicians and representatives of international institutions in their domestic affairs. Given the above, the emergence and dissemination of humorous texts result(ed) from people’s (unfulfilled) expectations concerning politicians’ roles and practices (see denigration jokes in Section 2) and from their tendency to compare their current financial and social conditions to earlier ones, when most of them had access to better state services and material resources (see exposure jokes in Section 2). All this will be discussed in more detail in the following Section.

4 Data analysis This Section includes the analysis of material coming from the Greek corpus (see Section 3) in terms of the thematic categories of political jokes identified by Raskin (1985: 222–246). Each subsection is dedicated to a different category, while the last one (Section 4.8) pertains to some examples which were not easy to subsume under any of Raskin’s categories.

4.1 Denigrating a political figure Raskin (1985: 223) observes that “jokes DENIGRATING A POLITICAL FIGURE constitute the most popular and universal type of political humor” (emphasis in the

Victor Raskin’s overlooked analysis of political jokes

145

original). In order to claim that a political leader, representative, etc. is unsuitable for the job s/he has undertaken or been assigned, such jokes highlight that person’s ignorance, incompetence, corruption, luxurious lifestyle, immorality, and/or unreliability. The following examples illustrate some of these themes:7 (1)

Πώς καταφέρνει ο Γιωργάκης να κάνει τόσο πολλές βλακείες σε μία μόνο μέρα; – Σηκώνεται νωρίς το πρωί! How does little George [i.e. George Papandreou, the then Prime Minister] manage to do so many stupid things in only one day? – He wakes up early in the morning!

(2)

Σαμαράς. Ο πρώτος Έλληνας πρωθυπουργός που απέδειξε ότι είναι ανίκανος να κυβερνήσει πριν κυβερνήσει. Samaras [i.e. Antonis Samaras, the then Leader of the Opposition]. The first Greek Prime Minister who proved that he is incapable of ruling before ruling.

(3)

Ο Τσίπρας πάλι στο CNN! Απόψε μιλάει στα greeklish. Tsipras [talks] again on CNN! Tonight he is speaking in Greeklish.

(4)

Όσο και να σπουδάζεις πάντα οι κόρες του Πολύδωρα θα έχουν τα απαραίτητα προσόντα.. No matter how hard you study, Polydoras’ daughters will always have the required qualifications. . .

(5)

O διευθυντής των φυλακών συγκεντρώνει ένα πρωί όλους τους κρατούμενους στο προαύλιο και τους λέει: – Αύριο να είστε καθαροί και ξυρισμένοι γιατί θα έλθει στη φυλακή ο Παπανδρέου με τον Βενιζέλο.

7 All the data presented here was translated by the author for the purposes of the present study. Some humour may have been lost on the way. Unconventional spelling was maintained in the Greek original texts, but was not reproduced in the English translations. Square brackets include additional explanatory material. It should also be noted that due to space limitations short jokes were preferred to longer ones (which were more than one page long in several cases). Short jokes also proved relatively easier to translate. The analysis provided here partly draws from earlier publications (Tsakona 2015, 2017a, 2017b, 2018a, 2018b) with modifications and more examples to fulfill the needs of the present study.

146

Villy Tsakona

Οπότε ένας κρατούμενος από το βάθος φωνάζει: – Τους πιάσατε? One morning the prison warden gathers all the prisoners in the yard and tells them: – Tomorrow I want you to be clean and shaved because [George] Papandreou and [Evangelos] Venizelos are going to be here. A prisoner from the back of the yard shouts: – Did you [finally] arrest them? The two Leaders from the then major parties, namely George Papandreou from the socialist PASOK (joke 1) and Antonis Samaras from the conservative New Democracy (joke 2), are represented as stupid and thus incapable of, and even dangerous for, ruling the country. The nickname Γιωργάκης ‘little George’ in joke (1) further reinforces a widely circulating stereotype referring to Papandreou’s allegedly limited political skills. In joke (3), the leader of the left–wing party of SYRIZA Alexis Tsipras is targeted for his “inadequate” knowledge and use of English in international contexts. The term Greeklish commonly refers (often in a derogatory manner) to a rather low–prestige variety of written Greek used mostly by youths who write the Greek language using the Latin alphabet (e.g. in digital communication) instead of the Greek one. Tsipras is stereotypically represented as using a mixture of Greek and English when he is expected to use standard English (cf. Spilioti 2017: 71–72; Vladimirou and House 2018); hence, it is implied that he cannot fulfil his duties properly. In joke (4), a prominent member of the conservative party of New Democracy, Vyron Polydoras, is accused of non–meritocracy and cronyism. As soon as he became Parliamentary Chair only for one day in June 2012 (because on the next day a second round of elections was announced and the parliament was dissolved), he managed to hire his daughter in the Greek parliament. Accusations of corruption and criminal behavior are implied in joke (5), where George Papandreou and Evangelos Venizelos from PASOK are humorously represented as criminals who should be behind bars. In general, in such jokes, Greek politicians are portrayed as corrupt individuals, who try to maintain their power any way they can, to preserve their own interests (including the interests of their parties and families), and to embezzle Greek people’s money. Such practices are represented as humorous (i.e. they create script oppositions) and are hence rejected. Politicians are implicitly and ideally expected to be honest, care for people’s interests and well–being, support meritocracy and transparency, respect their parties’ political values and decisions, and not be obsessed with power.

Victor Raskin’s overlooked analysis of political jokes

147

In political jokes, political figures are also denigrated through allusions to their sexual lives. “The political figure is not supposed to be thought of in sexual terms, and any attempt to present him or her as functioning in the capacity of a (potential) sexual partner is viewed as degrading and compromising” (Raskin 1985: 224). So, politicians’ sex life may become an object of ridicule, especially after a politician has him/herself attracted attention to it. Petros Tatsopoulos, a SYRIZA parliamentarian at that time, was called a “faggot” by a member of the extreme right-wing party of Golden Dawn. In his defense, Tatsopoulos publicly declared that this could not be accurate as he has slept with half the (female) population of Athens. This statement led to the production of several jokes, including the following two: (6)

Ο Τατσόπουλος αύριο θα πηδάει στα Κάτω Πετράλωνα όσες το επίθετό τους αρχίζει από Α μέχρι και την Αναγνωστοπούλου. Tomorrow Tatsopoulos will fuck in Kato Petralona [i.e. a district in Athens] those women whose surnames start with A [alphabetically] until the one called Anagnostopoulou.

(7)

Ο Τατσόπουλος όταν έχει αϋπνίες μετράει Αθηναίες. When Tatsopoulos cannot sleep, he counts Athenian women.

Jokes (6–7) denigrate Tatsopoulos for his statement, as it seems unexpected and inappropriate to hear a parliamentarian to publicly disclose (rather exaggerated) accounts of his/her sex life as a response to political opponents. In joke (6), Tatsopoulos is humorously represented as being utterly organized and busy in his effort to find time to have sexual intercourse with enough Athenian women so as to maintain his record. Joke (7) humorously exploits the practice of counting sheep when suffering from insomnia to undermine Tatsopoulos’ (alleged) sexual prowess. In such jokes, it is usually implied that politicians are not serious or trustworthy enough to represent the citizens in the parliament. Denigration is also achieved through humorously representing politicians as unpopular or unknown or even through wishing them dead. Normally, politicians are popular and respected individuals “wished a long and prosperous life for the benefit of his or her people” (Raskin 1985: 225). The following examples confirm Raskin’s suggestions: (8)

1 φλιτζάνι καφέ, αξία: 5EURO 1 ποτό στο μπαρ, αξία: 10EURO 1 λίτρο βενζίνη, αξία: 1,70EURO 1 γιαούρτι στα μούτρα ενός κουστουμάτου, ψεύτη πολιτικού, αξία: ΑΝΕΚΤΙΜΗΤΗ!

148

Villy Tsakona

1 cup of coffee, price: 5 euro 1 drink at the bar, price: 10 euro 1 liter of gas, price: 1.7 euro 1 yogurt at the face of a well–dressed, lying politician: priceless! (9)

Η προσευχή ενός απλού Έλληνα. . . «Αγαπητέ Θεέ τον τελευταίο χρόνο μου πήρες: τον αγαπημένο μου τραγουδιστή Νίκο Παπάζογλου τον αγαπημένο μου συνθέτη Μανώλη Ρασούλη τον αγαπημένο μου ηθοποιό Θανάση Βέγγο. . . Απλά ήθελα να σου θυμίσω ότι οι αγαπημένοι μου πολιτικοί είναι: ο Παπανδρέου, ο Παπακωνσταντίνου, ο Πάγκαλος, ο Βενιζέλος . . .. από βδομάδα θα σου πω και τους άλλους!!!!» The prayer of a common Greek. . . “Dear God last year you took away from me: my favorite singer Nikos Papazoglou my favorite composer Manolis Rasoulis my favorite actor Thanasis Veggos. . . I just wanted to remind you that my favorite politicians are: Papandreou, Papakonstantinou, Pangalos, Venizelos. . . next week I will give you more names!!!!”

Joke (8) constitutes a parody of the Mastercard “Priceless” advertisements and explicitly suggests that it is priceless, namely particularly satisfying, to throw yogurt at the face of “lying” politicians. Such a gesture constitutes an act of protest and denigration against people who are perceived as deceitful and worthless. Instead of a product or something that could be paid by a Mastercard and please us, the punch line involves a more violent type of “pleasure”.8 Parody is also involved in joke (9), where a prayer turns into a curse for prominent members of the socialist party PASOK. George Papandreou was its Leader and the Prime Minister at that

8 The product prices may also allude to the incongruous increase of prices during the crisis due to heavy taxation, thus this joke involves some kind of exposure as well (see Section 4.6).

Victor Raskin’s overlooked analysis of political jokes

149

time, George Papakonstantinou the Minister of Finance who signed the first Memorandum of Understanding between Greece and its creditors, Evangelos Venizelos was the Minister of Finance later on, and Theodoros Pangalos has been one of the most prominent members of PASOK since the 1980s and has been Minister in several PASOK governments. All of them are implicitly represented as the main responsible people for the crisis and they are jokingly wished dead, so that people can get revenge for all their suffering during the crisis. The “prayer” script is opposed and eventually replaced by the “curse” one to create humor. Finally, it should be noted that, in the corpus examined here, there do not seem to be any jokes representing politicians as unknown or not recognized by their voters. This may be due to the fact that contemporary media cover politicians’ activities and all sorts of political events, thus enhancing people’s access to how politicians look like and act.

4.2 Denigration of a political group or institution In the political jokes of this category, the political group or institution referred to “is not what it purports to be” (Raskin 1985: 227). Similarly to the jokes of the previous category, political groups or institutions are represented as failing in their roles and as unworthy of the power invested in them. In the corpus examined here, Greek goverments, members of the Greek parliament, and politicians may be collectively targeted, as shown in the following examples: (10) Βλέποντας αυτούς που εξέλεξαν οι Έλληνες στη Βουλή, σκέφτομαι πως δεν πρέπει να μας διώξουν από την ευρωζώνη αλλά από τον πλανήτη. When I see those elected in the parliament by the Greek people, I think that we should not be expelled from the Eurozone but from the planet. (11) Ψήφισε Αλί Μπαμπά. Έχει μόνο 40 κλέφτες. Vote for Ali Baba. He has only 40 thieves. (12) Η κυβέρνηση πασχίζει να διατηρήσει την καλή εικόνα της χώρας στο εξωτερικό! Δηλαδή τι; Όταν έχει επεισόδια στέλνει καρτ ποστάλ από τη Μύκονο; The government strives to maintain the country’s positive image abroad! By doing what exactly? By sending post–cards from Mykonos whenever there are riots?

150

Villy Tsakona

(13) Οι αρχηγοί θα συμφωνήσουν μόνο μεθαύριο, στη σύσκεψη επιχορήγησης των κομμάτων για τις νέες εκλογές. The political leaders will only reach consensus the day after tomorrow, during the discussion concerning the state funding for the parties for the next elections. Joke (10) implies that Greek parliamentarians are the worst and most dangerous choices Greek people could make for their parliament – and perhaps for the whole planet. In addition, joke (11) humorously represents Greek parliamentarians as thieves stealing Greek people’s money and property. Thus Greeks are advised to vote for Ali Baba instead, who only has 40 thieves, while the Greek parliament has 300 members/thieves (see also joke 5 in Section 4.1). Whenever there was/is social unrest and violence in the streets of Greek cities (particularly Athens) as a result of of protests against the austerity measures, the events were/are covered in both Greek and non-Greek media and it was feared that such coverage could deter tourists from visiting Greece. Joke (12) refers to the efforts made by the Greek government(s) and media to minimize the impact of such reports and to persuade tourists that Greece is a beautiful and safe place to visit (hence the reference to the popular, cosmopolitan island of Mykonos). The “unreal” script of the “good image” contrasts with the “real” one of riots and upheaval. Joke (12) denigrates all the leaders of the Greek parliamentary parties for not being able to reach consensus in any of the important problems Greece is facing. The only issue they appear to agree on is how much funding they will receive from the state budget to sponsor their pre-election campaigns (see also Section 4.1 on the representation of specific political figures). The “normal” script is that politicians would agree on important political and financial issues for the benefit of the Greek people (e.g. through forming coalition governments) and would not waste the limited financial resources of the state on pre-election campaigns. Expressions such as αυτούς ‘those’ (joke 10), η κυβέρνηση ‘the government’ (joke 12), and οι αρχηγοί ‘the leaders’ (joke 13), confirm Raskin’s (1985: 227) observation that the people mentioned in the jokes of this category are “devoid of their individualities” as part of the denigration process. Furthermore, the fact the word κλέφτες ‘thieves’ (joke 11) unambiguously (at least for the Greek readers) refers to Greek parliamentarians is also indicative of the low prestige and unreliability attributed to Greek parliamentarians by the Greek people (see also Koniordos 2011: 53).

Victor Raskin’s overlooked analysis of political jokes

151

4.3 Denigration of a political idea or slogan Further denigration of political figures or groups is achieved through ridiculing their political ideas or discursive choices. In the corpus of political jokes examined here, it seems that no slogans were targeted via humor. Instead, politicians’ statements or widespread political discourses become part of script oppositions and are thus framed as humorous/ridiculous. For example, early in the crisis, and when unemployment rates started to increase rapidly, the then Prime Minister George Papandreou emphatically declared that a main aim and commitment of his in such difficult times would be not to leave any family without at least one employed person. This was broadly perceived as an indication of Papandreou’s insensitivity and lack of touch with citizens’ reality, as the cost of living and the taxes had already started to increase rapidly, and families could not be supported by a single salary. The following jokes denigrate Papandreou’s statement: (14) Γνωρίζω οικογένεια με 2 συνταξιούχους και 2 εργαζόμενους. που κάνω καταγγελία; I know a family with 2 retired people [receiving pensions] and 2 employed ones. Where do I file a complaint? (15) Ήταν ένας Άγγλος, ένας Γάλλος κι enas ergazomenos. . . There was an English man, a French man, and an employed man. . . (16) Ποια είναι η διαφορά ενός εργαζόμενου από μία πίτσα; Η πίτσα μπορεί να θρέψει οικογένεια. What is the difference between an employed man and a pizza? A pizza can feed a family. In joke (14), Greek people appear to be ready to turn against each other and file complaints to the Greek authorities when they see a family with more than one person receiving a salary or a pension, thus ironically targeting the then Prime Minister for “not doing his job properly” (see also Section 4.1). Joke (15) humorously recontextualizes the opening lines of tripartite ethnic jokes to imply that employed persons are fictional characters and do not actually exist in crisisridden Greece: they are a joke. Finally, joke (16) underscores the fact that a single salary cannot feed (i.e. support) a family living in crisis-ridden Greece.9

9 Joke (16) could also belong to the jokes exposing shortages (see Section 4.6).

152

Villy Tsakona

The following jokes ridicule a statement by the then Minister Michalis Chrisochoidis when the first Memorandum was signed by the Greek government, the European Central Bank, the International Monetary Fund, and the European Commission (in spring 2010; see Section 3). Although Chrisochoidis was a prominent member of that government, he publicly admitted that he voted for the Memorandum without having read (any part of) it. His statement attracted public attention and was heavily criticized as cynical by the media and the citizens who accused him of frivolity and shamelessness. It was implied that, if Michalis Chrisochoidis were suitable for the job, he would have read the whole Memorandum text: (17) Δεν διάβασα το Μνημόνιο γιατί ένα διαστημόπλοιο προσγειώθηκε στο σαλόνι. I didn’t read the Memorandum because a spaceship landed in our living room. (18) Δεν διάβασα το Μνημόνιο γιατί ήταν γραμμένο στα εγγλέζικα. I didn’t read the Memorandum because it was written in English. In joke (17), Chrisochoidis’ excuse for not having read the first Memorandum is strongly reminiscent of children’s false excuses for not doing their homework. Such an inappropriate response highlights the irresponsibility attributed to him and clearly points to his unsuitability for the job (see Section 4.1). In joke (18), the Greek word εγγλέζικα ‘English’ was mostly used a couple of decades ago; nowadays it is perceived as belonging to low style and carrying a negative evaluation of the English language (e.g. as a redundant, alien code, used mostly by the elite or snobbish people). The joke targets the then Minister for not speaking English and/or for referring to it in a derogatory manner (and hence for being inadequate in his role; see also joke 3 in Section 4.1). When the crisis began, parliamentarians belonging to ruling parties often stated that the country should do whatever it takes to avoid “becoming like Argentina”, meaning to avoid the grave financial problems Argentina was facing at that time. In other words, the Argentinian crisis was alluded to as the worst possible scenario, but Greece, according to these politicians, was not there, yet. This was a way to justify the austerity measures imposed on the Greek people and eventually to sugar the pill (if possible) for them. The following joke recontextualizes such discourses: (19) Φορολογικό σύστημα Αμερικής, φορολογία Σουηδίας, μισθοί Βουλγαρίας, κοινωνικές παροχές Ζιμπάμπουε αλλά ευτυχώς δεν γίναμε Αργεντινή.

Victor Raskin’s overlooked analysis of political jokes

153

[We got] the American tax system [i.e. a very strict one], Swedish [i.e. very high] taxes, Bulgarian [i.e. very low] salaries, Zimbabwean [i.e. of very low quality, non–existent] social benefits, but thank God we have not become Argentina. The political jokes in my corpus also satirize widespread political discourses circulating among Greeks, for example, nationalist ones: (20) Και επειδή η Ελλάδα έδωσε τα Φώτα σε ολόκληρο τον κόσμο ήταν ανάγκη να της στείλουν το λογαριασμό 2500 χρόνια μετά; And just because Greece enlightened the whole world [a metaphor for the global impact of Ancient Greek culture], was it necessary to send it the bill 2500 years later? Based on a pun on the Greek words φωτίζω ‘enlighten, bring light to, shine a light on’ and λογαριασμός ‘bill’, the joke juxtaposes the Greek debt bill (the “real” script) and a domestic bill for electricity (the “unreal” script) by engaging the idea of the importance attributed to Ancient Greek civilization. The debt is humorously represented as a bill Greek people have (but cannot afford) to pay for their “important” contribution to humanitarian and other values which are supposed to originate in Ancient Greece. The Greek verb φωτίζω is often metaphorically used to describe the influence of the Ancient Greek civilization around the world (or at least on the Western world). This metaphor is common in Greek nationalist discourses capitalizing on the high international prestige of Ancient Greek culture and values. Even though no political slogans appear to be humorously targeted in the corpus examined here, politicians’ statements and widespread political discourses do become the object of ridicule in these jokes, thus denigrating those who put them into circulation and/or those who accept them as valid.

4.4 Exposure of national traits Raskin (1985: 230) characterizes this category of political jokes as the closest one to ethnic humor. Similarly to ethnic jokes, which are usually premised on well-known stereotypes about ethnic groups, the political jokes of this category target an ethnic group for being inadequate in their role and/or for behaving in an inappropriate manner. In the corpus examined here, it is Greeks who are mostly targeted for acting improperly, for example, by showing indifference for their own country:

154

Villy Tsakona

(21) Η χώρα διαλύεται από την αδιαφορία. Αλλά τι με νοιάζει εμένα; The country is collapsing due to [its people’s] indifference. But why should I care? In the implied “normal” script, Greek people are supposed to care for their country, but they do not, as the punch line But why should I care? reveals (i.e. the “abnormal” script). Greek people also blame themselves for frivolity and for living beyond their means; in other words, they represent themselves as responsible for their current financial problems. They also accuse themselves of apathy and lack of resistance against the austerity measures: (22) Οι νεοέλληνες ξοδεύουν χρήματα που δεν έχουν για να αγοράσουν αντικείμενα που δε χρειάζονται ώστε να εντυπωσιάσουν ανθρώπους που δεν εκτιμούν. Greeks spend money they do not have to buy things they do not need so as to impress people they do not appreciate. (23) Μετανάστες με i-phone δεν θα έχουν ξαναδεί στην Γερμανία. . . It will be the first time they see immigrants with i-phones in Germany. . . (24) Δεν είναι ότι δεν είμαστε επαναστάτες. Απλά είμαστε άτυχοι που δεν είναι όπλο ο καναπές. It is not that we are not rebels. We are just unlucky that the couch is not a weapon. (25) Σε αυτή τη χώρα ο μόνος τρόπος να γίνει επανάσταση είναι να φτάσει ο freddo τα 10 ευρώ. In this country the only way to cause a revolution is to charge 10 euro for a freddo [i.e. a kind of iced coffee particularly popular among Greeks]. Greek people are portrayed as wasting their money to satisfy their vanity (joke 22) and as spending money on gadgets even though they cannot actually support themselves and are forced to become immigrants (joke 23). They are also represented as lazy and self–indulgent (i.e. lying on the couch all day or drinking freddo in coffee shops; jokes 24–25), hence they do not care for, and do not fight against, the austerity measures imposed on them. In joke (26), Greeks are represented as living beyond their means and enjoying the simple pleasures in life (in this case, food) more than other people and perhaps in an exaggerated and hence humorous manner. In the same joke, Europeans are also collectively (and stereotypically) represented as richer but more frugal and perhaps stingy:

Victor Raskin’s overlooked analysis of political jokes

155

(26) Οι Ευρωπαίοι με μισθό 2.500 ευρώ ο καθένας παραγγέλνουν: – 4 σουβλάκια – 1 χωριάτικη – 2 μπίρες Οι Έλληνες με βασικό μισθό 500 ευρώ ο καθένας παίρνουν: – 2 μερίδες γύρο χοιρινό – 1 τζατζίκι – 1 χωριάτικη – 1 σαγανάκι – 1 λουκάνικο – 2 πατάτες – 2 αναψυκτικά light – 2 μπύρες – Και ψωμί στα κάρβουνα αν γίνεται. . .. Και γυρνάει ο ένας και λέει στον άλλον: –Λες ρε φίλε να πάρουμε και τηγανητά κολοκυθάκια μήπως δε μας φτάσουν; Europeans with a salary of 2,500 euro each order: – 4 souvlaki10 – 1 Greek salad – 2 beers Greeks with a basic salary of 500 euro each order: – 2 portions of pork gyros – 1 tzatziki – 1 Greek salad – 1 fried cheese – 1 sausage – 2 fried potatoes – 2 diet sodas – 2 beers – And bread grilled on the coals, if possible. . . And they tell each other: – Dude, shall we also get some fried zucchini in case all this is not enough? Thus, both Greeks and Europeans are targeted through humor, each group for different reasons. Greeks are represented as eating too much and probably more than they can afford, while Europeans appear to eat less although, as it is

10 Souvlaki, pork gyros, tzatziki, Greek salad, fried cheese, and fried zucchini are popular Greek food, also often tasted by tourists visiting Greece.

156

Villy Tsakona

implied, they can afford a richer meal and life-style, in general. Such humor is based on widespread stereotypes circulating among the jokers and their recipients.

4.5 Exposure of political repression Raskin’s (1985: 232) description clearly states that this category of political jokes refers to state violations of political rights and lack of freedom among citizens, while the regime denies such restrictions: Every repressive regime tries to pass itself as being based on the clearly expressed preference of the people. It views itself as, and/or pretends to be, the answer to what the oppressed population wants for government. As a result, while crushing the smallest expressions of dissidence or resistance, it usually also suppresses all information about arrests and other forms of political terror which it imposes on the population. This type of political humor EXPOSES THE REPRESSIEVE [sic] NATURE of the regime [. . .] by alluding to the suppressed script of arrest and/or terror (Raskin 1985: 232, emphasis in the original).

Hence, such political jokes expose the political oppression the political regime wishes to sweep under the propaganda carpet. Such exposure jokes have not been found in my corpus and it is not hard to understand the reasons why: the oppression Greek people feel does not involve their civil rights and liberties. In general, they are allowed to express their complaints and indignation with the current sociopolitical and financial situation in Greece: the public circulation of jokes such as the ones examined here is part of their ability to do so. The closest jokes to this category are those referring to the social benefits and public services Greeks citizens have been deprived of due to the crisis. Among other things, they appear to be missing satisfactory public services (e.g. health, education, transportation), jobs, satisfactory working conditions, etc. The following examples are illustrative: (27) Αν ακούσετε τίποτα για καμιά θέση εργασίας ενημερώστε με σας παρακαλώ. . . ψάχνω δεύτερη δουλειά γιατί πρώτη δεν βρίσκω με τίποτα!! If you hear anything about a job post, please let me know. . . I am looking for a second job, because I can’t find a first one no matter what!! (28) – Συγνώμη αφεντικό, αλλά έχουμε να πάρουμε μισθό 4 μήνες. – Δεκτή η συγνώμη. – I am sorry, boss, but we haven’t been paid for 4 months. – Apology accepted.

Victor Raskin’s overlooked analysis of political jokes

157

(29) Στην τρίτη λυκείου θα πρέπει να προσθέσουν μάθημα «Πως να ζήσετε στο εξωτερικό». . . In the third class of Lyceum [i.e. the final year of Greek secondary education] they should add a course on “How to live abroad [as an immigrant]”. . . (30) Ο χρόνος είναι ο καλύτερος γιατρός. Γιαυτό στο ΙΚΑ σου κλείνουν ραντεβού για μετά από 4 μήνες. Time is the best doctor. This is why at IKA [i.e. the then largest social security organization in Greece] they give you an appointment for 4 months later. (31) 25 ευρώ η είσοδος στα νοσοκομεία. Με ποτό ή χωρίς; 25 euro [is] the entrance fee at the hospitals. With or without a drink? (32) Συνομιλία μητέρας–γιού στην Ουγκάντα το 2011: –Aμπντούλ διάβασε τα μαθήματά σου παιδί μου. Τα παιδάκια στην Ελλάδα δεν έχουν ούτε βιβλία.. A mother–son conversation in Uganda in 2011: –Abdul dear, do your homework. Children in Greece do not even have textbooks. . . Such jokes focus, among other things, on the difficulty to find a job (joke 27); on the fact that even employed people may not be properly paid because their employers cannot afford to provide salaries or prefer to keep the money for themlselves – in the latter case, they do not even feel the need to apologize for it, but put the blame on the employees (joke 28); on the fact that young people may be forced to leave Greece to support themselves, so Greek education needs to prepare them for this (joke 29). Furthermore, due to recent cuts in public expenditure, the public health system does not work properly and doctor appointments are scheduled with significant delays (joke 30); or the access to public hospitals is too expensive, especially for unemployed people (joke 31); and, in Greek schools, students do not have textbooks (joke 32). It should be noted here that in Greece school textbooks are given to all students of Greek public schools for free at the beginning of each academic year. However, in September 2011, when the academic year began, the books were unavailable, hence teachers and students had to work with photocopied materials for a few months. It therefore seems that political oppression has been replaced in this corpus by financial oppression and its consequences for Greek people’s standards of living.

158

Villy Tsakona

4.6 Exposure of shortages Relevant to the previous category are jokes exposing lack of goods and material resources: “[e]very repressive political regime depicts itself as an excellent provider for all the needs of its population [. . .]. The SHORTAGE jokes are therefore based on the opposition of the script of ‘plenty’, of no shortage, and its more realistic negation” (Raskin 1985: 234, emphasis in the original). Shortage jokes are particularly common in the corpus examined here, but their script opposition is not related to state propaganda denying the shortages. Instead, it stems from Greek people’s comparison of their current financial and material situation (i.e. the “real” script) to that of the recent past, that is, before the eruption of the crisis (i.e. the “now unreal” script). This is illustrated in the following examples: (33) Σε λίγο θα πηγαίνουμε σούπερ μάρκετ, δεν θα αγοράζουμε τίποτα και θα κάνουμε μόνο like. Soon we will go to the super market, we will not buy anything and we will only “like” the products [i.e. as we do on Facebook]. (34) Έλεγα να πάρω σκύλο στο σπίτι αλλά τι φταίει το δόλιο να κρυώνει; Θα πάρω πιγκουίνο. I thought of getting a dog at home but why should the poor animal be cold? I’ll get a penguin instead. (35) Οι νέοι λογαριασμοί της ΔΕΗ θα έχουν φωσφορούχα γράμματα για να διαβάζονται στο σκοτάδι. . .. The new electricity bills will be written in fluorescent letters, so that they can be read in the dark. . . (36) Πάει ένας τύπος σε ένα μπαρ στεναχωρημένος και λέει στον μπάρμαν «βάλε μου ρε φίλε 5–6 ποτάκια, να σου πω τι έχω». Τα βάζει ο μπάρμαν και τον ρωτάει «τι έχεις ρε άνθρωπε. . .». –«Βάλε ρε φίλε άλλα 5–6 ποτάκια να σου πω. . .». Τα ξαναβάζει ο μπάρμαν, «θα μου πεις τώρα;». –«Μόνο 5 ευρώ». A guy walks into a bar looking upset and says to the bartender: “Pour me 5–6 drinks, pal, and I’ll tell you what’s wrong with me”. The bartender pours the drinks and asks him: “What’s wrong with you, man?. . .”

Victor Raskin’s overlooked analysis of political jokes

159

“Pour 5–6 more drinks, pal, and I’ll tell you. . .” The bartender pours again: “Are you going to tell me now?” “[I have] only 5 euro”. (37) Πάω στο ΑΤΜ. Βάζω τηλεκάρτα. «Λάθος κάρτα» γράφει η οθόνη. Βάζω την κανονική, πατάω «υπόλοιπο» και μου γραφει «Βάλε καλύτερα την τηλεκάρτα». I go to the ATM. I insert my phone card. “Wrong card” says on the screen. I insert the proper card, I ask for the amount left in my account and it says: “You’d better insert the phone card again”. The most significant shortage for Greeks appear to be that of money. The cutoffs in salaries and pensions, the increased cost of living, and the unusually high rates of unemployment have resulted in limited financial and hence material resources. In the jokes above, Greek people are represented as having no money to go shopping at the super market (joke 33), to pay for house heating (joke 34), to pay for electricity (joke 35),11 to buy themselves one or more drinks (joke 36); their bank accounts are empty (joke 37). Hence, through such jokes Greek people express their complaints and indignation for various goods they used to take for granted but now miss in their everyday lives. The jokes examined in Sections (4.5–4.6) are not literally exposure jokes as they do not expose anything. Instead, they talk about situations and events that are well-known to the Greek people, as the majority of Greeks experience them or witness other people suffering from them. These jokes refer to the consequences of the Greek governments’ inability to provide decent conditions of living and employment to a considerable part of the population. Such inadequacy and its consequences do not, however, constitute “forbidden” topics of discussion, as it would have been in an oppressive regime, but instead prevail in both the public and private spheres as major problems that need to be dealt with. Rather then exposing shortages, these jokes highlight them and frame them as unexpected and inappropriate.

11 See also joke (20) in Section (4.3).

160

Villy Tsakona

4.7 Exposure of specific political situations In this category, “[the] standard script evokes a particular political situation, usually of an unflattering kind for the government, and the joke is based on the opposition of the official or desirable script excluding the compromising situation and the negation of the script recognizing the situation in question” (Raskin 1985: 235). Once again, such jokes seem to be related more to oppressive regimes which try to hide embarrassing or inappropriate events or situations rather than to less authoritative political systems where the media provide access to information on various aspects of the sociopolitical and financial life of a (usually democratic) state. Examining the Greek corpus, I have come to the conclusion that Greek crisis jokes recontextualize events and situations which are well-known to people living in Greece either because these events and situations are part of their everyday experiences or because they learn about them from the media. All the jokes analyzed so far illustrate this point – including joke (12) about the goverment’s effort to project a false positive image for Greece and joke (13) about political leaders’ greed for state funding for their parties (see Section 4.2). It therefore appears that political jokes of this category are most probably typical of less open societies and more politically repressive regimes than the Greek one.

4.8 Other topics and categories? A careful search of the corpus under scrutiny has brought to the surface a few jokes that would not be easily classified in any of the above-mentioned categories. These concern, on the one hand, the effects of the crisis on Greek people’s mood and feelings and, on the other, the ways the media frame the Greek debt crisis and its consequences. Some of these jokes follow: (38) Αν εξαιρέσουμε την πραγματικότητα, όλα τα άλλα πηγαίνουν υπέροχα. If we do not consider reality, everything else is going great. (39) Έβαλα στο google, Happy new year και μου έβγαλε, “not available in your country”. I googled the phrase “Happy New Year” and it said “not available in your country”.

Victor Raskin’s overlooked analysis of political jokes

161

(40) Η κατάσταση είναι τόσο σοβαρή που το Star έχει ειδήσεις. The situation [in Greece] is so critical that Star Channel has news [to report]. (41) Έχω μια περίεργη αισιοδοξία σήμερα. Θα δω ειδήσεις να μου περάσει. I feel strangely optimistic today. I will watch the TV news and get over it. (42) Σκέφτομαι πόσο σοβαρά προβλήματα έχω.. μετά βλέπω ότι αυτός στο master chef δεν πέτυχε τη σωστή οξύτητα στη σαλτσα.. Ηρεμώ. I think of how serious my problems are. . . Then I see that that guy at Master Chef [i.e. a cooking contest on TV] didn’t manage to get the perfect acidity in his sauce. . . And I calm down. Jokes (38–39) depict the pessimism and disappointment people feel due to the enormous socioeconomic changes they experience and suffer from. Such changes are implicitly evaluated as unexpected and, of course, unpleasant. The Greek media, on the other hand, appear to present the events of the crisis in an exaggerated manner, perhaps in an effort to convince the audience about the necessity of the austerity measures imposed on them. So, in joke (40), the TV news of Star Channel, which used to be full of stories about celebrities and their life-styles, has started to report on current political events in Greece and on the sociopolitical changes taking place due to the debt crisis. In general, TV news on the Greek media seem to terrorize rather than inform Greek people, thus making them feel unhappy and hopeless (joke 41; see also jokes 39–40). Sometimes, however, the Greek media underestimate the gravity of the consequences of the crisis for Greek people and broadcast shows that focus on “trivial” matters. In joke (42), Master Chef is a popular cooking contest, where the candidates try to create and/or execute original and complicated haute cuisine dishes, and sometimes their failures are projected as too important, especially in front of an audience of unemployed people or for people whose financial and material resources have significantly decreased. Needless to say, a corpus of political jokes coming from a different cultural and political context could include more or other categories of such jokes, depending on the contextual particularities and the respective social changes. The jokes examined in this Section could be perceived as both exposure jokes highlighting the bad psychological state of Greeks (jokes 38–39, 41–42) and denigrating ones attacking the Greek media for providing either exaggerated versions of the news or shows that project specific life-style choices as part of Greek people’s lives, although they are not (jokes 40–42).

162

Villy Tsakona

5 Discussion and conclusion The present study has attempted to re-evaluate one of the most neglected aspects of Victor Raskin’s work: the analysis of political jokes already included in his (1985) book. The analysis of a corpus of contemporary political jokes referring to the current debt crisis in Greece and its repercussions reveals that the categories of political jokes identified by Raskin can indeed be applied as a useful heuristic tool to investigate political jokes coming from different linguocultural and sociopolitical contexts from the ones initially studied by Raskin (1985). More specifically, the two main categories proposed by Raskin (1985: 222; see Section 2) are attested in the Greek corpus as well: denigration jokes targeting politicians, political institutions, and political ideas for being inadequate and inappropriate; and exposure jokes bringing to the fore unpleasant and undesirable circumstances under the ridiculed regimes. While the denigration jokes analyzed here (Sections 4.1–4.3) are remarkably similar to the ones analyzed by Raskin (1985: 223–230), there seem to be interesting but not inexplicable differences between the exposure jokes of the two corpora. The examples concerning political repression and material shortages provided by Raskin (1985: 232–235) mostly pertain to authoritative regimes (e.g. communist ones), where people were deprived of civil rights, freedom of speech, as well as food, fuels, and other material resources necessary on a daily basis. In my corpus (see Sections 4.5–4.6), financial oppression seems to have taken the place of political oppression: Greek people are deprived of public services and jobs due to significant decreases in public expenditure, among other things. Furthermore, what they lack is money: while all material resources (food, drinks, fuels, etc.) are available, people cannot afford them anymore. Many of them have become unemployed, others’ salaries or pensions have been significantly reduced, and the high cost of living and heavy taxation do not allow them access to goods they once had access to. In addition, in Raskin’s (1985) jokes, humor stemmed from the fact that the shortages were denied by the repressive regime, thus the official “unreal” script about the lack of shortages was negated by the “real” script of shortages. In the Greek case, shortages are not denied by the Greek governments; the respective “real” script of shortages is opposed by the “now unreal” script of the socioeconomic and material conditions in Greece before the crisis. In this sense, the jokes of these categories are not literally exposure jokes; it would be more accurate to call them shortage jokes. The differences between the oppressive regimes alluded to in Raskin’s jokes and the Greek regime are also clearly reflected in the category of jokes exposing specific political situations (Section 4.7). Such jokes are not attested in the Greek corpus since all the situations and events alluded to in the Greek material has

Victor Raskin’s overlooked analysis of political jokes

163

not been officially denied by Greek governments. The content of such jokes draws upon Greek people’s everyday experience and their knowledge about political affairs. It could be suggested that such jokes do not expose the inadequacies and failures of the government but instead discuss and criticize them openly (see also Billig 2005: 204, 209–213; Tsakona and Popa 2011: 11–12). Such a difference reminds us of the fact that, in oppressive regimes, political jokes are usually not told in open, public spaces, but are recycled in private interactions among intimates. In non-oppressive regimes, however, there are neither severe restrictions on the media nor laws prohibiting citizens from expressing their dissent with the government and its policies. As Laineste puts it (2009: 394), “democracy is a relatively mild and easy environment for jokes”. As a result, political jokes and their particularities (e.g. content, script oppositions, methods of dissemination) could serve as indicators for the degree of openness (or even democratization) of a state (see among others Lindstrom 1980; Tunç 2001, 2002; Mascha 2011; Popa 2011; Tsakona and Popa 2011: 11–14). It is also interesting to note here that some jokes of the Greek corpus could not be subsumed under any of Raskin’s (1985) categories. Such jokes referred to Greek people’s emotional reactions to the changes caused by the austerity measures imposed on them, as well as to the ways the Greek media represent the crisis: either they offer exaggerated, terrifying accounts of it or they ignore its repercussions on Greek people’s lives. A more extended corpus of Greek crisis jokes also covering the years from 2014 to date could bring to the surface more “new” topics and categories related, for example, to the policies of the coalition governments between the left-wing part of SYRIZA and the right-wing national-conservative party of ANEL (January 2015–July 2019). In general, more applications of Raskin’s categories to political jokes coming from different languages, cultures, and eras may reveal more categories that have not been discussed by Raskin or the present study, thus enriching the classification in question. They could also account for striking thematic similarities between different corpora of political jokes, since many political jokes “are freely interchangeable from one country to another and from one epoch to another” (Raskin 1985: 222; see also Astapova 2015). Last but not least, the present comparison highlights the significance of context for investigating humor, even within a competence theory of humor such as the SSTH, which focuses on idealized interpretations of jokes offered by ideal speakers (in Chomsky’s 1965 sense; see Raskin 1985: 2–3, 51, 58–59). A semantic–script account of humor cannot but take into serious consideration (at least) the sociopolitical context of the production and reception of jokes, as this context is necessarily evoked by the opposed scripts which are responsible for the creation of the humorous effect. Contextual differences will result in different script oppositions and simultaneously in speakers’ activation of different

164

Villy Tsakona

background knowledge to process humor (especially knowledge concerning the characteristics of political regimes, as shown in this study). The exploitation of Raskin’s (1985) classification of political jokes allows us to trace and interpret such differences – and this could inspire further applications of and elaborations on his work on political jokes. Sets of data originating in sociopolitical communities more distant to the Western world (e.g. in Asian or African countries) could give us the opportunity to explore even greater differences in political jokes and the respective scripts, practices, and cultures. Acknowledgments: The author would like to thank Argiris Archakis, Jan Chovanec, and Marianthi Georgalidou for their helpful suggestions for this paper.

References Astapova, Anastasiya. 2015. Why all dictators have moustaches: Political jokes in contemporary Belarus. Humor: International Journal of Humor Research 28(1). 71–91. Billig, Michael. 2005. Laughter and ridicule: Towards a social critique of humor (Theory, Culture and Society). London: Sage. Chomsky, Noam. 1965. Aspects of the theory of syntax. Cambridge: MIT Press. Ellinas, Antonis A. 2015. Neo–Nazism in an established democracy: The persistence of Golden Dawn in Greece. South European Society and Politics 20(1). 1–20. Ellinas, Antonis A. & Iasonas Lambrianou. 2014. Political trust in extremis. Comparative Politics 46(2). 231–250. Hatzidaki, Ourania & Dionysis Goutsos. 2017. The discourses of the Greek crisis. In Ourania Hatzidaki & Dionysis Goutsos (eds.), Greece in crisis: Combining critical discourse and corpus linguistics perspectives (Discourse Approaches to Politics, Society and Culture 70), 1–41. Amsterdam: John Benjamins. Koniordos, Sokratis. 2011. Living on borrowed money: On the social context and responses of the current Greek crisis. Economic Sociology: The European Electronic Newsletter 12(3). 48–57. Koutsoulelou, Stamatia. 2017. “Crisis is written all over me”: Greek songs in times of crisis. In Ourania Hatzidaki & Dionysis Goutsos (eds.), Greece in crisis: Combining critical discourse and corpus linguistics perspectives (Discourse Approaches to Politics, Society and Culture 70), 263–289. Amsterdam: John Benjamins. Laineste, Liisi. 2009. Conclusion. In Arvo Krikmann & Liisi Laineste (eds.), Permitted laughter: Socialist, post–socialist and never–socialist humor, 371–406. Tartu: ELM Scholarly Press. Lindstrom, Naomi. 1980. Social commentary in Argentine cartooning: From description to questioning. The Journal of Popular Culture 14(3). 509–523. Mascha, Efharis. 2011. Mocking Fascism: Popular culture and political satire as counter–hegemony. In Villy Tsakona & Diana Elena Popa (eds.), Studies in political humor: In between political critique and public entertainment (Discourse Approaches to Politics, Society and Culture 46), 191–213. Amsterdam: John Benjamins.

Victor Raskin’s overlooked analysis of political jokes

165

Pappas, Takis S. & Eoin O’Malley. 2014. Civil compliance and “political Luddism”: Explaining variance in social unrest during crisis in Ireland and Greece. American Behavioural Scientist 58(12). 1592–1613. Popa, Diana Elena. 2011. Political satire dies last: A study on democracy, opinion formation, and political satire. In Villy Tsakona & Diana Elena Popa (eds.), Studies in political humor: In between political critique and public entertainment (Discourse Approaches to Politics, Society and Culture 46), 137–165. Amsterdam: John Benjamins. Raskin, Victor. 1985. Semantic mechanisms of humor (Studies in Linguistics and Philosophy 24). Dordrecht: D. Reidel. Santa Ana, Otto. 2009. Did you call in Mexican? The racial politics of Jay Leno immigrant jokes. Language in Society 38(1). 23–45. Spilioti, Tereza. 2017. Radio talks, pranks, and multilingualism: Styling Greek identities at a time of crisis. In Janus Mortensen, Nikolas Coupland & Jacob Thøgersen (eds.), Style, mediation, and change: Sociolinguistic perspectives on talking media (Oxford Studies in Sociolinguistics), 51–76. Oxford: Oxford University Press. Tsakona, Villy. 2015. “The doctor said I suffer from vitamin € deficiency”: Investigating the multiple social functions of Greek crisis jokes. Pragmatics 25(2). 287–313. Tsakona, Villy. 2017a. «Δημοκρατία είναι 4 λύκοι και ένα πρόβατο να ψηφίζουν για φαγητό»: Αναλύοντας τα ανέκδοτα για τους/τις πολιτικούς στην οικονομική κρίση [“Democracy is 4 wolves and 1 sheep voting for food”: Analyzing jokes about politicians in the financial crisis]. In Thanasis Georgakopoulos, Theodossia–Soula Pavlidou, Miltos Pechlivanos, Artemis Alexiadou, Jannis Androutsopoulos, Alexis Kalokairinos, Stavros Skopeteas & Katerina Stathi (eds.), Proceedings of the 12th International Conference on Greek Linguistics, Volume 2, 1035–1049. Berlin: Romiosini. http://www.cemog.fu–berlin.de/ en/icgl12/offprints/tsakona/icgl12_Tsakona.pdf (accessed 17 October 2018). [in Greek] Tsakona, Villy. 2017b. “This is not a political party, this is Facebook!”: Political jokes and political (mis)trust in crisis–ridden Greece. European Journal of Humor Research 5(4). 136–157. Tsakona, Villy. 2018a. Intertextuality and cultural literacy in contemporary political jokes. In Arie Sover (ed.), The languages of humor: Verbal, visual, and physical humor (Bloomsbury Advances in Semiotics), 86–104. London: Bloomsbury Academic. Tsakona, Villy. 2018b. Intertextuality and/in political jokes. Lingua 203. 1–15. Tsakona, Villy & Diana Elena Popa. 2011. Humor in politics and the politics of humor: An introduction. In Villy Tsakona & Diana Elena Popa (eds.), Studies in political humor: In between political critique and public entertainment (Discourse Approaches to Politics, Society and Culture 46), 1–30. Amsterdam: John Benjamins. Tsakona, Villy & Diana Elena Popa. 2013. Editorial: Confronting power with laughter. European Journal of Humor Research 1(2). 1–9. Tunç, Asli. 2001. GIRGIR as a sociological phenomenon in Turkey: The transformation of a humor magazine. Humor: International Journal of Humor Research 14(3). 243–254. Tunç, Asli. 2002. Pushing the limits of tolerance. Functions of political cartoonists in the democratization process: The case of Turkey. Gazette: The International Journal for Communication Studies 64(1). 47–62. Vladimirou, Dimitra & Juliane House. 2018. Ludic impoliteness and globalization on Twitter: “I speak England very best” #agglika_Tsipra, #Tsipras #Clinton. Journal of Pragmatics 134. 149–162.

Władysław Chłopicki

Joke construction and joke structure Abstract: This paper deals with the thorny but fascinating process of (re)constructing jokes, the dynamic process of retelling, which produces a joke structure that is often different from “canonical”. In order to study this, I have exposed a small group of native speakers of Polish and a few native speakers English of both genders to a doctor joke in the written form. Then they were asked to retell a joke they have been exposed to. The retellings were recorded, transcribed and analysed as to the joke structure (e.g. the ways of introducing the playframe, the presence of jab lines and the punch line), hearer-speaker relations, and other aspects of joke-telling. The aim of the study is to supplement Victor Raskin’s original account of joke structure with that of joke construction as outlined by him too (1985: 139–147), although regarded as marginal to his central systemic account. Moreover, joke imagery, understood in terms of image schemata, was taken into account in the analysis of the speakers’ joke performance, playing the role of humour enhancers. Keywords: joke performance, doctor joke, macroscript, playframe, imagery

1 Script approach – the language system brushes the performance aside The script-based semantic theory of humour in its original format, later referred to by Victor Raskin as Semantic Script Theory of Humor (SSTH), made several important assumptions as to the necessary and sufficient features as well as other aspects of jokes. By jokes he meant (and I share his definition here) simple canned stories with punch lines, which are easy to remember and tell. The now classical Main Hypothesis (Raskin 1985: 99) claimed most significantly that texts can be attributed the feature of “carrying jokes”; in other words, if a speaker chooses an appropriate sequence of words that are compatible with two different scripts, “opposed in a special sense”, and structured into a setting and a closing punch line (cf. Attardo et al. 1994), a joke will result. An important corollary is that a more accurate account labels the two scripts in question macroscripts. This is significant because in fact all the lexical items in the joke evoke (or rather can be taken to refer to) at least one script (but usually more) – these scripts can Władysław Chłopicki, Jagiellonian University, Poland https://doi.org/10.1515/9781501511707-009

168

Władysław Chłopicki

consequently be labelled micro-scripts (although usually they are not referred as such). The consecutive words of the joke act as disambiguators of the microscripts, and one macroscript emerges, clashing then in the punch line with another macroscript, so far concealed, but compatible with part of the text. The perception of texts as “carriers” of jokes (see also Aharoni 2018 for a recent treatment of jokes as shifting meaning to their “carriers”) has farreaching consequences as it assumes that humour1 lies in the text of the joke, which is at the same time referred to as the “stimulus” – an apparent contradiction. Of course, the humour formula proposed by Raskin (1985: 5–6, 14–19) includes other important factors too, such as the speaker, the hearer, their experiences and psychologies, the situation and the sociology of the humour act, but then he deliberately leaves them out in most of his study, and only discusses some of them in the short but significant section entitled “Joke construction” (Raskin 1985: 138–147 – see 1.2 for discussion). This agrees with his major assumption that the analysis of humour is aimed at matching the idealised “humour competence” of the native speaker – the view that echoes the notion of language competence advanced in Noam Chomsky’s generative linguistics of the 1960s (cf. Chomsky 1965). The presence of this echo is not surprising given the fact that at the time Raskin’s study was written (early 1980s) the generative paradigm was still holding strong in linguistics. In other words, Raskin has always been a system-oriented researcher, claiming the limited role of pragmatic phenomena in linguistic interpretation and preferring to see the default speaker, hearer or situation (the above mentioned elements of his humour formula) as part of semantics (contextual as it may be called), and thus seeing them as part of the broad language system, and supporting in principle the semantic-pragmatic boundary. He would brush aside the purely pragmatic phenomena, such as the psychology and experience of the speaker and hearer or the sociological (cultural) aspects of the humour act and not see them as factors that would be helpful in answering the question: “What makes people laugh?” and “What is the nature of humour?”, which are in his view variants of the same major question: “What is humour?” (Raskin 1985: 8). Consequently, he originally aimed at “the assignment of the feature of funniness to texts”, which (in the ideal world) “should coincide with the native speaker’s judgement as to the funniness of texts” (Raskin 1985: 58), and the goal of his theory was “to model the semantic competence of the native speaker in its relevant manifestations – just as the goal of any linguistic theory is to model its linguistic

1 Humour is used as an umbrella term, not including funniness though, since the latter is dependent on the audience (but see also Raskin’s comments on funniness below).

Joke construction and joke structure

169

competence as a whole” (Raskin 1985: 59). At the same time the approach echoes the structuralist assumptions of the existence of emic vs etic distinctions: “As such, [the SSTH is] concerned with the necessary and sufficient conditions for a text to be funny and not with whether a given person in a given situation will find a text more or less funny” (Attardo and Raskin 2017: 58), thus stressing the idealisation of the text, the situation and its participants.

1.1 Elements of joke construction The above-mentioned section on “Joke construction”, written by Raskin in 1985 (138–147), displayed the same necessary and sufficient conditions requirement. Raskin assumes there are four, essentially text-centred elements necessary for the joke to occur – a switch to a non-bona-fide mode,2 the text itself, two scripts which are opposed, and a trigger which carries out the switch. He does emphasize the correlation of the text with the joke situation, though, by saying that the text should contain all the above elements, unless they are present in the actual situation of the speaker. As the joke-telling situation can contain an “obvious” or “potential” trigger, or in fact make one of the scripts manifest (as an impulse to make a joke), the speakers should adjust their joke performance by focusing on humour techniques (e.g. exaggeration or garden path, cf. Attardo 2001), not so much as on trying to uncover the hidden second script. All this sounds convincing and yet some aspects of variation in joke telling are missing from the account– hence the present attempt to bring them to the foreground. Aware of the need to account for joke pragmatics, Attardo (2008: 115) rightly drew attention to the fact that “the SSTH incorporates at the very core of its theoretical apparatus a whole battery of pragmatic devices and concerns” as well as a “pragmatic component”, and argued that overall humour is non-cooperative and, via a Gricean implicature, potentially communicative. He also drew attention to the attempts within the broadly script-based approach to develop the audience theory of humour (cf. Carrell 1997, who advanced the concept of a humour community, or Żygulski 1985, who in in Polish study had coined the term laughter community). Indeed, “we need to ask ourselves what is the repertoire of choices that Ss [speakers] have and what significance is attached to each of them (and/or its absence). For example, laughter after a joke expresses, on H’s [hearer’s] part, some degree of agreement with S [speaker]

2 See Shilikhina (2017) for an interesting study of various types of non-bona-fide modes (notably she distinguishes humorous, ironic/sarcastic and absurd modes).

170

Władysław Chłopicki

that the occasion was appropriate for joking . . . Withholding laughter may therefore be seen as rejection . . . and disapproving” (Attardo 2008: 1163). What is more, Attardo (2017: 96, 103) distinguishes between performance as the range of choices that speakers need to make, taking into account humorous text, context, repertoire, beliefs of the speaker and systems of belief, and the delivery of the jokes per se, which includes prosody (pitch and volume, pauses), laughter and smiling (2017: 103–104). I am largely concerned here with the first type of performance, although some of the comments below do verge on the latter.

1.2 Role of performance as seen today Along with the shift towards performance in many schools of linguistics, including various pragmatic and cognitive approaches, the ground has been set for the performance analysis of joke-telling; it has also become clearer that the role of pragmatic and cognitive aspects of the joke-telling situation have been underemphasized so far and that humour researchers should perhaps underemphasise if not abandon modelling and idealization of the competences of native speakers and focus on the actual delivery of humour. Yet Ritchie (2018:8) asks a pertinent question: ‘Why still examine jokes?”, and sees the answer in the establishing of a solid foundation for the study of humour more generally. He sees a joke as “a culturally defined and carefully crafted artefact, designed to create a particular effect on the audience” that relies on “human propensities in a few areas, including perception, cognition and emotions” (Ritchie 2018:9). Effectively he proposes to examine more jokes in order to “expand the descriptive coverage” and thus develop a sounder hypothesis about “jokehood” (Ritchie 2018: 184–185). Compatibly, Oring (2017: 199), as a folklorist, considers jokes as “extraordinary arrangements of words and ideas” that are “art – verbal art – as are folktales, ballads, toasts, proverbs, and jump-rope rhymes”. Generally, he appreciates the joke as an art form, emphasizes the jokes’ complexity and claims they are worth literary analyses. More significantly for our present perspective and compatibly with Ritchie, he develops the notion of performance as “an act of communication that is framed and displayed for an audience”. And so, in his view, performers should take responsibility for the effective performance, which is “a function of place, time, situation, and personnel” and emerges “within the field of

3 See also Bell 2017 for the study of failed humour and Marsh 2015 for the treatment of humour support and unlaughter.

Joke construction and joke structure

171

dynamic interaction between these variables”, including also monitoring the performance and necessary feedback that allows the performers to see themselves “through the eyes of their audience” (Oring 2017: 200). What is crucial here is that he sees a straightforward connection between the personality of a joke teller and the way they prefer to tell jokes: the two types of joke tellers he distinguishes are those who consider joke telling a theatrical performance – more extrovert, dominating personalities – and those who like to focus on the verbal and on narrative – the more distanced commentators on the human action. In this connection it is worth drawing attention to the related research on standup, where the personality of the performer is considered the most important aspect, side by side with direct communication with the audience and the sense of immediacy enhanced by the use of present tense (cf. Double 2005). In some contrast, Kuipers (2006) in her sociological study of jokes draws attention to the decline of (canned) jokes as a traditional form of humour and explains the process in terms of social context of joke-telling – the working class environment of alcohol and smoke-filled strip clubs, where simple jokes were told for easy appreciation, as opposed to the middle class milieu, where sophistication and breaking boundaries of taste was expected and as a result new forms of humour were preferred. Still jokes do persist, not necessarily in the working class environment, and they are worth examining, especially in the conversational context, as small forms of conceptual art told in a play frame (Fry 1963) or non-bona-fide mode (Raskin 1985), containing structural elements such as jab lines or punch line (Attardo 2001) and other aspects of joketelling. Significantly, several aspects of joke imagery need to be taken into account while analysing the speakers’ joke performance, as they seem to complement the performance aspect of the jokes. In particular, elsewhere (cf. Chłopicki 2009), I have analysed (largely visual) image schemata, including some elements of force dynamics (such as UP/DOWN, PATH, LINK, MATCHING, RESTRAINT, BLOCKAGE, ENABLEMENT, COUNTERFORCE, BALANCE etc.; see Croft and Cruse 2005), which can be taken to be humour or enjoyment enhancers (e.g., Triezenberg 2004, Sover 2014).

2 Joke experiment Having all of the above perspectives in mind, I have conducted a small joke experiment that involved the retelling of a joke. The aim of the experiment was to test for the presence of various aspects of joke telling which were not present in

172

Władysław Chłopicki

the original Raskin’s account but are considered significant if the joke performance is to be successful. They include the speaker/hearer relation, the situation in which the joke is told, the specific introduction of the play frame (and specifically, the non-bona-fide mode), joke imagery and the hearer’s feedback. In order to make the testing of these factors possible (most of them are discussed in individual subsections in section 3 below), I have asked the respondents to read the joke first, remember it and retell it to someone, without looking at the text. The joke I have selected was one of the doctor jokes I have analysed in my study of doctor jokes (see Chłopicki 2019). One of the reasons was that doctor jokes have a general appeal as they tend to strike the universal chord of the human anxiety about health, regardless of the language they are told in, and tend not to be offensive.4 1. A man returns from a trip to Shanghai and is feeling very ill. He goes to see his doctor and is immediately rushed to the hospital to undergo a series of tests. The man wakes up after these tests in a private room at the hospital and the phone by his bed rings. “This is your doctor,” says the voice on the phone. “We have the results back from your test and . . . I’m sorry, you have an extremely contagious deadly disease known as G.A.S.H.” “G.A.S.H?” replies the man. “What in the hell is that?” “It’s a combination of Gonorrhea, AIDS, SARS and Herpes,” explains the doctor. “My goodness, Doc!” screams the man in a panic, “what are we going to do?” “Well, we’re going to put you on a strict diet of pizza, pancakes, quesadillas and pita bread,” says the doctor matter-of-factly. “Will that cure me?” “Well, no,” says the doctor, “but it’s the only food that we can slide under the door.” (http://mistupid.com/jokes/page066.htm) The apparently independent Polish version was as follows (here it is provided in back translation, with the original cited in the Appendix): 2. The State Clinical Hospital no 4 in Lublin. The mobile phone rings in an isolation room. A man gets up from his bed and reaches for the sideboard. ‘Mr Kowalski?’, a voice in the receiver is heard. ‘Yes’. ‘The head of the infectious diseases ward, Professor Kwiatkowski, speaking. Sir, I now have the results of your tests. You are ill with diphtheria, borreliosis, cholera, bacterial dysentery, typhoid, bubonic plague, tuberculosis, syphilis, hemorrhagic enteritis,

4 The general appeal may of course (but does not have to) be restricted for the audience who are doctors and currently patients, but this would require a separate study (see also the comment of one of the respondents below).

Joke construction and joke structure

173

whooping cough, chickenpox, glanders, measles, foot-and-mouth disease, tularemia, gonorrhea, anthrax, trichinosis, pediculosis, rabies and AIDS. ‘Oh my God’, cries the patient. ‘But you will surely treat me?’ ‘Yes, yes, of course, we will. We have even selected a special pancake diet for you already.’ ‘And the diet will help me?’ ‘It will help or not. How do I know?’ the irritated professor replies. ‘But pancakes are the only dish that can be fitted into the opening under the door.’ (source of the original joke in Polish: https://dow cip.net/dowcip/panstwowy-szpital-kliniczny-nr-4-w-lublinie-5806) Among the many doctor jokes in circulation, the one above seemed suitable for the study as it is not too long and thus hard to remember, and also not too short and thus cannot easily be retold verbatim without having to read the text. The list of diseases from the Polish version is obviously editable and there is no need to cite them all to achieve the comic effect. In contrast, the fictitious (though not unlikely) disease acronym from the English version is a useful mnemonic device, even though the four diseases are not difficult to remember even without the acronym. The Polish joke does not mention any reason the unfortunate patient acquires the series of diseases – the ill-fated visit in Shanghai in the English version – and takes the hearer right into the middle of the scenario: already in the close ward of a hospital. The English version has other typical cultural features5 – private room in a hospital connotes the patients choice, while medical isolation room from the Polish version suggests lack of choice – being allocated to it by doctors. Significantly, the relationship between the doctor and the patient is different too – it seems more equal in English (“What in the hell is that?” asks the English patient and addresses the doctor as Doc; he also asks the doctor “what are we going to do?” using the first person plural, assuming some sort of joint action), while the Polish doctor is characteristically more dominant and less caring, not only by offering the patient a stricter and monotonous diet (just pancakes), but also by getting irritated with the patient’s incredulous question: “And the diet will help me?”. On the other hand, the joke is not too complicated and is relatively easy to remember due to the visual imagery of flatness and narrowness present in the punchline (“sliding under the door”, “fitting into the opening”). Notably, the quesadillas and pita bread were not included in the Polish joke (they are not likely to be recognized for their shape by most Poles, and even pizza did not sound dietary enough to fit the joke), while pancakes

5 The significant cultural differences between the jokes are not the focus of the article, but still they are worth pointing out, not the least as evidence of their independent origin.

174

Władysław Chłopicki

(naleśniki) is a dish known nationally and perceived as compatible with a health-enhancing diet. Respondents were asked to read the text of joke 1 (or its independent version in Polish, joke 2), remember the gist of it, and then find a friend or acquaintance and tell it to them, recording their performance. They were also asked to make the joke-telling situation natural and, as evidence, to record a few moments before the joke and after it to incorporate the opening of the playframe and the audience’s feedback. Thus in my introduction to the online survey (I used the Survey Monkey system – see the full text of the survey in the Appendix), I made a few comments or recommendations which aimed at limiting the expected performance problems. I was expecting that the respondents might consider that the survey was intended to assess their personal joke telling skills or might fear that the task would be too complicated. I also wanted to avoid the situation where they would just read the text of the joke instead of re-telling it or use notes to help them remember details, thus changing the nature of the performance. I was also wary that the respondents would expect the selected joke to be a side-splitter and if they found it not to be very amusing or even slightly offensive, they would be unwilling to re-tell it. In order to avoid these problems, the respondents were encouraged to tell the joke even though they did not consider it very funny, just in order to entertain their friends of family members. For this purpose they were also encouraged to embellish (enhance) the joke if they felt like it in order to improve the experience. What is more, I left them the choice to inform their audience before the telling or only afterwards that the joke would be recorded (although I emphasized that the latter was preferable). I also encouraged them to fill in the survey even without doing the actual recording if they were unwilling to do the latter. The rationale of this setup included the performance-related conviction that the construction and wording of jokes are crucially dependent on the speaker and hearer relations as well as on the situation of telling, not to mention the culture of telling jokes or lack of it (in sociological terms; cf. Kuipers 2006). This is going beyond Raskin’s assumption that the joke text would need to be adjusted by dropping certain elements already present in the context. Thus respondents were discouraged from recording the joke without the audience (an obviously much easier thing to do than actually telling it to someone). Another important point concerned the perceived relation between the oral and written form of the joke; the latter is considered secondary to the former, following the assumptions that have prevailed among linguists since the time of Franz Boas and other American structural linguists. The written form, like all the jokes listed in Raskin (1985), is the idealisation and abstraction from the

Joke construction and joke structure

175

“real” spoken form as it ignores the circumstances of the telling, just like the speaker’s (esp. native speaker’s6) sentence is the idealisation and it abstracts from his actual performance in a specific, potentially deficient or emphatic, hearer-oriented utterance.

3 Results and discussion I have received 14 survey responses altogether and 12 recordings, including 9 Polish recordings and 3 English ones (2 Australian and one from a non-native speaker from Belarus), as 1 respondent from the US and 1 from the UK used the opportunity to just fill in the survey and not to send their recordings leaving interesting comments which I quote below. The low turnout, in my view, was, among other reasons, the result of my demand for respondents to record the joke to a real audience and send it via email as well as of the decline in joketelling as a social activity in general. With regard to the age of Polish respondents, it varied from 10 to 60, all of the speakers being women. Among the English-speaking respondents they were between 21 and 60 years of age, 3 being females and 2 males. The audience were family members in 8 cases, there were also 3 friends (actually 7, counting the presence of extra hearers in one joke), and 1 neighbour. 7 times the audience were told of the recording in advance, 5 times they were told only afterwards, and twice no recording was made. 14 comments on the experience were sent and they concerned the assessment of the joke itself (“peculiar joke”, one speaker said; ‘I’ve heard better”, said another) or their own inability to remember the joke accurately (“I was not able to list all the diseases mentioned in the joke”). One respondent mentioned another similar joke that came to her mind, others expressed their preference for shorter jokes, for conversational humour, or for listening to rather than telling jokes. A few respondents emphasized some problems with joke telling as an activity as well as pointed out the embarrassing social situation (sense of artificiality) that involved both the joke telling itself, the recording of it, and the speakers’ lack of experience with it (one person admitted she “has probably not

6 Linguistics, esp. cognitive linguistics, has long departed from Chomsky’s early assumption (cf. Chomsky 1965) that the native speaker’s judgement is the basis for the sentence grammaticality and acceptability. Now the key factors seem to be the cognitive preferences and restrictions of speakers regardless if they are native or not.

176

Władysław Chłopicki

told anyone any joke before” and was afraid of embarrassment, and another actually recorded herself attempting to start twice and succeeding only the third time). Some respondents also tried to explain the lack of hearer’s feedback (“four people are laughing and the target did not get the joke. An interesting experience that survey was”; “Told the joke to my long suffering wife who is used to hearing corny jokes (dad jokes) hence the complete lack of response following”). Two people in their 60s–70s, who did not retell the joke, offered two comments, one pointing out the potential offensiveness of the joke and the other the incoherence of it (the latter clearly did not understand the joke): I found the joke amusing and something that could be told almost anywhere, except to anyone suffering an incurable illness. The only problem with it is that you could find yourself telling it to someone who decided it WAS offensive. For example, a friend of mine once told me a joke about someone refusing an operation which would cure a terminal illness. She told the joke in a pub and someone nearby objected very strongly, having recently had a member of her family in hospital with cancer. It is a good joke but you must know your audience. (UK) The punch line seems to be unrelated to the main premise of the joke (US)

Now with regard to the joke versions recorded by respondents, the variety of adopted approaches is striking and worth briefly commenting on. The responses varied in the length of the recorded jokes (from 105 to 213 words, and from 0.48 to 1.57 minutes in Polish; and from 164 to 307 words and from 1.11 to 2.15 minutes in English). What is responsible for the variety is the size of the preceding and following dialogue. The nature of the experiment led to the occurrence of a relatively limited number of interruptions and efforts to co-create, although backchannelling is present in many of the dialogues, not only at the beginning or at the end, but also throughout (esp. in joke 8, which is the most natural of the Polish conversations and as such it is quoted and discussed below). Short bursts of quiet laughter while the joke is being told are audible in many cases (probably signifying the presence of jab lines, but this would require a separate study). The pragmatic speaker/hearer situation in all the jokes is presented in Table 1 below, which also includes the joke situation and the hearer’s reaction. Below the specific performance-related aspects of the data are discussed in some detail, with the exception of the speaker/hearer relation, which (as can be seen from the table above) was largely friendship and kinship, with the two exceptions: one person told the joke to her dog, and one person talked to her neighbour (joke 8, discussed below in some detail).

Family members (hearer – male)

Family members

 Friends

 friends, same speaker as in 

 friends, same speaker as in  and 

Joke 

Joke 

Joke 

Joke 

Joke 

Speaker/hearer relation

Regular

Regular

One friend (male) uninformed of the recording (several females present do know) – a practical joke of sorts on him

Performed, artificial setup added (two tellers perform for the sake of a young girl)

Regular, no playframe recorded, hearer informed afterwards

The joke situation

Table 1: Pragmatics of the joke performances. Other comments

The joke partly read from the text, esp. the long list of diseases. The teller seemed ironic with regard to some of the names of exotic diseases mentioned in the written text (extra jabline)

The joke largely read from the text, esp. the long list of diseases. The speaker enjoyed the joke at the expense of the male

The hearer who completed the survey does not like telling jokes, and prefers listening to them

(continued )

Long laughter twice (s and s) at the Joke shortened (including the list of end, and the comment – ‘That was diseases) compared to  and  terrible”

Long laughter (s) at the end, with the comment ‘a very good joke’ (perhaps ironic)

The male hearer does not get the joke, the others (females) laugh twice for  to  seconds, and he comments, laughs, trying to save face

laughter twice during the performance, laughing voice at the end

Laughter (s) at the end, repetition of The speaker would prefer a shorter key phrase joke

The hearer’s reaction

Joke construction and joke structure

177

 relatives

Speaker (female) and her dog

neighbours

Family members

Family members

Family members

Family members

Joke 

Joke 

Joke 

Joke 

Joke 

Joke 

Joke 

Speaker/hearer relation

Table 1 (continued )

Natural setup, the female hearer told of the recording afterwards by the speaker (male)

Regular

Somewhat artificial setup prior to the joke, Belarussians speak English

Regular

Visit to tell the joke, hearer uniformed of the recording, natural setup

Unusual setup, an obviously silent hearer, the speaker’s voice changed into a ‘dog voice’

Regular

The joke situation

The hearer does not laugh because, as the speaker claims, she is used to corny jokes instead

Failure to start three times, selfdeprecating laughter

Laughter of s the speaker openly self-critical after her performance

Failure to get the joke, until hint given, but then no laughter follows

Somewhat embarrassed laughter at the end by both sides during profuse and somewhat jocular apologies

None

Laughter (s) at the end

The hearer’s reaction

wife washing dishes while listening

The speaker admittedly not used to telling jokes

Speaker not used to telling jokes, esp. in English

The punch line shortened (no opening on the door mentioned), which might have confused the hearer

Embarrassment of the speaker

Speaker admittedly does not like telling jokes

Speaker seems to rely somewhat on the text

Other comments

178 Władysław Chłopicki

Joke construction and joke structure

179

3.1 The joke telling situation Based on the analysis of the transcripts some speaker preferences as to joke experience become apparent, e.g. 4 speakers (all women) explicitly voiced their dislike or lack of experience as to canned joke telling, which is in line with Kuipers’ (2006) claim. Some speakers expressed their dislike of joke-telling, which might have resulted in poorer performances: the performances involved reading part of the text (contrary to the instructions); restarting it three times due to two outbreaks of embarrassed laughter; omitting part of the punch line, which resulted in joke failure; or even, in one extreme case, choosing their dog as the audience. The latter case (joke 7) is interesting since it displays some “dog talk” with unnaturally high pitch utterances and repetitions, as well as some unusual emphasis and whispering voice that move the joke closer to the genre of a fairy tale told to a child. The joke involved some performance failure too since following the patient’s question “Will it help?’, the doctor replies “Well, I hope it does. And besides this is the only dish we can slide in the opening under the door”, thus obviously the joke teller counts that the dog will not notice the slip.

3.2 Introduction of the playframe Generally, introducing the non-bona-fide mode in the data occurs altogether in 9 out of 12 cases (except in joke 1, 3 and 9). In four cases it involves the speaker announcing the start of the joke by saying: ‘Watch out, I am telling a joke’ (3); ‘I am recording, watch out’ (4) or simply ‘Then I am starting’ (5) or ‘Are you ready?’ (11); in three other cases, the speaker asks some form of a question paired with a request, preceded by a vocative address in two cases, making the request somewhat more urgent, although in 7 the urgency of the vocative is questionable, the addressee being a dog: ‘Asia, I have a question, can I tell you a joke?’ (8); ‘Ok, you want to hear a joke?’ (12); ‘Brunek, I will tell you a joke, do you want to hear? (7). The two artificial setups (in both cases the speakers were obviously involved in some scheming against the hearers) are worth commenting on too in terms of introducing the non-bona-fide mode. In joke 2, a young girl is asked to produce the line “The mosquitoes bit you so much”, which is meant to prompt the story of a man called Kowalski (one of the most popular surnames in Poland) who had it even worse in the hospital. After she somewhat unwillingly does so, the whole joke seems designed to surprise and amuse her, although it is not her who laughs at the end. In joke 10 in turn, the setup involves the

180

Władysław Chłopicki

speaker saying that the hearer (her husband) always wanted her to tell him a joke and now she has one, whereupon he comments sarcastically that it sounds ominous, with the speaker being a little bit put off – she asks why, to which he replies that she rarely initiates a joke that is why he is surprised.

3.3 Features of joke text and joke imagery Apart from the above remarks, two other interesting aspects of the joke content need be emphasized. Even though the joke text was included by Raskin among his necessary conditions for the joke to occur, he did not discuss some aspects of joke performance, such as changing the text wording for various reasons as well as visual imagery. First, in the Polish joke speakers variously deal with the long list of diseases, too long of a list being a challenge to tellers, unless they read them all (21 altogether, listed in full in jokes 3 and 4). The numbers of diseases actually included in the actual performance vary from 4 to 8 diseases (6 or 7 prevailing – 4 cases), many of those not mentioned in the original joke (rubella, psoriasis, yellow fever, small pox, contagious erythema, sick liver, sick adrenal glands etc); this is certainly a creative aspect the joke on the part of the speakers, who invent more and more absurd and unlikely diseases that the poor patient suffers from in order to contribute to the funniness of this jab line. Even in the case of the GASH acronym in the English joke, which involves just four diseases, names get changed too, and herpes gets exchanged for hepatitis (12). Some metalinguistic comments also follow the shortened list in the Polish jokes: ‘they know all the diseases, he lists them lists them lists them right?’ (9); ‘even’ is added before the last disease in (2) to mark the exaggerated nature of the list; and ‘something something’ follows the list in (5) to make up for the deficiency of the speaker’s memory. Speaker 7 is ingenious as, talking to her dog, she develops the list into clauses (as if she was talking to a child, making sure it understands): ‘Both the head will be hurting you because you have a brain tumor, and the prostate will not be too good either, and all other symptoms connected with the sick liver, kidneys, stomach”, and ends with an ironic comment: “well and your general condition is simply great, well’. Notably, the prosody of the list of diseases often involves an emphasis and pause at the outset (as if the speaker was getting ready to perform) and pauses between the names. One speaker (teller of joke 3 and 4) seems to have the need to express playful although ironic distance to the joke by repeating the name of some more exotic diseases she is reading (with rising intonation), as if what she has just read has surprised her.

Joke construction and joke structure

181

Second, the implied imagery of the punch line (implied flatness of the dish) facilitates the comprehension of the joke, especially if the opening under the door is mentioned too. In one case the opening seems to be in the door, not under it (joke 7), which is not very likely and thus incongruous in itself. The mention of the sliding movements can be considered a jab line too. Here are the variants of the punch line, including the PATH and IN/OUT as well as ENABLEMENT and even MATCHING image schemas in various configurations (cf. Croft and Cruse 2004: 45 for full list and discussion). – ‘jedyna potrawa, która zmieści się/ może się zmieścić w szparze’ [the only dish that will/can fit in the opening] (1, 3, 6, 8) [ENABLEMENT, MATCHING] – ‘tylko naleśniki zmieszczą się w szparę/szparze [only the pancakes will fit in the opening] (2) [ENABLEMENT, MATCHING] – ‘jedyne, co mieści się w szparze’ [the only [thing] that fits in the opening] (5) [ENABLEMENT, MATCHING] – ‘jedynie danie, jakie możemy panu wsunąć przez szparę w drzwiach [the only dish we can slide you through the opening in the door] (7) [PATH, IN/ OUT, ENABLEMENT, MATCHING] – ‘the only food that we can slide under the door for you’ (10) [PATH, IN/ OUT, ENABLEMENT, MATCHING] Once ‘szpara’ (opening) is omitted by the speaker, the joke (9) falls flat, with neither PATH, IN/OUT nor MATCHING schemas being present, thus the ENABLEMENT does not seem to have a clear purpose: – ‘jedyna potrawa, która zmieści się pod drzwiami’ [the only dish that will fit under the door] (9) [ENABLEMENT]. In 11 (the jokes from Australia told in English), however, the joke seems to have worked in spite of this lack (it is followed by 2.5 seconds laughter) and in spite of the fact that a ‘thing’ of unspecified shape is fitted under the door in the punch line (dish under the door would be an even more bizarre image). Still the teller is the speaker who starts the joke three times before she finally succeeds, so the laughter at the end might also be more of a coping laughter than that of amusement. In contrast, 12 succeeds, even though the opening is not mentioned, but there the movement of feeding is emphasized and the image schemata evoked are more numerous: – ‘This is the only thing we can fit under the door’ (11) [ENABLEMENT] – ‘That’s all that they can feed under the door’ (12) [PATH, IN/OUT, ENABLEMENT, MATCHING]

182

Władysław Chłopicki

3.4 Hearer’s feedback The non-bona-fide mode is closed and the bona-fide mode is restored in diverse ways too: by laughter only (11); by an incredulous repetition of the key word or phrase (‘pancake diet’, 1 and ‘pancakes’ 3) by the hearer in an emphatic voice; by the hearer assessing the performance (‘good’ (1), ‘a very nice joke’ (4) ‘terrible, this was terrible’ (5), ‘Jeesus’ (8)’; see also Norrick 1994 for detailed analysis of such comments); by another metapragmatic comment (‘but what should I say?’ (3), ‘the end, right?’(6), ‘I don’t get it’ (9), “Aha, I see’) (12)); by the long apology (8) or self-assessment of the speaker (10 ‘That was kind of a little bit long in my performance’) followed by the call for hearer assessment (‘but at least you enjoyed it a little, didn’t you?’). In one case the speaker closes the mode explicitly and somewhat ironically herself – ‘end of story’ (4).

3.5 One joke delivery example One joke performance is worth special attention as it contains all the elements of a potentially successful, spontaneous performance that is delivered by the speaker without previous announcement that the conversation was going to be recorded. The speaker and hearer (both in their early 20s) have known each other for 18 months, they study and work together, and enjoy each other’s company, although, as the speaker declared, they do not spend too much leisure time together. As the speaker also said they share a sense of humour, although the speaker, contrary to her neighbour, does not actively tell narrative jokes so much as short quips (she admits she does not have a talent for that7). They both enjoy and share situational humour, anecdotes, visual humour, often involving puns, and generally their humour is more a spontaneous nature.8 The performance starts with the speaker knocking at the door of her neighbour’s room (they share an apartment and the neighbour has just come back from work) and asking her whether she can tell her a joke, and also apologising for the intrusion.

7 The speaker seems a dominating performer in the sequence, but at the same time she denies she likes telling narrative jokes, which seems an exception to Oring’s (2017) two types of joke tellers, mentioned above. 8 This information was provided to me by the speaker via email correspondence.

Joke construction and joke structure

183

Joke performance 8 (Polish) Visiting a neighbour (knocking at the door) S1 (fem) Dzień dobry. ‘Good morning’ S2 (fem) Cześć ‘Hi’ S1 Asiu. Mam pytanie, mogę ci opowiedzieć dowcip? ‘Asia, I have a question, can I tell you a joke?’ S2 No pewnie, ‘Sure’ S1 [przepraszam w ogóle, że] ‘I am sorry that’ S2 [No siadaj, pewnie] ‘Well sit down, sure’

Following that she tells her that she has heard “a funny joke” (this sets the non-bona-fide mode). The neighbour co-creates the performance almost right away by asking whether it is important if the joke’s action is happening specifically at the hospital clinic number 4 in Lublin, which the teller denies and they both laugh; the teller is then prompted to explain that she wanted to “fit in the content”, not admitting that this had actually been part of the original joke she was asked to read. S1 Słyszałam a taki śmieszny dowcip. A więc to się dzieje w szpitalu klinicznym numer 4 w Lublinie, nie wiem dlaczego ‘I have heard such a funny joke. And so this is happening in a hospital clinic number 4 in Lublin.’ S2 A to jest ważne? ‘And is that important?’ S1 Yy niezbyt ‘Not very’ S1/S2 (laughter) S2 dobra ‘ok.’ S1 ale tak chcę się wpasować [w treść] ‘But I want to fit in the content’ S2 [dobra] ‘ok’

Then the joke continues and it is enhanced by the teller adding a number of exotic diseases and other conditions that the patient is afflicted by that were not mentioned in the original joke (she emphasizes the sexually transmitted diseases, animal diseases as well as bad kidney and liver conditions). The laughing voice of the teller (it appears three times including the occurrence

184

Władysław Chłopicki

immediately prior to the punchline) signals her enjoyment of the situation although perhaps it also serves the purpose of alerting the interlocutor to the jab line and punch line status of particular lines. The hearer reacts only once with brief laughter in the middle of the telling, while her reaction is more intense following the punch line (the latter does mention the opening under the door into which the pancake is fitted, thus evoking the PATH, IN/OUT, ENABLEMENT and MATCHING image schemas), when she utters a shocked “Jeesus”. S1 no to później dzwoni telefon i pacjent odbiera (.) no i słyszy ‘Dzień dobry, czy to pan Kowalski?’ ‘Tak, tak oczywiście pan Kowalski’ (laughing voice) No i z drugiej strony ‘Panie Kowalski, tutaj ordynator szpitala. Mam właśnie pana wyniki badań. I ma pan (.) kiłę, rzeżączkę, AAIDS (laughing voice), dur brzuszny, łuszczycę, ‘Well then later a phone rings and the patient picks up and hears ‘Good morning, is that Mr Kowalski?’ yes, yes of course, Mr Kowalski. Well then from the other end ‘Mr Kowalski, here is the hospital director. I have now your test results, sir. And you have syphilis, gonorrhoea, AIDS, typhoid, psoriasis’ S2 (short laugh) [inaudible] S1 chore nadnercza, chorą wątrobę i chore nerki.’ ‘O mój Boże, panie doktorze, i co z tym robimy? Jaki jest sposób leczenia?’ ‘A więc (.), panie Kowalski. Przepisze panu specjalną dietę naleśnikową.’ Panie doktorze, ale jak to dietę naleśnikową? Co to (.) pomoże?’ ‘ Hmm możliwe, że nie, ale naleśnik to jedyna (laughing voice) potrawa, która zmieści się w szparze pod drzwiami. ‘sick adrenal glands, sick liver, and sick kidneys’. ‘Oh my God, doctor, what shall we do with it? What is the treatment?’ ‘Well, Mr Kowalski, I will prescribe you a special pancake diet.’ ‘But doctor, how come the pancake diet? Will that help?’’Hmm, maybe not, but the pancake is the only dish that will fit in the opening under the door’. S2 Jeezuu. ‘Jeesus’

Now the teller closes the non-bona-fide sequence with a relatively long apology (she apologizes three times), perhaps prompted by the fact that she was not so comfortable telling a long narrative joke, but also the fact that, as she herself admitted, she tends to apologize a lot in general. The neighbor is surprised by both the telling and the recording and she laughingly teases the joke teller (“Are you kidding?” (lit. What are you babbling?), “How could you?”). S1 Asia (laughter), a teraz cię (laughter) bardzo przepraszam za opowiedzenie tego dowcipu, ale (laughter) moja mama (laughing voice) mi podesłała (laughter) taką ankietę, którą mój znajomy (.) prowadzi do badań ‘Asia, and now I apologize to you for telling you this joke, but my mother sent to me such a survey which my acquaintance runs for research’ S2 no no? ‘Well well?’

Joke construction and joke structure

185

S1 yy i musiałam to nagrać jak to opowiadam (laughter) ‘And I had to record it how I am telling it’. S2 co ty gadasz? ‘Are you kidding?’ S1 przepraszam cię bardzo (laughter) ‘I am very sorry’. S2 no jak mogłaś? (laughing voice) ‘Well how could you?’ S1 no nie chciałam tego mówić wiesz, że nagrywam, przepraszam cię ‘Well I did not want to tell you, you know, that I was recording, I am sorry’.

4 Conclusions Overall, the study has shown the variety of options open to speakers when performing a joke, some of which would not be easy to predict, notably entertaining a child by organizing a joke performance for her and asking her to perform, telling the joke to a dog, visiting a neighbour (who is also a friend) purely under the pretext telling of a joke, telling a joke to a sick wife while she is washing dishes, or half-reading the written text when joke-telling. In discussing all these cases I have drawn attention to factors which were underemphasized or ignored in the original Semantic Script Theory of Humour. Firstly, the identifying of speaker-hearer relation is crucial when actually deciding to tell the joke and selecting the type of performance (joke-telling situation) and then adjusting the playframe, text, its imagery, prosody and other cues (like the laughing voice9). In this study it was family members or friends who prevailed among joke tellers and hearers, but the influence of the relationship between joke teller and hearer on the shape of the performance was best exemplified in the one fully transcribed joke performance discussed above (this concerned both the preceding and following sequences as well as the jab lines, although this correlation would require a separate study). Secondly, the playframe (non-bona-fide mode in this case) was introduced in various ways in 9 out of 12 jokes (although in two cases the recording did not include it, and in one case – the dog listener – it was obviously not needed), including rather standard ways, except the few cases where the teller wanted to announce the beginning of the recording. Thirdly, the variability of the principal jab line of the joke (the list of diseases) was something expected, and the speaker’s

9 The limited scope of the paper did not allow me to analyse this part of the performances, but excellent research in this area has recently been pursued e.g. by Attardo et al. (2011).

186

Władysław Chłopicki

choices reflected their attempts to embellish the joke in various ways (including the adding of various diseases they considered funnier). Fourthly, the visual imagery evoked by the punchline was found useful in the analysis because a correlation was discovered between the presence of particular image schemas and the successful performance of the joke. Fifthly, the hearer’s feedback took rather standard forms with a few unusual, seemingly critical reactions (“Jeesus”, “Terrible, this was terrible”) or the repetition of the key word of the punch line in an attempt to save face in the situation when the target hearer did not quite understand the joke. Finally, it should be stressed that a couple of the retellings failed due the speakers’ self-admitted lack of experience and dislike of joke-telling, which only confirms Kuipers’ claim about the decline in the frequency of this social pastime. The study thus testifies to the fact that humour does not lie so much in the text of the joke, idealized as it is claimed to be by Raskin, as in its performance. In terms of the title of this study, joke construction (rather than joke structure) refers to the process of joke performance, including various factors that influence it, such as speaker/hearer relations, joke situation, the ways the playframe is introduced and ended, speakers’ choices as to the form of jab lines and joke imagery that influence the actual overlap of two macroscripts, as well as the hearer’s feedback. Some of these aspects do require separate studies that would include more joke material and cross-cultural data and provide systematic evidence.

Bibliography Aharoni, Ron. 2018. Shifting from meaning to its carrier: A common denominator for three strains of humour. The European Journal of Humour Research 6(3): 13–29. Attardo, Salvatore. 2001. Humorous texts: A semantic and pragmatic analysis. Berlin, New York; Mouton de Gruyter. Attardo, Salvatore. 2008. A primer for the linguistics of humor. In Victor Raskin (ed.), The primer of humour research. 101–156. Berlin, New York: Mouton de Gruyter. Attardo, Salvatore. 2017. The GTVH and humorous discourse. In: Władysław Chłopicki & Dorota Brzozowska (eds.), Humorous discourse. 93–105. Boston, Berlin: Mouton De Gruyter. Attardo, Salvatore et al. 1994. The linear organization of jokes: analysis of two thousand texts.Humor: International Journal of Humor Research 7(1): 187–198. Attardo, Salvatore & Victor Raskin. 1991. Script theory of humor (re)visited. Joke similarity and jokerepresentation model. Humor: International Journal of Humor Research 4(3–4). 293–347. Attardo, Salvatore, Lucy Pickering & Amanda Baker. 2011. Prosodic and multimodal markers of humor in conversation. Pragmatics and Cognition 19(2). 224–247.

Joke construction and joke structure

187

Attardo, Salvatore & Victor Raskin. 2017. Linguistics and humor theory. In: Attardo, Salvatore (ed.), The Routledge handbook of language and humor. 49–63. New York, London: Routledge. Bell, Nancy. 2017. Failed humor. In: Attardo, Salvatore (ed.), The Routledge handbook of language and humor. 356–370. New York, London: Routledge. Carrell, Amy T. 1997. Humor communities. Humor: International Journal of Humor Research 10 (1). 11–24. Chłopicki, Władysław & Dorota Brzozowska (eds.). 2017. Humorous discourse. Boston, Berlin: Mouton De Gruyter. Chłopicki, Władysław. 2009. Perceptual imagery in humour processing. In: Elżbieta Chrzanowska-Kluczewska & Grzegorz Szpila (eds.), In search of (non)sense. 179–205. Cambridge: Cambridge Scholars. Chłopicki, Władysław. 2019. What do doctors advise patients in jokes and why? Humor: International Journal of Humor Research 32 (3). 475–497. Chomsky, Noam. 1965. Aspects of the theory of syntax. Cambridge, Mass: MIT Press. Croft, William & Alan D. Cruse. 2004. Cognitive linguistics, Cambridge: Cambridge University Press. Davies, Christie. 2011b. Jokes and targets. Bloomington, Indianapolis: Indiana University Press. Double, Oliver. 2005. Getting the joke: The art of stand–up comedy. London: Methuen. Fry, William. 1963. Sweet madness. A study of humor. Palo Alto, CA: Pacific Books. Hurley, Matthew M., Dennett, Daniel C., & Adams, Reginald B. Jr. 2011. Inside jokes: Using humor to reverse engineer the mind. Cambridge, MA: MIT. Kuipers, Giselinde. 2006. Good humor, bad taste. A sociology of the joke. Humour research 7. Berlin, New York: Walter de Gruyter. Marsh, Moira. 2015. Practically joking. Logan: Kegan Paul. Norrick, Neal. 1994. Conversation joking. Humor in everyday talk. Bloomington: Indiana University Press. Oring, Elliott. 2017. Joking asides. The theory, analysis and aesthetics of humor. Logan: Utah University Press. Raskin, Victor. 1985. Semantic mechanisms of humor. Dordrecht: Reidel. Raskin, Victor (ed.). 2008. The primer of humour research. Berlin, New York: Mouton de Gruyter. Raskin, Victor. 2017. Humor theory. What is and what is not. In Władysław Chłopicki & Dorota Brzozowska (eds.), Humorous discourse. 11–22. Boston, Berlin: Mouton De Gruyter. Ritchie, Graeme. 2018. The comprehension of jokes. A cognitive science framework. New York, London: Routledge. Shilikhina, Ksenia. 2017. Metapragmatic markers of the bona-fide and non-bona-fide modes of communication. In Władysław Chłopicki & Dorota Brzozowska (eds.), Humorous discourse. 107–130. Boston, Berlin: Mouton De Gruyter. Sover, Arie. 2014. Humour and enjoyment reducers in cinema and theatre comedy. The European Journal of Humour Research 2(3). 86–97. Taylor, Julia. 2017. Ontological Semantic Theory of Humor in a context of humorous discourse. In: Władysław Chłopicki & Dorota Brzozowska (eds.), Humorous discourse. 205–218. Boston, Berlin: Mouton De Gruyter. Triezenberg, Katrina E. 2004. Humor enhancers in the study of humor literature. Humor: International Journal of Humor Research 17(4). 411–418.

188

Władysław Chłopicki

Tsakona, Villy. 2017. Genres of humor. In Salvatore Attardo (ed.), The Routledge handbook of language and humor. 489–503. New York, London: Routledge. Żygulski, Kazimierz. 1985. Wspólnota śmiechu. Studium socjologiczne komizmu. Warszawa: PIW.

Appendix Joke telling survey This short survey serves the purpose of linguistic research. It is anonymous and only some basic demographic data are required. It is intended as objective and it does not aim at judging your joke telling skills and habits. Your participation is voluntary and most appreciated. It will help me to write an academic article – if you like to receive the bibliographic information about the article when it appears do let me know. The survey consists of three steps. The first step is described on the next page and involves reading a joke (maximum 5 minutes). The second step requires telling the joke to a person of choice and recording your own performance (only audio recording is needed) (5 minutes including the setup). The third step is most essential – it involves returning to the survey, sending the file with the recording to my email address (preferably the mp3 or mp4 format) and completing the survey (5 minutes altogether) Thank you in advance for your cooperation. Władysław Chłopicki Jagiellonian University Kraków, Poland Now please move on to the instruction on the next page Please read the joke below and remember it. The joke is not very complicated and it will be easy for you to remember it. Do not make any notes and do not read the joke– it does not matter whether you deliver the joke in the same words – just remember the gist and tell it to someone in your own words. The joke is inoffensive and no one should feel bad about it. It does not matter whether you like the joke a lot or just a little – tell it to someone in an attempt to entertain them. I do hope you will like the joke though, at least a little; you can also and are encouraged to embellish the joke to make it better. If you feel you cannot complete this task, I understand, of course. Then please move to the end of the survey, answer a few questions and fill in the comment space if you like.

Joke construction and joke structure

189

If you can do it, however, and I hope you will be able to, please read the joke below and remember the gist of it. THE JOKE A man returns from a trip to Shanghai and is feeling very ill. He goes to see his doctor and is immediately rushed to the hospital to undergo a series of tests. The man wakes up after these tests in a private room at the hospital and the phone by his bed rings. “This is your doctor,” says the voice on the phone. “We have the results back from your test and . . . I’m sorry, you have an extremely contagious deadly disease known as G.A.S.H.” “G.A.S.H?” replies the man. “What in the hell is that?” “It’s a combination of Gonorrhea, AIDS, SARS and Herpes,” explains the doctor. “My goodness, Doc!” screams the man in a panic, “what are we going to do?” “Well, we’re going to put you on a strict diet of pizza, pancakes, quesadillas and pita bread,” says the doctor matter-offactly. “Will that cure me?” “Well, no,” says the doctor, “but it’s the only food that we can slide under the door.” [The original Polish version used in the Polish survey: Państwowy Szpital Kliniczny Nr. 4 w Lublinie. W izolatce rozlega się dzwonek telefonu komórkowego. Mężczyzna podnosi się z łóżka i sięga do szafki. – Czy pan Kowalski ? – rozlega się głos w słuchawce – Tak. – Mówi ordynator oddziału zakaźnego prof. Kwiatkowski. Drogi panie mam już wyniki pańskich badań. Jest pan chory na błonicę, boreliozę, cholerę, czerwonkę bakteryjną, dur brzuszny, dżumę, gruźlicę, kiłę, krwotoczne zapalenie jelit, krztusiec, ospę wietrzną, nosaciznę, odrę, pryszczycę, tularemię, rzeżączkę, wąglik, włośnicę, wszawicę, wściekliznę i AIDS. – Boże – krzyczy pacjent – ale chyba będziecie mnie leczyć ? – Aaa? Tak, tak, oczywiście, będziemy. Już nawet wyznaczyliśmy panu specjalną dietę naleśnikową. – A ta dieta mi pomoże? – Pomoże albo nie. Skąd mam wiedzieć? – odpowiada zdenerwowany profesor – Ale naleśniki to jedyna potrawa, która zmieści się w szparę pod drzwiami] Now, please tell it (in English, of course) to a friend, acquaintance or a family member. Please try to make the joke telling situation sound natural. You can tell your friend beforehand that you are going to record the joke performance – you can also tell them afterwards (this is preferable). Please record a few moments prior to telling the joke and a few moments following the telling, so that the conversation situation is complete.

190

Władysław Chłopicki

After you have recorded the joke, return to the survey and move on to the last page. Please send your file to the following email address – w.chlopicki at uj.edu.pl (this survey system does not allow audiofiles to be attached – you can send your file from any gmail or another anonymous address – there is no need to provide your name). After you have done it please fill in the following questions. 1. 2. 3. 4.

What is your age? 10–20, 21–40, 41–60, over 61 What is your gender? male, female What is your nationality/country of residence? Is English your native language? Yes, no, other (please specify) [not in the Polish version of the survey] 5. Whom did you tell the joke? a close friend, an acquaintance, a person of less than 18 years of age, a family member, other (please specify) (several options possible) 6. When did you tell your friend or acquaintance that you were recording the joke? beforehand, afterwards, other (please specify) 7. Please comment on your experience. You can also comment on the joke itself.

Béatrice Priego-Valverde

‘Stop kidding, I’m serious’: Failed humor in French conversations Abstract: This article investigates humor as non-bona-fide communication, in the case of humor produced by the recipient, while the main speaker is engaged in a serious storytelling, i.e. bona-fide communication. Taking into account the various interactional constraints weighing in on participants’ roles and actions (humor included), and considering that these constraints are even more active in serious storytelling, the aim of this article is to investigate the switch from BFC into NBFC as a reason of failed humor. To do so, 105 instances of failed humor were collected from three audio and video recorded conversations taken from the CID corpus. Among them, 56 instances of failure are due to a switch from BFC into NBFC. And among those, 38 were produced in a storytelling context. Keywords: failed humor, conversation, (non) bona-fide communication, storytelling; (non) cooperation

1 Introduction The goal of this article is to apply Raskin’s (1985, 1992a, 1992b) distinction between bona-fide communication (henceforth, BFC) and non-bona-fide communication (henceforth, NBFC) to failed humor in conversation. More precisely, it is to investigate the consequences of a switch from BFC to NBFC, from the hearer, both on the humor itself and on the interaction. Concerning (N)BFC, this study will focus on one specific case highlighted by Raskin (1985: 102) i.e. when the speaker produces humor when the hearer does not expect it. Raskin suggests that such a mismatch between the two modes of communication may not only lead to a comprehension issue (1985: 101), but also to an “infelicitous speech act” (Raskin 1992a: 87). For the present study, the positions of speaker and hearer will be inverted: we analyzed the humorous utterances produced by the participant who is initially in the position of hearer. In other words, we analyzed the introduction of a NBFC by the hearer in a speaker’s BFC. Moreover, only the cases of switch from BFC into NBFC which led to infelicity were analyzed, eliminating those which led instead to a comprehension issue. Béatrice Priego-Valverde, Aix Marseille Univ, CNRS, LPL, Aix-en-Provence, France

https://doi.org/10.1515/9781501511707-010

192

Béatrice Priego-Valverde

Analyzing humor in conversation implies taking into consideration the various constraints conversation puts on participants. Moreover, focusing on storytelling in conversation, which obeys a different organizational principle than that of the conversation in which it is embedded, if only by being an asymmetrical activity, implies the postulate that the specific organization of storytelling impacts both the production of humor and its reception. Indeed, such an asymmetrical activity distributes specific interactional roles to the participants: a teller (or main speaker) and a recipient (the hearer), with related actions. Considering humor as a NBFC, this study investigates the consequences of the introduction, by the recipient, of a humorous utterance, into a serious storytelling delivered by the teller. This study will focus on the teller’s negative reactions. Three different types of responses are highlighted: 1) serious response to the utterance; 2) ignoring the utterance; and 3) explicit rejection of the utterance. They are presented as a continuum according to the degree of non-cooperation the speaker displays faced with the recipient’s humorous utterance(s). The responses are analyzed using an interactional approach, which will show that such a switch from the recipient, from a BFC into a NBFC, has two kinds of consequences: on the humor itself, which can then fail; and on the interaction which can then become conflictual.

2 Theoretical background In this section, an overview of Raskin’s work on (non)-bona-fide-communication, storytelling and failed humor will be presented.

2.1 Bona-fide communication vs non-bona-fide communication The notion of non-bona-fide communication is presented in Raskin’s (1985) chapter 4 of Semantic mechanisms of humor. This chapter is undoubtedly the one which has the most deeply impacted the linguistic field of Humor Studies because it is devoted to the notion of Script Opposition, the cornerstone of Raskin’s Semantic Script Theory of Humor. In this chapter, as a sort of preamble (an interruption, in Raskin’s words: 1985: 100) in order to understand the Script Opposition, Raskin presents joke telling as a non-bona-fide communication (1985: 100–104) in contrast to the bona-fide communication introduced by Grice (1975).

‘Stop kidding, I’m serious’: Failed humor in French conversations

193

As presented by Grice (1975), BFC corresponds to a serious register of communication entirely governed by truth and relevance, two fundamental elements allowing participants to make and give meaning to what they say and hear. More precisely, this serious mode of communication is governed by the “Cooperative Principle” (Grice 1975, henceforth CP) constituted by 4 conversational maxims (Maxim of Quantity, Maxim of Quality, Maxim of Relation and Maxim of Manner, Grice: 1975: 45–47). These maxims prescribe a way of speaking that is effective and that offers the hearer a basis to interpret what is being said. In both cases, these maxims highlight the expectations weighing in on the participants’ talk. According to the CP, being a cooperative participant means to respect these maxims. Thus, it can be seen as a framework of production and reception which, paradoxically, may be considered a necessary and helpful straitjacket: Make your conversational contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged. One might label this the Cooperation principle. (Grice 1975: 45).

Raskin insists that BFC is entirely tuned to coherence; that is, on the necessity for participants, “to give meaning to” something: Bona-fide communication is governed by the ‘co-operative principle’ introduced by Grice (1975). According to this principle, the speaker is committed to the truth and relevance of his text, the hearer is aware of this commitment and perceives the uttered text as true and relevant by virtue of is recognition of the speaker’s commitment to its truth and relevance [. . .], (Raskin 1985: 100–101).

By virtue of the same need for coherence which makes a conversation successful, Raskin opposes to the Gricean CP, a “cooperative principle for the nonbona-fide communication” (1985: 103), constituted by the same 4 maxims but applied to joke telling in order to make joking successful: (i) (ii) (iii) (iv)

Maxim of Quantity: Give exactly as much information as is necessary for the joke Maxim of Quality: Say only what is compatible with the word of the joke Maxim of Relation: Say only what is relevant to the joke Maxim of Manner: Tell the joke efficiently (Raskin 1985: 103)

In other words, both bona-fide communication and non-bona-fide communication are presented through their components (the maxims) in order to make a message – serious or humorous – successful. In this respect, both BFC and NBFC are cooperative (Raskin 1992a: 87). Furthermore, the reason why Raskin proposes a cooperative principle of non-bona-fide communication is precisely to highlight

194

Béatrice Priego-Valverde

the cooperative aspect of humor or joke telling. In any case, Raskin considers NBFC as subordinate to BFC which would be primary (Raskin 1998: 106). NBFC is thus the counterpart of BFC, imposing on the participants the same kinds of constraints with the same goal of coherence; that is, the success of joke telling for the former, and of serious communication for the latter. However, Raskin goes further. Considering that these two registers can be present in a communication, he reminds us that speaker and hearer may or may not be in the same mode of communication. He considers 4 possibilities: (110) (111)

(i) (ii) (i) (ii)

The speaker makes the joke unintentionally The speaker makes the joke intentionally The hearer does not expect a joke The hearer expects a joke

(Raskin 1985: 100)

Mixing these four possibilities results in eight different situations where speaker and hearer can agree or not on the mode of communication in which they are engaged or in which they want to be engaged. In this article, I will focus on one specific situation (110ii–111i) i.e., the situation where “the speaker throws a joke on the hearer unexpectedly for the latter” (Raskin 1985: 102). According to Raskin, in (110ii), humor is produced intentionally in order not to convey any information, but rather, to create a certain effect, i.e. “to make the hearer laugh” (Raskin 1985: 101). In (111i), the hearer does not expect any joke because he is engaged in BFC. In this latter case, Raskin explains what the consequences of a joke on the hearer’s perception may be: [. . .] the hearer does not expect a joke and will initially interpret the speaker’s utterance as conforming to the requirements of bona-fide communication. After his attempts to interpret the utterances within this mode fail, he will have to look for an alternative way to interpret it, and this will bring him into the joke-telling mode because, in our culture, joke telling is a much more socially acceptable form of behavior than, for instance, lying and a more frequent form of behavior than, for instance, play acting (Raskin 1985: 101).

As shown by Raskin, not understanding or mis-understanding an unexpected humorous utterance is unquestionably one of the reasons for failed humor.1 And unquestionably too, this not understanding or mis-understanding can be explained, at least partially, by the fact that NBFC does not respect the four maxims of the Cooperative Principle. Once again, the Cooperative Principle of

1 Although they are not analyzed here, some instances appear in the data on which this article is based.

‘Stop kidding, I’m serious’: Failed humor in French conversations

195

NBFC is relevant to highlight the necessity each participant has to find a meaning to the current conversation. However, delivering a joke properly does not guarantee its success. As Raskin said (1985: 103): “[. . .] the hearer of the joke can fail to get it even if the speaker provides all the necessary ingredients and follows all the maxims.” According to him, the reason why a joke can fail despite the respect of the maxims is the presence or absence of two opposite and compatible scripts in the joke: [. . .] the maxims [of the cooperative principle for the NBFC], while shedding some lights on the semantics of humor, do not really provide an explicit account of the semantic mechanisms of humor. The latter are, of course, based on the scripts and combinatorial rules of the scriptbased semantic theory and on the relation of script oppositeness. (Raskin 1985: 103–104).

Indeed, Raskin’s main hypothesis is the full – or part – compatibility of a text with two opposite scripts (Raskin 1985: 99). Both this compatibility and this oppositeness are presented as “necessary and sufficient conditions for a text to be funny” (Raskin 1985: 99). Doing so, Raskin makes the Script Opposition the cornerstone of his semantic theory of humor, and of humorous competence. Moreover, Raskin (1992a) increases the importance of the notion of NBFC by adding two important elements. First, considering more deeply the case in which participants can be in different modes of communication, he evokes the case where one participant voluntarily switches from BFC into NBFC, thus creating an infelicitous speech act: “Such a deviation from cooperation cannot be sustained for a long time, and it will be totally infelicitous as a speech act if nobody cooperates with the speaker on it” (Raskin 1992a: 87). Without naming it as such, Raskin evokes a potential conflict between the two modes of communication. Second, confirming the importance which has to be given to NBFC, Raskin considers the introduction of a fifth maxim to his joke telling Cooperative Principle: JT CP may have other maxims as well. Thus, one maxim unparalleled in BF MC CP is the maxim of mode switching which says something like, “Signal to the hearer that you are switching to JT MC.” (Raskin 1992a: 87).

Summing up Raskin’s words focusing on the importance of the particularities of BFC and NBFC, we can see the following: In his early work (1985), Raskin considers (i) Simultaneously, the similarities and specificities of the two modes of communication: while BFC is tuned to truth and relevance, NBFC is tuned toward a certain effect on the hearer (basically, laughter) (ii) Both are tuned to coherence in order to be successful but in different ways

196

Béatrice Priego-Valverde

(iii) The necessity of coherence explains the presence, in both modes of communication, of a cooperative principle based on the same maxims. One is applied to the serious mode of communication, and the other is applied to the joke telling (iv) The case where two participants can be engaged in different modes of communication (110ii–111i) is considered (v) But the only possible consequence considered of such coexistence is an impact on the comprehension of humor by the hearer, not the consequences on the interaction itself (vi) Which leads Raskin to insist on the importance of the Script Opposition More recently, Raskin (1992a) deepens the analysis of humor as a NBFC: (i) He increases the impact that the switch from BFC into NBFC may have on the success of humor (ii) He deepens the analysis of the (110ii–111i) case considering that such a mismatch can lead to a “infelicitous speech act” (iii) This possibility leads Raskin to consider introducing a fifth maxim in the cooperative principle of NBFC Such a work leads Raskin (1998) to ‘advocate’ for a revaluation of NBFC which is in any case subordinated to BFC which would be primary. However, although highly attractive, the idea according witch BFC and NBFC are equal has to be qualified. Needless to say, both are highly organized (in their own way) and considering both equals is a principled position meaning that humor is, in any case, less important than serious discourse. While Raskin’s position is totally acceptable, the data presented here shows another picture. Participants, while they are in conversation, do not always accept to switch into NBFC while they are engaged in BFC. In other words, when two different modes of communication appear simultaneously, only one can remain possible: Since the bona fide and non-bona fide modes require different sets of interpretation rules, the dialogue cannot be continued successfully (or, at least, without leading to an interpersonal conflict) until a common mode is established. (Shilikhina 2018: 71).

Analyzing the coexistence of these two modes of communication in genres other than conversations, Shilikhina (2018) shows that both have to be framed enough to be recognized as belonging to one or the other mode, and both have to be negotiated and might be rejected. This notion of negotiation converges with Raskin’s (1992b) study on “non casual language”. He points out that while BFC is an idealization, the notion of cooperation “seems to be a more essential and more realistic phenomenon than BF.” (1992b:24). And indeed, focusing on

‘Stop kidding, I’m serious’: Failed humor in French conversations

197

the mismatch between the two modes of communication, the present article also highlights the importance of the notion of cooperation per se. To sum up, focusing on Raskin’s work on humor as NBFC and applying it to conversations, I will analyze instances of failed humor in conversation. To this end, I will simultaneously consider two different but interconnected phenomena: humor as NBFC on the one hand, and the conversation itself through the interactional constraints it imposes on the participants, on the other hand. More specifically, I will analyze instances of failed humor appearing in a specific conversational activity: storytelling. If a conversation is generally considered a symmetrical activity, storytelling is not. It is an asymmetrical activity where participants have specific roles – of teller and recipient. I will show that not only do they have these specific roles, but moreover, they sometimes want to keep them as such. Leaving apart the fact that a voluntary switch from BFC into NBFC can lead to a comprehension issue, I will follow Raskin (1992a) in saying that such a switch can lead to an “infelicitous speech act”. Even if one may consider that the more this switch is framed as such, the more it may be accepted by the speaker, the data also show that it is not always sufficient. In this article, I will suggest that, a voluntary switch from BFC into NBFC: (i) Can create more than an “infelicitous speech act” including a real interactional conflict between the participants (Priego-Valverde 2003) (ii) Has indeed to be signaled, but this signal does not guarantee its acceptance a. Because the signal is not clear enough b. Because, even in the presence of a clear signal, the other participant does not still accept the switch In other words, agreeing with the fact that the presence of two compatible and opposite scripts present in an utterance are necessary to make it humorous, it is not sufficient. In conversation, I suggest that another element is necessary for humor to be successful: its legitimacy.

2.2 Storytelling in conversation Since the seminal work in Conversation Analysis (Sacks et al. 1974), the interactional machinery of a conversation has been analyzed in depth. Equally,the literature about storytelling in conversation is more than substantial. Besides the general organization of storytelling, (see Labov and Waletzky 1966; Jefferson 1978; Norrick 2000; Stivers 2008; Mandelbaum 1989, Mandelbaum 2013; Selting 2000; Bavelas et al. 2000), specific kinds of storytelling have also been studied

198

Béatrice Priego-Valverde

such as complaints (see Mandelbaum 1991/1992; Selting 2010, Selting 2012) and humorous storytelling (see Sacks 1974; Selting 2017). In this section, I will focus on some relevant work on storytelling in conversation.

2.2.1 General storytelling organization Various structural particularities of storytelling that differ from the organization of the conversation in which the storytelling appears have been highlighted. First,2 because storytelling is considered an “extended telling” (Schegloff 2007), or a “large project” (Selting 2000), it has a specific sequential organization which diverges from the canonical adjacency pair and “during which participants are oriented to a suspension3 of normal turn-taking rules” (Stivers 2013: 200). The turn-taking system relevant to storytelling is analyzed more in depth by Mandelbaum (2013: 493): [. . .] the telling of a story usually requires extended turn-at-talk on the part of the teller, and a passing up of the opportunity to take turns, on the part of the recipient – a suspension of the ordinary turn-taking arrangement of conversation that guarantees a speaker only one turn-taking unit of talk [. . .].

In other words, because participants are in storytelling, the teller may legitimately require the right to extend his/her own turn while the recipient may have the right to pass on his/her turn. Second, the goal of a storytelling is different from the rest of the conversation in which it is inserted, and more particularly, from other adjacency pairs: while an adjacency pair is usually defined in terms of actions (such as question/answer), a storytelling is oriented to the establishment of a stance toward an event (Stivers 2013). These two specificities combine to create a third one: storytelling is an asymmetrical activity (Stivers 2008; Guardiola and Bertrand 2013) which not only impacts the conversational turn-taking system, but allocates to the participants specific interactional roles with related actions.

2 It is worth noting that while all the structural specificities of the storytelling are necessarily presented here successively, they are in fact highly interrelated. 3 My emphasis

‘Stop kidding, I’m serious’: Failed humor in French conversations

199

2.2.2 Interactional roles of teller and recipient Following Stivers (2008: 34), the teller has to keep the floor “until story completion” in order to convey a stance for the recipient, while the recipient has to align (i.e. to react in a way which facilitates the ongoing narrative) and to affiliate (i.e. to adopt or support the teller’s stance) with the teller. This is the reason why the participants are divided in teller (or “main speaker”, Guardiola and Bertrand 2013) and recipient. Cooperation between teller and recipient But if participants’ roles and actions are specific, they are also deeply complementary: “[. . .] both participants actively participate and work together to construct the narrative.” (Guardiola and Bertrand 2013: 2). From the teller side, conveying a specific stance toward the event s/he relates is not only a duty, it also provides help both for him/herself and for the recipient. The teller has to deliver a stance in enough of a recognizable way to ensure, as far as possible, both understanding and agreement from the recipient, i.e. his/her alignment and affiliation: The teller’s stance, and the extent to which this is conveyed, is a crucial resource for recipients, as it makes available the teller’s expectations regarding how the events of the storytelling are to be responded to. (Mandelbaum 2013: 499).

Concomitantly, recipient’s responses are “crucial” (Mandelbaum 2013: 501) for the progress of the teller’s storytelling. Recipient’s responses can be thus divided into two categories. While a “preferred response” (Jefferson 1978) “mirrors the stance that the teller conveys (Stivers 2008: 33), a dispreferred response may disrupt the ongoing telling. In other words, storytelling is co-constructed. Non-cooperation between teller and recipient Cooperation between teller and recipient promotes the progress of the storytelling. Each participant is, in some way fulfilling his/her own role and duty. By contrast, a non-cooperation between participants may endanger the ongoing activity. This non-cooperation may come from either the teller or the recipient. While the teller may not deliver enough information for the recipient to follow and/or to give him/her the possibility to participate, the recipient may also disrupt the storytelling. A recipient’s disruption can be due to a disaligned answer such as competing for the floor, and / or a disaffiliative answer, i.e. not adopting the teller’s stance (Stivers 2008).

200

Béatrice Priego-Valverde

Mandelbaum (1991/1992), highlights various recipient’s devices allowing him/her to disalign and/or disaffiliate from the teller, which the author labels “disattending.” Thus, acting on the structure itself of the storytelling, the recipient can point out a detail of the teller’s discourse to focus on it and to make it the new topic. S/he can also react by talking about him/herself instead of responding to the teller’s story. Finally, acting more on the teller’s stance, the recipient can react proposing another stance. All these disattended answers are a sign of non-cooperation that have to be managed by the teller: In the course of working this out, the original speaker of the disattended matter has the option of pursuing it, or of dropping it, perhaps in favor of pursuing something else, or pursuing it later. (Mandelbaum 1991/1992: 135).

2.2.3 Serious storytelling vs humorous answer The common point of many of these studies, at least of the recent ones, is the essential role of the notion of stance perceived as a reference point thanks to which a storytelling may be considered successfully achieved or not. On the one hand indeed, the notion of stance characterizes storytelling as a conversational activity in its own right, with its specific organization: it makes it a particular activity tuned toward its delivery and not toward actions. On the other hand, this notion of stance dictates participants’ actions and duties: it has to be successfully delivered by the teller and agreed on or adopted by the recipient. However, while this essential role of the stance has been carefully analyzed and demonstrated, and disaffiliative answers have also been studied, one specific case concerning storytelling has received less attention thus far. It concerns the teller’s serious stance to which the recipient opposes a humorous stance rather than adopting a serious one. To the best of my knowledge, when this case is considered, the humorous answer is positively perceived, i.e. as an “embellishment” of the story (Goodwin 1997: 80). The present article investigates this specific case of a recipient’s humorous answer to a serious storytelling and highlights the possibility that humor is not always considered an embellishment by the speaker.

‘Stop kidding, I’m serious’: Failed humor in French conversations

201

2.3 Failed humor in conversation While researchers in Humor Studies agree that failed humor was neglected for a long time, they note that important work has been conducted on failed humor since the 1990s. These studies can be divided into three categories. The first category concerns studies about reactions to humor – or teasing and/or irony – (among others, Hay 1994, Hay 2001; Eisterhold et al. 2006; Attardo 2002; Bell 2009). In such studies, without always explicitly naming humor as failed, negative reactions to it have been highlighted, making, de facto, humor which triggered such answers a form of failed humor. In these studies, failed humor is not the observable phenomenon per se. The second category of studies consists of transposing a framework of successful humor to failed humor, and thus determining the reasons for the failure. As one example, Bell (2015) offers a reinterpretation of Hay’s model of successful humor (2001). Thus, “humor may fail when it is a.) not recognized, b.) not understood, c.) not appreciated, d.) not fully agreed with, and e.) not engaged with.” (Bell 2015: 28). This reinterpretation has the merit of integrating interactional reasons to failures. But here again, failed humor is addressed through successful humor. Finally, the third category of studies focuses on the reasons of the failure, but considers failed humor as an observable phenomenon per se. Various models are proposed, listing many possible reasons for the failure. Focusing on humor in interaction, Hay (1995 and 2001) lists eight reasons why humor can fail, from the analysis of the humorous item per se, to the impact such an item may have on the ongoing interaction (such as disrupting it). More recently, Bell and Attardo (2010) have proposed another framework, which was revisited by Bell (2015). These two last models, probably the most complete thus far, integrate different linguistic levels (such as semantic, pragmatic, lexical), associated to the analysis of humor per se (through the notion of incongruity) and taking also into account social and interactional functions of failed humor. In her revision of the Bell and Attardo (2010) model, Bell (2015) integrates the question of the humorous frame, which can be considered NBFC. However, although essential for my own analysis in the present paper, the frame is considered in this model as an element which can inhibit comprehension and not as an element which can be refused. However, this question of acceptance or refusal of the frame has been highlighted by Priego-Valverde (2003) and PriegoValverde (2009) within a Bakhtinian approach of (failed) humor. Thus, in Priego-Valverde (2003) and Priego-Valverde (2009) the main reason advanced to explain the failure of humor was the fact that sometimes, because of the way humor is delivered, it might render a speaker’s utterance so opaque that the

202

Béatrice Priego-Valverde

hearer does not even know if s/he is faced with humor or not, i.e., s/he does not know which frame they are in. Although these studies inevitably include some gaps, taken together they offer a robust image of what failed humor is. Consequently, in this article, I do not want to attempt to propose another model. I will just pull some strands out of this previous work and will link them to some interactional constraints weighing in on the participants of a conversation. For this reason, I will deliberately exclude the comprehension level – which is however essential in failed humor – in order to focus on one interactional reason of failure: the more or less obvious rejection of humor because it is seen as non legitimate or parasitic. To sum up, in this article, I will analyze instances of failed humor in conversation, and more precisely, instances of failed humor when participants are engaged in a serious storytelling. This specific setting simultaneously takes into account various elements of storytelling in conversation: (i) That conversation is, in principle, a symmetrical interaction, that allows the hearer to feel free to intervene into the speaker’s talk. All the more so when all the participants are good friends (see Section 3) (ii) But storytelling, seen as a conversational activity, is an asymmetrical one. The conversational rules are thus different, as well as the participants’ actions. Because a participant is the main speaker (or teller) and the other the recipient (or hearer), the former has the right to keep the floor longer, while the latter has to produce preferred responses at specific moments. These specific responses must be aligned and affiliated, and hence, facilitate the ongoing storytelling (iii) If the recipient provides a disaligned and/or disaffiliative answer, the ongoing storytelling might be disrupted. Applied to the examples of my data, i.e. a humorous intervention from the hearer, while the speaker in engaged in a serious storytelling: (i) The hearer does not provide a preferred response and he thus disaligns with the teller, allocating himself a turn which does not belong to him, which may create a conflict over control of the floor (ii) He also disaffiliates, adopting another stance or, switching in another mode of communication. In doing so, he does not collaborate to the ongoing activity (iii) He takes the risk, not only to disrupt the ongoing storytelling produced by the teller, but, as it is the case in the data, to see his humorous attempt more or less obviously rejected for two reasons. First, because he left his role of hearer, granting him the right to be a speaker, or at least, a cospeaker of the telling. Second, because he also imposed a different frame of the telling, avoiding adopting the serious teller’s stance.

‘Stop kidding, I’m serious’: Failed humor in French conversations

203

3 Corpus and methodology 3.1 Description of the corpus The “CID” – Corpus of Interactional Data – (Bertrand et al. 2008) includes 8 onehour French face-to-face interactions. They were recorded in an anechoic room at Aix-Marseille University, France. Each speaker was wearing a microphone and both were filmed by a single camera. All the participants were French native speakers and members of the university (students, scholars, or staff members). Two tasks were performed: half of the participants had to tell unusual stories while the others had to relate stories about professional conflicts they had experienced. Despite this protocol and the setting, this data is considered as very close to natural conversation because of the context: the duration of each recording (one hour) allowed the participants to deviate from the initial task; they were familiar with the anechoic room and, most importantly, the participants in each pair were close friends outside of the university. For the present article, three interactions taken from the corpus were studied; two about conflicts (EB_SR; AP_LJ) and one about unusual stories (AG_YM). All involved only male participants.

3.2 Annotation scheme methodology used Each audio signal was automatically segmented in Inter-Pausal Units (IPU), i.e. blocks of speech bounded by silent pauses over 200 ms, and time-aligned to the speech signal. An orthographic transcription was then provided at the IPUlevel to include all phenomena occurring in spontaneous speech (such as hesitations, repeats, etc.). The transcription used was the Enriched Orthographic Transcription (OTIM project, Blache et al. 2009). Using Praat (Boersma and Weenink 2009), systematic annotations related to the different linguistic domains were also provided. Only manual orthographic transcription and automatic detection of laughter were used in this study. The location of humorous instances was manually annotated and information related to the humor was shown on seven different tiers: – Tiers 1 and 2: Participants’_IPU – Tier 3: The whole humorous sequence – Tier 4: Successful or failed humorous item(s) produced by Participant 1, in the humorous sequence – Tier 5: Successful or failed humorous item(s) produced by Participant 2, in the humorous sequence

204

Béatrice Priego-Valverde

– Tier 6: Humorous items produced in storytelling specifically – Tier 7: Reasons of failure (non-understanding, switch from BFC into NBFC) In order to investigate failed humor due to a switch from BFC into NBFC by the recipient while the teller is engaged in a serious storytelling, three methodological restrictions were enforced. Specifically, only the following instances of humor were analyzed: (i) failed humorous items produced in a storytelling; (ii) by the recipient and not by the main speaker; (iii) only when the reason of the failure identified was a switch from BFC into NBFC.

4 Description of quantitative data In this section, a quantitative description of the data will be presented.

Figure 1: Number of failed and successful humorous items by interaction.

Figure 1 highlights the very high number of occurrences of humor in these onehour conversations. This is due to a methodological choice. In order to investigate both the way the speaker reacts to recipient’s humor and the interactional consequences of such a humorous item, all the humorous items were extracted and not only the humorous sequences, as a whole.

‘Stop kidding, I’m serious’: Failed humor in French conversations

205

More importantly, this figure shows a high degree of variation concerning the total number of humorous items produced in each interaction. AP_LJ produce much more humor than the two other pairs regardless of the fact that AP_LJ were telling conflict narratives rather than unusual stories. This result may be explained by the nature of the relationship between the participants. While all are friends who know each other outside the university, AP and LJ are the two participants who knew each other the most. They have been very close friends and coworkers for many years and share a lot of non-work-related activities. This figure shows also high variation concerning the amount of failed humor. Here again, AP_LJ produce much more failed humor than the other participants. The difference is clearer in figure 2 which indicates the percentage of the humorous productions in each interaction.

Figure 2: Percentage of failed and successful humorous items by interaction.

In this figure, the high variability of the results appears more clearly. A third of the humorous items produced by AP_LJ fail. This percentage represents half of their own humorous productions which are much higher than the other participants’ production. The impact of the nature of their relationship is here questionable. Indeed, this result suggests that if the level of mutual acquaintance may be correlated to the amount of humor produced, it is not also correlated to the success of humor. In other words, a high level of acquaintance not only

206

Béatrice Priego-Valverde

does not guarantee the success of the humor, but at least in these data, it leads to a very high percentage of failure. Rather, the participants appear to act as if good friends would feel authorized to reject humor or to show explicitly their lack of understanding. In other words, because of their closeness, they appear to act as if the importance of maintaining “face” (Goffman 1967) is thus diminished (Priego-Valverde 2003). This last interpretation seems to be confirmed by figure 3 which shows the percentage of failed humor because of a switch into a NBFC.

Figure 3: Percentage of failed humor due to a switch into NBFC.

The reasons for failed humor other than a switch from BFC into NBFC are beyond the scope of this paper and are not discussed in detail. Mainly, they concern understanding issues. Figure 3 shows that while the results are more homogeneous among the three interactions, once again, AP_LJ stand out from the other pairs. While the reasons for failure in the cases of AG_YM and EB_SR concern more questions of understanding, with AP_LJ, the main reason is a switch from BFC into NBFC. In other words, the problem between them is less the fact they do not understand humor, but rather that they reject it. Such a result, highlighting the switch as the main reason, seems to confirm that the more participants know each other, the less significant the maintenance of face becomes. However, figure 4 highlights another possible reason: the kind of conversational activity in which participants are engaged.

‘Stop kidding, I’m serious’: Failed humor in French conversations

207

Figure 4: Percentage of failed humor due to NBFC, produced within or outside of storytelling.

This figure shows various results. First, there is a major difference between the distribution of failed humor according to the type of conversational activity where it appears. EB_SR did not produce any failed humor in storytelling. All the instances of failed humor in their exchanges are produced in other kinds of activity, mainly, argumentation. Second, the results concerning AG_YM’s production is more homogeneous. While their failed humorous items are more frequently produced outside of storytelling, the results remain quite balanced. Here again, the most remarkable results concern AP_LJ. The majority of their production of failed humorous items is produced in storytelling. Such a result seems to confirm previous analysis of storytelling (see Section 2.2 as an asymmetrical activity where participants are distributed as main speaker (the teller) and recipient, with related roles, duties and actions. It seems that if the main speaker is engaged in a serious storytelling, the recipient cannot easily switch into another mode of communication which has not been chosen by the teller. As will be shown in the next section, such a switch would be considered disaligned and disaffiliative. In other words, even in order to produce humor, it would be better – not to say expected – for the recipient to stay in the teller’s chosen mode.

208

Béatrice Priego-Valverde

5 Sequential analysis of some examples In this section, only instances of failed humor appearing in two conversations will be analyzed. This is because when EB_SR produce failed humor (much less than the other pairs), none of their instances are produced in storytelling. Two examples from AG_YM will be analyzed, and the conversation the most frequently represented in this analysis will be AP_LJ. First, because, as shown in the previous section, these two participants are the most prolific, both in humor and in failed humor. Second, because it is between them that the switch into NBFC creates interactional conflicts. The six examples of failed humor analyzed here are produced by the recipient while the teller is engaged in a serious storytelling. In other words, the recipient not only switches into a NBFC while the speaker is in a BFC, disaffiliating with the speaker, but also disaligns, taking upon himself to produce a speech turn at a moment when the speaker’s activity (storytelling and more, specific moments of the storytelling) does not allow it. In front of such “non-authorized” answers, the speaker reacts differently: from awareness of the hearer’s utterance, but in a serious mode, through unawareness, to an explicit rejection of it. Thus, the examples will be presented through a continuum according to the degree of non-cooperation the speaker displays when reacting to the recipient’s humorous item(s).

5.1 Humor utterances answered seriously In this sub section, two examples of failed humor will be analyzed. In both cases the main speaker responds through a “po-faced” answer (Drew 1987): i.e., the disaffiliative turns are taken into account, but they are responded to in a serious way; that is, the same serious mode in which the speaker is already engaged. In this case, the hearer’s utterances are taken into account, integrated into the speaker’s talk (since he reacts) but their humorous dimension is disregarded. Whether the main speaker recognizes the hearer’s humorous intention as such or not, he does not display any form of recognition (Drew 1987). That is the reason why these two examples are considered as failed humor. Example 1: The student / L’étudiant) 1 AG: et + assez assez agé tu (v)ois il d(e)vait avoir la [quarantaine, quarantaineu] donc c’(é)tait 2 l’étudiant tu vois qui venait faire euh 3 YM: attardé mh @

‘Stop kidding, I’m serious’: Failed humor in French conversations

209

4 AG: pas attardé si tu veux mais tu vois qui gen(re) à quarante ans en inde / il avait besoin de 5 faire euh un euh un certificat ou un truc il venait là 6 YM: ouais 1 2 3 4 5 6

AG: and + quite old you know he should be in his forties so he was a student you know who came to er YM: retarded mh @ AG: not retarded if you want but you know who kinda at 40 years old in India he needs to have a er a er a certificate or something he came there YM: yeah

AG is recalling student memories. When he was student, he lived for a short time in a foreign country, renting an apartment with other foreign students from various kinds of countries and of all ages. In (1), AG is in the beginning of his storytelling, the orientation phase (Labov and Waletsky 1966) where he only sets the scene. In this orientation phase, AG talks about one of the students who was around 40 years old. Line (3), YM, the recipient, latches onto the student’s age and qualifies him as “retarded”. In doing so, he plays on the double meaning of the French word “attardé” (“late” and “retarded”). Framing his utterance as obviously humorous (he laughs at the end), YM signals he is switching into a NBFC and thus, that the meaning “retarded” has to be chosen. But, as Stivers (2008) shows, at the beginning of a storytelling the preferred responses produced by the recipient are more or less feedback only produced in order to help the progression of the storytelling, until “story completion.” By producing, at that moment, a commentary on AG’s character and moreover, in a humorous mode, YM disaligns and disaffiliates with AG. In (4–5), AG shows this disalignment when he reacts seriously to YM’s humorous utterance. By refusing its humorous dimension, he confirms he is, and he wants to stay in a BFC. YM’s utterance is also highly disaligned. Pointing out a detail in AG’s story (Mandelbaum 1991/1992), YM clearly disrupts the ongoing telling. AG is indeed forced to interrupt his telling, to “excuse” the Indian student and to justify why, at around 40, he is still a student. In (6), with a feedback response, YM accepts both AG’s explanation and the assertion of his serious mode of communication. With YM “back into line”, the storytelling can continue. Example (2) shows a similar pattern, but one in which the recipient’s humorous utterance is produced as an overlap.

Example 2 : Italians are good searchers / ils fouillent bien les Italiens 1 LJ :il casse et en plus il a un peu tendance à s’acharner quoi moi je me souviens y avait un 2 fouilleur italien qui était venu 3 AP : mhm

210

4 5 6 7 8 9 10 11 12

Béatrice Priego-Valverde

LJ : euh bon il était spécial et euh ben pas AP : ils fouillent bien les italiens hein AP : ils ont une réputation de fouilleurs LJ : pas trop parce qu’ils ont des ouvriers euh payés AP : @ d’a LJ spécialisés LJ : su pu sur les chantiers de fouille LJ : qui sont là pour faire ça quoi AP: mhm

1 LJ: he breaks and in addition he tends to persist so I do remember there was an Italian 2 searcher who came 3 AP: hm 4 LJ: err well he was special and err well not 5 AP: they search well the Italians don’t they 6 AP: they have a reputation of searchers 7 LJ: not that much because they had paid workers err 8 AP: @ ok 9 LJ: specialized 10 LJ: on the excavation sites 11 LJ: who are there to do that so 12 AP: mhm

When he was younger, LJ, the main speaker, studied Archeology. Used to going to various excavation sites, he is telling a story about one of his previous experiences and, more specifically, he is talking about an Italian man who was there at the same time than him. In lines (1 to 4), after having set the scene and, more particularly, the Italian character of his story (not transcribed here), LJ is engaged in the complication phase (Labov and Waletsky 1966) where, as main speaker, he explains the situation to the recipient. Similarly to the orientation phase, the complication phase allows the teller to present the event, before the climax (Selting 2017). In (3), AP aligns with LJ by producing afilliating feedback; however, in (5–6), he overlaps with LJ’s disfluent utterance in order to produce two humorous items. AP’s disalignment and disaffiliation are inseparable here: AP overlaps twice on LJ’s talk, while LJ has difficulties with his own talk. AP neither helps the speaker to find his words nor stays silent, but instead switches into a NBFC. Moreover, AP’s utterances are irrelevant and are totally meaningless. One may think that such a switch is produced both to make fun of the Italian character depicted as “special”, and to destabilize LJ. As in the previous example, LJ reacts seriously to AP’s utterances in line 7. And again, this serious answer in order to going back to a BFC, forces the speaker to interrupt the progress of his telling. After a laugh, AP produces feedback showing that he has abandoned the NBFC he introduced.

‘Stop kidding, I’m serious’: Failed humor in French conversations

211

In these two examples, the recipients want to switch into a NBFC while the speakers are not only engaged in a BFC but in a storytelling. Thus, in both cases, their humorous interventions disaffiliate with the speakers’ talk. Indeed, the speakers at those moments give no sign that the hearers should interpret their stories as humorous. The recipients’ humorous utterances can thus be considered by the speakers as illegitimate and disaffiliative because not only do they display a (humorous) stance that the tellers do not want to display, but also because the recipients choose the teller’s stance in their stead. Moreover, the hearers’ utterances are also highly disaligned. The fact that they claim a speech turn too early forces the tellers to justify their words which leads to a disruption of the ongoing narrative. In both examples, the recipients’ utterances are parasitic for the progression of the storytelling and their humorous dimension are illegitimate, or at least, considered as such by the main speakers. Consequently, in both examples, producing a “po-faced” answer, i.e. treating the hearers’ humorous interventions as if they were relevant in a BFC while they are clearly a sign of NBFC, allows the speakers to going back faster to a BFC and to their storytelling, thus limiting the damage.

5.2 Humorous utterances ignored In this sub section, two examples are analyzed. Contrary to the previous examples where the meaning of the hearers’ utterances was negotiated in a BFC, here, the hearers’ humorous utterances are simply ignored. Example 3 : Great idea / Super idée 1 AP : ouais je sais plus il s’était passé une histoire avec euh un mec {qui, qu’il} connaissait qui 2 avait soit disant branché par internet euh euh euh une saoudienne euh quoi 3 LJ : mh mh 4 AP : et soit disant la f famille a été m mise au courant je sais p 5 LJ : super idée @ 6 AP : je s je peux pas te dire euh comment pourquoi etc 7 AP : mais le mec il a eu des problèmes quoi 8 LJ: ah ouais 1 2 3 4 5 6 7 8

AP: yeah i don’t remember something happened with err a guy he knows who pretended to have chatted a saudi girl up through internet err err err a saudi girl you see LJ: mh mh AP: and supposedly the family has been notified I don’t know LJ: great idea @ AP: I can’t tell you err how why etc AP: but the guy he had problems LJ: ah yeah

212

Béatrice Priego-Valverde

In this excerpt, AP, the main speaker, is telling a story he heard few years before about a man who contacted a Saudi woman via the internet which did not please to the woman’s family. At that moment, AP is in the complication phase of his story, presenting the situation to LJ before delivering the climax, i.e. the fact that the man incurred serious consequences. But LJ (5) does not wait for the climax to comment on the situation. In fact, he does not even wait for the end of AP’s utterance. Overlapping on the word “family” (4), he produces an ironic and denigrating assessment about the fact that the woman’s family heard about the story, blaming in some ways, the responsible party for this denunciation. In doing so, he disaligned with AP, producing a more elaborated turn than a feedback response would be at such a phase of the storytelling. The fact that LJ regularly stages himself as an “expert” of Saudi Arabia because he lived there when he was younger (which will be more obvious in example 5), may explain such utterance. It may also explain that he guessed (even produced in an antiphrastic ironical comment) the follow up of AP’s story, (i.e. the “problems” the character had). Precisely because he guessed the follow up of the story before the teller had delivered it, LJ’s utterance is disaligned and may explain the fact it is ignored in order to let the speaker say what he has to say. In addition, LJ also disaffiliates with AP. The ironic utterance he produces, framed moreover with a laugh, is a signal of his switch into NBFC while AP, as shown in (7) with the word “issues”, considers the story to be serious. That is probably the reason why AP not only does not respond on LJ’s utterance, but ignores it, finishing his own utterance (4, 6–7) in a serious mode of communication, until he produces the climax of his story, in line (7). As in the previous examples, LJ produces feedback, signaling both the return in a BFC and the fact that he takes into account AP’s talk. Here again, by virtue of the constraints dictated by the activity itself and the effect on the participants’ actions, the recipient’s humorous utterance failed. Because it was neither the right moment, nor the right mode of communication, the recipient’s humor was considered illegitimate by the main speaker. Example 4 : You insist / Tu y tiens 1 YM : § je vais renégocier § mais ça m(e) paraissait pas [juste, justeu] quoi de ouais de de 2 sortir mille cinq cent euros 3 YM : (l)ors que j’a(v)ais euh 4 AG : non en francs 5 YM : pas euh francs @ 6 AG : @ t’y tiens 7 YM : (l)ors qu(e) j’a(v)ais pas utilisé l’[EMA, eèma] machin bidule donc euh 8 AG: ouais ouais ouais ouais 1 YM: §i’m going to negotiate again§ but it seemed to me unfair to yeah to to pay 1500 euros 2 YM: while I had err

‘Stop kidding, I’m serious’: Failed humor in French conversations

3 4 5 6 7

213

AG: no in francs YM: not err francs @ AG: @ you insist YM: while I didn’t use the EMA [a machine] thingy so err AG: yeah yeah yeah yeah

This extract corresponds to the end of YM’s storytelling. The climax has been delivered: YM had to fund himself for the few months he spent in a foreign university while he was in a PhD program. In line (1), YM, the main speaker, displays a specific stance about the event he is reporting: it is “unfair”. This extract corresponds to the evaluation phase (Labov and Waletsky 1966), which is the more symmetrical phase of this asymmetrical activity (storytelling). Both participants finally share the same information about the event reported and the stance the teller wanted to convey, thus they can both become co-speakers in order to comment the point of the story. But YM, in line (1), is mistaken. Instead of saying “francs”, the French currency at the moment when the event related happened, he says “euros”. Consequently, instead of affiliating with YM, AG rectifies the term and proposes “francs” (3). In doing so, he further disaligns with YM in that he forces him to interrupt himself and to pronounce the right currency (4). At that moment, YM does not seem to be disturbed; he laughs and frames as selfdisparaging his own mistake. Here, contrary to all the examples seen above, AG stays in BFC. Moreover, the fact that AG is points out YM’s mistake makes YM laugh. With his laugh, YM introduces a NBFC. In line (5), AG, the recipient, continues in this NBFC, targeting YM and his mistake directly. He also frames his utterance as humorous. But at this time, his utterance, although respecting the humorous frame initiated by YM, is ignored (6). YM chooses to go back to evaluate his story, to develop the stance he displayed, and to do it in a serious mode of communication. The reaction is totally accepted by the hearer (7). This example is different from the others in many ways. First, it corresponds to the evaluation phase of the storytelling; that is, where both participants share the same information about the event reported and the right to comment on it. From teller and recipient, they can become co-speakers. This fact can explain why AG’s repair (3), although disaligned and disaffiliated, is accepted. This fact can also explain why the speaker himself introduces NBFC whereas he was serious during his telling. Until then, the recipient’s utterance accepted by the speaker is not a humorous one (3), but a repair. By contrast, and this is the second major difference with the other examples, when AG, the recipient, produces a humorous utterance (5), although he only follows YM’s lead, this utterance is totally ignored. Various reasons can explain this failure: (i) YM is explicitly the target of the humorous utterance, (ii) AG talks in a very

214

Béatrice Priego-Valverde

low tone of voice, as if he did not want to disrupt the ongoing telling, and (iii) YM wants to go back to his story and justify the stance he delivered (the injustice), he thus wants to go back in BFC, even if he was the instigator of the NBFC. Whatever the reason is, this example suggests that even when the participants are distributed in a more symmetrical position, the teller remains the main speaker.

5.3 Humorous utterances explicitly rejected In this last sub section, two examples will be analyzed. Here, the humorous utterances are explicitly rejected by the main speaker. In the first example, the humorous dimension, i.e. the switch from the hearer into NBFC, is clearly rejected. Example 5 : It’s European terrorism (C’est du terrorisme européen) 1 LJ : ils ont passé un film crypté de canal plus, un film de cul 2 AP : @ 3 AP : @ 4 LJ : euh @ mais attends mais c’était incident 5 AP : @ * 6 AP : § c’est du terrorisme euh européen § 7 LJ : non c’es c’es c’ r rigole pas c’était incident diplomatique euh limite quoi 8 AP : putain tu m’étonnes euh 9 LJ : les mecs ils l’ont très mal pris 1 2 3 4 5 6 7 8 9

LJ: they programmed an encrypted movie from canal + a porn film AP: @ AP: @ LJ: err @ but wait but it was incident AP: @ AP: § it’s european err terrorism § LJ: no it’s it’s don’t laugh it was a diplomatic incident err kind of so AP: fuck you surprise me err LJ: the guys they took it very badly

LJ, the main speaker, spent a large part of his childhood in various countries of the Middle East when he was a child. As mentioned in example (3), because of that, throughout the conversation, he tells many different stories about these countries. In addition, he presents himself as an “expert” who knows and understands these countries. In this excerpt, he is telling a story about Arab Emirates. LJ is engaged into a long storytelling where he relates the fact that many people there have a

‘Stop kidding, I’m serious’: Failed humor in French conversations

215

satellite dish in order to watch foreign channels and more particularly, French channels. Line (1) is the climax of his story: one day, in this country which is depicted as highly religious earlier in the conversation, a French TV channel transmitted, by mistake, a porn movie. At that time, LJ does not frame his utterance as humorous. However, considering it as such, or at least highly incongruous, AP, the recipient, laughs (2–3), thus switching into a NBFC. His laughter is briefly followed by the teller’s laughter in return (4). In this way, LJ acknowledges how the situation he has depicted can be incongruous and accepts the recipient’s switch into NBFC. But immediately after having laughed, in the same utterance (4), he postpones this NBFC (“but wait”) in order to produce his own evaluation: the story was serious enough to be called “incident”. But in (5), AP laughs again, refusing the teller’s return into a BFC and persisting in NBFC. AP not only laughs but also in an overlap with serious beginning of LJ’s explanation, produces a humorous utterance based on exaggeration: programming a porn movie in such a country is associated with terrorism (6). Doing so, he stays in NBFC and furthermore, he denigrates the Arab Emirates, making fun of their religiousness. Comparing the broadcast of a movie, even if pornographic, to terrorism is directly targeting the Arab Emirates. It is so exaggerated that it becomes absurd; as it is absurd to consider such an event serious. At that point, because the climax has been delivered (1), the participants are engaged in the evaluation phase of the storytelling, which should be more symmetrical and give more opportunities to participate to the recipient. However, AP’s utterance is highly disaligned. He overlaps LJ who was going back to BFC in order to express how serious the incident was. But AP does not listen to him and persists in his nonbona-fide denigrating mode of communication. He thus also disaffiliates with LJ’s evaluation: it is serious and not funny at all. Because his attempt to postpone NBFC failed in line (4), LJ, (7), explicitly rejects the NBFC introduced and maintained by the recipient. His utterance begins with an explicit “no” as a sign of a return to serious communication (Schegloff 2001) and follows with a rejection of NBFC: “don’t laugh”. In doing so, LJ rejects both AP’s humor and what such humor implies: denigration. LJ, still in line (7) elaborates on the kind of incident: as serious as it as though it had been a “diplomatic incident”. This utterance allows LJ to reject the recipient’s humor and his maintenance in NBFC at the same time. Moreover, by evaluating his story as serious, LJ reframes it. In doing so, he recovers his position of main speaker, and as legitimate that he claims an image of “expert” of Middle East countries. And indeed, in line (8), AP abandons both the NBFC he has introduced and tried to maintain and his very punctual role of co-speaker, which was refused by LJ. Finally, LJ can develop in (line 9) the reasons why it was a very serious incident.

216

Béatrice Priego-Valverde

Even though this excerpt is short, due to the recipient’s giving up on the attempt at humor, it is however conflictual, as the overlap and the explicit rejection from the teller show clearly. This conflict concerns two different but related aspects of the recipient’s utterances: the fact that they introduce NBFC and the fact that AP wants to become a co-speaker, evaluating the event reported by the main speaker. On the one hand, because the participants are engaged in the evaluation phase of the storytelling, AP, the recipient, wants to collaborate and become a co-speaker. However, his utterance is not only produced in a simple overlap, he overlaps on a return from LJ to BFC. He thus disrupts the teller’s ongoing talk in two different ways: by talking at the same time and by wanting to maintain a mode of communication already rejected by LJ. AP’s participation and humor is thus considered illegitimate by the speaker. AP’s humor is all the more illegitimate in that it is quite aggressive and denigrating. On the other hand, this example shows also that, even in an evaluation phase, participants are still in a storytelling, i.e. an asymmetrical activity. Consequently, faced with an utterance that is considered disaligned and disaffiliative, the teller can claim his role of main speaker and can decide whether or not the hearer’s utterances are legitimate or not. In other words, as main speaker, the teller can take things in hand at anytime. The last example is also a conflictual one, but a longer one because, this time, the recipient resists the attempt at change for a longer time. Example 6 : the excavation site / le chantier de fouille 1 LJ : en t’es t’es quand même bien handicapé y a des chantiers école 2 LJ : que d où tu payes euh bonbon je crois qu’y en a un à Lattes, 3 AP : tu payes pour fouiller 4 LJ : tout ça 5 LJ : ouaistututupayesouaismaisatusorst’asuneespècedepasundiplômejesaispasmais 6 enfindc’c’est§j’aifouilléàmachinet§ 7 AP : @ tu payes pour faire le manœuvre 8 AP : * super 9 LJ : et là bon c’est c’est 10 AP : @ § j’ai un diplôme de fouilleur § 11 LJ : si tu veux c’est c’est fait euh 12 AP : @ § j’ai tenu une pioche pendant une semaine § 13 LJ: c’est un chantier école c’est-à-dire que t’as des cours t’as des cours sur la céramique euh 14 AP : ah hum hum hum 15 AP : mh mh ah ouais OK ouais 16 LJ : euh tout le bordel quoi enfin et c’est c’est quand même complet et et c’est hard mais par 17 contre euh 18 AP : ah ouais ouais d’accord ouais ouais c’est pas que creuser ouais

‘Stop kidding, I’m serious’: Failed humor in French conversations

217

19 20 21 22 23 24 25 26 27 28 29 30

LJ : je crois que j’y suis pas passé mais j’en ai entendu parler de ce site là c’est euh LJ : c’est pelé y a rien enfin si tu veux LJ : c’est pas un je crois que c’est pas un site très intéressant quoi tu creuses et tu creuses tu creuses et @ AP : @ c’est le site où y a rien à trouver quoi LJ : non si mais disons que p m par rapport à des sites où t’as y t’as y te reste des bouts de monuments des trucs comme ça AP : le site des bleus AP : mh mh LJ : euh je crois c’est LJ : hum hum c’est un peu ingrat quoi AP : eh ouais d’accord ouais ouais

1 2 3 4 5 6 7 8 9 10 11 12 13

LJ: in you’re you’re anyway handicapped enough there is a school excavation sites LJ: from which you pay a lot I think there is one in Lattes AP: you pay to search LJ: everything LJ: yeah you you you pay yeah but you fishe you have a sort of not a diploma I don’t know but well it’s §I searched at§ AP: @ you pay to be a worker AP: great LJ: and there well it’s it’s AP: @ §I’ve a diploma of searcher§ LJ: if you want it’s it’s been done err AP: @ §I’ve hold a pick during one week§ LJ: it’s a school excavation site meaning you have classes you have classes about ceramics err AP: ah hm hm hm AP: hm hm ah yeah ok yeah LJ: err all the stuff well and it’s it’s anyway full and and it’s hard but however err AP: oh yeah yeah okay yeah it’s not only digging yeah LJ: I think I haven’t been there but I heard about this site it’s err LJ: it’s peeled there is nothing well if you want LJ: it’s not a I guess it’s not a very interesting site well you dig and you dig and you dig and @ AP: @ it’s the site where nothing is to be found no LJ: no yes but let’s say that in comparison with sites where you have some rests pieces of monuments stuff like that AP: the site for rookies AP: mh mh LJ: err I think it’s LJ: hm hm it’s a little ungrateful well AP: eh yeah ok yeah yeah

14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

The participants are talking about the same topic as in example (2): the excavation sites LJ knows because he studied Archeology before Linguistics. In this excerpt, there is never any explicit rejection of the recipient’s humor as was the

218

Béatrice Priego-Valverde

case in the previous example (5). All the instances of AP’s humor fail because they are answered in a serious way or, because they are ignored. Thus, this example could have been presented in the two previous sub sections. However, both the recipient’s insistence in switching to NBFC and the speaker’s resistance, make this sequence highly conflictual. Consequently, the way the teller answers the recipient’s humorous utterances will be only mentioned in order to focus on the interactional conflict between both participants. The extract begins in the orientation phase of LJ’s storytelling: he is explaining the functioning of school excavations sites, explaining that the students have to pay to be there. But LJ does not just mention in passing that these sites charge the students, he also says that they are expensive (2). From this perspective, although he is in the orientation phase, LJ produces a personal comment and assumes a stance toward what he is telling. AP immediately plays on it (3), framing his utterance at that moment, not clearly humorous – which will be done on turn (7) – but highlighting the fact that it is incongruous by the association of paying and searching i.e. working. In other words, AP entered in the breach opened by LJ. But, pointing out one element of the situation exposed by the teller, such utterance disaligns with LJ because it disrupts the ongoing storytelling. And indeed, LJ has not finished setting the scene. That is probably the reason why he answers in a serious way in lines (5 and 6), repeating and confirming that students have to pay to search, and justifying that point by the fact they receive a sort of diploma at the end, i.e. they pay for something presumably valuable. Taking the recipient’s humorous utterance seriously, he does not take into account its humorous side, which allows him to continue telling his story in BFC. Nonetheless AP persists; in lines 7 and 8, he rephrases his own utterance, but this time, he clearly framing it as humorous (he laughs and makes an ironical comment: “great”). Such insistence added by the ironical comment makes AP’s utterances highly denigrating, without knowing if the target of his denigration is the school which organizes such digs or LJ himself who experienced some of these digs when he was a student. Whatever the reason, LJ (9) ignores AP’s utterances and attempts to pursue in own story. But AP refuses it. Overlapping on LJ, in (10 and 12), he ridicules LJ’s words repeating “diploma” and in fictitious reported speech (see Guardiola and Bertrand 2013; Bertrand and Priego-Valverde 2011; Priego-Valverde 2018), stages himself as having such a diploma. Here, not only does AP insist in maintaining the NBFC he has introduced while LJ was in BFC and to which LJ wanted to go back, but also, he clearly teases LJ. In line (11), LJ tries again to stay in the BFC he introduced first, answering to AP’s teasing in a serious way (“if you want”). His “strategy” seems to work for a while because AP (14–15), produces some feedback answers to finally take into account what is

‘Stop kidding, I’m serious’: Failed humor in French conversations

219

being said in a serious way. Furthermore, he acknowledges the fact that students do other things than just digging (17). Interestingly, while AP has finally accepted the BFC LJ wanted to install by any means, LJ himself switches into a NBFC (lines 20–21), producing the verb “dig” three times, as if he wanted finally also to denigrate such charged work. Paradoxically, while AP follows him in the new NBFC the speaker introduced (22 and 25), LJ returns to his previous reactions: AP’s humorous utterance is answered seriously in (22) and just ignored in (25) in order to go back to a serious evaluation of this kind of sites (27–28). This return to BFC is finally accepted by AP. The conflict between both participants is remarkable for various reasons. The first one and the more obvious is its duration due both to the recipient’s instance on switching into a NBFC and the speaker’s resistance by staying in a BFC. Moreover the conflict between the two modes of communication is not the only one. An interactional conflict between the two participants and the respective roles they have or claim is also present. Because they are engaged in the orientation phase of the storytelling, LJ is the main speaker and seems to want to keep this role. All the kinds of answers he produced to the recipient’s utterances highlight this fact: whether he ignores, responds in a serious way, or overlaps the recipient’s talk, he keeps talking in the way he has decided. Interestingly, even when he finally frames himself his story as humorous, he refuses the recipient’s follow up into a NBFC. He acts, as if the problem was not only the switch into another mode of communication, but the fact that the hearer has an opinion about an event he knows more than him: it is his story and moreover, he is a sort of “expert”, the only one thus legitimate to judge, even negatively, the archeology policies. In other words, this conflict between BFC and NBFC concerns more the recipient’s disalignment whose utterances interrupt the ongoing storytelling and forces the teller to justify the usual archeology policies, than his potential disaffiliation: indeed, they both think that paying to search seems, at least, incongruous. The conflict between the two participants is thus more interactional and more related both to their interactional roles in a storytelling activity and to the image of expert claimed by the main speaker.

6 Concluding remarks This article focused on humor as non-bona-fide communication, produced by the recipient, while the speaker is engaged in a serious storytelling, i.e. a bonafide communication. Taking into account the various interactional constraints

220

Béatrice Priego-Valverde

weighting in on participants’ roles and actions (producing humor included) and considering that these constraints are even more active in serious storytelling, the aim of this article was to investigate the switch from BFC into NBFC as a reason of failed humor. The quantitative results have shown a wide difference between the 3 conversations. AP_LJ standing out at any point: (i) they produce much more humor than the other pairs, (ii) they also produce much more failed humor, (iii) the switch from BFC into NBFC is the main reason of the failure of humor (more than a half), and (iv) their failed humorous items appear more than the other pairs in storytelling. Such results tend to confirm the very active constraints the storytelling imposes on participants. Moreover, one could consider that the more humor is produced in serious storytelling, the more it may fail. The first reason is the interactional roles of teller and recipient in story telling: producing humor while a participant has the role of recipient implies to take a potential illegitimate (and sometimes, even parasitic) speech turn. The second reason is that this potential illegitimacy is all the more so as humor produced by the recipient implies also a switch from BFC chosen by the teller into NBFC. Furthermore, AP_LJ’s results raise a question: the impact on humor, and more precisely, on its reception, of the nature of the relationship between the participants. Indeed, if all the participants are friends, AP and LJ are the closest ones. They are intimate friends. Consequently, considering that the more two participants know each other, the more humor may succeed is not confirmed here, and it is even overturned. Two reasons could explain this result. Firstly, one could say that the nature of the conversational activity in which the participants are engaged would counterbalance their close relationship, as if the constraints of storytelling were heavier than their friendship. The second reason would concern the paradoxical effects of such relationship: the more participants know each other, the more they feel authorized to treat negatively the humor produced, as if the stake of face was diminished. Needless to say, a recipient’s humor may succeed, and in fact does, even when produced in storytelling. Focusing on failed humor because of a switch into a NBFC has highlighted to what extent, in serious storytelling, a humorous utterance produced by the recipient may be illegitimate or even parasitic. In each case, the humorous utterances are both disaligned and disaffiliative. They are disaligned because, at a minimum, they are produced in a phase of the storytelling where they are not expected (orientation, complication). At the maximum, they are disaligned because they disrupt the ongoing telling. All of them are disaffiliative because through them, on the one hand, the recipient, instead of adopting the teller’s stance, displays another stance, a humorous one and, on the other hand, imposes a humorous frame, i.e. a way to interpret the

‘Stop kidding, I’m serious’: Failed humor in French conversations

221

telling, while the main speaker has not yet displayed his own. In sum, when a humorous utterance is produced by the recipient, a switch from BFC into NBFC is intrinsically disaffiliative. However, it is worth noting that the humorous utterances produced in example (4) are different: they are produced in the evaluation phase of the storytelling, i.e. in a phase developed after the delivery of the climax and where, usually, both participants can participate. The example shows that, if indeed YM’s humor is ignored, it is also produced by the recipient himself in a low tone of voice, as if he did not want the speaker to take it into account, as if he did not want to interrupt the teller. Focusing on failed humor has also highlighted three negative reactions from the main speaker to the recipient’s humor, highlighting thus his non-cooperation presented as a continuum: (i) humor is answered in a serious way. Here, the utterance is taken into account and responded to, but not its humorous aspect. Such answer is the most cooperative negative reaction. Furthermore, it allows the speaker to stay in the BFC he has introduced. (ii) Humor is purely and simply ignored, i.e. the main speaker keeps his speech turn as if the recipient has not participated. Here, if the speaker’s cooperation is more than questionable, his “strategy” remains effective: he can follow up his storytelling and, if the recipient’s humor is considered illegitimate, it is not parasitic. Finally, (iii) humor is explicitly rejected. This reaction is undoubtedly the less cooperative. If the three reasons may be considered a conflict between the two modes of communication, which leads itself to an interactional conflict between the participants, the third kind of reaction is the more obvious type of conflict. All the examples show that the recipient’s humor is perceived by the speaker as, at least illegitimate and sometimes parasitic because it may disrupt the ongoing storytelling. In other words, the sequential analysis of some examples confirms not only the significant restrictions the storytelling imposes on participants’ roles and duties, but also how much such constraints might be strict. However, example (5) is different: as example (4), the humorous utterance is produced during the evaluation phase, i.e. after the climax. But here, the climax (the TV channel in the Emirates broadcasts, by mistake, a porn movie), does not seem to be the central point of the teller’s story. The central point seems to be the stance he wants to deliver, i.e. this broadcast was a very serious event. In other words, for the teller, presenting himself as an “expert” of Middle East countries was more important than the story itself. This shows that the storytelling rules can be also violated by the speaker and that some individual agendas (here, the image of “expert”) can legitimate such a violation. In other words, the three speaker’s reactions (increasingly non-cooperative) highlight the fact that the teller claims all the time for his role of main

222

Béatrice Priego-Valverde

speaker and for the correlative rights he has. In this regard, the recipients’ cooperation is more questionable. On the one hand, switching to a NBFC is clearly non-cooperative, both because they are recipients and because they change the mode of communication installed by the main speaker. But, on the other hand, most of the time, except in example (4), they frame their humorous utterance as such, in order to help the speaker to follow them in this new mode of communication. In this regard, they are cooperative. The fact that, most of the time, the recipients frame as humorous their utterances while they still fail tends to confirm the suggestion (see Section 2) according to which signaling a switch into a NBFC from the recipient may remain insufficient for it to be accepted, and hence, successful. Such a result confirms also that, firstly, the question of the legitimacy of humor is essential in conversation, and secondly, that this legitimacy is allowed by the main speaker. This legitimacy will be analyzed in more detail in (Priego-Valverde in preparation), but the current findings tend to confirm that the interactional constraints of the conversation or of the conversational activity in which humor appears have to be necessarily taken into consideration as a fundamental element allowing a better comprehension of conversational humor. They show that conversation is not only a setting where humor appears, but one of its intrinsic characteristics which justifies the terms of conversational humor or interactional humor. Finally, this study does not answer the question whether BFC is more important than NBFC, even if each time, BFC finally prevailed over NBFC. However, it highlights the fact that two different modes of communication cannot coexist. Obviously further work is needed. On the one hand, the data on storytelling need to be enlarged in order to confirm or not the significance of the storytelling constraints. On the other hand, the results presented here need to be applied to other conversational activities such as argumentation. If the switch into NBFC is less problematic, this further work could confirm that storytelling is a highly constraining activity. On the contrary, if it appears that in another conversational activity, this switch remains as non-cooperative as in storytelling, the conclusion would be different and remarkable: it could signify that, in the end, conversation is not as cooperative as one would want to think, since it would share some constraints with non-cooperative modes, such as argumentation.

Conventions of transcription @ : + §word§ underlined word

laughter vocalic extension silent pause reported speech overlap

‘Stop kidding, I’m serious’: Failed humor in French conversations

in bold *

223

humorous utterance undetermined noise

Acknowledgments: The author wishes to thank Salvatore Attardo and Lucy Pickering for their editorial suggestions for the present paper.

References Attardo, Salvatore. 2002. Humor and Irony in Interaction: From Mode Adoption to Failure of Detection. In L. Anolli, R. Ciceri, & G. Riva (eds.), Say not to Say: New perspectives on miscommunication, IOS Press. 166–185. Bavelas, Janet B., Linda Coates and Trudy Johnson. 2000. Listeners as co-narrators. Journal of Personality and Social Psychology, 79. 941–952. Bell, Nancy. 2009. Responses to failed humor. Journal of Pragmatics 41. 1825–1836. Bell, Nancy. 2015. We are not amused. Failed humor in interaction. Berlin, Boston, Munich: Mouton de Gruyter. Bell, Nancy and Salvatore Attardo. 2010. Failed humor: Issues in non-native speakers’ appreciation and understanding of humor. Intercultural Pragmatics 7–3. 423–447. Bertrand, Roxane and Béatrice Priego-Valverde. 2011. Does prosody play a specific role in conversational humour?. Pragmatics and Cognition 19. 333–356. Bertrand, Roxane, Philippe Blache, Robert Espesser, Gaëlle Ferré, Christine Meunier, Béatrice Priego-Valverde, and Stephane Rauzy. 2008. ‘Le CID – Corpus of Interactional Data – Annotation et Exploitation Multimodale de Parole Conversationnelle. Traitement Automatique des Langues 49. 105–134. Blache, Philippe, Roxane Bertrand, and Gaëlle Ferré. 2009. Creating and exploiting multimodal annotated corpora: the ToMA project. In M. Kipp, J.C. Martin, P. Paggio, and D. Heylen. (eds.). Multimodal Corpora. From Models of Natural Interaction to Systems and Applications, Berlin, Heidelberg: Springer-Verlag. 38–53. Boersma, Paul and David Weenink. 2009. Praat:DoingPhonetics by Computer. (Version5.1.05) [Computer program]. Available on line at: http://www.praat.org/ Drew, Paul. 1987. Po-faced receipts of teases. Linguistics 25. 219–253. Eisterhold, Jodi, Salvatore Attardo, and Diana Boxer. 2006. Reactions to irony in discourse: evidence for the least disruption principle. Journal of Pragmatics 38. 1239–1256. Goffman, Erving. 1967. Interaction Ritual. New York: Anchor. Goodwin, Margaret H. 1997. Byplay: Negotiating evaluation in storytelling. In Gregory R. Guy, Crawford Feagin, Deborah Schiffrin and John Baugh (eds.). Towards a social science of language: paper in honor of William Labov. Vol.2: 77–102. New-York, Amsterdam: John Benjamins. Grice, H. Paul. 1975. Logic and conversation. In Peter Cole and Jerry L. Morgan (eds.) Syntax and semantics, Vol. 3. 41–58. New-York: Academic Press. Guardiola, Mathilde and Roxane Bertrand. 2013. Interactional convergence in conversational storytelling: when reported speech is a cue of alignment and/or affiliation. Frontiers in Psychology 4. 1–17. Hay, Jennifer. 1994. Jocular abuse in mixed gender interactions. Wellington Working Papers in Linguistics 6. 26–55.

224

Béatrice Priego-Valverde

Hay, Jennifer. 1995. Gender and humour: beyond a joke. Master thesis. New Zealand: Victoria University of Wellington. Hay, Jennifer. 2001. The pragmatics of humor support. Humor 14(1). 55–82. Jefferson, Gail. 1978. Sequential aspects of storytelling in conversation. In Jim Schenkein. (ed.), 219–248. Studies in the organization of conversational interaction. New York: Academic. Labov, William and Joshua Waletzky. 1966. Narrative analysis: oral versions of personal experience. In June Helm. (ed.), Essays on the Verbal and Visual Arts: Proceedings of the 1966 Annual Spring Meeting of the American Ethnological Society, 12–44. Seattle: University of Washington Press. Mandelbaum, Jenny. 1989. Interpersonal activities in conversational storytelling. Western Journal of Speech Communication 53(2). 114–126. Mandelbaum, Jenny. 1991/1992. Conversational Non-Cooperation: An Exploration of Disattended Complaints. Research on Language and Social Interaction 25. 97–138. Mandelbaum, Jenny. 2013. Storytelling in conversation. In Jack Sidnell and Tanya Stivers. (eds), Handbook of Conversation Analysis, 492–507.Wiley-Blackwell. Norrick, Neal. 2000. Conversational narrative: Storytelling in everyday talk. Amsterdam : John Benjamins. Priego-Valverde, Béatrice. 2003. L’humour dans la conversation familière: description et analyse linguistiques, France : L’harmattan. Priego-Valverde, Béatrice. 2009. Failed humor in conversation: a double voicing analysis; In Neal Norrick & Delia Chiaro, D. (eds.). Humor in interaction, 165–183. Amsterdam, Philadelphia: John Benjamins. Priego-Valverde, Béatrice. 2018. Sharing a laugh at others: Humorous convergence in French conversation. European Journal of Humour Research 6 (3). 68–93. Raskin, Victor. 1985. Semantic Mechanisms of Humor. Dordrecht, Holland: D. Reidel Publishing Company. Raskin, Victor. 1992a. Humor as a Non-Bona-Fide Mode of Communication. In E. L. Pedersen, (ed). Proceedings of the 1992 annual meeting of the Deseret Language and Linguistic Society. 87–92. Provo, UT: Brigham Young University. Raskin, Victor. 1992b. Using the powers of language: non-casual language in advertising, politics, relationships, humor, and lying. In E. L. Pedersen (ed.), Proceedings of the 1992 Annual Meeting of the Deseret Language and Linguistic Society. 17–30. Provo, UT: Brigham Young University. Raskin, Victor. 1998. The sense of humor and the truth. In Willibald Ruch (ed.). The Sense of Humor. Explorations of a personality Characteristic. 95–108. Berlin, New-York: Mouton de Gruyter. Sacks, Harvey. 1974. An Analysis of the Course of a Joke’s Telling in Conversation. In Explorations in the Eth– nography of Speaking. Richard Bauman and Joel Sherzer, (eds). 337–353. Cambridge: Cambridge University Press. Sacks, Harvey, Emanuel Schegloff, and Gail Jefferson. 1974. A Simplest Systematic For The Organization Of Turn Taking For Conversation. Language 50 (3). 696–735. Schegloff, Emanuel. 2001. Getting serious: Joke -> serious ‘no’. Journal of Pragmatics 33. 1947–1955. Schegloff, Emanuel. 2007. Sequence organization in interaction: A primer in conversation analysis. Vol. 1. Cambridge: Cambridge University Press.

‘Stop kidding, I’m serious’: Failed humor in French conversations

225

Selting, Margret. 2000. The construction of “units” in conversational talk. Language Society 29. 477–517. Selting, Margret. 2010. Affectivity in conversational storytelling: An analysis of displays of anger or indignation in complaint stories. Pragmatics 20 (2).229–277. International Pragmatics Association. Selting, Margret. 2012. Complaint stories and subsequent complaint stories with affect displays. Journal of Pragmatics 44. 387–415 Selting, Margret. 2017. The display and management of affectivity in climaxes of amusing stories. Journal of Pragmatics 111. 1–32. Shilikhina, Ksenia. 2018. Discourse markers as guides to understanding spontaneous humor and irony. In Villy Tsakona & Jan Chovanec. (eds.), The Dynamics of Interactional Humor. 57–75. Amsterdam, Philadelphia: John Benjamins. Stivers, Tanya. 2008. Stance, Alignment, and Affiliation During Storytelling: When Nodding Is a Token of Affiliation. Research on Language and Social Interaction 41(1). 31–57. Stivers, Tanya. 2013. Sequence organization. In Jack Sidnell & Tanya Stivers. (eds), Handbook of Conversation Analysis: 191–209.Wiley-Blackwell.

Part 3: Ontological semantics

Julia Taylor Rayz

Scripts in the Ontological Semantic Theory of Humor Abstract: The Ontological Semantic Theory of Humor (OSTH) is the latest development in the script-based humor families. The first theory, Script-based Semantic Theory of Humor (SSTH) introduced scripts and proposed that a text contains a joke if there are two scripts in the text that overlap and opposed. The second theory, the General Theory of Verbal Humor (GTVH), added five knowledge resources to the SSTH, namely Logical Mechanism, Situation, Target, Narrative Strategy, and Language. The Script Overlap/Oppositeness was left from the SSTH as the coarsest grain knowledge resource. Neither the SSTH nor the GTVH relies on computational implementation, and the knowledge resources, including scripts are often selected based on the intuition of the annotating expert. The OSTH does not introduce any novelty in terms of humor processing, but provides a foundation for processing text of a joke in a more consistent manner. This paper uses the OSTH to analyze some examples introduced by Raskin. The paper shows why different levels of scripts can still be considered the same, and what methods can be used to streamline the annotations of the scripts. Keywords : ontological semantics, scripts, humor The Ontological Semantic Theory of Humor is the latest development in the scriptbased humor families. The first theory, the Script-based Semantic Theory of Humor (SSTH, Raskin 1985) introduced scripts and proposed that a text contains a joke if there are two scripts in the text that overlap and oppose. The second theory, the General Theory of Verbal Humor (GTVH, Attardo & Raskin 1991), added five knowledge resources to the Script Overlap/Oppositeness1 described by the SSTH, namely the Logical Mechanism, Situation, Target, Narrative Strategy, and Language. Script Overlap/Oppositeness was left from the SSTH as the coarsest grain knowledge resource. Neither the SSTH nor the GTVH relies on computational

1 The terms Script Oppositeness and Script Opposition refer to the same phenomenon and will be used interchangeably throughout this paper Julia Taylor Rayz, Purdue University https://doi.org/10.1515/9781501511707-011

230

Julia Taylor Rayz

implementation, and the knowledge resources, including scripts are often selected based on the intuition of the annotating expert.2 The Ontological Semantic Theory of Humor (OSTH) does not introduce any novelty in terms of humor processing, but provides a foundation for processing the text of a joke in a more consistent manner. In order to fully appreciate the OSTH one needs to understand Ontological Semantics (Nirenburg & Raskin 2004), which describes a way of processing natural language text from a meaning perspective. While the desire to process text based on its meaning may seem obvious now, it was not always the case that scholars considered it a desideratum, especially considering that meaning-based processing is still far from achieved. This paper will use OSTH to analyze some jokes used as examples in Raskin (1985). It will start with examining the famous Doctor/Lover joke in terms of ontological semantics. It will continue with less famous examples from the book and concentrate on those that are described using different scripts, depending on the chapter in the book and the level of analysis involved. It will show why different levels of scripts can still be considered the same, and what methods can be used to streamline the annotations of the scripts. Finally, it will discuss implications of ontological precision on SO interpretation.

1 Why do we need an ontology? Ontologies can be looked at as a method of avoiding disagreements between people. According to Guarino, an ontology’s purpose is “understanding, clarifying, making explicit and communicating people’s assumptions about the nature and structure of the world,” (2008) and, one may add, about theories that describe phenomena in said world. An ontology then is a way of explicitly stating information about the observed world (what people perceive every day), about the inferences in this world, such as stereotypes that are so common in humor domain, and drawing boundaries of what is possible and impossible in this world. As with any formal system, correctness of information is important. Additionally, to avoid misunderstandings, it is desirable for an ontology to have high precision with the respect to the conceptualization of the phenomena. In other words, not only should an ontology be correct with the respect to the phenomena in question, it

2 The author is grateful to Salvatore Attardo for pointing out that it is possible that an obviously lexical script, which is right in the text, is evoked, and then the choices are eliminated.

Scripts in the Ontological Semantic Theory of Humor

231

should describe as little as possible outside of these phenomena. This is not an easy task to do when one attempts to describe everything that is available for a discussion in human conversation, and thus, inevitably, an ontology that can be useful for humor processing slides into lower-than-desired precision, thus allowing misunderstandings and disagreements among people that may use this ontology as their source of knowledge. The same, of course, applies to a computer that is unlikely to differentiate (based on its experience) between correct and incorrect information. Raskin (1985) proposes an ontological semantic base for the theory that he described (p. 57): “[t]he goal [. . .] is to account for the meaning of every sentence in every context it occurs. The theory does not incorporate our entire knowledge of the world and does not claim that it is possible to do so.” However, it does take into account context that is needed to understand every sentence. The context, in this case, is proposed to be handled by scripts – knowledge that is available to whatever entity is processing the text, and is activated by lexical senses. In other words, these scripts are lexical scripts, with each word potentially activating a number of such scripts. Raskin hints at the difficulty of constructing well-defined and well-structured scripts: “the scripts are constructed from elementary level up in order to avoid the typical lexicographic circularity.” (p.85) Scripts capture the knowledge needed to understand jokes (or text for that matter), and are based on words that are part of the lexicon. The meanings of words are characterized in the lexicon and are determined “by use” of the words. Combinatorial rules are used to “combine the meanings of the words which make up the sentence [. . .] into the semantic interpretations, or simply meaning, of the whole sentence.” (p.79) Ambiguous sentences will produce multiple semantic interpretations, paraphrases should produce the same semantic interpretations. Production of such interpretation is not a trivial task, and Raskin outlines a wish list of what they should do to make the theory work as desired. The scripts outlined in the book are shortcuts to real scripts: “they themselves utilize the material of other scripts” (p.85) which is not explicitly shown in the examples provided. However, even glancing at such simplified scripts, one can recognize what will become ontological concepts in them. Moreover, lexical entries of the scripts would be eventually described as words and senses of the lexicon in the Ontological Semantics (Nirenburg & Raskin 2004), and combinatorial rules would be nothing but modules of ontological semantic engine, sometimes referred to as STAn (Hempelmann et al. 2010, Raskin et al. 2010, Taylor et al. 2010) POST, or many others depending on implementation. This combination of formally defined senses, concepts, rules and modules of

232

Julia Taylor Rayz

OSTH provides a firm foundation and a verification platform for the theory introduced in SSTH.

2 Doctor/Lover through Ontological Semantic Theory of Humor The famous Doctor/Lover example3 of Raskin (1985) is based on the fact that there is readily available shared knowledge that can be processed by the listener of a joke. Raskin goes into great details identifying this information. For example, a (lexical) script of doctor is defined with an emphasis on the fact that after getting a degree from a medical school, a doctor receives patients either in a hospital or a doctor’s office, with the purpose of curing them. He also defines a lexical script of a lover, with views that were common at the time of publication of the SSTH. In particular, the main actors in the script are two adults of opposite gender that meet regularly in some secluded location with the purpose of “mak[ing] love.” If one were to translate these scripts to the later released Ontological Semantics (a foundation of the Ontological Semantic Theory of Humor), the ontologically anchored lexical senses of a doctor and lover would look something like this: doctor-n1 pos: n morph: NNS +s syn-struc: NP var0 sem-struc: DOCTOR_MD doctor-n2 pos: n morph: NNS +s syn-struc: NP var0 sem-struc: DOCTOR_PHD lover-n1 pos: n morph:

3 For the benefit of non-humor scholars (Raskin, 1985, pp. 32 or 100): “Is the doctor at home?” the patient asked in his bronchial whisper. “No,” the doctor’s young and pretty wife whispered in reply. “Come right in.”

Scripts in the Ontological Semantic Theory of Humor

233

NNS +s syn-struc: NP var0 sem-struc: LOVER

It is the ontology itself that defines lovers, doctors, patients, and many more (it should be noted that the ontology is language independent, and the definitions should work for multiple languages, although cultural differences would normally be adjusted elsewhere), and lexical senses use an anchoring mechanism of sem-struc4 to create the connection. The definitions of lexicon-evoked ontological concepts, very roughly sketched, would look something like this: DOCTOR_MD IS-A PROFESSIONAL-OCCUPATION AGENT-OF-ROLE-IN DEFAULT DIAGNOSE-ILLNESS AGENT-OF-ROLE-IN DEFAULT TREAT-ILLNESS LOCATION DEFAULT MEDICAL-INSTITUTION SEM BUILDING RELAXABLE-TO PHYSICAL-LOCATION PAST-AGENT-OF-ROLE-IN SEM RECEIVE-DIPLOMA TOPIC DEFAULT MEDICINE DOCTOR_PHD IS-A PROFESSIONAL-OCCUPATION

... PAST-AGENT-ROLE-IN SEM RECEIVE-DIPLOMA TOPIC DEFAULT PHILOSOPHY PROFESSIONAL-OCCUPATION IS-A SOCIAL-ROLE PAST-AGENT-ROLE-IN SEM RECEIVE-DIPLOMA LOVER IS-A SOCIAL-ROLE EXPERIENCER-OF-ROLE-IN DEFAULT MAKE-LOVE CARDINALITY

>1

LOCATION SEM PHYSICAL-LOCATION SECRECY VALUE HIGH PATIENT IS-A SOCIAL-ROLE BENEFICIARY-OF-ROLE-IN TREAT-ILLNESS

It is worth noting that all of these concepts are so called SOCIAL-ROLES. In other words, a DOCTOR is a role that a particular person plays in a particular situation. This person’s responsibilities, while participating in this “role,” is diagnosing and treating diseases, as outlined above. Similarly, a PATIENT is a role in which a person participates in a particular situation. Both of these people (doctor and/or patient) can play other roles at the same time for applicable situations. For example, a person that has a SOCIAL-ROLE DOCTOR_MD can at the same time

4 In Ontological Semantics a sem-struc is a field of a sense in a lexicon that describes the meaning of a sense using concepts and properties defined by the ontology.

234

Julia Taylor Rayz

have a SOCIAL-ROLE HUSBAND; a person that has a SOCIAL-ROLE PATIENT, can be a LOVER at the same time as well. It is also possible that our PATIENT and LOVER is also somebody else’s HUSBAND, but we won’t complicate this situation more than absolutely necessary. The first sentence of the Doctor/Lover is relatively straight-forward: a PATIENT, activated by the lexical sense of patient-n1 triggered by the noun patient, is participating in a COMMUNICATIVE-EVENT, namely ASK, that requires two people. The PATIENT inquires whether the DOCTOR_MD, activated by the lexical sense of doctor-n1 triggered by the noun doctor, is at his RESIDENCE, triggered by the lexical sense of home-n1, which is not described here. At the same time, another sense of a noun doctor, that of DOCTOR_PHD, activated by the lexical sense of doctor-n2, will be triggered. From the first sentence, it is not clear whether doctor is a medical doctor or a doctor of philosophy (as described by the ontology and lexicon), but since “bronchial whisper” is involved, we can assume that the medical doctor will result in a higher level of activation. While it is clear that this DOCTOR_PHD activation is not needed for this joke processing (at least as far as human understanding is concerned), it should be noted that this other sense is acceptable and one could argue, if one really tried, that a background opposition or background incongruity (Attardo et al. 2002, Samson & Hempelmann 2011) should be acknowledged (medical doctor vs doctor of philosophy, possibly, on the grounds that their occupations are very different). The “bronchial whisper” component of the first sentence will connect the speaker’s body part, BRONCHI, – activated by lexical sense bronchial-adj1 – to the speaker himself, who happens to be a PATIENT. Since a patient typically needs to be treated (see ontological definition above), and a body part is mentioned, it is reasonable to try to link this body part to an illness, which is exactly what an OST processing engine would do. Inflammation of BRONCHI is BRONCHITIS, which can cause LARYNGITIS, which in turn cause vocal chords to swell, thus causing voice to become hoarse, leading to WHISPERing. This is a convenient interpretation, confirmed by the noun whisper. The first sentence could be summarized as SEEKing_MEDICAL_HELP script (possibly more accurately than DOCTOR or PATIENT), however, the label itself is of little importance. What is important is that there is nothing in the text of the first sentence that makes a human or a computer suspect that the most likely interpretation will be questioned. What is also important is that there are two people involved in a conversation, one of them with a clearly identified SOCIAL-ROLE, and that he is whispering. The second sentence introduces the other participant of the conversation, the doctor’s young and pretty wife. While there is no incongruity present within these two sentences, a second script begins to emerge. First of all, another

Scripts in the Ontological Semantic Theory of Humor

235

SOCIAL-ROLE,

that of a WIFE, is introduced. The person in this social role is answering a door (doorbell?) in her own house, thus, potentially activating some domestic (not professional) script. The WIFE is young and pretty, giving more weight to a non-professional script. She participates in a COMMUNICATIVE-EVENT (in this case RESPOND) with the man who inquires about her HUSBAND at the same volume level as he does. Notice that if we accept the script of a regular communication between a HOUSEWIFE and a person that rings the bell of the house, then the SOCIAL-ROLE of this person becomes a VISITOR, not his other SOCIAL-ROLE of PATIENT. This trivial observation of a person changing social roles depending on the situation is important for dismissing or emphasizing some of the oppositions within scripts. A mother that drives her child to school has a SOCIAL-ROLE MOTHER relative to the CHILD, and a DRIVER relative to the CAR. A doctor performing a surgery can be a SURGEON and a PARENT, although apparently not relative to the same person, as one learns from a surgeon puzzle.5 The question to ask here is whether this difference in social roles, while being the same person, would qualify as overlap and oppositeness, or it is just a hint that should be taken in an exploration of the situations. From a computational point of view, the social role hint is significant as it outlines the main actions or events that correspond to the roles themselves and one can explore these events. However, the question of exactly what (objects, events, properties, etc.) should overlap and oppose to count as a clean and honest SO remains to be answered. This detour into social roles takes us to the third sentence: a WIFE (domestic script activated) invites a VISITOR to her house after quietly conversing with him, and confirming that her HUSBAND – the DOCTOR – who her VISITOR – the PATIENT – supposedly came to see, is away from home. Raskin states (p. 105) “as soon as the appropriate script, LOVER, is evoked, all the previously odd pieces fall nearly into place.” This is, undoubtedly, so. The question is, how does this script get evoked, and what happens when it does not. Let us consider the latter question first. So far we have a script of a social VISIT following a conversation and a script of unsuccessfully SEEKing-MEDICALHELP. Using a conservative analysis of the overlap, there is one person that is a salient actor in both scripts, and one person that is a salient actor in one of the two scripts. The scripts themselves have no connection to each other, other than the people involved. The oppositeness is in the roles that these people play but

5 A father and his son are in a car accident. The father dies instantly, and the son is taken to the nearest hospital. The doctor comes in and exclaims: “I can’t operate on this boy.” “Why not?” the nurse asks. “Because he’s my son,” the doctor responds. How is this possible? (http://www.bigriddles.com/riddle/doctor-cant-operate)

236

Julia Taylor Rayz

only if one considers that a patient should not socially visit his doctor’s wife.6 Yet, he does, which means that they know each other. Assuming that a computer also learned that there is gender bias on the internet (Bolukbasi et al. 2016, Garg et al. 2018) and that it should ignore it at any cost, the words young and pretty will not activate the LOVER script, but rather be used as a description of the WIFE’s age and appearance. What we are left with is a man (who is a VISITOR and a PATIENT) is inside a DOCTOR’s house with his WIFE, while not being able to receive medical help, which was the stated purpose of the visit. This brings us to a clearly unsatisfactory opposition: received medical attention (stated goal) vs. did not (implied result) which is perfectly acceptable to a computer as a clean opposition. In other words, what happens when the LOVER script does not get evoked is that it is quite possible that a computer – which we use for theory verification (Ritchie 2001) – would still see it as a joke, with a superficial and very partial resolution (Rothbart & Pien 1976) of incongruity. Let us now consider the former question: how can this script get evoked? First of all, a computer should know stereotypes and respect them, at least for the humor processing. A question of how such stereotypes could be learned raises its head. From the ontological semantic point of view, not only should they be learned, but they should also be inserted somewhere into the knowledge base that can be accessible for script construction. A dismissive answer to such a problem is to say that such scripts are created though a manual process of analyzing common situations in jokes, and thus a script of ADULTERY (ideally, with a young and pretty wife being an experiencer) will be present in the ontology. We will take this easy way out with a full realization that it is a herculean task (see ontological coverage in Section 4). It should be noted that once an ADULTERY script is hypothesized, it is easy to see that it would be accepted by the computer, provided that the necessary components of the script are: one of the spouses meeting with a third person while the other spouse is away. Semisecluded setting and lack of noise is a plus. The question is how does one hypothesize to test this script out many, many others. Going back to our immediate problem of activation of the script LOVER or ADULTERY, it would be nice if one could identify a SOCIAL-ROLE that is present in the text of the joke that could be linked to such script, similarly to a PATIENT being linked to SEEKing_MEDICAL_HELP. The descriptions for such SOCIAL-ROLEs that are present in the text of the joke are: a whispering pretty young wife

6 This knowledge is not likely to be in the basic ontology, but let us assume that a computer might somehow learn something.

Scripts in the Ontological Semantic Theory of Humor

237

whose husband is away and a whispering man. Ideally, a role would be described by a single concept, however, there is no single concept that exists for either of these descriptions. Even if one takes into account recent advancements in word embedding and sense embedding, one may be tempted to find a word such that a vector representation of this word is similar to the combination of pretty young wife. While such combination is not found in the Google News corpus,7 it is possible that it may be found in a different corpus that may cover a domain of adultery or jokes. If such concept could be found, and it would be quite common for the adultery script, such script could be activated. At a present time word embedding of Google News corpus report that a combination of a whispering pretty young wife is similar to a mother (makes perfect sense why she whispers when her baby is asleep!) rather than lover, but we will leave this discussion for a later time. Another point worth making in the process of script processing is that while we have access to knowledge that is outside of the immediately activated concepts (through the ontology), unless they are needed to link the components of the joke, we can assume that they are discarded. Thus the scripts themselves should have a minimal footprint, as close to that of the joke text as possible. The first two sentences activate the scripts SEEKing-MEDICAL-HELP and house VISIT. The third sentence – if successfully interpreted – would activate the script of ADULTERY/LOVER, which takes place during the VISIT. The rest of the components of each of the scripts, for example their purpose, is not activated by the joke. The overlap can still be found in the activated components – the actors of the scripts are the same as in the analysis when ADULTERY was not discovered, and the opposition is rather similar too: received medical attention (stated goal) vs. did not (as real purpose of the VISIT is ADULTERY). At a coarser grain size, it is possible to represent these two scripts in terms of what was originally stated as a purpose of the visit and what was achieved: real/non-real (Raskin 1985, p. 108) or actual/non-actual (ibid, p. 111) through sex-non-sex feature (ibid, p.127). At a finer grain size, we can talk about this joke in terms of receiving medical help or not. These might be more acceptable oppositions as they are clearly on the same semantic axis (Attardo 1997, Attardo et al. 2001). Some may argue that such final representation does not account for everything that happens in the joke – the analysis is incomplete. It

7 Google News dataset, containing about 100 billion tokens, started to be used in the early word embedding papers. Word2vec vectors trained on Google News are publicly available, making it a convenient and very large data source that is frequently used.

238

Julia Taylor Rayz

is true that it may be incomplete, but not because of the grain size that one uses to represent an oppositeness. In order to check whether the analysis is complete one needs to carefully examine what scripts were activated (not the labels, but the each of the components): there is a huge different in the correctness of joke analysis between the two scenarios – in shorthand – below, also shown in Figures 1 and 2:

Figure 1: Two scripts that are commonly labeled as DOCTOR/LOVER without activation of adultery.

Figure 2: Two scripts with activation of adultery.

Scripts in the Ontological Semantic Theory of Humor

239

SEEK-MEDICAL-HELP AGENT PATIENT RESULTS-IN VISIT AGENT MAN

= PATIENT

BENEFICIARY WIFE SEEK-MEDICAL-HELP AGENT PATIENT RESULTS-IN VISIT AGENT MAN

= PATIENT

BENEFICIARY WIFE PURPOSE ADULTERY AGENT MAN

= PATIENT

BENEFICIARY WIFE

This means that while people analyze jokes for each other by providing verbal shortcuts to the necessary explanation – unless we assume that the shortcuts themselves are indicators of the salient explanation that the speaker is attempting to convey – this explanation may or may not represent an understanding of the actual scripts involved. In the case of DOCTOR/LOVER joke, the selection of grain size of the tags (or names, or annotations) of the scripts, by a human or a computer, points us to the fact that the scenario with ADULTERY was activated. The beauty of OSTH is that it is supposed to remove the guesswork of tag naming. Since the theory assumes the presence of world knowledge represented for computational use, whether some knowledge is activated or not becomes immediately clear when one examines the results of the found scripts as demonstrated in Figures 1 and 2 (notice the presence or absence of the ADULTERY concept in the graph). Granted, at the present level of OSTH publications, one needs to be well versed in Ontological Semantics to puzzle out what actually happened in these activated scripts, however, one can be optimistic that with the advancements in human friendly computing, this information would become more “naturally” available.

3 Grain size of SOs The

DOCTOR/LOVER

joke could arguably be used with any profession, not just that can offer professional advice and whose office could be at home. For example, a man could be guardedly walking at night, trying to stay unnoticed, to an attorney’s house. The rest of the conversation could mimic that of a DOCTOR,

240

Julia Taylor Rayz

DOCTOR/LOVER

joke. The man whispers, as he looks around, and asks if the attorney is at home. If the man looks like he needs an attorney, the rest of the scripts would work exactly the same, at least from the ontological processing point of view:

SEEK-LEGAL-HELP AGENT CRIMINAL RESULTS-IN VISIT AGENT MAN

= CRIMINAL

BENEFICIARY WIFE PURPOSE ADULTERY AGENT MAN

= CRIMINAL

BENEFICIARY WIFE

On the surface of it, we are dealing with two different pairs of scripts, DOCTOR/ LOVER and ATTORNEY/LOVER, yet, if one takes into account that SEEKing-LEGALHELP and SEEKing-LEGAL-HELP are likely to have the same ancestor, SEEKingPROFESSIONAL-HELP, and both concepts for DOCTOR and ATTORNEY are children of PROFESSIONAL-OCCUPATION, it can be clearly seen that we are dealing with exactly the same oppositions (see Figure 3).

Figure 3: Comparison of original DOCTOR/LOVER joke with any professional role as a substitute for DOCTOR.

Scripts in the Ontological Semantic Theory of Humor

241

It is clear from the example above that depending on what script annotation is chosen for both jokes, they could be looked at as identical if one only uses annotation to denote scripts, without looking at the script body. On the other hand, it may be useful to compare similar jokes based on the scripts without considering other resources by building a hierarchy of SOs. High level scripts are described by SSTH (p.111), namely actual/non-actual (“actual situation in which the hero of the joke finds himself or, somewhat more generally, in which the joke is actually set, and non-actual, non-existing situation which is not compatible with the actual setting of the joke”), normal/ abnormal (“normal, expected state of affairs and opposes it to the abnormal, unexpected state of affairs”), and possible/impossible (“possible, plausible situations and a fully or partially impossible or much less plausible situations”). All three of these pairs are the immediate subtypes of real/unreal dichotomy. Moreover, each of these oppositions summarizes and evaluates what happens in the jokes. In order to evaluate whether something can happen according to the text of the joke, one needs to somehow compare it with the general knowledge of what is possible and expected in real (actual) world. We will follow the sex-related feature of the real/unreal dichotomy to determine the result of our evaluation, since this feature is what DOCTOR/LOVER is based on. The following joke is first introduced by Raskin (p.25), quoting Rapp (1951, pp. 49–50), as an anti-pretense example: (1)

An aristocratic Bostonian lady hired a new chauffer. As they started out on their first drive, she inquired: “What is your name?” “Thomas, ma’am,” he answered. “What is your last name?” she said. “I never call chauffeurs by their first names.” “Darling, ma’am,” he replied. “Drive on – Thomas,” she said.

Two more joke, quoting Legman (1975) are first introduced as an example of an anti-clerical joke (ibid, p. 308) and suppression/repression humor (ibid, p. 200) respectably: (2)

The Archdeacon has got back from London, and confides to his friend the doctor, “Like Saint Peter, I toiled all night. Let us hope that like Saint Peter I caught nothing.”

242

(3)

Julia Taylor Rayz

An elderly client in a whore-house insists on having a special girl, who happens to be occupied. ‘But what’s she got that all my other girls haven’t got?’ asks the madam. ‘Patience,’ says the old man.

The last joke that we will consider in this section is joke by Pocheptzov (1974, p.41): (4)

An English Bishop receives the following note from the vicar of a village in his dioceses: “Milord, I regret to inform you of my wife’s death. Can you possibly send me a substitute for the weekend?”

While introducing script oppositeness, the following scripts are loosely identified for each of the jokes (pp. 108–110): – employee vs. lover (1) – The lady calls the driver by his family name vs. The lady calls the driver “darling” – church vs. sex (2) – The archdeacon was involved in debauchery vs. The archdeacon was involved in honest toil – sex vs. impotence (3) – In a whore-house, a client seeks a partner for sex vs. In a whore-house, a client seeks a partner for (almost) no sex – church vs. sex (4) – The vicar wants a substitute for himself vs. The vicar wants a substitute for his late wife Upon classification into the three coarse grain sizes of script oppositeness, joke (1) was classified as possible, plausible/impossible; jokes (2) and (4) were classified as actual/non-actual; joke (3) was classified as normal/abnormal. In addition to this classification, one can consider the sex-related feature of the jokes. This feature too can be split into several types: i. Sexual/non-sexual opposition: overt, unspecified: opposition of non-sexrelated script with a sex-related script, yet any specific sexual script is unavailable for the joke (p.149) ii. Sexual/non-sexual opposition: overt, specified: overt opposition of a nonsexual script with a specific sexual script (p.154) iii. Implied sexual/non-sexual opposition and an overt non-sexual opposition (p.160) iv. Implied sexual/non-sexual opposition is reinforced with specific sexual opposition associated with one of the standard sexual scripts (p.165)

Scripts in the Ontological Semantic Theory of Humor

243

Specific sexual scripts could further be listed: PROWESS, EXPOSURE, SEXUAL IGNORANCE, FORBIDDEN SEX, GENITAL SIZE, etc. If one considers this classification, jokes (1) and (4) fall under (i), where no specific sexual script is available; joke (2) involves an overt opposition of a sexual script with non-sexual script; joke (3) falls under (iv) with a standard script of PROWESS. Adding three subcategories of real/non-real dichotomy to sex-related features, it is possible to construct a 3x4 matrix, where each of the four jokes occupy their own spot (no two jokes share the same matrix position). Yet, earlier loose descriptions of the scripts showed that two of the jokes have the same script oppositions, namely, church vs. sex. This seems to indicate that the coarser grain size of classification, which should cluster more jokes together based on their high-level features, produced fewer overlaps between these jokes than a finer grain size of classification, which should pay attention to lower level description, and thus, produce fewer overlapping jokes. In other words, if church vs. sex jokes really have the same scripts and features, not just the labels, it should be reflected in the script dichotomy with sex-related feature matrix, which was not. Ideally, there should be a mechanism that takes a text of a joke and produces a representation that is compatible with any level of analysis. Yet, it would be doing so in such a way where an explanation to each step is possible. The Ontological Semantic Theory of Humor attempts to provide such mechanism. Let us analyze the employee vs lover joke (1). Roughly, the joke can be described in the following way, consistent with ontological description: A woman of superior social position, residing in Boston, hired a person to drive her automobile. The person, being a driver, started to drive the car to an unspecified location, with a woman taking a role of a passenger. The woman asked the driver’s name. The driver responded “Thomas,” which is a male first name. The woman asked the driver his last name. The woman stated that she does not refer to employees by their first names. The man responded that his last name was “Darling”, which is also an affectionate form of address to a beloved person. The woman gave an order to continue driving and addressed the driver as “Thomas,” which was his first name.

The same information can be described in ontological forms. An ontologybased text meaning representation of the 1st sentence of joke is below: HIRE-EMPLOYEE AGENT

SEM HUMAN GENDER VALUE FEMALE

BOSTON

RESIDENCE

VALUE

SOCIAL-STATUS

VALUE HIGH

PATIENT SEM HUMAN HAS-SOCIAL-ROLE SEM DRIVER

244

Julia Taylor Rayz

An ontology-based text meaning representation of the 2nd sentence of the joke is below: DRIVER IS-A PROFESSIONAL-OCCUPATION AGENT-OF-ROLE-IN DEFAULT DRIVE INSTRUMENT

DEFAULT CAR

DRIVE AGENT-ROLE-IN

DEFAULT DRIVER

INSTRUMENT

DEFAULT CAR

DESTINATION

VALUE UNKNOWN

BENEFICIARY

SEM HUMAN

HAS-SOCIAL-ROLE SEM PASSENGER GENDER VALUE FEMALE

An ontology-based text meaning representation of the 3rd sentence of the joke is below: ASK AGENT SEM HUMAN GENDER VALUE FEMALE BENEFICIARY-ROLE-IN SEM DRIVER THEME SEM NAME

An ontology-based text meaning representation of 4th sentence, with some information from the lexicon/onomasticon/proper name dictionary is below: RESPOND PRECONDITION

SEM ASK

AGENT-ROLE-IN

SEM DRIVER

BENEFICIARY

SEM HUMAN

GENDER VALUE FEMALE

“Thomas” Thomas – n1 pos: n syn-struc: NNP var0 sem-struc: HUMAN(HAS-FIRST-NAME(VALUE(“Thomas”))) THEME VALUE

An ontology-based text meaning representation of the 5th and 6th sentences, corresponding to the lady’s response:

Scripts in the Ontological Semantic Theory of Humor

245

ASK AGENT SEM HUMAN GENDER VALUE FEMALE BENEFICIARY-ROLE-IN SEM DRIVER THEME

SEM LAST-NAME

INFORM AGENT SEM HUMAN GENDER VALUE FEMALE BENEFICIARY-ROLE-IN SEM DRIVER TOPIC SEM REFERENCE THEME-ROLE-IN SEM EMPLOYEE HAS-FIRST-NAME SEM FIRST-NAME EPISTEMIC VALUE

0

An ontology-based text meaning representation of 7th sentence, corresponding to the driver’s response, and some information from the lexicon: RESPOND AGENT-ROLE-IN BENEFICIARY GENDER THEME

SEM DRIVER SEM HUMAN VALUE FEMALE SEM HUMAN

“DARLING” darling – n1 pos: n syn-struc: NP var0 sem-struc: SOCIAL-ROLE(EXPERIENCER-OF-ROLE-IN(SEM(LOVE))) HAS-LAST-NAME VALUE

An ontology-based text meaning representation of the punchline: REQUEST AGENT SEM HUMAN GENDER VALUE FEMALE BENEFICIARY-ROLE-IN SEM DRIVER HAS-NAME VALUE

“THOMAS”

TOPIC SEM DRIVE PHASE VALUE CONTINUE

A graphical representation is shown in Figure 4. It can be seen from the figure that the scripts oppose in handling of the word “darling,” which can be used a form of endearment (as indicated in the ontology-based lexicon sense of darling – n1) or as a last name of a person – an interpretation that is forced on the reader by the text of the joke, but that is easily compatible with the ontological knowledge. It is also seen from the text meaning representation (or Figure 4)

246

Julia Taylor Rayz

ask precondition respond precondition ask precondition inform precondition reference

beneficiary-role-in

agent theme

topic

female

high

Boston

professional-occupation

human

is-a employee

theme

agent-role-in respond prebeneficiaryagent epistemic condition role-in beneficiaryagent request 0 role-in agent-roletopic beneficiary in beneficiarytheme agent hire-employee drive theme role-in instrument agent beneficiary agent-role phase themehas-name -in role-in patient continue car Thomas driver human residence has-social- gender socialhas-social-role is-a is-a status role passenger

beneficiary

has-name has-firstname

has-first -name

has-last -name

hasvalue

name is-a is-a

last-name

first-name

has-value Darling is-a social-role experiencer-ofrole-in love

Figure 4: Graphical representation of joke (1). Grey rectangles show approximate script activation, slightly darker rectangles show representation of people involved in the conversation. White ovals represent ontological concepts.

that there is quite a bit going on in the interaction between the lady and her chauffeur, including pointing out to him that she never calls her chauffeur (and possibly any other employees) by their first name. The script of HIRE-EMPLOYEE is quite prominent in the interpretation, although it is a background knowledge script that is evoked only through the ontology. Nevertheless, it points us to the SOCIAL-ROLE concept of EMPLOYEE, which is an annotation that is given to one of

Scripts in the Ontological Semantic Theory of Humor

247

the scripts when Raskin loosely sketches them. The sense of the word “darling” that corresponds to the term of endearment is defined somewhat similarly to that of a LOVER: SOCIAL-ROLE(EXPERIENCER-OF-ROLE-IN(MAKE-LOVE)). The only difference is between the events MAKE-LOVE and LOVE, and while it can be argued that they are a world apart, they are very much connected in the ontological representation. Thus, the annotation of the second script is justified as well. Let us now look at the sex-related feature of this joke. According to Raskin, this joke follows the first category of the feature classification, namely, when “opposition of non-sex-related script with a sex-related script, yet any specific sexual script is unavailable for the joke.” As one does not see anything in the text that would activate any specific sexual script in the ontology (see Figure 4 or text meaning representation), this is indeed the case. It is more difficult to classify this joke as possible, plausible/impossible, at least from looking at its representation without any extra computation being involved. It is likely that this classification can be arrived from frequencies of the events described, rather than definitions and properties of objects and events in the described world, which is what an ontology typically provides. It is also possible that a reasoning mechanism can get us to that classification: since the lady stated that she does not call her employees by their first name, she is unlikely to do so. Yet, she is forced into this unsatisfactory solution since it is preferred to being misunderstood. If one ventures into the logic proper, there is a type of logics called modal logic that extends classical logics to express modality. Ontological semantics introduced modality features that somewhat resemble modalities of modal logic, although formal interpretation of these modalities have not been extensively developed. It is thus possible with the text-meaning representations described above to arrive at all versions of scripts mentioned in SSTH as an interpretation of joke (1). Joke (2) is much shorter than joke (1) and can be paraphrased in the following way, consistent with the ontological representation described below: The Archdeacon returns from a trip to London, and secretly tells his friend who is a doctor, that similarly to Saint Peter he worked hard all night. He hopes that similarly to Saint Peter, he did not catch anything.

Just like joke (1), this joke relies on multiple meanings of the words, each of which is compatible with the opposing scripts. However, unlike the previous joke, this joke relies on knowledge of the quotes from the Bible, not just ontological representation of the world. It is possible, however, to detect the two scripts with some superficial knowledge of Christianity, namely, with the knowledge of who Saint Peter is. This knowledge is captured by basic ontological semantics. The interpretation is made much easier if one allows the

248

Julia Taylor Rayz

knowledge of the event in question, interpretation of which could be derived if the Bible would be one of the texts that ontological semantics processed. The text meaning representation is also depended on the fact that in this case the expression “caught nothing” means didn’t catch any sexually transmitted disease. This interpretation is reinforced by an ontological connection from any DISEASE to DOCTOR who cures diseases, and who happens to be the recipient of secret information reported by the archdeacon. The activation of this sense is not as straight-forward as activation of the sense of a lover in the previous joke, but arguably easier than activation of the sexual connection in the DOCTOR/ LOVER joke that relies only on the world-knowledge based inference. MOVE ORIGIN VALUE LONDON AGENT-ROLE-IN

ARCHDEACON

INFORM AGENT-ROLE-IN SEM

ARCHDEACON

BENEFICIARY SEM HUMAN HAS-SOCIAL-ROLE SEM DOCTOR HAS-SOCIAL-ROLE SEM FRIEND FRIEND-OF SEM

ARCHDEACON

TOPIC SEM WORK DURATION VALUE NIGHT DIFFICULTY VALUE HARD SIMILAR-TO SEM WORK AGENT VALUE “ST. PETER” TOPIC SEM HOPE THEME SEM CATCH AGENT-ROLE-IN SEM

ARCHDEACON

THEME DEFAULT SEXUALLY-TRANSMITTED-DISEASE EPISTEMIC VALUE

0

SIMILAR-TO SEM CATCH AGENT VALUE “ST. PETER” EPISTEMIC VALUE

0

THEME SEM FISH

One could argue that the two actual scripts that are activated by this joke are FISHING and SEX, not CHURCH and SEX. On the surface of the text meaning interpretation, it might be the case, however the activation of the script of FISHING is impossible without knowledge of the events in the Bible, which brings us back to the script of RELIGION or CHURCH. One may ask, where is the actual opposition, what makes FISHING be opposing to SEX? The answer is nothing – and they are likely to be quite compatible in general – unless one looks at the context of the text in question. Again, within ontological specifications or from representing

Scripts in the Ontological Semantic Theory of Humor

249

various text of the Christianity-related domain, a system should learn – possibly unwisely – that sex-related topics typically do not appear in the context of this domain, and if they do, it is likely to be a negative scenario. Moreover, participating in an event that can result in a sexually transmitted disease is likely to be perceived as a sin, where is any action of Saint Peter, is the opposite of it. Thus, through reasoning, one can arrive at a clean opposition of SIN vs. VIRTUE. Another joke that is tentatively annotated by Raskin having CHURCH vs. SEX scripts is joke (4). This joke can be paraphrased in the following way, consistent with the ontological representation described below: An English Bishop receives a note from a vicar. The vicar is assigned to a village in the dioceses overseen by the bishop. The note informs the bishop of the vicar’s wife’s death. The vicar asks the bishop for a substitute for the weekend.

The joke is based on the fact that it is not clear whether the vicar is asking to substitute himself for the services over the weekend or to substitute the services of his late wife for the weekend. The joke only works if the interpretation of substituting the wife is the first interpretation that a system detects. Luckily, in order to fill the topic of substitution the system would have to refer to the previous sentence, and the wife is mentioned after the vicar, thus resulting in the desired selection for processing. The selection of the vicar for substitution follows shortly after, assuming that all possible interpretations are considered. The resulting representation is shown below and sketched in Figure 5. RELAY-INFORMATION AGENT SEM VICAR BENEFICIARY SEM BISHOP INSTRUMENT SEM NOTE THEME-INFORMATION SEM INFORM AGENT SEM VICAR BENEFICIARY SEM BISHOP TOPIC SEM DEATH EXPERIENCER-ROLE-IN SEM WIFE EXPERIENCER-OF-ROLE-IN SEM MARRIAGE EXPERIENCER SEM VICAR CAUSE-FOR SEM REQUEST TOPIC SEM REPLACE DURATION VALUE WEEKEND BENEFICIARY SEM WIFE BENEFICIARY SEM VICAR VICAR IS-A PROFESSIONAL-OCCUPATION AGENT-OF-ROLE-IN DEFAULT RELIGIOUS-SERVICE GEO-LOCATION SEM DIOCESES HEADED-BY SEM BISHOP

250

Julia Taylor Rayz

Figure 5: Ontological representation of joke (4). Shapes of grey correspond to different scripts.

Similar to DOCTOR/LOVER joke, the SOCIAL-ROLEs here activate the scripts of interest, namely RELIGIOUS-SERVICE and MARRIAGE. The text meaning representation explicitly states that the VICAR performs RELIGIOUS-SERVICEs, and thus is it for this purpose he needs to be replaced (relying on the unstated knowledge that regular religious services are on the weekend). It is not explicitly stated for what services a WIFE could be replaced, and one should derive it from the knowledge of the world, which, hopefully, is captured in the ontology. However, if it is not, we are left with an interpretation that contains an unsatisfactory opposition, namely of gender (wife vs. husband), which would still result in a joke, according to the theory. If the ontological knowledge is complete, including containing information that SEXUAL-ACTIVITY is part of MARRIAGE, one would arrive to the CHURCH/SEX opposing scripts. This by itself does not mean

Scripts in the Ontological Semantic Theory of Humor

251

that the script interpretation would be complete – one would need to link wifesubstitution to sex outside of wedlock to achieve a CHURCH/SEX opposition, which may require common sense reasoning, which in turn is mostly absence in the present day computing systems. Nevertheless, in a perfect world, computational analysis of jokes (4) and (2) have distinct similarities: not only do they have CHURCH/SEX opposition, the reason for this opposition is due to churchmen appearing to have sex outside of wedlock. The differences in the scripts are also clear: in joke (2) the archdeacon is aware of his actions, hoping that there would be no consequences for them; in joke (4) the vicar is asking his superior for what could be interpreted to result in unspecified sexual activity (or hiring a housekeeper for the weekend).

4 Ontological quality: Are we there yet? The details of interpretation of both scripts depend on the conceptualization of the phenomena. However, which phenomena should be conceptualized remains an open question. Arguably, it is impossible to thoroughly describe every detail of every situation, potentially because it is easy to overlook their existence. It is easy to adjust an ontology to represent text in jokes (2) and (4) if one understands the phenomena touched on by these jokes. Misunderstanding of it could result in inaccurate conceptualization, which, in turn, could result in inaccurate ontology. Such inaccuracies are difficult to evaluate as the fact that an ontology correctly represents knowledge in 100 jokes does not mean that it would correctly represent the world described in 101st joke. According to Guarino (2009), an ontology should have a high precision (“non-intended models are excluded”) and maximize correctness (“negative examples are excluded”). Some of the unsatisfactory interpretations of opposing scripts in both DOCTOR/LOVER and REPLACINGVICAR/REPLACINGWIFE jokes may be there precisely because of the low precision of the ontological models of CHURCH and SEX, eventually resulting in low correctness. It is tempting to look for correlations between the number of scripts that a system sees and precision or correctness. One should be careful, however, to differentiate between scripts that are interpreted through varying grain size of information and scripts that are interpreted when the grain size is fixed. For example, we described several seemingly opposing script annotation in joke (2), such as: FISHING/SEX, CHURCH/SEX, ACTUAL/NON-ACTUAL, going up or down the grain-size hierarchy. However, all of them rely on the absence or presence of (un)permissible sex. Joke (4), on the other hand, had an additional opposition,

252

Julia Taylor Rayz

namely gender of people to be replaced, that was of the same grain size as the activity that these people participated in. The gender opposition is easier to discover computationally as fewer inferences have to be made (at the same basic level). Reducing the number of such background oppositions from allowable ones may constrain the model further and increase the precision. Such reduction is not easy in practice. As an example, let us consider joke (3), which is paraphrased below: An elderly client in a brothel insists on a particular employee, who is occupied. The employee’s supervisor inquires what skill the girl possesses that others do not. The client responds that the girl possesses patience.

The following knowledge is required to represent the joke for computational detection: – Brothel – is a place where sexual activity is performed for payment – females who perform the said activity are typically employed – Female employees are supervised by another female, referred to as madam – Female employees may have regular clients, who are typically male – Sexual activity – Changes in arousal are experienced with age, measured by intensity of stimuli and time of stimulation – The refractory period increases with age Age related information about sexual activity may be received form the medical domain texts (for the sake of accuracy), however, it immediately increases the ontological coverage of the world, which is likely to impact precision, as the sample size of tested jokes would remain proportionally (to the domain increase) small. It is possible that in order to claim sex-feature coverage by the ontology that supports the OSTH, the medical domain is a must. However, as we have seen, for any higher-level script pair (SEX vs. CHURCH) coverage of knowledge of the second domain is also required. If one were only interested in what is theoretically possible, one would evaluate the OSTH only on those domains that are well covered. In other words, if the ontology represents a model of the phenomena well, one could evaluate the OSTH for the scripts that the ontology covers. If the scripts are available to the ontology, but overlap and oppositeness are not evoked upon reading a text of a joke, then it would indicate that the theory does not do the job. However, if the scripts are not present, at least theoretically, no conclusion

Scripts in the Ontological Semantic Theory of Humor

253

can be reached about the humor theory itself. On the other hand, if such constraints are created, practically, very few jokes would be able to be evaluated due to poor coverage of the domain. A possible solution for it is create ontology coverage gradually and granularly. Such granular perspective would mean that if one is interested in sexrelated humor (which could be looked at as a category of its own (Ruch 1992)) one would need to look more carefully at components of the jokes. For sexrelated scripts to be processed, one would expect to cover ontological knowledge relevant to social interaction (relevant for jokes (2), (3), and (4)) as well as biological processes of the human body (relevant for joke (2) and (3)). Sexrelated scripts are only part of the equation, and one need to represent knowledge for the other part. For example, if one is interested in covering more examples in the church vs. sex SO, the ontology has to be enriched with the domain of religion. If religious texts are covered by the ontological knowledge, one can expect joke (2) to be recognized. If the hierarchy of Anglican ministry is covered by the ontological knowledge, then one could expect joke (4) to be recognized (see Figure 6). It should be noted that while only scripts were the topic of this paper, the OSTH is capable of handling other resources that are outlined in the GTVH. For example, targets, language, situation and narrative strategy could be derived though ontological representation. Similar to the hierarchy of scripts, it would be possible to derive a hierarchy of some of these knowledge resources that would be helpful in closer comparison of jokes. Similar restrictions would need to be obeyed with processing these resources: only those that are known to the ontology that is used for text processing could be expected to be discovered during text processing. In summary, the Ontological Semantic Theory of Humor, whenever used, should be used for what it is capable of doing with the resources that are available to it. If the resources are augmented to analyze a given subset of jokes, such analysis can be done on a much more detailed level than the SSTH due to the nature of the foundation of OSTH, namely Ontological Semantics. It is also possible to use an OSTH analysis to avoid disagreements about where overlaps or oppositions are within the scripts, provided that one agrees on the knowledge that is available, prior to analyzing the jokes. Finally, the annotations of scripts (labels) play very little role in OSTH processing, and are mostly useful for determining what knowledge is needed for an expected analysis.

Figure 6: Interaction of various script-related knowledge components.

254 Julia Taylor Rayz

Scripts in the Ontological Semantic Theory of Humor

255

Acknowledgements: The author is grateful to Kanishka Misra for working with word embeddings for Doctor/Lover joke, and to Salvatore Attardo for editorial suggestions.

References Attardo, Salvatore 1997. The semantic foundations of cognitive theories of humor. Humor: International Journal of Humor Research 10(4). 395–420. Attardo, Salvatore, Hempelmann, Christian F., & DiMaio, Sara. 2002. Script oppositions and logical mechanisms: Modelling incongruities and their resolutions. Humor: International Journal of Humor Research 15(1). 3–46. Attardo, Salvatore & Raskin, Victor. 1991. Script theory revis(it)ed: joke similarity and joke representation model. Humor: International Journal of Humor Research 4(3/4). 293–347. Bolukbasi, Tolga, Chang, Kai-Wei, Zou, James Y., Saligrama, Venkatesh & Kalai, Adam T. 2016. Man is to computer programmer as woman is to homemaker? Debiasing word embeddings. In Daniel D. Lee and Masashi Sugiyama and Ulrike V. Luxburg and Isabelle Guyon and Roman Garnett. (Eds.) Advances in Neural Information Processing Systems 29 (NIPS 2016) 4349–4357.Neural Information Processing Systems Foundation, https://pa pers.nips.cc Garg, Nikhil, Schiebinger, Londa, Jurafsky, Dan & Zou, James. 2018. Word embeddings quantify 100 years of gender and ethnic stereotypes. Proceedings of the National Academy of Sciences 115(16) E3635–E3644 Guarino, Nicola. 2008. Introduction to applied ontology and ontological analysis. First Interdisciplinary Summer School on Ontological Analysis, Trento, Italy. Guarino, Nicola, Oberle, Daniel & Staab, Steffen. 2009. What is an Ontology? In: Staab, Steffen., Studer, Rudi. (eds.) Handbook on Ontologies, 2nd ed.1–17. Springer, Heidelberg Hempelmann, Christian F., Taylor, Julia. M., & Raskin, Victor. 2010. Application-guided ontological engineering, In Hamid A. Arabnia, David. de la Fuente, Elena B. Kozerenko, and Josè A. Olivas. Eds. Proceedings of the 2010 International Conference on Artificial Intelligence, ICAI 2010, July 12–15, 2010, Las Vegas NE. Nirenburg, Sergei. & Raskin, Victor. 2004. Ontological Semantics. Cambridge, MA: MIT Press. Raskin, Victor. 1985 Semantic Mechanisms of Humor. Dordrecht/Boston/Lancaster: D. Reidel Raskin, Victor, Hempelmann, Christian F., & Taylor, Julia M. 2009. How to understand and assess and theory: The evolution of the SSTH into the GTVH and now into the OSTH. Journal of Literary Theory 3 (2)285–311. Raskin, Victor, Hempelmann, Christian F., & Taylor, Julia M. 2010. Guessing vs. Knowing: The Two Approaches to Semantics in Natural Language Processing, In Aleksandr E. Kibrik. Ed. Proceedings of Annual International Conference Dialogue 2010, Moscow, Russia, May 2010. Ritchie, Graeme. 2001. Current directions in computational humour. Artificial Intelligence Review 12(2),119–135 Rothbart, Mary K. & Pien, Diana 1976. Elephants and Marshmallows: A Theoretical Synthesis of Incongruity-Resolution and Arousal Theories of Humour. In Chapman, Antony J., and Hugh C. Foot (eds.), It’s a Funny Thing, Humour. Oxford: Pergamon, 37–40.

256

Julia Taylor Rayz

Ruch, Willibald. 1992. Assessment of appreciation of humor: Studies with the 3WD humor test. Advances in Personality Assessment 9, 27–75. Samson, Andrea & Hempelmann, Christian F. 2011. Jokes and cartoons with and without background incongruity: Does more required suspension of disbelief affect humor perception? Humor: International Journal of Humor Research 24(4) 167–186 Taylor, Julia M., Raskin, Victor, Hempelmann, Christian F., & Attardo, Salvatore. 2010. An unintentional inference and ontological property defaults. In Proceedings of SMC 2010: IEEE International Conference on Systems, Man, and Cybernetics. Istanbul, Turkey, October 10–13.

Vladik Kreinovich and Olga Kosheleva

Which fuzzy logic operations are most appropriate for ontological semantics: Theoretical explanation of empirical observations Abstract: In several of their papers, Victor Raskin and coauthors proposed to use fuzzy techniques to make ontological semantics techniques more adequate in dealing with natural language. Specifically, they showed that the most adequate results appear when we use min as an “and”-operation and max as an “or”operation. It is interesting that in other applications of fuzzy techniques, such as intelligent control, other versions of fuzzy techniques are the most adequate. In this chapter, we explain why the above techniques are empirically the best in the semantics case. Keywords: Ontological semantics, fuzzy logic, “and”-operation, “or”-operation, distinguishability

1 Formulation of the problem Use of fuzzy degrees in ontological semantics (OS): a brief (and simplified) reminder. In several of his papers, V. Raskin et al. proposed to use fuzzy techniques to make ontological semantics a more adequate description of natural language; see, e.g., Taylor & Raskin 2010, 2016a/b. For example, according to Raskin et al., from the purely logical viewpoint, there are many correct answers to a query “Who can drive a car?”. For example, “a man”, “a woman”, “an engineer” are all logically correct answers. However, from the natural language viewpoint, the above answers sound weird. From the natural language viewpoint, the most natural answer is “an adult.” This answer is the most natural because it is the most general. It is desirable to utilize this idea in computer-based natural language processing algorithms. One of the main difficulties in such use of computers is that computers have been originally designed to process numbers, not words, and computers

Vladik Kreinovich, Olga Kosheleva, University of Texas at El Paso, 500 W, University, El Paso, USA https://doi.org/10.1515/9781501511707-012

258

Vladik Kreinovich and Olga Kosheleva

are still much better in processing numbers than in processing textual information. Thus, to utilize the above idea in a computer-based system, it is desirable to describe different degrees of generality by numbers. Notions from natural language form a natural hierarchy, ranging from the most general notions to the most specific ones. The most general notions have the highest level of generality. Let us denote this level by an integer L. Slightly less general ones – barely distinguishable from the most general ones – fill the next, ðL − 1Þ-st level of generality; then comes the ðL − 2Þ-nd level, etc., until we get to the lowest level of generality. It is reasonable to select the integer L so that the lowest level of generality is level 0 – in other words, it is reasonable to select, as L, the number of levels in the corresponding hierarchy minus one. To each notion, we can thus assign a value v from 0 to L that describes the degree to which this particular notion is general. The idea of a degree to which some property is satisfied is well known in computer science. This idea can be traced to the 1960s pioneering works of Lotfi Zadeh who realized that many imprecise (“fuzzy”) natural-language concepts like “general” (or “small”) do not represent well-described dichotomy into general–not-general, small–not-small. Instead, each object can be described by a degree to which this object satisfies the given property – e.g., a degree to which a notion is general, or a degree to which an object is small. This realization led to a development of efficient technique for processing such degrees, technique known as fuzzy logic; see, e.g., Belohlavek et al. 2017; Klir & Yuan 1995; Mendel 2017; Nguyen & Walker 2006; Novak & Moc̆kor̆ 1999; Zadeh 1965. These degrees can be viewed as a natural generalization of simple dichotomic notions (i.e., notions which can be either true or false). For such simple notions: – “true” is usually represented in a computer as the number 1, and – “false” is usually represented as the number 0. From this viewpoint, it is natural to re-scale the degrees so that: – the largest possible degree – corresponding to maximum confidence – becomes 1, while – the smallest possible degree becomes 0, so that intermediate degrees corresponds to numbers from the intervals ð0, 1Þ. For original degrees ranging from 0 to some number L, the most natural way to achieve such a re-scaling is to divide the corresponding degree by L. This is exactly how degrees are usually assigned in fuzzy logic: to describe to what extent some value of a quantity of interest is small, we ask the expert to mark its degree of slowness of a Likert-type scale – e.g., on a 0 to 10

Which fuzzy logic operations are most appropriate for ontological semantics

259

scale. If the expert marks his/her degree by a value 7 on a 0 to 10 scale, we take 7=10 = 0.7 as the degree to which, to this expert, the corresponding object is small. Need for “and”- and “or”-operations with fuzzy degrees. One of the main applications of fuzzy logic is utilizing imprecise (fuzzy) expert knowledge in automated computer systems, e.g., in systems designed for control. To achieve this objective, we need to translate expert knowledge into computerunderstandable numerical terms. The main ideas of fuzzy logic enable us to provide a numerical value for simple properties like “pressure is low”. However, expert rules often use more complex conditions, e.g., “and”- and “or”-combinations of simple properties. For example, an expert can recommend a certain control strategy for situations in which the pressure in a chemical reservoir is low and the temperature is low, and another strategy for the opposite situation which the pressure is high or the temperature is high. In the ideal world, we should do, for each such combination, what we did for individual properties: ask the expert, for each pair of pressure and temperature, to indicate to which extend the complex property “pressure is low and temperature is low” is satisfied for the two given values. However, realistically, this is not possible: there are too many such combinations, and it is not feasible to ask the expert’s opinion about all of them. Because of this impossibility to explicitly elicit the corresponding degree from the experts, we need to estimate these degrees based on whatever information we have – i.e., based on the known degrees to which individual statements hold. In other words, we need to be able, given our degrees of confidence a and b in statements A and B, to provide a degree to which the “and”-combination “A and B” is satisfied. The algorithm that computes this estimate is known as an “and”-operation (or, for historical reasons, a t-norm); see, e.g., Belohlavek et al. 2017; Klir & Yuan 1995; Mendel 2017; Nguyen & Walker 2006; Novak et al. 1999; Zadeh 1965. We will denote such an algorithm by f& ða, bÞ. Similarly, we must be able, given our degrees of confidence a and b in statements A and B, to provide a degree to which the “or”-combination “A or B” is satisfied. The algorithm that computes this estimate is known as an “or”operation (or, for historical reasons, a t-conorm) Belohlavek et al. 2017; Klir & Yuan 1995; Mendel 2017; Nguyen & Walker 2006; Novak et al. 1999; Zadeh 1965. We will denote such an algorithm by f_ ða, bÞ. Which “and”- and “or”-operations should we choose: general idea. There are many possible “and”- and “or”-operations. Depending on the objective, in different situations, different pairs of “and”- and “or”-operations are optimal; see, e.g., Kreinovich 1998, Kreinovich & Kumar 1990, Kreinovich et al.

260

Vladik Kreinovich and Olga Kosheleva

1992; Nguyen & Kreinovich 1998, Nguyen et al. 1993; Nguyen et al. 1994, Nguyen & Walker 2006; Smith & Kreinovich 1995; Tolbert 1994. For example: – if we want to select the most stable control, then it is best to select f& ða, bÞ = minða, bÞ and f_ ða, bÞ = minða + b, 1Þ; – if we want to select the smoothest control, then it is best to select f& ða, bÞ = a · b (or f& ða, bÞ = maxða + b − 1, 0Þ) and f_ ða, bÞ = maxða, bÞ; – if we want to select a control which is, on average, the least sensitive to measurement uncertainty with which we know inputs, then we should select f& ða, bÞ = a · b and f_ ða, bÞ = a + b − a · b; – if we want a control which is the least sensitive in the worst case, then we should select f& ða, bÞ = minða, bÞ and f_ ða, bÞ = maxða, bÞ; the same pair should be selected if we want a control which is the easiest (and fastest) to compute, etc. In different application areas, different pairs of operations most adequately describe expert reasoning: e.g., different pairs are most appropriate in describing expertise of medical doctors and expertise of geoscientists; see, e.g., Buchanan & Shortliffe (1984). Which “and”- and “or”-operations are most adequate for ontological semantics: an empirical fact that needs explaining. According to Taylor & Raskin (2010, 2016a/b), for ontological semantics, the most adequate selection is f& ða, bÞ = minða, bÞ and f_ ða, bÞ = maxða, bÞ. Why? How can we explain this empirical fact? The main objective of this paper is to provide such an explanation.

2 Analysis of the problem Possible degrees. As we have mentioned earlier, in the ontological semantics, the corresponding degrees d 2 ½0, 1 are obtained after we divide an integer – the ordinal number of the corresponding level of generality – by the ordinal number L of the level containing the most general notions. As a result, we have the following set S of possible values of the degree of generality:   1 2 1 S = 0, , , . . . , 1 − , 1 . L L L Thus, we are interested in operations f& ða, bÞ and f_ ða, bÞ that transform values a, b 2 S into a new value c 2 S. By using mathematical notations, we can say that we need to find functions f& : S × S ! S and f_ : S × S ! S.

Which fuzzy logic operations are most appropriate for ontological semantics

261

What properties should these functions satisfy? First property: “and” and “or”-operations must be conservative. As we have mentioned earlier, degree 0 usually corresponds to “false” and degree 1 usually corresponds to “true”. Thus, it is reasonable to require that when each of a and b is either 0 or 1, our “and”- and “or’’-operations become the usual “and”- and “or”-operations of the 2-valued propositional logic: f& ð0, 0Þ = f& ð0, 1Þ = f& ð1, 0Þ = 0, f& ð1, 1Þ = 1, f_ ð0, 0Þ = 0, f_ ð0, 1Þ = f_ ð1, 0Þ = f_ ð1, 1Þ = 1. Second property: monotonicity. Our degree of belief in A& B cannot be larger than our degree of belief in each of the original statements. So, we must have f& ða, bÞ ≤ a and f& ða, bÞ ≤ b. Similarly, our degree of belief in A _ B cannot be smaller than our degree of belief in each of the original statements. So, we must have a ≤ f_ ða, bÞ and b ≤ f_ ða, bÞ. Third property: “and”- and “or”-operations must be consistent with distinguishability. According to the original definition of degree, the neighboring degree Li and i +L 1 are barely distinguishable from each other. To be more precise, there is no intermediate degree between then. So, if we slightly increase the lower degree or slightly decrease the higher degree, they will become truly indistinguishable. In this sense, pairs of neighboring values are “barely distinguishable” in the sense that they are limits of indistinguishable pairs. If we denote indistinguishability by ≈ and “barely distinguishable’’ by ⁓, then we can say that a⁓a′ if and only if there exist values an ! a and a′n ! a′ for which an ≈ a′n for all n. Intuitively, if a is indistinguishable from a′ and b is indistinguishable from b′, then, e.g., f& ða, bÞ should be indistinguishable from f& ða′, b′Þ. So: – if a⁓a′ are limits of indistinguishable pairs an ≈ a′n , b⁓b′ are limits of indistinguishable pairs bn ≈ b′n , – then, since an is indistinguishable from a′n and bn is indistinguishable from b′n , we conclude that f& ðan , bn Þ should be indistinguishable from f& ða′n , b′n Þ. In the limit n ! ∞, we conclude that the limit values f& ða, bÞ and f& ða′, b′Þ should be barely distinguishable: f& ða, bÞ⁓f& ða′, b′Þ. Now, we are ready to formulate our main result.

262

Vladik Kreinovich and Olga Kosheleva

3 Definitions and the main results Definition 1. Let L be a positive integer, and let S be the set   1 2 1 0, , , . . . , 1 − , 1 . L L L We say that values a, b 2 S are barely distinguishable and denote it by a⁓b if they either coincide or are neighbors in the above sequence. Comment. One can easily check that a⁓b , ja − bj ≤

1 . L

Definition 2. By an OS-“and”-operation, we mean a function f& : S × S ! S that satisfies the following properties: – f& ð1, 1Þ = 1; – f& ða, bÞ ≤ a and f& ða, bÞ ≤ b for all a and b; – if a⁓a′ and b⁓b′, then f& ða, bÞ⁓f& ða′, b′Þ. Comment. Note that we did not use all the above properties – e.g., we did not use the fact that f& ð0, 1Þ = 0, neither we require commutativity or associativity, as is usually done with “and”-operations in fuzzy literature. We did not use these additional properties because, as the following result shows, even without these properties, we can uniquely determine the corresponding “and”-operation. The three properties that we did use are necessary: as we show a few lines later, without even one of these properties, the following result will not hold. Proposition 1.

The only OS-“and”-operation is minða, bÞ.

Comments. – For readers’ convenience, all the proofs are placed in a special (last) Proofs section. – This result explains the empirical fact – that minimum is the most adequate “and”-operation for operational semantics. – All three properties of an OS-“and”-operation are needed for Proposition 1 to hold: – if we do not require that f& ð1, 1Þ = 1, then we can have f& ða, bÞ = 0 for all a and b; – if we do not require monotonicity, then we can have f& ða, bÞ = 1 for all a and b; and – if we do not require consistency with distinguishability, then we can have f& ð1, 1Þ = 1 and f& ða, bÞ = 0 for all other pairs ða, bÞ.

Which fuzzy logic operations are most appropriate for ontological semantics

263

– A similar result explains why maximum turned out to be the most adequate “or”-operation. Definition 3. By an OS-“or”-operation, we mean a function f_ : S × S ! S that satisfies the following properties: – f_ ð0, 0Þ = 0; – a ≤ f_ ða, bÞ and b ≤ f_ ða, bÞ for all a and b; – if a⁓a′ and b⁓b′, then f_ ða, bÞ⁓f_ ða′, b′Þ. Proposition 2.

The only OS-“or”-operation is maxða, bÞ.

Comment. All three properties of an OS-“or”-operation are needed for Proposition 2 to hold: – if we do not require that f_ ð0, 0Þ = 1, then we can have f_ ða, bÞ = 1 for all a and b; – if we do not require monotonicity, then we can have f_ ða, bÞ = 0 for all a and b; and – if we do not require consistency with distinguishability, then we can have f_ ð0, 0Þ = 0 and f_ ða, bÞ = 1 for all other pairs ða, bÞ.

4 Auxiliary result: From binary to n-ary operations In practice, we often need to combine more than two statements. In fuzzy logic, it is usually assumed that we combine them one by one: for example, if we need to “and”-combine a, b, and c, then: – first, we combine a and b into f& ða, bÞ, and then – we combine the result of a-and-b combination with c, resulting in f& ðf& ða, bÞ, cÞ. Instead, we can explicitly define n-ary “and”- and “or”-operations for every n. Definition 4. Let n ≥ 2 be an integer. By an n-ary OS-“and”-operation, we mean a function f& : Sn ! S that satisfies the following properties: – f& ð1, . . . , 1Þ = 1; – f& ða1 , . . . , an Þ ≤ ai for all i; – if ai ⁓a′i for all i, then f& ða1 , . . . , an Þ⁓f& ða′1 , . . . , a′n Þ. Proposition 3.

The only n-ary OS-“and”-operation is minða1 , . . . , an Þ.

Definition 5. Let n ≥ 2 be an integer. By an n-ary OS-“or”-operation, we mean a function f_ : Sn ! S that satisfies the following properties:

264

Vladik Kreinovich and Olga Kosheleva

– f_ ð0, . . . , 0Þ = 0; – ai ≤ f_ ða1 , . . . , an Þ for all i; – if ai ⁓a′i for all i, then f_ ða1 , . . . , an Þ⁓f_ ða′1 , . . . , a′n Þ. Proposition 4.

The only n-ary OS-“or”-operation is maxða1 , . . . , an Þ.

5 Proofs Proof of Proposition 1 1 . Let us first prove that for every a 2 S, we have f& ða, aÞ = a. Specifically, we will prove, by induction over k, that for all k, we have   k k k =1− . f& 1 − , 1 − (1) L L L Indeed, due to conservativeness, this property holds for k = 0, so we have the induction base. Suppose now that the property (1) holds for k, let us prove that it holds for k + 1 as well. Due to monotonicity, we have   k+1 k+1 k+1 ≤1− ,1− . (2) f& 1 − L L L On the other hand, we have  1−

   k+1 k ⁓ 1− . L L

Thus, due to consistency with distinguishability, we have     k+1 k+1 k k k f& 1 − ⁓f& 1 − , 1 − =1− . ,1− L L L L L

(3)

1 We know that a⁓b means ja − bj ≤ , thus L a≥b−

1 . L

Therefore, the property (3) implies that     k+1 k+1 k 1 k+1 ≥ 1− − =1− ,1− . f& 1 − L L L L L From (2) and (4), we conclude that

(4)

Which fuzzy logic operations are most appropriate for ontological semantics

265

  k+1 k+1 k+1 =1− f& 1 − ,1− . L L L The induction step is proven. Thus, indeed, f& ða, aÞ = a for all a. 2 . Let us now prove that f& ða, bÞ = minða, bÞ for all a and b. In Part 1, we have already proved it for the case when a = b, so to complete the proof, it is sufficient to consider the case when a≠b. We will prove it for the case when a > b; the case a < b is similar. Specifically, for each k, we will prove, by induction over , ≥ 0, that we have   k k−, k−, = . (5) f& , L L L For , = 0, we have already proved it in Part 1 of this proof. Let us now assume that we have already proved the equality (5) for a given value ,; let us prove that it is true for , + 1 as well. Indeed, from monotonicity, it follows that   k k − ð, + 1Þ k − ð, + 1Þ ≤ . (6) f& , L L L On the other hand, since k − ð, + 1Þ k − , ⁓ , L L the consistency with distinguishability implies that     k k − ð, + 1Þ k k−, k−, ⁓f& , = f& , . L L L L L

(7)

As we have mentioned in Part 1 of this proof, a⁓b implies a≥b−

1 . L

Therefore, the property (7) implies that   k k − ð, + 1Þ k − , 1 k − ð, + 1Þ ≥ − = . f& , L L L L L

(8)

From (6) and (8), we conclude that   k k − ð, + 1Þ k − ð, + 1Þ = f& , . L L L The induction step is proven. Thus, indeed, f& ða, bÞ = minða, bÞ for all a and b.

266

Vladik Kreinovich and Olga Kosheleva

Proof of Proposition 2. One can easily check that a function f_ ða, bÞ is an OS“or”-operation if and only if its “dual” def

f& ða, bÞ = 1 − f_ ð1 − a, 1 − bÞ is an OS-“and”-operation. Since min and max are duals to each other, Proposition 2 follows from Proposition 1. Proof of Proposition 3 is similar to the proof of Proposition 1: – first, similarly to Part 1 of that proof, we prove, by induction, that f& ða, . . . , aÞ = a for all a; – then, to prove that f& ða1 , . . . , an Þ = minða1 , . . . , an Þ, similarly to Part 2 of that proof, we start with a = maxða1 , . . . , an Þ and then step-by-step decrease each input by L1 , proving each time that the desired equality if preserved. Proof of Proposition 4. One can easily check that a function f_ ða, bÞ is an n-ary OS-“or”-operation if and only if its “dual” def

f& ða1 , . . . , an Þ = 1 − f_ ð1 − a1 , . . . , 1 − an Þ is an n-ary OS-“and”-operation. Thus, Proposition 4 follows from Proposition 3. Acknowledgments: This work was supported in part by the National Science Foundation grant HRD-1242122 (Cyber-ShARE Center of Excellence). The authors are greatly thankful to Victor Raskin for valuable discussions.

References Belohlavek, Radim, Joseph W. Dauben, & George J. Klir. 2017. Fuzzy Logic and Mathematics: A Historical Perspective, Oxford University Press, Oxford/New York, NY: Oxford UP. Buchanan, Bruce G. & Edward H. Shortliffe. 1984. Rule Based Expert Systems: The MYCIN Experiments of the Stanford Heuristic Programming Project. Reading, MA: AddisonWesley. Klir, George J. & Bo Yuan. 1995. Fuzzy Sets and Fuzzy Logic, Upper Saddle River, NJ: Prentice Hall. Kreinovich, Vladik. 1998. From semi-heuristic fuzzy techniques to optimal fuzzy methods: mathematical foundations and applications, In: Basil K. Papadopoulos & Apostolos Syropoulos (eds.), Current Trends and Developments in Fuzzy Logic, Proceedings of the First International Workshop, Thessaloniki, Greece, October 16–20, 1998, 1–62.

Which fuzzy logic operations are most appropriate for ontological semantics

267

Kreinovich, Vladik & Sundeep Kumar. 1990. Optimal choice of – and – operations for expert values, Proceedings of the 3rd University of New Brunswick Artificial Intelligence Workshop, Fredericton, New Brunswick, Canada. 169–178. Kreinovich, Vladik, Chris Quintana, Robert N. Lea, Olac Fuentes, Anatole Lokshin, Sundeep Kumar, Inna Boricheva, & Leonid Reznik. 1992. What non-linearity to choose? Mathematical foundations of fuzzy control, Proceedings of the 1992 International Conference on Fuzzy Systems and Intelligent Control, Louisville, Kentucky, 1992. 349–412. Mendel, Jerry M. 2017. Uncertain Rule-Based Fuzzy Systems: Introduction and New Directions, Cham, Switzerland: Springer. 2nd ed. Nguyen, Hung T. & Vladik Kreinovich. 1998. Methodology of fuzzy control: an introduction. In: Hung T. Nguyen and Michio Sugeno (eds.), Fuzzy Systems: Modeling and Control, Boston, MA: Kluwer. 19–62. Nguyen, Hung T., Vladik Kreinovich, & Dana Tolbert. 1993. On robustness of fuzzy logics, Proceedings of the 1993 IEEE International Conference on Fuzzy Systems FUZZ-IEEE’93, San Francisco, California, March 1993, 1. 543–547. Nguyen, Hung T., Vladik Kreinovich, & Dana Tolbert. 1994. A measure of average sensitivity for fuzzy logics”, International Journal on Uncertainty, Fuzziness, and Knowledge-Based Systems 2(4). 361–375. Nguyen, Hung T. & Elbert A. Walker. 2006. A First Course in Fuzzy Logic, Boca Raton, FL: Chapman & Hall/CRC, Novák, Vilém, Irina Perfilieva, and Jiří. Močkoř. 1999. Mathematical Principles of Fuzzy Logic, Boston, MA: Kluwer. Smith, Michael H. & Vladik Kreinovich. 1995. Optimal strategy of switching reasoning methods in fuzzy control. In Hung T. Nguyen, Michio Sugeno, Richard M. Tong,& Ronald R. Yager (eds.), Theoretical Aspects of Fuzzy Control,New York, N.Y.: Wiley.117–146. Taylor Rayz, Julia & Victor Raskin. 2010. Fuzzy ontology for natural language, Proceedings of the Annual Conference of the North American Fuzzy Information Processing Society NAFIPS’2010, Toronto, Canada, July 12–14, 2010. Taylor Rayz, Julia & Victor Raskin. 2016a. Conceptual defaults in fuzzy ontology, Proceedings of the 2016 Annual Conference of the North American Fuzzy Information Processing Society NAFIPS’2016, El Paso, Texas, October 31 – November 4, 2016. Taylor Rayz, Julia & Victor Raskin. 2016b. Logic of natural language: Through the eyes of ontological semantics, In: Yingxu Wang, Newton Howard, Bernard Widrow, Konstantinos N. Plataniotis, & Lotfi A. Zadeh (eds.), Proceedings of the 2016 IEEE International Conference on Cognitive Informatics and Cognitive Computing ICCICC’16, Stanford, California, August 22–23,2016, pp. 511–515. Tolbert, Dana. 1994. Finding “and” and “or” operations that are least sensitive to change in intelligent control, University of Texas at El Paso, Department of Computer Science, Master’s Thesis, 1994. Zadeh, Lotfi A. 1965. Fuzzy sets. Information and Control 8. 338–353.

M.L. Gavrilova

Decoding intricacies of human nature from social network communications Abstract: In the rapidly evolving area of cognitive informatics and data mining, behavioral profiling of online communications is one of the most dynamic areas of interest. It has numerous connections to other areas, such as linguistics, sentiment analysis, visualization, robotics, emotion recognition, e-learning, natural communication and humor recognition. This chapter explores the state-of-the-art machine learning methods for online pattern analytics and user behavior recognition. Typical applications are in domains of cybersecurity, virtual reality, human-computer interaction, e-commerce and e-education. User identification based on a combination of visual cues, social network activities, spatio-temporal communication patterns, online communication style, habits as well as linguistic patterns expressed in those activities are emerging research directions discussed in this chapter. Keywords: biometrics, social biometrics, user identification, semantic-script theory of humor

1 Introduction In modern society, abundant information about individuals is being stored, distributed, enriched and analyzed through a variety of means. It is thus not surprising that individual’s personal life, professional activities, circle of friends, interests and attachments may be discovered through a simple on-line search. Many corporations and businesses have taken advantage of this abundant online information through employing powerful data analytics, decision-making, artificial intelligence and pattern recognition techniques. In the academic domain, those fields received a boost from studying such aspects as the behavior of social network users, their verbal and non-verbal interactions, their connectivity and even their aesthetic preferences (Paul et al. 2015; Segalin et.al. 2014; Yampolskiy & Gavrilova 2012; Vinciarelli et al. 2012; McDuff et al. 2015; Popa et.al. 2012). An interesting outlet for such research has been recently introduced in the context of identification of an individual. While typical biometrics recognize person’s identity from physiological or behavioral biometric data (face, fingerprint, M.L. Gavrilova, University of Calgary, Calgary, AB, Canada https://doi.org/10.1515/9781501511707-013

270

M.L. Gavrilova

iris, gait, signature and so on (Jain et al. 2004)), social connections may also play an important role in an identity verification (Sultana et al. 2017a, Sultana et al. 2017b). These supplementary biometric patterns may include user’s online presence habits (time, topics, writing style), the nature of interaction with others (blogs, chats, texts, tweets), the content of interaction (topics, preferences, sentiment analysis), online shopping preferences and similar patterns (Drosou et al. 2015; Segalin et al. 2014; Sultana et al. 2015; Raskin 1984; Raskin 2008). This book chapter takes an in-depth look into intricacies of human nature as expressed through on-line communication. An analysis of emerging directions based on social ties, collaborative activities, professional networks, linguistic profiles and even aesthetic preferences will be provided in the context of human behavior recognition and understanding. A number of open questions will conclude this article. The answers to those questions will provide insights into the emerging research topics linking human and artificial entities in our interconnected world.

2 Relevant research on social interactions as biometric features Social interactions and behavior have been the focus of interest for researchers spanning the domains of social science, psychology, neuroscience, organizational behavior and targeted marketing for many years. However, it is clear that contextual, or auxiliary information can provide even more cues about someone’s identity, emotions or activities, than a direct analysis of typical features such as a physical appearance or a signature (Gavrilova & Monwar 2013). In addition, such information can be used for simulating human behavior and emotions in humanoid robots and avatars (Zhang et al. 2014; Boucenna et al. 2014). Another analogy that can be made is that the recognition process based on such contextual information resembles the function of the human brain, where past experiences, memories, interactions, emotional responses, and contextual information play as important role as a visual recognition during a process of human facial identification (Haxby et al. 2002; Isola et al. 2014). Figure 1 illustrates this concept. It depicts a framework of a biometric system where contextual features (e.g. social information, auxiliary data, writing style, and even humor perception) and soft biometric traits (aesthetic preference, age, gender, etc.) are utilized for enhanced person identification. In addition, a dynamic behavior biometric can be introduced, where the observation of certain traits can be carried on over a period of time (textual communications over an online

Decoding intricacies of human nature from social network communications

271

Figure 1: A generalized architecture of the biometric authentication system, based on the fusion of contextual, behavioral and physiological data (Sultana et al. 2014).

social network would fall under this category). For instance, the use of some common phrases during speech of online interactions may act as social behavioral biometric during authentication (Sultana et al. 2017a, Sultana et al. 2017b).

3 Social interactions in biometrics The following section presents an overview of emerging directions in utilizing a variety of social interactions for person recognition purposes.

3.1 On-Line communications as social biometrics In modern times, online presence and interactions may not only enrich, but even replace face-to-face human communication. This trend has already become so widespread, it has infiltrated all of our everyday activities. Grocery shopping and travel booking are done with a few clicks of the mouse. Medical advice is often sought online and prescriptions can be filled by online pharmacies. University applications, job postings and even promotions are administered through web

272

M.L. Gavrilova

applications. In the personal sphere, Facebook, Twitter, Instagram, Flickr and other social networks dominate the relationship and personal expression scene. Family gatherings, special events and milestones are being planned online. Figure 2 shows some examples of on-line domains where social communication may occur.

Figure 2: On-line application domains of social biometrics.

Wikipedia, Google Scholar, Facebook and other prominent social networks and web sites became the international hub of so-called ‘collective intelligence’. But what does it tell about us, as willing, or unwilling, participants in this global experiment? As it turns out, way more then we might think. Extraction of social behavioral information from on-line communications and their subsequent uses in biometric applications is one emerging directions of research, that focuses on the very nature of a human being. It turns out that humans are highly predictable in their behavioral patterns, habits and traditions. Even such seemingly stand-alone areas as aesthetic preferences and humor recognition can tell a lot about one’s identity (Paul et al. 2015, Azam and Gavrilova2016).

Decoding intricacies of human nature from social network communications

273

The concept of social biometric has wide implications on other areas of research, concerned with human identity and behavior. Those areas may include person recognition, border control, anomaly detection, risk estimation, behavior analytics, and situation awareness (Sultana et al. 2014, Yampolskiy and Gavrilova 2011). As well known, the traditional way of on-line user verification is by login, passcode, and/or security questions. However, knowing the user behavioral patterns may assist in replacing those authentication mechanisms with tools which are more natural or convenient for the user. As an example, information obtained from social media can assist in the generation of security questions. Moreover, seamless and continuous authentication can be done through analysis of browsing history, social network usage, friendship connections, discussion topics or temporal patterns of accessing various online services. Thus, idiosyncratic behavioral features or spatiotemporal information and browsing patterns can provide a rich foundation for new ways of user authentication. In addition, various on-line domain may be used to gather various domain specific information, which could be highly unique and thus usable as user authentication. For instance, humor identification (Raskin 1984, Raskin 2008, Attardo & Raskin 1991) could be used as one of those domain specific characteristics, and can be investigated in online blogs, comments, or even memes.

3.2 Aesthetic preferences as social biometrics One of the fascinating questions that puzzled scientist was whether human ethics, norms and standards of behavior are predetermined from birth by our unique gene codes, or they can be instilled, taught, enhanced and even completely changed through educational or medical interferences. Some emerging research in the field of neurology (Duch 2003) demonstrated that each of us has a unique way of thinking, where neurons form dynamic unique patterns that are very distinctive from person to person. However, no research so far has been done on trying to verify such ‘brain fingerprints’. Nonetheless, the outcomes of the brain activities, such as various preferences, can be in fact quite distinctive for different people. Each individual has their own taste in food, clothing, electronics, music, ethics and even perception of beauty. It was shown recently that such perceptions may be used to tell people apart and even to correctly predict their gender (Azam & Gavrilova 2016). Users of a popular social network, Flickr, can share their favorite images with others (Lovato et al. 2014). Recent works on automatic aesthetic evaluation of photographic images (Aydin et. al. 2015), aesthetic image classification

274

M.L. Gavrilova

(Xiaohui et al. 2014), on memorable images (Khosla et al. 2014) and on assessing the aesthetic quality of photographs (Marchesotti et al. 2011) demonstrated that there are some very distinct features responsible for human perception of aesthetics. A genetic algorithm-based approach was developed to train expert models on recognizing person’s gender, and to identify features that correlate specifically to female and male aesthetic preferences (Azam & Gavrilova 2016). Similarly machine learning-based approaches could be taken to recognize a given person’s favorite videos, animations, scents, or music collections. Even favorite fabrics or spaces can be used to uniquely identify an individual; the scope of research is practically limitless.

3.3 Linguistic and humor analysis as social biometrics Aesthetic preferences may help identify people, behavior, linguistic expressions and communication of people in a collaborative environment. People communicate through a variety of means. Sentiment analysis is the most frequently used term, in the sense of extrapolating the meaning and drawing some conclusions from opinions expressed in writing. The writing can be in the form of a scholarly work, a fiction, a blog, a tweet or a post in one of the on-line networks. It has been shown that, while the linguistic features alone might not be sufficient for correct author identification, the combination of such features with other information extracted from texts (time, date, domain) can be very revealing and unique (Sultana et al. 2014, Sultana et al. 2017a, Sultana et al. 2017b). We thus postulate that the use of humor in one’s writing can be similarly distinctive, and thus can be used as an additional social biometric feature in the context of a multi-modal biometric system. We shall base this assumption on the works of one of the most prominent visionaries of our time, and a worldrenowned expert on the theory and practice of humor, Prof. Victor Raskin. Before his groundbreaking work changed forever the direction of the humor research, there were many other theories of humor, based on physiological, neural and even spiritual and mystical premises (Buijzen and Valkenburg 2004, Meyer 2000). However, only Victor Raskin’s General Theory of Verbal Humor presents a clear and formal distinctions between serious and humorous communication, as well as the different types of humor (Raskin 1984, Raskin 2008, Attardo and Raskin 1991). As such, analysis of a context in which jokes were used, plays a tremendous role in his theory. Similarly to a contextual gait recognition, where the time of the day, the type of shoes, and the amount of lighting can have an effect on the overall person recognition, the contextual analysis of the writing is quite important. For example, in the General Theory of Verbal Humor, a Target (TA) of a

Decoding intricacies of human nature from social network communications

275

joke, is a person or thing made fun of; this can be used as one of the humor biometric features. Other concepts, such as Script Opposition (SO), Logical Mechanism (LM), and Narrative Strategy (NS) may also contribute to unique identifiable features, in addition to a more traditional language (LA) analysis. Additional research is unequivocally required in this area, to uncover the potential of merging the social behavioral biometric with the general theory of verbal humor.

4 Conclusions and open questions There are many research problems that remain unexplored in this exciting area of research. The social behavioral biometrics listed above can be utilized in a wide variety of application domains. The applications include various types of multi-modal biometric systems that can incorporate contextual, social, aesthetic or even linguistic models as one of their components. A measurement of a degree of engagement during collaborative activities based on non-verbal activities of the participants is another application. Marketing and prediction of online shopping behavior may be useful for business applications. Moreover, aesthetic preferences can be important for targeted advertisements, or during an election campaign. Medical diagnostics can benefit from understanding the body language of a patient, or to localize pain to a specific body part. For this, contextual and emotional information could be very helpful. Finally, gaming environments, humanoid robots and virtual worlds increasingly rely on a type of a social interaction that uses special awareness, a perception of the context, and most importantly social interactive features such as emotional cues or a language expression. Through researching those traits, new dynamic virtual environments and interactive collaborative spaces of the future can be constructed. To summarize the above discussion, this chapter explored the possibility of a new concept of biometric fusion using physiological, behavioral, social, linguistic and humor-based features. Methodologies on how biometric information can be extracted from social, contextual, temporal, and behavioral data of individuals were briefly surveyed. This chapter also presented an overview of the open problems, challenges, application domains and future research. Acknowledgement: The author would like to deeply acknowledge the insightful and humorous conversations with distinguished Prof. Viktor Raskin, that took place both at the SERIAS Centre and ICCI*CC Conferences since 2015. The author is also grateful to the NSERC DISCOVERY program grant RT731064, NSERC ENGAGE and MITACS ACCELERATE funding, for partial support of this project.

276

M.L. Gavrilova

References Attardo, Salvatore & Raskin, Victor. 1991. Script theory revis(it)ed: joke similarity and joke representation model. HUMOR: International Journal of Humor Research 4(3–4). 293–347. Aydın, Tunç Ozan, Aljoscha Smolic, and Markus Gross. 2015. Automated Aesthetic Analysis of Photographic Images, IEEE Transactions on Visualization and Computer Graphics 21(1). 31–42. Azam, Samiul & Marina Gavrilova. 2016. Soft Biometric: Give Me Your Favorite Images and I Will Tell Your Gender, 15th IEEE International Conference on Cognitive Informatics and Cognitive Computing (ICCI*CC), pp. 535–541. Boucenna, Sofiane, Philippe Gaussier, and Laurence Hafemeister. 2014. Development of first social referencing skills: Emotional interaction as a way to regulate robot behavior, IEEE Trans. Autonomous Mental Develop 6(1). 42–55. Buijzen, Moniek, and Patti M. Valkenburg. 2004. Developing a Typology of Humor in Audiovisual Media”. Media Psychology 6(2). 147–167. Drosou, Anastasios, Dimosthenis Ioannidis, Dimitrios Tzovaras, Konstantinos Moustakas, & Maria Petrou. 2015. Activity related authentication using prehension biometrics. Pattern Recognition 48(5). 1743–1759. Duch, Włodzisław. 2003. Coloring black boxes: visualization of neural network decisions. In 2003. Proceedings of the International Joint Conference on Neural Networks, vol. 3, 1735–1740. Gavrilova, Marina L., & Maruf Monwar. 2013. Multimodal biometrics and intelligent image processing for security systems, Hershey, PA: Idea Group International. Haxby, James V., Elizabeth A. Hoffman, & M. Ida Gobbini. 2002. Human neural systems for face recognition and social communication, Biological Psychiatry 51(1). 59–67. Isola, Phillip, Jianxiong Xiao, Devi Parikh, Antonio Torralba, & Aude Oliva.2014. What makes a photograph memorable?, IEEE Transactions on Pattern Analysis and Machine Intelligence, 36(7), pp.1469–1482. Jain, Anil K., Arun Ross, & Salil Prabhakar. 2004. An introduction to biometric recognition, IEEE Transactions on Circuits Systems and Video Technologies 14(1). 420–428. Khosla, Aditya, Atish Das Sarma, & Raffay Hamid. 2014. What makes an image popular? 23rd international conference on World Wide Web (WWW ‘14), pp. 867–876. Lovato, Pietro, Manuele Bicego, Cristina Segalin, Alessandro Perina, Nicu Sebe, & Marco Cristani. 2014. Faved! Biometrics: Tell me which image you like and I’ll tell you who you are, IEEE Transactions on Information Forensics and Security 9(3). 364–374. Marchesotti, Luca, Florent Perronnin, Diane Larlus, & Gabriela Csurka. 2011. Assessing the aesthetic quality of photographs using generic image descriptors, IEEE International Conference on Computer Vision (ICCV), pp. 1784–1791. McDuff, Daniel, Rana El Kaliouby, Jeffrey F. Cohn, & Rosalind W. Picard. 2015. Predicting Ad liking and purchase intent: large-scale analysis of facial responses to Ads, IEEE Transactions on Affective Computing 6(3). 223–235. Meyer, John C. 2000. Humor as a Double-Edged Sword: Four Functions of Humor in Communication. Communication Theory 10(3). 310–331. Paul, Padma Polash, Madeena Sultana, Sorin Adam Matei, & Marina Gavrilova. 2015. Editing behavior to recognize authors of crowdsourced content, IEEE International Conference on Systems, Man, and Cybernetics (SMC), pp. 1676–1681.

Decoding intricacies of human nature from social network communications

277

Popa, Mirela, Alper Kemal Koc, Leon JM Rothkrantz, Caifeng Shan, & Pascal Wiggers. 2012. Kinect sensing of shopping related actions, Constructing Ambient Intelligence: AmI 2011 Workshops, Amsterdam, The Netherlands, November16–18, 2011. Revised Selected Papers. Berlin: Springer 91–100. Raskin, Victor. 1984. Semantic Mechanisms of Humor. Dordrecht: D. Reidel. Raskin, Victor. 2008. The Primer of Humor Research, Berlin: Mouton de Gruyter. Segalin, Cristina, Alessandro Perina, & Marco Cristani. 2014. Personal aesthetics for soft biometrics: A generative multi-resolution approach, 16th International Conference on Multimodal Interaction (ICMI ‘14), pp.180–187 Sultana, Madeena, Padma Polash Paul, & Marina Gavrilova. 2014. A concept of social behavioral biometrics: Motivation, current developments, and future trends, Proceedings of the 2014 International Conference on Cyberworlds, IEEE, pp 271–278. Sultana, Madeena, Padma Polash Paul, & Marina Gavrilova. 2015. Social behavioral biometrics: An emerging trend, International Journal of Pattern Recognition and Artificial Intelligence 29(8), 1556013. Sultana, Madeena, Padma Polash Paul, & Marina Gavrilova. 2016. Identifying users from online interactions in Twitter, Transactions on Computational Science XXVI, pp. 111–124, Springer Berlin Heidelberg. Sultana, Madeena, Padma Polash Paul, & Marina Gavrilova. 2017a. Social Behavioral Information Fusion in Multimodal Biometrics, IEEE Transactions on Systems, Man, and Cybernetics: Systems (99). 1–12. Sultana, Madeena, Padma Polash Paul, & Marina Gavrilova. 2017b. Person Recognition from Social Behavior in Computer Mediated Social Context, IEEE Transactions on HumanMachine Systems, 47 (3). 356–367. Vinciarelli, Alessandro, Maja Pantic, Dirk Heylen, Catherine Pelachaud, Isabella Poggi, Francesca D’Errico, & Marc Schroeder. 2012. Bridging the gap between social animal and unsocial machine: a survey of social signal processing, IEEE Transactions on Affective Computing 3(1). 69–87. Wang, Xiaohui, Jia Jia, Jiaming Yin, and Lianhong Cai. L. 2013 Interpretable aesthetic features for affective image classification, 20th IEEE International Conference on Image Processing (ICIP), 3230–3234. Yampolskiy, Roman V., and Marina L. Gavrilova. 2012. Artimetrics: biometrics for artificial entities, IEEE Robotics & Automation Magazine 19(4). 48–58. Zhang, Yuzhe, Jianmin Zheng, and Nadia Magnenat-Thalmann. 2014. Example-guided anthropometric human body modeling, The Visual Computer 31(12). 1615–1631.

Part 4: Other applications

Dallin D. Oaks

A creative approach for linguistic funny business: Using linguistic paradigms and taxonomies Abstract: With the pressures that exist for professional writers of humor or wordplays to generate clever material and to do so in limited time frames, effective brainstorming strategies could be very useful. This article will demonstrate the value of using linguistic paradigms and taxonomies to brainstorm ideas for generating humor, particularly linguistic-based humor. As part of this consideration, some attention will be given to such linguistic matters as structural ambiguities, Gricean maxims, implicatures, speech acts, felicity conditions, adjacency pairs, deixis, and presuppositions. Keywords: ambiguity, taxonomy, paradigm, implicature, speech act

1 Introduction Professional writers of humor or wordplays, whether for jokes, comedy, or advertising, often face relentless pressure to produce creative and original material in very restricted time frames. Some of these professionals may even worry about the possibility of running out of ideas. Such writers have some strategies to mitigate these challenges, but they could probably benefit from some additional creative approaches. In this regard, linguistic-based approaches seem to be an especially promising area. As Raskin has noted, first in his Semantic Script Theory of Humor and later with Attardo in the General Theory of Verbal Humor, a crucial component in the construction of humor is the overlapping or opposition of semantic scripts or “chunk[s] of structured semantic information” (Raskin 1985: 81, 107–114; Attardo and Raskin 1991: 307–309, 328–329; cf. also Attardo 2017a: 127–128, 133). Sometimes the script opposition coincides with ambiguity in the structural categories of phonology, morphology, or syntax. And sometimes the script opposition is found in situational or discourse matters involving language concerns that are less directly tied to these structural categories.

Dallin D. Oaks, Brigham Young University https://doi.org/10.1515/9781501511707-014

282

Dallin D. Oaks

Developing linguistic-based humor depends on forms, structures, or linguistic issues that may be hard to anticipate or imagine, but one powerful strategy that is available for writers who need kernel ideas around which humor may be developed is to identify and examine linguistic paradigms and taxonomies. The term “paradigm” will be used here in a general sense that could include not only lists of grammatical or structural features such as a list of the inflectional suffixes in the language, but also a list of the periphrastic verb constructions (such as the present perfect, present progressive, etc.), and identified clause types. Within “taxonomies” I will include not only such structurally based classifications as parts of speech, or verb types, but also language frameworks that have been proposed to account for discourse or other language behavior, more specifically, frameworks such as Grice’s conversational maxims (Grice 1975) or speech act theory (cf., for example, Austin 1975; Searle 1969 and 1975; and Hurford, Heasley, and Smith 2007). In this article my focus will be on identifying humorous linguistic potentials, and thus I will not get into the matter of how a humorous potential must be contextualized within an effective script opposition in order for the humor potential to be successfully realized. Previous linguistic research has described relationships between humor and language patterns found at the phonological, morphological, or syntactic levels (cf. Pepicello and Green 1984; Bucaria 2004; Aarons 2012; and Oaks 2010a), including structural ambiguity (Oaks 2010a).1 And some research has looked at humor involving speech acts and implicatures (cf., for example, Hancher 1980; Aarons 2012; and Ferrar 1993). Aarons has explored how humor reveals things about our internal linguistic knowledge, as humor sometimes brings such knowledge to our conscious awareness (2012: 7–12). She provides useful data and analyses, and I have integrated a few of her examples here. Dubinsky and Holcomb use humor to a different end, for pedagogical purposes, “as a vehicle for introducing linguistic concepts and the various subfields in which they play a part” (2011: 1). But my overall approach and point is a different one, since I suggest actively utilizing these patterns, paradigms, and taxonomies in an applied strategy for identifying and predicting language possibilities for humor and then potentially generating humor around these possibilities. A creative approach that examines paradigms and taxonomies to anticipate and generate humor is not one that merely engages in analogous reapplications of humorous material we have already seen. Although the approach may sometimes generate wordplays and jokes that are similar in some ways to existing

1 This two-volume work sets forth various phonological, morphological, and syntactic features and combinations of the English language that allow structural ambiguity to occur.

A creative approach for linguistic funny business

283

humor, the similarity would not be because the humor was generated through a simple and deliberate modification of existing wordplays and jokes, but instead because the approach is itself productive in generating fresh instances of such humor.

2 Identifying humor potentials through grammatical paradigms Some years ago, Oaks provided a couple of examples showing how paradigms could reveal potentials for structural ambiguities, which are crucial to many types of wordplays.2 At that time the pronoun/determiner her and the word final [s] were specifically examined (Oaks 2010b). A brief overview of information within or related to that article will be useful here. A comparison of paradigms for the pronouns and determiners reveals the overlapping forms of the feminine object pronoun and the feminine possessive determiner. Note how the overlapping forms for her become immediately apparent when a combination of the paradigms for the feminine forms is set forth: Subject Object Possessive Determiner Possessive Pronoun

she her her hers

Of course for the possessive and object forms to be ambiguous and form part of a wordplay, they need a compatible syntactic environment (as well as a semantic/pragmatic context) where the possessive and object forms can both occur. But once again, a consideration of the paradigm of clause types in the language can reveal some significant possibilities. The set of English basic clause types includes, among others, the following:

2 In some ways this point, specifically with regard to structural ambiguity in English, is moot because Oaks 2010a has already outlined rather comprehensively the grammatical features of the English language that enable structural ambiguity. But although much of the work in that book is based on the observation and analysis of authentic textual examples of humor, the approach described here for identifying structural ambiguity potentials through structural paradigms is still usefully applicable, independent of his work, and could also be applied in other languages where a detailed grammatical description of structural ambiguity potentials doesn’t exist.

284

Dallin D. Oaks

SV (Subject + Verb) Example: The dog barked. SVC (Subject + Linking Verb [often a form of BE] + Subject Complement [Noun]) Example: The fence was a nuisance. OR SVC (Subject + Linking Verb [often a form of BE] + Subject Complement [Adjective]) Example: Their house was large. SVO (Subject + Verb + Direct Object) Example: We sang songs. SVOO (Subject + Verb + Indirect Object + Direct Object) Example: We gave the teacher a gift. (cf. Quirk et al. 1985: 720–722 for an outline of the basic clause types) Note that if we consult a taxonomy that lists the verbs that can be used in each of the clause types (verbs determine the types of clauses that may be created structurally, varying in their abilities with regard to the clause types they may form), then we can anticipate to some extent the kinds of combinations in which the form her may be used to create a structural ambiguity. Since the form her may be either an object form that occupies its own constituent type (it may occur as an object by itself, as in “I saw her”) or as a determiner to introduce a noun object (as in “I saw her cars”), then with the right kind of verb it may potentially occur in both an SVO or SVOO clause type. An examination of the verbs that may set up the different clause types reveals that verbs such as BRING, SEND, SELL, TAKE, MAKE, and FIND may set up both an SVO clause or an SVOO clause. With this information we may begin to anticipate that following one of the previously listed verbs with the form her and preceding a plural or non-count noun3 will likely result in a structural ambiguity between an SVO and SVOO clause type. This ambiguity type is evident in the following examples from Oaks: We brought her candy. The man sent her cards. The children sold her books. I took her medicine. (2010b: 53)

This kind of structural overlap explains why one famous Gracie Allen and George Burns comedy routine worked. The structural ambiguity type of course needed a clever contextualization for an effective and humorous script opposition:

3 The reason that plurals and non-count nouns are crucial for creating a structural ambiguity in this environment is because they don’t require a determiner like a, the, my, etc., which would otherwise clarify the constituent boundaries of the noun phrases. Notice for example how the SVO vs SVOO ambiguity of “The man sent her cards” is removed if we say, “The man sent her a card.”

A creative approach for linguistic funny business

285

George: (looking at Gracie, who is arranging a large vase of beautiful flowers) Grace, those are beautiful flowers. Where did they come from? Gracie: Don’t you remember, George? You said that if I went to visit Clara Bagley in the hospital I should be sure to take her flowers. So, when she wasn’t looking, I did. [italics are mine; D. O.] (cited in Oaks 2010a: 163)

In this example we have an ambiguity based in a structure that we could simplify for illustrative purposes as “I took her flowers.” The co-occurrence of several factors – the use of TAKE, the form her, and the plural flowers all work together for the wordplay to occur. A consideration of the related paradigms reveals how this joke could be constructed. The other part of the article (Oaks 2010b) shows how a comparison of the paradigms for English inflections and for contractions for the forms of BE, HAVE, and DO, sometimes within periphrastic verb constructions, and a consideration of word-final [s] endings could reveal overlapping morphological possibilities that might work into structural ambiguities.4 Such a comparison reveals that the word-final [s] sound can represent the following possibilities: A) a possessive form on a noun phrase. [The cat’s collar is pink] B) a contraction for is as the third person singular present tense form of BE as a main verb. [The cat’s happy] C) a contraction for is as the third person singular present tense form of BE as an auxiliary verb in the progressive verb construction. [The cat’s coming] D) a contraction for is as the third person singular present tense form of BE as an auxiliary verb in the passive construction. [The cat’s fed] E) a contraction for has as the third person singular present tense form of HAVE as an auxiliary verb in the perfect verb construction. [The cat’s come] F) a contraction for does as the third person singular present tense form of the auxiliary DO. [What’s the cat do?] G) a noun plural. [The cats played] H) a third person singular present tense inflection on the main verb. [The cat walks carefully] I) the final consonant sound of a noun stem. [The cat watched the bus] J) the final consonant sound of a verb stem. [The cat will miss the vacation] (This list, without the bracketed examples, is provided in Oaks 2010b: 42)

4 The article by Oaks treats the issue of the voicing of [s] resulting in [z]. In cases where the [z] results from a predictable phonological process such as occurring after or between voiced sounds, it will be considered here as a variant of the voiceless [s] (2010b: 42–43).

286

Dallin D. Oaks

One surprising finding was that each of these possibilities may be combined with any of the others for a structural ambiguity. In fact, these varying combinations can result in at least 45 different structural ambiguity types, just for the word final [s]. In looking at a sample of joke data involving structural ambiguity Oaks notes that at least some of these ambiguous possibilities with [s] have been used in jokes. For example, consider the relatively famous riddle, “What has four wheels and flies? Answer: A garbage truck.” This combines the possibility of G (a noun plural) and H (a third person singular present tense inflection on the main verb). Another joke involves an ambiguity that combines D (a contraction for is as the third person singular present tense form of BE as an auxiliary verb in the passive construction) and E (a contraction for has as the third person singular present tense form of HAVE as an auxiliary verb in the perfect verb construction): First Cannibal: Am I late for chow? Second Cannibal: Yes, everybody’s eaten. (joke example from Kohl and Young 1963: 31; analysis from Oaks 2010b: 43–48)

Although some combinations with [s] are potentially less productive of ambiguities than others, there are still probably some combinations whose potential productivity still remains to be more effectively utilized, and whose potential for use in humor may be more easily recognized through a consideration of the ambiguity types that their overlapping paradigms suggest.5 In fact, a conscious understanding of the possibilities suggested by this consideration of the paradigms could provide some real resourcefulness in constructing a variety of wordplays. Another set of insights about wordplay that can be revealed through comparing paradigms is evident in our verb constructions involving auxiliary verbs. We can see that both the passive (a form of BE + a past participle) and the progressive (a form of BE + a present participle) can be ambiguous with other structures, particularly the SVC clause type. When we examine the SVC clause type, we see that the linking verb is often a form of BE, followed by a subject complement that is an adjective. Add to this the fact that past participles like animated or outspoken and present participles like revolting or appealing can

5 A table showing the use of various ambiguities with [s] in humor appeared incorrectly in the printed version of Oaks 2010b, with each column being off by one position, but the online version appears correctly.

A creative approach for linguistic funny business

287

serve as adjectives, and the potential for ambiguity becomes evident. Consider the following ambiguities: Al Gore turned down a chance to be on The Simpsons. He explained, “I’ve never been animated and I’m not going to start now.” (From the comedian Conan O’Brien, cited in That’s Funny 1996: 108) He: There’s one thing about my mother-in-law. She’s outspoken. She: Not by anyone I know. (Allen 2000: 177) Farmer [who is being attacked by chickens]: The chickens are revolting. Wife [who can’t see her husband’s situation]: Finally, something we agree on! (From the movie, Chicken Run)6 Steve: Why aren’t you behind bars, as other criminals would be? Because your lawyer is appealing? Louis: He is not appealing. He is cute, but I wouldn’t call him appealing. Not to me at least. (Allen 1987: 271)

3 Identifying humor potentials through linguistic taxonomies So far, this paper has looked at paradigms related to forming wordplays using structural ambiguities. And we have seen that a consideration of paradigms can reveal the potential for some ambiguities. But linguistic humor is not limited to structural ambiguity. Nor does linguistic humor always come primarily from the overlapping of phonological, morphological, or syntactic forms. In what follows now, we shall move from the role of examining structural paradigms to the role of considering linguistic taxonomies (including proposed frameworks, classifications, and dichotomies) that have been proposed for looking at language and language interactions. Specific linguistic forms will still sometimes be important to the humor, but our method of discovery will not be so closely tied to structural paradigms.

6 I am grateful to Katie L. Baker, a student from my Spring 2002 term, for calling my attention to this example from the movie. A clip of this portion of the movie is available at https://www. youtube.com/watch?v=WCznIAxhMSc&t=35s.

288

Dallin D. Oaks

We shall begin our consideration of taxonomies by looking at the conversational maxims of quantity, quality, relation (relevance), and manner that H. P. Grice identified as integral to the cooperative principle employed in communication. As he pointed out, people try to provide an appropriate amount of information that is true, that is relevant, and that is communicated in a clear, reasonable, way (1975: 45–46). Of course, communication that violates one or more of these maxims can be the basis of humor (cf. Morreall 1983: 79–81) since its variance from the expected script of communication, if still sharing enough of the other communicative aspects of the situation under consideration, can set up an effective script opposition. The potential usefulness of considering the Gricean maxims to brainstorm humor possibilities seems evident in Attardo’s belief that “a large number of jokes involve violations of one or more of Grice’s maxims” (1990: 355). The departure from what is expected will also commonly involve a misleading implicature. More will be said about this later. If we just look at the four dimensions of quantity, quality, relation, and manner, we have four different ways of contrasting a discourse utterance with what would normally be expected. Each of these four dimensions presents numerous possibilities for humor. Aarons provides an example of a humorous scene from the comic movie, The Pink Panther Strikes Again, that shows a violation of the quantity maxim. In this scene the bumbling Inspector Clouseau enters a lobby of a hotel and sees a dog on the floor. Clouseau apparently intends to pet the dog but first asks the clerk at the desk, the only person in the room, whether his dog bites. Aarons, reports it this way: Clouseau: Does your dog bite? Hotel Clerk: No. Clouseau: [bowing down to pet the dog] Nice doggie. [dog bites Clouseau’s hand] Clouseau: I thought you said your dog did not bite! Hotel Clerk: That is not my dog. (The Pink Panther Strikes Again, 1976, as cited in Aarons 2012: 46)

As Aarons explains, the clerk has violated the maxim of quantity (in addition to the maxim of relation), for he has provided an insufficient amount of information when he is asked whether his dog bites (2012: 46). The number of possible situations in which an inappropriate amount of information could be provided is immense. There are various ways in which someone could provide less information than expected, as in the situation with Clouseau, or they could provide much more than would be called for. These kinds of violations could be framed differently in a variety of settings or situations.

A creative approach for linguistic funny business

289

The above example also illustrates well the role of conversational implicature when considering the Gricean maxims. Conversational implicatures come from our expectations that the Gricean maxims are being followed, or in the significance we ascribe to departures from those maxims. In this example, when Clouseau asks the clerk whether his dog bites, the clerk’s simple “no” answer and his failure to provide the essential information that the nearby dog is not his, conveys the implicature that the dog that Clouseau is asking about is the clerk’s dog and poses no danger for Clouseau. Now consider two other examples with implicatures growing out of the maxims of both quantity and relation: Pam: I fell off a sixty-foot ladder today. Melba: It’s a miracle you weren’t killed. Pam: Oh, I only fell off the first rung. (Phillips 1990: 84) I don’t like country music – but don’t mean to denigrate those who do. For those who like country music, denigrate means “to put down.” (Bob Newhart, as cited in Price 2005: 167)

In the joke about the woman who said she fell off the ladder, the inclusion of the unnecessary additional information that the ladder was sixty feet tall carries the implicature that she fell from the top or nearly the top of the ladder. Otherwise what would be the point of mentioning its height? Her implicature also grows from a violation of the relation maxim since the story is hardly worth telling in the first place when she only fell from the first rung. In the case of the comedian Bob Newhart’s statement about country music, his providing of a definition for denigrate, specifically to country music lovers, carries the implicature that country music listeners are not very bright or educated. His definition is unnecessary additional information and is thus a violation of the maxim of quantity and to some extent relation as well. Let’s consider one more example, an old Monty Python sketch that shows a violation of the maxim of relation. In this sketch an airplane captain and his first officer are in the cockpit and are very bored. After they tire of playing one kind of diversionary game, we find the following dialogue: Captain: I’m fed up with that game. Let’s play another game. I know what. First Officer: What? (The Captain picks up a microphone.) Captain: (over intercom) “Hello, this is your Captain speaking. There is absolutely no cause for alarm.” That will get them thinking. (First Officer reaches for the microphone.)

290

Dallin D. Oaks

Captain: No, no, no, no. Not yet, not yet. Let it sink in. They will be thinking, er, “What is there absolutely no cause for alarm about? Are the wings on fire?” (over intercom) “The wings are not on fire.” Now they are thinking, er, “Why should he say that?” So we say . . . (Steward enters the cockpit.) First Officer: Oh, how are we doing? Steward: (looks down the aisle) They’ve stopped eating; Looking a bit worried. Captain: Good. [The sketch continues.]7

The passengers in this sketch panic despite being told that things are fine with the plane. Apparently, for the Captain to provide reassuring reports seems irrelevant to them unless there is some kind of dangerous situation that has developed during their flight. In other words, the implicature of his message is that there is something wrong. Viewers of the skit find humor in the dynamic that develops as the passengers react to the violation of the relation maxim. From the examples I have provided so far, it should not be understood that implicatures result only from a violation of the Gricean maxims. Attardo explains that “implicatures can be derived by the following of the maxims and the CP [cooperative principle] as well as by their non-observance” (2017b: 176). But humor will often involve implicatures that come from some kind of violation of the maxims. I won’t take time here to illustrate violations of the other maxims (quality and manner). It should be noted, however, that the manner maxim contains multiple dimensions that may be violated such as clarity, brevity, and orderliness. The matter of clarity, of course, includes ambiguity and vagueness, which are rich potential sources of humor. The issue of implicature is more complex than I have presented here. Attardo, for example, shows that a thorough and accurate view of the cooperative principle and implicature should consider not only the “amount of cooperation, based on the CP [cooperative principle], that two speakers must put into the text in order to encode and decode its intended meaning,” but also the cooperation that they “must put into the text/situation to achieve the goals that the speaker (and/or the hearer) wanted to achieve with the utterance” (1997: 756). This includes taking into account the purposes behind the communication (1997: 759). For the goals of my paper, and the brainstorming approach it is

7 For a transcription of this sketch, see http://www.montypython.net/scripts/irritate-airline. php. A film clip of this scene may be viewed at https://www.youtube.com/watch?v= pVWCEKDDI-E.

A creative approach for linguistic funny business

291

intended to outline, however, my brief explanation of the Gricean maxims and implicature will probably suffice. Let’s consider now the matter of speech acts. Speech acts are functions that we perform with what we say. We don’t merely make statements or form questions. Thus an utterance might be an apology, a threat, a compliment, etc. The function of what we say is often termed the “illocutionary force” (Searle 1969). Aarons explains that “exploiting ambiguity in the interpretation of illocutionary force is one of the most productive joke-making mechanisms” (2012: 26; cf. also Hancher 1980). Speech acts may be direct or indirect. Direct speech acts are those with a form that very directly corresponds to the speech act or function they are intended to convey. Indirect speech acts are those speech acts in which the form doesn’t directly correspond to the intended function (cf. Yule 2017: 149).8 Perhaps the most commonly given examples of the difference between a direct and indirect speech act are evident in the request or “directive” for someone to pass the salt. If we say, “Pass the salt” or “Please pass the salt,” we have issued a request or directive that is a direct speech act. If, on the other hand, we use a form such as “Would you mind passing the salt?” or “Can you pass the salt?” or even “I need the salt,” we have used an indirect speech act. Humorous situations can be created when there is a confusion about whether a particular utterance is a direct or indirect speech act, especially when the difference in the two also corresponds with a difference in the type of speech act or function that is conveyed. In the movie What About Bob? we have a great illustration of a direct speech act that is interpreted as an indirect one, leading to humorous results. In one scene, Bob, an individual with severe personality disorders, intrudes into a nationally televised interview featuring Dr. Marvin (his therapist) and his latest book, essentially taking over the interview and making Dr. Marvin look foolish in front of a national audience. After this disastrous and embarrassing interview, Dr. Marvin laments the damage to his career. Bob, who is standing next to Dr. Marvin’s other family members and who was the cause of this

8 In this paper I will work with the notion that an interpretation of an utterance as either a direct or indirect speech act is mutually exclusive of the other unless we are able to interpret the utterance ambiguously in a particular context. Searle, however, seems to regard indirect speech acts as performing an illocutionary act that is in addition to a more basic or literal one. In speaking of indirect speech acts he says, “In such cases a sentence that contains the illocutionary force indicators for one kind of illocutionary act can be uttered to perform, in addition, another type of illocutionary act” (1975: 168). Moreover, he says, “The cases we will be discussing are indirect speech acts, cases in which one illocutionary act is performed indirectly by way of performing another” (1975: 168).

292

Dallin D. Oaks

debacle, tries to respond reassuringly to Dr. Marvin by telling him that he (Dr. Marvin) has done a great job in his interview. But Dr. Marvin responds to him by saying, “Get Out.” In this case Dr. Marvin intends a direct speech act: Bob must leave. But Bob interprets the utterance as an indirect speech act in which Dr. Marvin is declining a compliment. So Bob issues another reassuring comment about how well he thinks Dr. Marvin has done (saying, “No, we won’t get out. We won’t. You deserve it”). Then Dr. Marvin responds with a clarification of his previous speech act, this time raising his voice to make his intention clear as a directive for Bob to leave: “I mean, ‘Get Out’; [then shouting] ‘Get out!’”9 In this scene, Bob has mistaken a direct speech act for an indirect one. On the other hand, the movie The Princess Diaries has a scene in which an indirect speech act is mistaken for a more direct speech act, as a young American girl expresses amazement when learning for the first time from her grandmother that she (the granddaughter) is the royal heir to a European kingdom. Shocked by the news, the granddaughter blurts out, “Shut Up.” The grandmother at first thinks her granddaughter has used a direct speech act to disrespectfully tell her to be quiet.10 Aarons recounts the joke about a wealthy man at a golf club who is dissatisfied with the service he is getting. “After a few minutes of trying unsuccessfully to get a waiter’s attention, he marched up to the manager’s area and said, indignantly, ‘Do you have any idea who I am?’ The assistant manager replied, ‘Honestly, Sir, I’ve only just been employed here, so I don’t know if I can help you myself. But as soon as we get less busy, Sir, I’ll ask at the office and get right back to you and let you know’” (2012: 31). This joke relies on a possible ambiguity about whether the utterance is a direct speech act involving a literal yes/no question requesting information (a question taking a yes or no answer) or was instead a demand by the patron to be treated with greater deference. A similar joke, this time with a wh-question (a question beginning with a who, what, where, etc.), is evident in the following: “What does this mean?” yelped the patron in a Bowery restaurant. “There’s a cockroach in the bottom of my teacup.” “Listen, bud,” snapped the waiter, “if ya wantcha fortune told, go see a Gypsy!” (Moulton 1942: 57)

It would seem that rather than regard these ambiguous cases as isolated ones, we should ask ourselves whether there is a creative strategy that could be deliberately and methodically mined for other humorous effects. After all, there are

9 A video of this scene may be viewed at https://www.youtube.com/watch?v=0pKymngWgJw. 10 A video of this scene may be viewed at https://www.youtube.com/watch?v=5iJ6Q8BHg5s.

A creative approach for linguistic funny business

293

many speech acts, many of which (if not all of which) have direct and indirect ways of being expressed, some of which, as we have seen, will overlap with forms associated with a separate speech act. Such overlaps constitute a rich potential for developing humor. One strategic approach for humor, therefore, might be to consider some of the forms used for specific indirect speech acts and consider whether any of them are also used for direct ones or vice versa. For example, returning briefly to the examples above, we can imagine how a consideration of the expressions we use for expressing amazement could yield, “Wow,” “You’re kidding,” and “Shut up.” The latter expression, as an indirect speech act, has a counterpart in a very different and direct speech act. This awareness of an overlap in the forms between a direct and indirect speech act could provide the basic language play around which a humorous dialogue could be fashioned, yielding the kind of scene we see in The Princess Diaries. This same kind of strategic approach could have worked in generating the ambiguity in the previously mentioned scene from What About Bob? that uses the utterance “Get out!” A consideration of commonly used forms for declining compliments might have been able to provide the idea for using “Get out.” Of course, humor based on the difference between speech acts is not limited to confusions involving direct versus indirect speech acts. Often we use utterances whose forms are potentially ambiguous between two possible indirect speech acts. And the previously described strategy for speech acts could also be useful for identifying shared forms among indirect speech acts. But beyond this, an additional strategy involves a consideration of the felicity conditions that distinguish each speech act from another. Every speech act has a set of felicity conditions that are necessary for the particular speech act to be completed successfully. But some speech acts share some of their felicity conditions with other speech acts, and it is in the shared similarities as well as crucial differences between a pair of speech acts, that we can construct a humorous script opposition. For example, linguists have noted the similarity in the felicity conditions of promises and threats. Both relate to future acts; both relate to an act that is under the power of the speaker, etc. But as Searle noted, they differ in the crucial condition of what the speaker believes that the hearer wants to occur. A promise is made to commit to do something that the speaker believes the hearer wants, whereas a threat is a stated intention to do something that the speaker believes the hearer doesn’t want (1969: 58). The similarities between the sets of felicity conditions in the two speech acts and the important difference allow us to more easily construct humor around a juxtaposition in these two speech acts. We only need to make sure that a particular setting allows some ambiguity about what the perceived desire of the hearer would be (cf., for example, Dubinsky and Holcomb 2011: 88). A similar felicity condition difference involving a speaker’s

294

Dallin D. Oaks

state of mind becomes the basis for humor in the following joke, where there is a potential ambiguity about whether a diner in a restaurant is complaining (perhaps a directive) versus issuing a helpful warning: Diner: Waiter, your thumb is in the soup! Waiter: Oh, it’s all right, lady, it isn’t that hot. (Allen 2000: 221)

Hancher, who discusses some speech acts and felicity conditions in relation to humor, looks at two very important felicity conditions that Searle identified for directives: 1) the proposed act is something that the hearer can do, and the speaker believes that to be so, and 2) the proposed act of the hearer is a future one. Hancher provides an example of humor related to a violation of each of these two felicity conditions. The first violation example is found in a cartoon in the publication Punch that “shows an airline stewardess admonishing an about-to-be-sick passenger ‘I’m sorry, sir. No vomiting before take-off.’” He also reports on a humorous story told by Ira Gershwin about “a woman at a party, who belatedly exclaims to her escort: ‘Good grief, it’s three in the morning! What’s more, I’m giving you warning: I must be home by twelve o’clock!’” (1980: 23–24). We might also imagine other pairs of speech acts with overlapping felicity conditions. Such types of paired speech acts could include compliments versus insults, offers to help versus bribes, apologies versus explanations, expressions of gratitude versus complaints, etc. Now we shall consider another aspect of discourse known as “adjacency pairs.” Linguists have noted that some utterances are followed by another of a specific type. For example, a greeting requires another greeting; an invitation requires an acceptance or rejection of the invitation; a compliment requires an acknowledgement; etc. Of course the members of an adjacency pair are also speech acts, each with their own felicity conditions. But even without consciously considering the felicity conditions and merely considering the function of the two paired elements of the adjacency pair, we can get humorous results when one member of the adjacency pair is followed by another that doesn’t seem to fit. I once had a cantankerous professor about whom the following interaction was reported: One day one of his colleagues greeted him with “Good morning,” to which he responded, “Prove it.” In this regard, he treated a greeting as if it had been offered as a statement, and he challenged it. His response was a violation of the anticipated second member of the adjacency pair and was apparently viewed as sufficiently humorous that his colleagues and others retold the incident.

A creative approach for linguistic funny business

295

Another humorous example involving an unexpected element of an adjacency pair was used in a Citibank television commercial that was brought to my attention by one of my students. The commercial for Citibank credit cards was promoting a thank-you redemption network for the company’s customers. The commercial wished to call attention to the power of thanking (and thus the desirability of Citibank’s new rewards program), by contextualizing a thank-you in a memorably unusual and humorous situation: Woman 1: “Oh, You must be having a boy. When are you due, huh?” Woman 2: “I’m not pregnant.” Woman 1: “You’re not . . . ?” Woman 2: “No! Why [did you think I was pregnant]?” Woman 1: “Uhhh . . . thank you?” Woman 2: “Me?” Woman 1: “Yes. Thank you.” Woman 2: [gesturing for a hug from the other woman] “Come on.” Voiceover: “It’s amazing what a simple ‘Thank you’ can do. Introducing Thank you, from Citi, a new kind of rewards program.”11

Because we have various adjacency pairs in the language, each of which can be violated in a variety of ways, one strategy for developing humor could be to identify multiple adjacency pairs and determine which ones might be most effectively altered, and in which ways, for humorous effect. Even just compiling a list of adjacency pairs would be a great starting point for the initial brainstorming of humorous ideas. Under the discussion of taxonomies we could also consider the phenomenon of deixis or deictic expressions. The referent of a deictic expression, what it refers to, depends on the special circumstances in a given setting. For example, the word tomorrow can refer to different times, depending on when that word is expressed. Thus, knowing the time of utterance is crucial to knowing what tomorrow actually refers to. Similarly, knowing the referent of the pronoun I depends on knowing who is speaking at the time. Deictic expressions may be classified according to whether they involve time (such as now, then, yesterday, tomorrow, two years ago, etc.), place (here, there, this, that, to the left, etc.), or person (I, you, she, them, the man behind me, etc.). Of course this is not to say that all expressions of time, place, or person are deictic. If I refer to Salt Lake City, I have not used a deictic expression, because the referent of Salt Lake City doesn’t depend on where I as the speaker happen to be situated when I use the

11 Example collected by Caitlin Anderson, Fall 2004. A video of this commercial is currently available at https://www.youtube.com/watch?v=gPairMk345s.

296

Dallin D. Oaks

expression. But an expression like here certainly does. Aarons says that “jokes based on ambiguity in the use of deictics are legion” (2012: 48). Consider how the following joke plays off the general issue of time deixis, though the joke’s humor is not dependent on a specific deictic expression. Julie: What time is it? Counselor: Three o’clock. Julie: Oh, no! Counselor: What’s the matter? Julie: I’ve been asking the time all day. And everybody gives me a different answer. (Dubinsky and Holcomb 2011: 79)

On the other hand, a humorous passage from the television show, The Simpsons, specifically depends on the deixis of the pronouns I and you. In this passage a Smokey the Bear statue asks people to answer the question, “Only who can prevent forest fires?” The statue contains two buttons that may be pressed. One says you, and the other says me [Of course we all know that Smokey the Bear tells us that “Only you can prevent forest fires”]. But when Bart Simpson selects the answer you, the statue responds, “You pressed YOU, meaning me. This is incorrect. You should have pressed ME, meaning you.”12 Deictic humor can also involve the highly productive area of ambiguous pronoun reference (cf. Dubinsky and Holcomb 2011: 80). We can get some idea of this potential from one reported advertisement, where I have italicized the crucial pronoun: In a shoe advertisement for True Wash leather Keds, two filthy children are shown standing on the porch. The advertisement says, “Go ahead, just toss them in the washer and dryer. (The sneakers, that is.)” (Oaks 2010a: 271)13

A similar use of a pronoun ambiguity, this time with an indefinite pronoun rather than a personal pronoun, is found in the following joke, where I have once again italicized the pronoun: When they brought their first baby home from [the] hospital, the wife suggested to her husband that he should have a go at changing the baby’s nappy.

12 This example came to my attention through a University of Pennsylvania linguistics website, http://ling.upenn.edu/~beatrice/humor/simpsons.html, which gathered this example from a Linguist List posting by Heidi Harley. A video clip of the original Simpsons cartoon scene may be viewed at https://www.youtube.com/watch?v=vihwYGENbFg. The wording differs somewhat between Smokey the Bear’s final response in the original Simpsons cartoon and its reported form on the linguistics website. The final quoted part of this scene that I use is drawn from the version on the linguistics website. 13 This advertisement appeared in the magazine, Parents, Oct. 1993, 191.

A creative approach for linguistic funny business

297

“I’m busy,” he said. “I’ll change the next one.” So three hours later, she tried again: “Darling, will you change baby’s nappy?” “No,” he said. “I meant the next baby.” (Tibballs 2000: 43)

Deictic expressions, as the above discussion indicates, fall under different subsets with each subset containing multiple deictic terms around which a person can design different jokes. A consideration of the various subsets will provide some useful ideas for humor creation. Presuppositions are another linguistic area that deserves some significant attention. Presuppositions normally contain established, known, or agreed upon information. When we use them in ways that present an incongruity between what is in the presupposition and what we know or would assume to be the case, or is at odds with other things that we have claimed, we can get some humor effect. Morreall, in fact, discusses the humorous use of presuppositions as a type of “pragmatic incongruity” (1983: 77–78). One comedian who used to do an impersonation of President Richard Nixon would have his Nixon character, speaking of the Watergate scandal, say, “I didn’t do it, and I promise never to do it again!” (Helitzer 1987: 192). The contrast, of course, between his assertion that he didn’t do it and the word again, which introduces the presupposition that he in fact had done it, creates the humor. The word again is also used in the following joke that continues the topic of restaurants, which we have seen in a couple of the preceding jokes: A waiter brought a customer the steak he ordered with his thumb pressing down on the meat. The customer was appalled. “What are you doing putting your hand on my steak?” “Well,” replied the waiter, “you wouldn’t want it falling on the floor again.” (Tibballs 2000: 280)

In the matter of presuppositions, those individuals fashioning humor can benefit from the research that has identified specific constructions that trigger presuppositions (cf. for example, Levinson 1983: 181–185; Riley and Parker 1988: 332–341; Hickey and Orta 1990: 193–195; and Sedivy and Carlson 2011: 102–107). For example, the word again, which we have seen was useful in the two preceding jokes, is just one of a group of words called “iteratives,” which involve the notion of repetition. Other iteratives include also, too, still, etc. Together that class of expressions constitutes just one type of a variety of what are sometimes called presupposition “triggers.” Another type of trigger is the wh-question. Wh-questions, as previously noted, involve words such as who, what, where, when, why, which, and by extension also how, how many, etc. Thus if someone were to spot a friend about to enter a job interview and ask him in front of his prospective employer, “A new job? Why do you always get fired at every job after just a couple of

298

Dallin D. Oaks

weeks?” the damaging presupposition is that he always gets fired at every job after just a couple of weeks. If we realize that all the iteratives and wh-questions with their various forms and possible combinations constitute just two types of presupposition triggers and that there are multiple trigger types, with their own forms and possible combinations, then we get some idea of the immense creative potential for humor that we could have by considering the various triggers for presuppositions.

4 Additional taxonomies: Dichotomous categories Up to this point the taxonomies we have explored have primarily involved classifications and frameworks consisting of multiple categories around which humor may be constructed. This is the case whether we look at speech acts, Gricean maxims, adjacency pair types, deictic expressions, speech acts, and even the choice between direct and indirect speech acts since there are multiple types of speech acts to choose from when constructing an ambiguity along this dimension. But some classifications, at least as far as the creation of humor is concerned, involve dichotomies where we generally have only two options, and our interpretation is between one or the other choice. We can see this, for example, when we have to decide whether something is a set expression or not, a name or not, or a yes/no question versus a forced-choice question. Of course these are not the only possible dichotomies we could consider here, but they will serve to illustrate the potential usefulness that dichotomous categories as a subcategory of language taxonomies can have when constructing linguistic humor. We shall now consider some examples of these dichotomous categories. Let’s begin first with set expressions that could be alternatively interpreted as language constructions comprised of more independently arranged syntactic elements. There are actually three types of set expressions that will be considered here: idioms, compound nouns, and multi-word verbs. Idioms have a fixed and often relatively invariable form whose meaning is largely independent from its individual parts. In other words, the meaning of the idiom is frequently not directly derivable from the individual words within the idiom. One commonly discussed idiom is “kick the bucket.” It would be hard for someone not already familiar with the idiom and its meaning to see how the words in this expression add up to the meaning “die.” Some idioms are of course more semantically transparent. We might, for example, be able to recognize a metaphorical, historical, or literary basis for them. But they still require some

A creative approach for linguistic funny business

299

existing familiarity with the fixed nature and overall meaning of the idioms. In those cases where we recognize an idiom and wish to create humor, we often only need to come up with a situation that could also accommodate a more literal interpretation that adds up the words in the idiom. The potential use of idioms in humor is well illustrated in the following examples: Most of the time people get what’s coming to them – unless it’s mailed! (Berle 1989: 397) Joan: “Did you hear Erica is marrying her X-ray specialist?” Jane: “Well, she’s lucky. Nobody else could ever see anything in her.” (Copeland & Copeland 1965: 194) Anything Less Just Won’t Cut It. (Billboard advertisement for lawnmowers, as cited in Oaks 2010a: 457)

When trying to form humor around idioms, we could consider some of the idiomatic expressions that are commonly associated with specific scripts, topics, or discourse situations and which could be cleverly manipulated for humor. For example, let’s look at how we respond to someone’s expression of gratitude. Among the common idioms we use are “Don’t mention it” and “It was the least I could do.” With these idioms we could construct alternative, non-idiomatic ways of interpreting these idioms. Examples of this humorous potential that have been formed into jokes include the following: A bribe is a gift with which the giver says, “Thanks,” and the receiver says, “Don’t mention it!” (Berle 1989: 474) It was the least I could do. I always do the least I can do. (From Hawkeye Pierce’s character on the TV comedy, “M*A*S*H,” reported in Tibballs 2004: 349)

Other groups of set expressions, which are similar to idioms, at least in their potential for humor and perhaps in the way humor may be deliberately fashioned around them, are the compound nouns and multi-word verbs. Once we identify something as a compound noun or multi-word verb, with its set interpretation, it is often not so difficult to imagine an alternative interpretation that adds up the individual component parts differently. In the examples that follow, the first two involve compound nouns, and the next two involve multiword verbs: I was the Best Man at a wedding. I thought the title was a bit much. If I’m the Best Man, why is she marrying him? (Jerry Seinfeld, as reported in Brown 2005: 433) I visited an American supermarket. They have so many amazing products here. Like powder milk. You add water and you get milk. And powder orange juice. You add water and you get orange juice. Then I saw baby powder. And I said to myself, “What a country! I’m

300

Dallin D. Oaks

making my family tonight!” (Yakov Smirnoff, as reported in Tibballs 2004: 34. Smirnoff’s comedy routines played off his new immigrant status and language confusions.) Midwives help people out. (A bumper sticker, as cited in Oaks 2010a: 464)14 Why was the cannibal expelled from school? Because he kept buttering up the teacher. (Smart Alec’s 1987: 118)

Another dichotomy to consider is whether a term or utterance constitutes a name or not. Perhaps the most famous example of comedy built around this is in the Abbott and Costello routine of Who’s on First?15 In this routine, Abbott discusses baseball players with names such as Who, What, and I don’t know, but Costello is confused by these referring expressions and tries to interpret them as interrogative pronouns (Who, What), or in the case of I don’t know, as statements. As is noted in Oaks (2010a: 142), our society is used to accepting almost anything as a name. This is shown by the humorous and true story of the California child who arrived at kindergarten on the first day of class, and, in accordance with previous instructions given to the parents, bore a name tag identifying himself. The name displayed on his name tag was Fruit Stand, and all day long he was addressed with that name. This continued until the end of the school day when he was to return home and the name tag turned over to display, on the opposite side of the tag, his bus stop destination. It was at that point that the teacher discovered the name Anthony on the back side of the name tag (Oaks 2010a: 142).16 The funny thing about this situation is that it didn’t occur to the teacher that the name tag might have been turned the wrong way at the beginning of the day. Names are so individual, and no one wants to appear to challenge or question someone else’s name. A comedy writer can easily develop humor around whether or not a word or phrase in fact constitutes a name. This is evident in jokes from Stephen Wright and Bennett Cerf: I bought a dog the other day . . . I named him Stay. It’s fun to call him . . . “Come here, Stay! Come here, Stay!” He went insane. Now he just ignores me and keeps typing. (Wright, as cited in Dubinsky and Holcomb 2011: 78)

14 I am indebted to Lloyd Oaks for bringing this example to my attention. 15 Their comic routine may be viewed at https://www.youtube.com/watch?v=kTcRRaXV-fg. 16 The story about Fruit Stand is reported by Luanne Oleas and appears in the Reader’s Digest, Sept. 1992, 48. A friend, Lori Marett, informed me of this example.

A creative approach for linguistic funny business

301

An amateur musician was making horrendous sounds on his saxophone in the middle of the night when the outraged landlord burst into his apartment, yanked the instrument out of his hands, and roared, “Do you know there’s a little old lady sick upstairs?” “I don’t think I do,” admitted the amateur. “Would you mind humming the first few bars of it?” (Cerf n.d., vol. I: 335)

It should also be noted that the latter joke relies on a mistaken speech act on the part of the musician, who assumes the landlord’s utterance is a prelude to the landlord’s asking the musician to play a specific piece of music. The potential humor productivity with names is immense. Because almost any given word, phrase, or clause could in theory constitute a name, it is easy to construct settings in which one of these constituent types could allow for an unexpected interpretation as a name. Marketers have also captured attention for their products by playing around with what we might not normally imagine could be a name, as with the butter substitute, “I Can’t Believe It’s Not Butter.” Marketed names can be as unusual and as long or elaborate as a marketer thinks will effectively promote a particular product. Nilsen provides the example of one pudding product in England that bore the bizarre and lengthy name, “Congratulations dear, but exactly what does an assistant vice president do?” (1979: 141). Some questions, because of their structure, may be interpreted as both a yes/no question, or alternatively as a forced-choice question (a question in which we must respond by selecting one of the choices that are provided). In speech, when using a question like “Has he been to the play or the movie?” the intonation and whether or not a slight pause is used within the question will generally reveal the intended meaning, though for the deliberate uses of humor these cues can perhaps be mitigated. In writing, deciding between a forcedchoice or a yes/no question can be a real problem, because the spoken cues are absent and both types of questions use the same punctuation. The potential for confusion and humor is illustrated in the following incidents, one involving a job applicant, and another involving some small children: An applicant for a job with the federal government was filling out the application form. He came to this question: “Do you favor the overthrow of the United States government by force, subversion, or violence?” Thinking it was a multiple-choice question, he checked “violence.” (Johnson 1989: 173) Some children were acting out a wedding ceremony. The priest asked the “bride,” “Do you take him for better or worse?” “For better,” the little girl said quickly. The priest continued, “For richer or poorer?” “For richer,” stated the miniature bride. (Glavich 2002: 125)

302

Dallin D. Oaks

5 Conclusion The categories presented here are not a complete list of those linguistic paradigms and taxonomies that could be usefully examined for ideas when formulating linguistic-based humor. However, I hope that they can serve to show the potential usefulness that consulting paradigms and taxonomies could hold for those who wish to develop linguistic-based wordplays and other linguisticbased humor. Some challenges of course remain. Although some detailed lists or descriptions exist and can be used to outline the various grammatical structures or constructions creating structurally ambiguous wordplays (cf. Oaks 2010a) and presuppositions (cf., for example, Levinson 1983), thus assisting in the kind of humor generating strategies I have proposed here, I am not currently aware of detailed and wide-ranging lists that are publicly available for speech acts, felicity conditions, and adjacency pairs. When it comes to building humor around dichotomous categories such as idioms, compound nouns, and multi-word verbs, some lists and resources exist, though they may not currently be topically organized and searchable in ways that will be most useful and compatible with the humor approach outlined here. But even in the absence of useful, readily accessible, and comprehensive lists for some of the categories and frameworks I have mentioned in this paper, the discussion I have provided should be useful in assisting people to engage in fruitful observation and introspection, perhaps enabling them to compile some of their own working lists and to generate humor angles in the process. Moreover, with the development of computer corpora such as the Corpus of Contemporary American English (Davies 2008) and iWeb Corpus (Davies 2018), it is now possible to consult very large databases for collocates for any given word. For example, if I know I want to do a joke about horses, I could search the corpus and with a few keystrokes be reminded or informed of commonly co-occurring words such as rein, ride, hoof, wild, race, saddle, buggy, stable, etc. This would certainly aid in considering various possibilities for humor development. In any event, we have reason to be increasingly optimistic about the contribution that linguistic scholarship and research might provide, not just in the theoretical understanding of humor but perhaps even its creation and development. As linguistic descriptions and frameworks are further refined, as more detailed outlines and lists are prepared, and as linguistic computing becomes more advanced, we can expect that people using linguistic approaches to humor can become even more resourceful.

A creative approach for linguistic funny business

303

References Aarons, Debra. 2012. Jokes and the linguistic mind. New York & London: Routledge. Allen, Steve. 1987. How to be funny: Discovering the comic you. With Jane Wollman. New York: McGraw-Hill. Allen, Steve. 2000. Steve Allen’s private joke file. New York: Three Rivers Press. Attardo, Salvatore. 1990. The violation of Grice’s maxims in jokes. In Proceedings of the Sixteenth Annual Meeting of the Berkeley Linguistics Society 16 (1). 355–362. Attardo, Salvatore. 1997. Locutionary and perlocutionary cooperation: The perlocutionary cooperative principle. Journal of Pragmatics 27. 753–779. Attardo, Salvatore. 2017a. The general theory of verbal humor. In Salvatore Attardo (ed.), The Routledge handbook of language and humor, 126–142. New York: Routledge. Attardo, Salvatore. 2017b. Humor and pragmatics. In Salvatore Attardo (ed.), The Routledge handbook of language and humor, 174-188. New York: Routledge. Attardo, Salvatore, and Victor Raskin. 1991. Script theory revis(it)ed: Joke similarity and joke representation model. Humor: International Journal of Humor Research 4 (3/4). 293–347. Austin, J. L. 1975. How to do things with words. 2nd edn. J. O. Urmson and Marina Sbisà (eds.). Cambridge, MA: Harvard UP. Berle, Milton. 1989. Milton Berle’s private joke file. Milt Rosen (ed.). New York: Three Rivers Press. Brown, Judy (ed.). 2005. The comedy thesaurus. Philadelphia: Quirk Books. Bucaria, Chiara. 2004. Lexical and syntactic ambiguity as a source of humor: The case of newspaper headlines. Humor: International Journal of Humor Research 17 (3). 279–309. Cerf, Bennett. n.d. Bennett Cerf’s bumper crop of anecdotes and stories, mostly humorous, about the famous and near famous. 2 vols. Garden City, NY: Garden City Books. Copeland, Lewis, and Faye Copeland (eds.). 1965. 10,000 jokes, toasts & stories. New York: Doubleday. Davies, Mark. 2008. Corpus of contemporary American English. Available at https://corpus. byu.edu/coca. Davies, Mark. 2018. iWeb corpus. Available at https://corpus.byu.edu/iweb. Dubinsky, Stanley, and Chris Holcomb. 2011. Understanding language through humor. Cambridge: Cambridge UP. Ferrar, Madeleine. 1993. The logic of the ludicrous: A pragmatic study of humour. Unpublished diss. University of London. Glavich, Mary Kathleen. 2002. Catholic school kids say the funniest things. Mahwah, NJ: Paulist Press. Grice, H. P. 1975. Logic and conversation. In Peter Cole and Jerry L. Morgan (eds.), Syntax and semantics (vol. 3, Speech Acts), 41–58. New York: Academic Press. Hancher, Michael. 1980. How to play games with words: Speech-act jokes. Journal of Literary Semantics 9 (1). 20-29. Helitzer, Melvin. 1987. Comedy writing secrets. Cincinnati: Writer’s Digest Books. Hickey, Leo, and Ignacio Vázquez Orta. 1990. Old information and presupposition in advertising language. International Journal of Advertising 9 (3). 189–196. Hurford, James R., Brendan Heasley, and Michael B. Smith. 2007. Semantics: A coursebook. 2nd edn. Cambridge: Cambridge UP.

304

Dallin D. Oaks

Johnson, Eric W. (ed.). 1989. A treasury of humor: An indexed collection of anecdotes. New York: Ivy Books. Kohl, Marguerite, and Frederica Young. 1963. Jokes for children. New York: Hill & Wang. Levinson, Stephen C. 1983. Pragmatics. Cambridge: Cambridge UP. Morreall, John. 1983. Taking laughter seriously. Albany: State University of New York. Moulton, Powers. 1942. 2500 jokes for all occasions. Philadelphia: The Blakiston Co. Nilsen, Don L. F. 1979. Language play in advertising: Linguistic invention in product naming. In James E. Alatis and G. Richard Tucker (eds.). Language in public life (Georgetown University Round Table on Languages and Linguistics, 1979), 137–143. Washington, D.C.: Georgetown UP. Oaks, Dallin D. 2010a. Structural ambiguity in English: An applied grammatical inventory. 2 vols. London & New York: Continuum. Oaks, Dallin D. 2010b. Paradigms as predictors of structural word play potential in humor. Southern Journal of Linguistics 34 (2). 37–66. Pepicello, William J., and Thomas A. Green. 1984. The language of riddles: New perspectives. Columbus, OH: Ohio State UP. Phillips, Bob. 1990. The all-new clean joke book. Eugene, OR: Harvest House Publishers. Price, Steven D. (ed.). 2005. 1001 insults, put-downs, and come-backs. Guilford, CT: The Lyons Press. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech, and Jan Svartvik. 1985. A comprehensive grammar of the English language. London & New York: Longman. Raskin, Victor. 1985. Semantic mechanisms of humor. Dordrecht: D. Reidel Publishing Co. Riley, Kathryn, and Frank Parker. 1988. Tone as a function of presupposition in technical and business writing. Journal of Technical Writing and Communication 18 (4). 325–343. Searle, John R. 1969. Speech acts: An essay in the philosophy of language. Cambridge: Cambridge UP. Searle, John R. 1975. Indirect speech acts. In Peter Cole and Jerry L. Morgan (eds.), Syntax and semantics (vol. 3, Speech Acts), 59–82. New York: Academic Press. Sedivy, Julie, and Greg Carlson. 2011. Sold on language: How advertisers talk to you and what this says about you. Chichester, West Sussex, United Kingdom: Wiley-Blackwell. Smart Alec’s spooky jokes for kids. 1987. New York: Ballantine. That’s funny: A compendium of over 1,000 great jokes from today’s hottest comedians. 1996. New York: MJF Books. Tibballs, Geoff (ed.). 2000. The mammoth book of humor. New York: Carroll & Graf Publishers. Tibballs, Geoff (ed.). 2004. The mammoth book of zingers, quips and one-liners. New York: Carroll & Graf Publishers. Yule, George. 2017. The study of language. 6th edn. Cambridge: Cambridge UP.

John D. Battenburg

Tourism after the Arab Spring in Tunisia: An analysis of advertising campaigns Abstract: Although tourism in Tunisia accounted for 9.5 percent of the GDP prior to the January 2011 revolution, it declined to 6.5 percent by 2016. This study examines two advertising campaigns created to attract tourists to the country: the 2011 campaign after regime change and the 2015 campaign after the Bardo National Museum and Sousse hotel terrorist attacks. Social media postings, signs, advertisements, and brochures in tourism campaigns constitute “communicative acts.” The purpose of such acts is to inform and persuade. The 2011 campaign attempted to lure tourists through the use of humor while the 2015 campaign used logic by pointing out that similar attacks have occurred in the West. Analysis of these campaigns reveals insights into the language of tourism with a country responding to various crises. Keywords: Tunisia, media, advertising, tourism

1 Introduction Prior to the revolution that toppled President Zine El Abidine Ben Ali in January 2011, tourism in Tunisia accounted for 9.5 percent of the GDP. In 2016, however, tourism represented only 6.5 percent of the GDP (while still employing one in five Tunisians either directly or indirectly). This study analyzes various advertising campaigns created to attract tourists to this Mediterranean country’s sun, sand, and surf. Although it has sought to diversify tourism to attract more affluent visitors, Tunisia has always appealed to the budget traveler seeking inexpensive package deals. Two campaigns targeting potential tourists are examined: the 2011 campaign after regime change and the 2015 campaign after the Bardo National Museum and Sousse hotel terrorist attacks. Social media postings, signs, advertisements, and brochures in tourism campaigns constitute what Van Leeuwen (2004) refers to as “communicative acts” (rather than speech acts focusing on spoken language). The purpose of these communicative acts is to inform and persuade. With advertising campaigns for much of the Arab world, though, an additional challenge has been to attract visitors who may be skeptical about visiting the region due to safety concerns. https://doi.org/10.1515/9781501511707-015

306

John D. Battenburg

This paper is organized as follows. First, the research questions posed and the methodology employed are discussed. Second, several studies on the language of tourism are introduced. Third, crisis repair and strategies for attracting tourists to the Arab world are explored. Next, recent Tunisian tourism advertising campaigns are scrutinized. Finally, concluding remarks are offered about the need for analysis of the language of tourism, particularly with destinations that have experienced upheaval.

2 Research questions and methodology This study considers several related research questions. How has Tunisia sought to attract tourists with advertising campaigns after the Arab Spring? What were the differences between the 2011 campaign following regime change and the 2015 campaign following the Bardo National Museum and Sousse hotel terrorist attacks? And why was the 2011 campaign viewed as more effective than the 2015 campaign? This qualitative analysis focuses on various linguistic and visual features of the 2011 and 2015 advertising campaigns. It takes into account strategies suggested by Avraham and Ketter (2016) to restore a destination’s image and attract tourists. This study also relies on participant observation as the researcher has lived in Tunisia and in other Arab countries prior to and after the Arab Spring.

3 Research on the language of tourism While the 2011 campaign attempted to lure tourists through the use of humor and exotic depictions, the 2015 campaign appealed to solidarity with citizens from other countries who experienced terrorist attacks. Both language and images were employed in these multimodal texts. Research in the past several decades focusing on the language of tourism provides tools for studying these campaigns. Cook (2001) in The Discourse of Advertising suggests issues that must be considered about the identity of the communicator, the type of society and situation, the medium, and the interrelationship of the communicative acts. It is important to view language and context holistically rather than in isolation when studying tourism advertising campaigns. Dann (1996) points out in The Language of Tourism: A Sociolinguistics Perspective, “amazingly, no one has comprehensively analyzed this language as a phenomenon in its own right. Certainly there have been some studies

Tourism after the Arab Spring in Tunisia: An analysis of advertising campaigns

307

which have alluded to the linguistic features of tourism promotion, but none so far brought them together and systematically examined tourism as a language per se” (33). More recent research on semantics, pragmatics, and discourse analysis in the language of tourism appear in studies by Ip (2008), Mocini (2009), and Persson (2012). Analysis of campaigns for locations that have experienced human or natural disasters is also found in Avraham and Ketter’s (2016) Tourism Marketing for Developing Countries: Battling Stereotypes and Crisis in Asia, Africa and the Middle East.

4 Crisis repair within tourism Avraham and Ketter (2016) state, “Tackling prolonged negative place images is crucial for developing tourism in Africa, the Middle East, Latin America, East Europe and Asia. Although these destinations differ greatly, in the eyes of many potential tourists they all suffer from weak place images, negative stereotypes and problematic perceptions” (3). These authors also distinguish between two types of negative destination repairs: short-term, unexpected crises versus long-lasting upheavals. The former can be dealt with by restoring an image while the latter are more difficult to address. The problem, of course, is that several short-term crises may be perceived as representing a systemic long-term problem, signaling overall instability. “When a destination’s image is threatened by repeated crises,” observe Avraham and Ketter, “developing countries must constantly communicate with crisis-related messages in order to protect their image and maintain their position in the global tourism arena” (2). Al-Hamarneh and Steiner (2004) propose four types of tourism in the Arab world: European/North American leisure tourism, Arab-oriented leisure tourism, multi-ethnic tourism, and multi-ethnic cultural and pilgrim tourism. It is noteworthy that the attacks in Tunisia and, in response, the advertising campaigns themselves targeted mostly Europeans. Before the Arab Spring, Tunisian tourism sought to attract “the butcher from Germany” and similar tourists much more than exclusive visitors representing a higher socio-economic group. The goal, of course, with more recent Tunisian tourism advertising campaigns was to recapture the 3.5 million or so tourists (or half of the 7 million) who no longer visited the country after 2010. Watson (2011) quotes various tourists who were reluctant to travel to Tunisia following the Arab Spring. One Belgium visitor stated “’A lot of people told me, ‘Oh don’t go to Tunisia, you are going to see a lot of people from the army with guns and you can’t leave the hotel.’ But there’s no problem here. The Tunisians are glad to see you, they’re

308

John D. Battenburg

happy there are tourists.’” A British vacationer also indicated, “’We didn’t want to come because of the problems, but everything started to settle down so the tour operator didn’t stop us coming.’”

5 Strategies to attract tourists Avraham and Ketter (2016) observe that certain strategies have been used to attract tourists to Middle Eastern countries in which crises have occurred. For example, both Egypt and Turkey chose the “ignore the crisis strategy” following terrorist attacks in the 2000s and with the 2011 Jasmine Revolution in Egypt. Egyptian officials continued this ploy by initially refusing to identify events on October 21, 2015 in the Sinai and on May 19, 2016 over the Mediterranean as airline accidents. A second strategy is to downplay the crisis and indicate that all has returned to “business as usual.” Egyptian Tourism Minister Mounir Fakhry Abdel Nour declared at a June 12, 2011 press conference during the Arab Spring, “For the return to normality, it requires effort for the world to know that everything is back to normal in the country and rest assured of the maximum security” (Masciullo 2011). A third strategy is to limit the crisis, with officials indicating that most areas are safe. In Egypt, for example, the tourism minister pointed out that Tahir square represents only “one square kilometer in downtown Cairo” (CNN Wire 2013). Or with the attack on a synagogue in Djerba, Tunisia, Avraham and Ketter (2008) cite an official claiming “There is no terrorism in Tunisia,” and one attack “is not the end of the world” (137). Another multi-faceted approach suggested by Avraham and Ketter (2016) is to soften the “hard” image through highlighting geographic wonders, films and festivals, cultural events, and food fairs. Abu Dhabi adopted this course of action during the Arab Spring by branding yellow taxis in New York City’s Times Square. Its aim, according to tourism promoter Mubarak Al Nuaimi, was “to take the taste, sound, sight, touch and scents of Abu Dhabi to one of the centers of New York’s action” (Sambidge 2012). The Formula One Grand Prix races in Bahrain and Abu Dhabi also softened “hard” images by showcasing safe destinations with “business as usual.”

6 The 2011 campaign after regime change Yet a final option noted by Avraham and Ketter (2016) is to “ridicule the stereotypes” and thereby nullify or mock the negative stereotypes. Tunisia employed

Tourism after the Arab Spring in Tunisia: An analysis of advertising campaigns

309

this strategy in its 2011 campaign with slogans proclaiming “They say that in Tunisia some people receive heavy-handed treatment,” “They say Tunisia is a real hot spot,” and “They say Tunisia is nothing but ruins.” Each of these slogans was accompanied by an image: in the first a woman is receiving a spa massage and the side of her breast is exposed; in the next a bare-chested man is sitting on a chair at the beach; and in the third an image of the ancient columns at the archeological site of Carthage appears. Dann (1996) comments, “where photographs are featured, almost without exception they appear in tandem with a verbal message” (188). The 2011 slogans had parallel sentence structures, all beginning with the ambiguous referent “they.” Upon seeing the advertisements, potential tourists were apt to immediately question: “Who are they?” Obviously, one goal of the advertisements was to persuade viewers to be non-conformist and different from the “they” or the crowd. Marmaridou (2000) analyzes language having minimal linguistic clues. She argues, citing Levinson (1987: 5) “if we are socially close, we don’t need to be explicit” (247). By violating Grice’s maxim of manner, Persson (2012) states there is an “implied intimacy between the sender and the reader” (30). While the pronoun ”they” appears in the slogans, it is clear that the advertisers are suggesting that the parties share some hidden knowledge. In Semantic Mechanisms of Humor, Raskin (1985) explores both grammatical and lexical triggers (73). Although pronouns such as “they” often serve as grammatical recursion triggers meant to signal deixis, it is significant that no antecedents appear in the 2011 slogans. Moreover, no lexical semantic triggers are included that provide additional semantic information. The target audience is left to wonder the identity of “they.” Syrine Cherif, creator of this campaign promoting Tunisian tourism, maintained that the words rather than the images were audacious and that the campaign was in “good fun,” yet certain Tunisians took issue with the sexualized messages and/or the suggestion that the loss of lives (with over 300 people killed) or livelihoods should be trivialized with a comparison to “ruins.” Cherif also stated that these advertisements were humorous and reflected “a fresh and free speaking” – what she referred to as the “New Tunisia” (IPT Staff 2011). It is noteworthy, however, that much of the pre-revolutionary appeal to tourism in Tunisia focused on sexual openness. After achieving independence in 1956 with the presidency of Bourguiba and later with Ben Ali, Tunisia was known as a Muslim and Arab country promoting the rights of women. It was also a country where sex accompanied sun, sand, and surf. And thus Tunisian tourism developed a reputation for the “4 s’s.” These advertisements attempted to attract foreign tourists, yet they also alienated certain Tunisians.

310

John D. Battenburg

Nicolas Courant, creative director for the Memac Ogilvy bureau in Tunis, declared, “’We wanted to create a debate, it is a way to drive awareness . . . there is a perception in Europe of there still being a security problem in Tunisia.” “We wanted to do something with Tunisian humor. It is a nice way to show how things have changed . . . . It is ok to live here, we are living here’” (Shalkh 2011). In support of the provocative advertisement campaign, Cherif also noted, “’We identified the main obstacle to incoming tourism as the fear of post-revolutionary Tunisia, so we decided to face the issue directly’” (Miller 2011). Such advertisements appeared on billboards on London double-decker buses and within Paris metro stations. These scenes were similar to the 2008 Israeli campaign cited by Avraham (2009) to attract tourists by featuring a commercial with a young female tourist strolling along the beach when she spots a handsome young man seated on the sand. She continues looking back at the young man and then bumps into a pole: the voice over then proclaimed, “Indeed, Israel can be a dangerous place.” This commercial illustrates a key concept within Raskin’s (1985) SemanticScript Theory of Humor. Specifically, script opposition is apparent within this text because “dangerous” is compatible with the semantic scripts of both bumping into a pole and becoming a target of terrorism. While the former is a normal everyday occurrence, the latter is an abnormal life-threatening experience. Attardo (2007) explains that such verbal humor reveals how scripts overlap while they are also opposed (108). With these 2011 advertisements we are reminded of Weightman’s (1987) observation that tourism promotion attempts to “mystify the mundane” and “amplify the exotic” (229). It is difficult to gauge the success of the 2011 advertisements, but tourism increased by 25.3 percent in 2012. Clearly Tunisia also benefitted from relative political stability after the overthrow of President Ben Ali. The advertisements were meant to address the fear felt by potential tourists. Journalist Muna Khan (2011) asked, “What is it that makes us uncomfortable about a country using its political past to sell its future?” She concluded, “Sometimes it helps to light up a little. Especially if that lightening up will help secure us the bucks needed to rebuild and rehabilitate an economy.” The use of humor to manage anxiety in the face of potential danger is increasingly analyzed by researchers. Long and Greenwood (2013) observe that creativity and positive thinking may serve as “adaptive benefits of humor” (496). They write, “Humor production may be particularly relevant to staving off death anxiety, not only because it typically is a culture-bound phenomenon, and hence useful for reaffirming one’s place in society, but humor has also been identified as a psychologically useful coping mechanism that enables individuals to remain resilient in the face of aversive life circumstances” (494).

Tourism after the Arab Spring in Tunisia: An analysis of advertising campaigns

311

Ridiculing the stereotype and humor was also used in March 2015 when the Tataouine region was wrongly reported by some Western media of harboring pro-ISIS fighters. Reminding readers of the first Star Wars film, Tunisian social media poked fun at purported ISIS connections with slogans including “ISIS troops were deployed in town,” “Sharia started to be enforced,” and “ISIS produced new propaganda videos.” Images of characters such as Darth Vadar dressed up in Star Wars costumes accompanied these slogans. Although such depictions appeared on Facebook, they did not lead to a large online following.

7 The 2015 campaign after terrorist attacks An “unofficial and spontaneous” Facebook campaign igniting more controversy, however, was launched by Tunisian digital designer and businessman Selim Ben Hadj Yahia in July 2015 after the attacks on tourists at the Bardo Museum (on March 18, 2015 with 22 killed) and at the Sousse beach resort (on June 26, 2015 with 39 killed). Clearly, ridiculing the stereotypes or using humor would be ineffective. By asking “How can I help my beautiful country that is dying,” Yahia (2015) sent out what he referred to as “a call to support Tunisia.” He continued in his statement to BuzzFeed France, “I admit it came out of frustration, because we suffer from this tragedy and I think 7% of the country’s GDP comes from tourism . . . . And in Tunisia, apart from the horror and the tragedy we’re living, we’re also thinking ‘shit, the tourists are not going to come anymore.’ This is one of the things that’s going on in the collective Tunisian psyche.” “I wanted to remind people of two obvious things: 1) You should come back to Tunisia, it’s not a war zone. 2) We are a land of peace and we need your support.” The 2015 “Would You Stop Visiting?” campaign portrayed vivid images of attacks in New York, London, Glasgow, and Paris along with the slogan “Support Tunisia–Land of Peace.” Whereas the 2011 campaign represented what George (2010) refers to as a “mythical landscape” with exotic depictions, the 2015 campaign aimed to counterbalance the effects of terrorist attacks by pointing out that similar attacks have occurred in the West. Persson (2012) indicates concerning the language of advertising, “paralanguage, images and lexical choices must work together, reinforcing one another, rather than creating some form of contradiction” (35). With the 2015 “Would You Stop Visiting?” advertisements, though, a contradiction was apparent. While images of the terrorist attacks in the West shocked the reader or evoked sympathy, the slogans also reminded potential tourists of safety concerns in Tunisia.

312

John D. Battenburg

The 2015 advertising campaign essentially pointed out what many experts have conceded: “very few destinations are now safe from the threat of war or terrorism, as contagion effects reverberate around the globe and erode the confidence of international tourists” (Carlsen and Hughes 2008: 147). But this is not the message tourists desire. The advertisements equating terrorist attacks in Western Europe with similar attacks in Tunisia did not reassure those with safety concerns. According to Mocini (2009), isotopy–recurring text having the same semantic field and thereby reinforcing the interpretation of a theme–commonly appears in the language of tourism. The isotopy in the 2015 advertising campaign is of argumentation. Mocini points out stages including admission, exemplification, contrast, and surprise that are employed in constructing an argument. The text “unfolds logically and persuasively to offer substantial support for the initial analogy” (156). By pointing out that other cities continue to attract tourists after experiencing terrorist attacks, Yahia attempted to use argumentation to appeal to potential tourists. This strategy was ineffective for several reasons: first, the images of the attacks in the West reinforced fear of terrorism; second, the attacks in Tunisia targeted tourists while those in the West were indiscriminate, and finally, there was a belief that the perpetrators attacking the West were from Arab countries such as Tunisia.

8 Conclusion Some journalists questioned the effectiveness of the “Support Tunisia–Land of Peace” advertising campaign with headlines including “Images of Global Terror Attacks Used to Draw Support for Tunisia” (New Age Islam, 2015) and “Images of 7/7 and 9/11 Used to Promote Tunisian Tourism” (Porter 2015). Many other individuals also responded negatively to the campaign. Overall, it is clear that logic plays little role in persuading individuals to visit a particular country, and this campaign reminded potential tourists of the possible dangers rather than allying their fears. While contributing to the growing body of literature on the language of tourism, this study about tourism advertisements after the Arab Spring in Tunisia considers both effective and ineffective ways in which a country has responded to crises. It also highlights the need for those working on tourism advertising campaigns to better understand how to craft communicative acts to attract visitors to destinations requiring repair.

Tourism after the Arab Spring in Tunisia: An analysis of advertising campaigns

313

In order to promote the tourism industry, advertisers use text and images that constitute a specific form of communication. Dann (1996) comments, “the language of tourism attempts to persuade, lure, woo and seduce millions of human beings, and, in so doing, convert them from potential into actual clients” (2). Analysis of tourism advertising campaigns in Tunisia after regime change and terrorist attacks provides insights into how a country seeks to restore its image after experiencing various crises.

References Al-Hamarneh, Ala & Christian Steiner. 2004. Islamic tourism: Rethinking the strategies of tourism development in the Arab world after September 11, 2001. Comparative Studies of South Asia, African and the Middle East 24(1). 173–182. Attardo, Salvatore. 2007. A primer for the linguistics of humor. In Victor Raskin (ed.), Primer of Humor Research, 101–156. Berlin, Germany: Mouton de Gruyter. Avraham, Eran & Eran Ketter. 2008. Media strategies for marketing places in crises: Improving the image of cities, countries, and tourist destinations. Oxford, England: Butterworth Heinemann. Avraham, Eran & Eran Ketter. 2016. Tourism marketing for developing countries: Battling stereotypes and crisis in Asia, Africa and the Middle East. London, England: Palgrave Macmillan. Avraham, Eli. 2009. Marketing and managing nation branding during prolonged crisis: The case of Israel. Place Branding and Public Diplomacy 5(3). 202–212. Carlsen, Jack & Michael Hughes. 2008. Tourism market recovery in the Maldives after the 2004 Indian Ocean tsunami. Journal of Travel & Tourism Marketing 23(2–4). 139–149. CNN Wire. 2013. Egypt’s tourism minister: Webcams will prove we’re safe. CNN, 7 February. (accessed 8 May 2018). Cook, Guy. 2001. The discourse of advertising. London, England: Routledge. Dann, Graham M. S. 1996. The language of tourism: A sociolinguistic perspective. Oxon, England: CAB International. George, Jodie. 2010. Writing the mythical landscape: An analysis of the poetic devices used in tourism’s promotional literature. Transnational Literature 3(1). 1–12. http://hdl.handle. net/2328/15097 (accessed 10 June 2018). Ip, Janice Yui Ling. 2008. Analyzing tourism discourse: A case study of a Hong Kong travel brochure. LCOM Papers 1. 1–19. IPT Staff. 2011. Sexy Tunisian tourism ads turn heads in Europe. International Business Times, 20 June. http://www.ibtimes.com/sexy-tunisian-tourism-ads-turn-heads-europe-292063 (accessed 3 June 2018). Khan, M. 2011. Getting the groove back in Tunisia: Sex and its many uses in luring tourists. Al Arabiya News, 27 June. http://english.alarabiya.net/articles/2011/06/27/155069.html (accessed 5 November 2017). Levinson, Stephen C. 1987. Minimization and conversation inference. In Jef Verschueren & Marcella Bertuccelli Papi (eds.), The Pragmatic Perspective, 61–129. Amsterdam, Netherlands: Benjamins.

314

John D. Battenburg

Long, Christopher R. & Dara N. Greenwood. 2013. Joking in the face of death: A terror management approach to humor management. Humor: International Journal of Humor Research 26(4). 493–509. Masciullo, Mario. 2011. Tourist flow to Egypt on the recovery. eTN: Global Travel Industry News, 12 June. http://www.eturbonews.com/65736/middle-east-aims-expand-micesector/53?no_redirect=true (accessed 30 September 2017). Marmaridou, Sophia. 2000. Pragmatic meaning and cognition. Philadelphia, PA: Benjamins. Miller, David E. 2011. Tunisia risks travel controversy with travel advertisements. Arab News, 17 June. http://www.arabnews.com/node/380980 (accessed 22 April 2018). Mocini, Renzo. 2009. The verbal discourse of tourist brochures. Annali della Facoltà di Lingue e Letterature Straniere dell’Università di Sassari 5. 153–164. New Age Islam Staff. 2015. Images of global terror attacks used to draw support for Tunisia. New Age Islam: Mapping an Agenda for the Twenty-First Century, July 6. http://www.newageislam.com/islamic-world-news/new-age-islam-news-bureau /images-of-global-terror-attacks-used-to-draw-support-for-tunisia/d/103786 (accessed 8 January 2018) Persson, Lotta. 2012. The language of tourism: How the tourism industry promotes magic. Halmstad, Sweden: Halmstad University BA thesis. Porter, Lizzie. 2015. Images of 7/7 and 9/11 used to promote Tunisian tourism. The Telegraph, 7 July. http://www.telegraph.co.uk/travel/destinations/africa/tunisia/articles/Images-of -77-bombings-used-to-promote-Tunisia-tourism/ (accessed 10 May 2018). Raskin, Victor. 1985. Semantic mechanisms of humor. Dordrecht, Netherlands: D. Reidel. Sambidge, Andy. 2012. New York Times Square to get Abu Dhabi makeover. HotelierMiddle East.com 13 May. http://www.hoteliermiddleeast.com/14222-new-yorks-times-sq-to-getabu-dhabi-makeover/. (accessed 22 June 2018). Shalkh, Thair. 2011. Risqué Tunisian tourism campaign woos holidaymakers. London CNN 16 June. http://edition.cnn.com/2011/WORLD/europe/06/16/tunisia.jasmine.revolution. adverts/index.html (accessed 22 November 2017). van Leeuwen, Theo. 2004. Ten reasons why linguists should pay attention to visual communication. In Philip LeVine & Ron Scollon (eds.), Discourse and Technology: Multimodal Discourse Analysis, 7–19. Washington, D. C.: Georgetown University Press. Watson, Katy. 2011. Tunisia tourism industry suffers amid Libya violence. BBC News 8 April. http://www.bbc.com/news/business-12993175 (accessed 5 June 2018). Weightman, Barbara A. 1987. Third world tour landscapes. Annals of Tourism Research 14. 227–239. Yahia, Selim Ben Hadj. 2015. Would you stop visiting New York? Buzzfeed 29 June.Translated from French by M. Telling. https://www.buzzfeed.com/ceciledehesdin/would-you-stopvisiting-new-york-asks-a-facebook-campaign-af (accessed 8 June 2018).

Names Index Aarons, Debra 282, 288, 291–292, 296, 303 Abbott, Valerie 37, 300 Abelson, Robert 13–14, 20, 41 Adams, Reginald B. Jr. 187 Aharoni, Ron 168, 186 Al-Hamarneh, Ala 307, 313 Alatis, James E. 304 Alexiadou, Artemis 165 Allen, Steve 284, 287, 294, 303 Alvermann, Donna E. 36–37 Anderson, Richard C. 12, 35–37, 295 Andor, József 21, 37 Androutsopoulos, Jannis 165 Angleitner, Alois 136 Apresjan Juri D. 90 Apte, Mahadev 102, 105, 111 Arabnia, Hamid A. 255 Astapova, Anastasiya 163–164 Atkins, Beryl T. 19, 21, 39 Attardo, Salvatore 1–6, 11–12, 14, 16, 18, 20–22, 24, 26, 28, 30, 32, 34, 36–38, 40, 94, 96–98, 101, 111–112, 116, 119, 134–135, 167, 169–171, 185–188, 201, 223, 229–230, 234, 237, 255–256, 273–274, 276, 281, 288, 290, 303, 310, 313 Austin, John L. 282, 303 Avraham, Eli 306–308, 310, 313 Avraham, Eran 306–308, 310, 313 Aydin, Tunç Ozan 273, 276 Azam, Samiul 272–274, 276 Baker, Amanda 186, 287 Bales, Michael E. 26–27, 38 Bartlett, Frederic Charles 12–13, 17, 20, 28, 35, 38–39 Baugh, John 223 Bavelas, Janet B. 197, 223 Bell, Nancy 93–94, 97–98, 101, 106, 109, 111, 170, 187, 201, 223, 235 Belohlavek, Radim 258–259, 266 Berle, Milton 299, 303 Bertrand, Roxane 198–199, 203, 218, 223 Bertuccelli Papi, Marcella 313 https://doi.org/10.1515/9781501511707-016

Bicego, Manuele 276 Billig, Michael 103, 111, 163–164 Bjarkman, P. C. 2, 6 Blache, Philippe 203, 223 Black, John B. 37, 122, 276 Boas, Hans C. 21, 38, 174 Bobrow, Daniel G. 41 Boersma, Paul 203, 223 Boguslavsky, Igor M. 44, 54, 90 Bolukbasi, Tolga 236, 255 Bondarko, Alexander V. 43, 54 Boricheva, Inna 267 Boucenna, Sofiane 270, 276 Bourdieu, Pierre 99, 105, 111 Boxer, Diana 223 Brachman, Ronald J. 16, 24–25, 38, 40 Bransford, John D. 36, 38 Brewer, William E. 40 Brown, Ann L. 29, 31, 36, 38, 50–52, 102, 113, 299, 303 Brown, Judy 29, 31, 36, 38, 50–52, 102, 113, 299, 303 Brown, Roger 29, 31, 36, 38, 50–52, 102, 113, 299, 303 Bruce, Bertram C. 40, 266 Brzozowska, Dorota 6, 113, 186–187 Bucaria, Chiara 282, 303 Buchanan, Bruce G. 260, 266 Buela-Casal, Gualberto 134 Buijzen, Moniek 274, 276 Carlsen, Jack 312–313 Carlson, Greg 297, 304 Carrell, Amy T. 116, 169, 187 Carretero-Dios, Hugo 120, 134 Ceccato, Silvio 14, 38 Cerf, Bennett 300–301, 303 Chafe, Wallace L. 19, 52, 54 Chang, Kai-Wei 255 Chapman, Antony J. 255, 267 Charniak, Eugene 17–18, 38 Chiaro, Delia 109, 111, 224 Chłopicki, Władysław 4, 6, 113, 167–190 Chomsky, Noam 163–164, 168, 175, 187

316

Names Index

Coates, Linda 223 Cohn, Jeffrey F. 276 Cole, Peter 39, 54, 223, 303–304 Collins, Allan M. 24, 38, 41 Cook, Guy 306, 313 Cook, Walter A. 19, 38 Copeland, Faye 299, 303 Copeland, Lewis 299, 303 Coser, Rose Laub 93, 111 Coupland, Nikolas 165 Craik, Kenneth 38, 125–126, 129, 131–132, 134 Crisp, Peter 74, 76 Cristani, Marco 276–277 Croft, William 33, 35, 38, 171, 181, 187 Cruse, D. Alan 33, 38, 171, 181, 187 Csurka, Gabriela 276 Dann, Graham M. S. 306, 309, 313 Dauben, Joseph W. 266 Davies, Christie 96–97, 106, 112–113, 187 Davies, Mark 302–303 de la Fuente, David 255 Dennett, Daniel C. 187 DeSoucey, Michaela 95, 104, 110, 112 D’Errico, Francesca 277 DiMaio, Sara 37, 255 Donaldson, Wayne 38 Donohue, W. A. 7, 54 Double, Oliver 71, 104, 171, 187, 209, 224, 276, 310 Drew, Paul 169, 208, 223 Drosou, Anastasios 270, 276 Dubinsky, Stanley 282, 293, 296, 300, 303 Duch, Włodzisław 273, 276 Dunn, Jonathan 3, 55–58, 60, 62, 64–68, 70, 72, 74–76 Dunsmore, Kailonnie 39–40 Dux, Ryan 21, 38 Dziobek, Isabel 135 Edwards, Derek 12, 38 Eisterhold, Jodi 201, 223 Ellinas, Antonis A. 143, 164 Ellis, Donald G. 7, 54

Ellsworth, Michael 40 Espesser, Robert 223 Feagin, Crawford 223 Ferré, Gaelle 223 Ferrar, Madeleine 282, 303 Fillmore, Charles J. 1, 6, 12–13, 17–22, 26, 29–30, 33–35, 37–39 Fine, Gary Alan 95, 104, 110, 112, 115, 119, 281, 290 Fontenelle, Thierry 21, 39 Foot, Hugh C. 33, 173, 255, 289 Ford, Thomas 98, 111–112 Fried, Mirjam 21, 39, 155 Fry, William 99, 171, 187 Fuentes, Olac 267 Garg, Nikhil 236, 255 Gaussier, Philippe 276 Gavelek, James R. 39–40 Gavrilova, Marina L. 5, 269–270, 272–274, 276–277 Geckeler, Horts 20, 39 Geeraerts, Dirk 39 Georgakopoulos, Thanasis 165 George, Jodie 39, 50, 76, 140, 145–146, 148–149, 151, 266, 284–285, 304, 311, 313 Glavich, Mary Kathleen. 301, 303 Goatly, Andrew 74, 76 Gobbini, M. Ida 276 Goetz, Ernest T. 37, 41 Goffman,Erving 206, 223 Goodwin, Margaret H. 200, 223 Goutsos, Dionysis 143, 164 Gray, Jeanette 135 Green, Thomas A. 33, 35, 80, 282, 304 Greenbaum, Sidney 304 Greenwood, Dara N. 310, 314 Grice, H. Paul 45, 54, 192–193, 223, 282, 288, 303, 309 Gross, Markus 276 Guardiola, Mathilde 198–199, 218, 223 Guarino, Nicola 230, 251, 255 Guy, Gregory R. 32, 77, 99, 158, 161, 211, 214, 223, 313

Names Index

Hafemeister, Laurence 276 Hale, Adrian 93, 112 Hall, Edward T. 39, 101, 112, 266–267 Hamid, Raffay 255, 276 Hancher, Michael 282, 291, 294, 303 Hatzidaki, Ourania 143, 164 Haugeland, John 13, 39–40 Haxby, James V. 270, 276 Hay, Jennifer 13, 17, 39, 98, 100, 106, 112, 201, 223–224 Hays, David G. 13, 17, 39 Heasley, Brendan 282, 303 Hehl, Franz-Josef 117, 119–120, 133–136 Heintz, Sonja 119–120, 123, 125–126, 129, 131, 133–134, 136 Helitzer, Melvin 297, 303 Helm, June 224 Hempelmann, Cristian F. 7, 37, 54, 96–97, 112–113, 115, 117, 119, 134–136, 231, 234, 255–256 Heylen, Dirk 223, 277 Hickey, Leo 297, 303 Hoffman, Elizabeth A. 276 Hofmann, Jennifer 119, 124, 131, 134–136 Holcomb, Chris 282, 293, 296, 300, 303 House, Juliane 24, 88, 146, 159, 165, 235–237, 239, 242, 284, 304 Howard, Newton 267 Huber, Oswald 136 Hughes, Michael 312–313 Hurford, James R. 282, 303 Hurley, Matthew M. 187 Ioannidis, Dimosthenis 276 Ip, Janice Yui Ling 307, 313 Isola, Phillip 270, 276 Jain, Anil K. 270, 276 Jakobson, Roman 43, 45, 54 Jefferson, Gail 197, 199, 224 Jia, Jia 277 Johnson, Christopher R. 39, 40 Johnson, Marcia K. 36, 38 Johnson, Eric W. 301, 303 Johnson, Mark 55, 76 Johnson, Stephen B. 26–27, 38

317

Johnson, Trudy 223 Johnston, Elizabeth B. 12, 39 Jurafsky, Dan 255 Kalai, Adam T. 255 Kaliouby, Rana El 276 Kalokairinos, Alexis 165 Kaszniak, Alfred W. 135 Ketter, Eran 306–308, 313 Köler, Gabriele 124, 136 Khan, Muna 310, 313 Khosla, Aditya 274, 276 Kibrik, Aleksandr E. 255 Kirchner, Jennifer 133, 135 Klein, Sheldon 13, 39 Klir, George J. 258–259, 266 Kobozeva, Irina M. 2, 27, 43–44, 46, 48, 50, 52, 54 Koc, Alper Kemal 277 Kohl, Marguerite 286, 304 Koniordos, Sokratis 143, 150, 164 Kosheleva, Olga 257–258, 260, 262, 264, 266 Kotthoff, Helga 106–107, 112 Koutsoulelou, Stamatia 144, 164 Kozerenko, Elena B. 255 Kramer, Elise 110, 112 Krasny, Karen A. 36, 39 Kreinovich, Vladik 5, 257–260, 262, 264, 266–267 Kuipers, Giselinde 3, 93–96, 98–100, 102, 104, 106–108, 110–112, 171, 174, 179, 186–187 Kumar, Sundeep 259, 267 Labov, William 197, 209–210, 213, 223–224 Laineste, Liisi 139, 163–164 Lakoff, George 32–34, 39, 55, 74, 76 Lambrecht, Knud 32, 39 Lambrianou, Iasonas 143, 164 Lampert, Martin D. 134 Larlus, Diane 276 Larsen, Gwen 135 Lassila, Ora 11, 17, 39 Lea, Robert N. 267 Leech, Geoffrey N. 45, 54, 304 Lefcourt, Herbert M. 125, 129, 135 Lehrer, Adrienne 39

318

Names Index

Levesque, Hector J. 16, 38, 40 Levinson, Stephen C. 297, 302, 304, 309, 313 Li, Charles N. 54 Lianhong, Cai L. 277 Lindstrom, Naomi 163–164 Lizardo, Omar 98, 112 Lokshin, Anatole 267 Long, Christopher R. 1–2, 5, 26, 49, 83, 100, 102, 116, 127, 145, 147, 173, 175–177, 180, 182, 184, 195, 201, 214, 300–301, 307, 310, 314 Lovato, Pietro 273, 276 Lytinen, Steven L. 14, 26, 39 Magnenat-Thalmann, Nadia 277 Mandelbaum, Jenny 197–200, 209, 224 Mandelbrot, Benoit B. 27, 40 Mann, William C. 44, 54 Marchesotti, Luca 274, 276 Marco, Cristani 136, 276–277 Marmaridou, Sophia 309, 314 Marsh, Moira 101, 112, 170, 187 Martin, Rod 98, 111–112, 125–126, 129, 134–135, 223 Mascha, Efharis 163–164 Masciullo, Mario 308, 314 Matei, Sorin Adam 276 McDuff, Daniel 269, 276 McGhee, Paul E. 116, 129, 135 McGuinness, Deborah 11, 17, 39 McVee, Mary B. 36, 39–40 Mel’čuk, Igor A. 25, 40, 79, 90 Mendel, Jerry M. 258–259, 267 Mervis, Carolyn B. 21, 34, 40 Meunier, Christine 223 Meyer, John C. 274, 276 Middleton, David 12, 38 Miller, David E. 310, 314 Minsky, Marvin 13, 16–18, 20–21, 28, 40 Mocini, Renzo 307, 312, 314 Moc̆kor̆, Jirì 258, 267 Monwar, Maruf 270, 276 Morgan, Jerry L. 38, 40–41, 54, 223, 303–304 Morreall, John 101, 111–112, 288, 297, 304 Mortensen, Janus 165 Moulton, Powers 292, 304 Moustakas, Konstantinos 276

Nelson, Arvalea J. 134 Nguyen, Hung T. 258–260, 267 Niewiadomski, Radoslaw 134 Nijhold, Anton 136 Nikiforidou, Kiki 21, 39 Nilsen, Don L. F. 116, 301, 304 Nirenburg, Inna 54 Nirenburg, Sergei 2, 7, 13, 31, 40, 47, 54–55, 57, 71, 76, 230–231, 255 Norrick, Neal 182, 187, 197, 224 Novák, Vilém 258–259, 267 Oaks, Dallin D. 6–7, 135, 281–286, 288, 290, 292, 294, 296, 298–300, 302, 304 Oberle, Daniel 255 Oliva, Aude 255, 276 Olivas, José A. 255 O’Malley, Eoin 134, 165 Oring, Elliott. 96, 101, 103, 112, 170–171, 182, 187 Orta, Ignacio Vàzquez 297, 303 Ouwerkerk, Jaap 112 Paducheva, Elena V. 83, 90 Paivio, Allan 37, 39–41 Pantic, Maja 277 Papadopoulos, Basil K. 140, 266 Pappas, Takis S. 143, 165 Parikh, Devi 276 Parker, Frank 297, 304 Paul, Padma Polash 54, 135, 187, 223, 269, 272, 276–277 Pavlidou, Theodossia-Soula 165 Pearson, P. David 35–37 Pechlivanos, Miltos 165 Pedersen, E. L. 224 Peirce, Charles Sanders 17, 22–24, 40 Pelachaud, Catherine 277 Pepicello, William J. 282, 304 Peréz, Cristino 134 Perfilieva, Irina 267 Perina, Alessandro 276–277 Perronnin, Florent 276 Persson, Lotta 307, 309, 311, 314 Petrou, Maria 276 Petruck, Miriam R. L. 21, 39–40 Philip LeVine 314

Names Index

Phillips, Bob. 289, 304 Picard, Rosalind W. 276 Pickering, Lucy 186, 223 Pien, Diana 236, 255 Plataniotis, Konstantinos N. 267 Platt, Tracey 117, 120, 131, 134–136 Poggi, Isabella 277 Popa, Diana Elena 139, 163–165, 269, 277 Popa, Mirela 139, 163–165, 269, 277 Porter, Lizzie 312, 314 Prabhakar, Salil 276 Price, Steven D. 84, 148, 289, 304 Priego-Valverde, Béatrice 4, 28, 191–192, 194, 196–198, 200–202, 204, 206, 208, 210, 212, 214, 216, 218, 220, 222–224 Proyer, René T. 126, 131–132, 134–136 Puhlik-Doris, Patricia 135 Quillian, M. Ross 13, 16–17, 20–24, 30, 38, 40 Quintana, Chris 267 Quirk, Randolph 284, 303–304 Radcliffe-Brown, Alfred 102, 113 Raskin, Victor 1–7, 11–13, 24, 26, 28–31, 40, 43, 46–47, 54–55, 57, 71, 76, 78, 93–98, 101, 107, 111, 113, 115, 119, 130–135, 139–145, 147, 149–151, 153, 155–165, 167–169, 171–172, 174, 180, 186–187, 191–197, 224, 229–232, 235, 237, 241, 247, 249, 255–257, 260, 266–267, 270, 273–277, 281, 303–304, 309–310, 313–314 Rauzy, Stephane 223 Rayz, Julia 7, 31, 96, 113, 229–230, 232, 234, 236, 238, 240, 242, 244, 246, 248, 250, 252, 254, 256, 267 Reed, Stephen K. 28, 40 Reynolds, Ralph E. 37 Reznik, Leonid 267 Riemann, Rainer 136 Riley, Kathryn 297, 304 Ritchie, Graeme 170, 187, 236, 255 Rodden, Frank A. 132, 135 Rosch, Eleanor 21, 34, 40

319

Ross, Arun 26, 32, 40, 276 Ross, Haj 26, 32, 40, 276 Rothbart, Mary K. 236, 255 Rothkrantz, Leon JM 277 Ruch, Willibald 3–4, 96, 98, 112–113, 115–120, 122, 124, 126, 128–136, 138, 224, 253, 256 Ruddell, Robert B. 37 Rumelhart, David E. 35–36, 40 Ruppenhofer, Josef 20–21, 40 Sacks, Harvey 197–198, 224 Sadoski, Mark 11, 29, 36, 39, 41 Saligrama, Venkatesh 255 Sambidge, Andy 308, 314 Samson, Andrea C. 120, 133, 136, 234, 256 Santa Ana, Otto 139, 165 Sarma, Atish Das 276 Schallert, Diane L. 37 Schank, Roger C. 13–16, 20–21, 25, 28, 41 Scheffczyk, Jan 40 Schegloff, Emanuel 198, 215, 224 Schiebinger, Londa 255 Schiffrin, Deborah 223 Schmid, Hans-J?rg 33, 41 Schroeder, Marc 277 Scollon, Ron 314 Searle, John R. 63–65, 67–68, 70, 75–76, 282, 291, 293–294, 304 Sebe, Nicu 276 Sedivy, Julie 297, 304 Segalin, Cristina 269–270, 276–277 Selting, Margret 197–198, 210, 225 Shalkh, Thair 310, 314 Shan, Caifeng 277 Shilikhina, Ksenia 169, 187, 196, 225 Shively, Rachel 93, 100–101, 113 Shortliffe, Edward H. 260, 266 Sidnell, Jack 224–225 Simmons, Robert F. 13, 39 Skopeteas, Stavros 165 Sloutsky, Vladimir M. 41 Smith, Michael B. 303 Smith, Edward E. 37 Smith, Michael H. 260, 267 Smith, Moira 93, 103, 113 see Marsch, Moira

320

Names Index

Smolic, Aljoscha 276 Somers, Harold L. 78, 90 Sover, Arie 165, 171, 187 Sowa, John F. 17, 24, 41 Spasovski, Ognen 136 Spilioti, Tereza 146, 165 Spinath, Frank M. 136 Spiro, Rand J. 40 Staab, Steffen 255 Stathi, Katerina 165 Steen, Gerard J. 57, 76 Steiner, Christian 307, 313 Stivers, Tanya 197–199, 209, 224–225 Strand, Michael 112 Studer, Rudi 255 Sugeno, Michio 267 Sulejmanov, Filip 120, 136 Sullivan, Karen 74, 76 Sultana, Madeena 270–271, 273–274, 276–277 Svartvik, Jan 304 Syropoulos, Apostolos 266

Valkenburg, Patti M. 274, 276 van Dijk, Wilco 112 van Leeuwen, Theo 305, 314 Vinciarelli, Alessandro 269, 277 Vladimirou, Dimitra 146, 165 Wagner, Lisa 136 Waletzky, Joshua 197, 224 Walker, Elbert A. 258–260, 267 Wang, Xiaohui 277 Wang, Yingxu 267, 277, 304 Watson, Katy 307, 314 Weber, Marco, W 120, 136 Weenink, David 203, 223 Weightman, Barbara A. 310, 314 Weir, Kelly 135 Weiser, Irwin 6–7, 40 Widrow, Bernard 267 Wierzbicka, Anna 26, 41 Wiggers, Pascal 277 Woods, William A. 17, 41 Xiao, Jianxiong 276

Taylor, Julia 5, 7, 96, 113, 135, 187, 229–232, 234, 236, 238, 240, 242, 244, 246, 248, 250, 252, 254–257, 260, 267 Tesnière, Lucien 18, 41 Thøgersen, Jacob 165 Thomson, Sandra A. 44, 54 Tibballs, Geoff 297, 299–300, 304 Tolbert, Dana 260, 267 Torralba, Antonio 276 Trier, Jost 19–20, 41 Triezenberg, Katrina E. 54, 171, 187 Tucker, G. Richard 304 Tulving, Endel 38 Tunç, Asli 163, 165 Tzovaras, Dimitrios 276 Ungerer, Friedrich 33, 41 Unrau, Norman 37 Urbain, Jerome 134

Yahia, Selim Ben Hadj 311–312, 314 Yampolskiy, Roman V. 269, 273, 277 Yin, Jiaming 277 Young, Frederica 126, 157, 177, 179, 224, 232, 234–237, 281, 286, 292, 304, 310 Yuan, Bo 258–259, 266 Yule, George 291, 304 Zadeh, Lotfi A. 5, 7, 258–259, 267 Zhang, Yuzhe 270, 277 Zheng, Jianmin 277 Zou, James 255 Zvegintzev, Vladimir A. 7 Żygulski, Kazimierz 188 Zysset, Stefan 136

Subject Index abstractness 63, 65, 76 adjacency pair 198, 281, 294–295, 298, 302 adverb 44 advertising 6, 224, 281, 303–307, 309, 311–313 allusion 147 ambiguity 6–7, 72–73, 121, 281–287, 290–294, 296, 298, 303–304 amusement 98, 102–103, 131, 134, 181 anteriority 2, 43–44, 47, 53 appreciation 100–102, 111, 115, 117, 120–123, 128, 130, 132–134, 136, 171, 223, 256 biometrics 269, 271–277 bona fide communication 98, 191 (see also non-bona-fide) canned joke 100, 106, 179 cheerfulness 115, 124, 126, 128, 133 clause type 282–284, 286 cognitive semantics 46 cohesion 47 collocate 302 common knowledge 47 communicative acts 305–306, 312 complex sentence 43 compound noun 298–299, 302 computational humor 4, 111, 136 computer corpora 302 conceptual dependencies 11, 13–16, 21, 25 conceptual mapping 55–56, 61–62, 70–71, 75 conceptual metaphor 55–57, 71, 74 conjunction 43–44, 52, 80 connector 2, 30, 43–45, 47–54 construction 4, 12, 34, 38, 40, 44, 74, 76, 79–80, 85–86, 117, 134, 136, 167–169, 171, 173–175, 177, 179, 181, 183, 185–187, 189, 225, 236, 281–282, 285–286, 297–298, 302 context 1, 4, 13–14, 17–18, 28–29, 32, 47–48, 50–51, 53, 57, 73, 78–80, 83–85, 88, 94–95, 101, 109, 112, 139–140, 142–144, 146, 161–164, 170–171, 174, 187, 191, https://doi.org/10.1515/9781501511707-017

203, 231, 248–249, 269–270, 274–275, 277, 283, 291, 306 conversational maxim 45, 193, 282, 288 cooperation 188, 191–193, 195–197, 199–200, 208, 221–222, 224, 290, 303 Cooperative Principle 45, 193–196, 288, 290, 303 coordinate construction 44 coreference 52 correlative construction 44 crisis repair 306–307 culture 20, 30, 93, 95–96, 100, 102, 104–105, 107–110, 112, 140, 153, 163–165, 174, 194, 310 cybersecurity 269 deictic expression 295–298 deixis 281, 295–296, 309 denigration 139, 141, 143–144, 147–151, 162, 215, 218 dichotomous category 298, 302 direct speech act 291–293 discourse 1, 4, 6–7, 12, 28, 36–37, 39, 43–45, 50–51, 53–54, 113, 151–153, 164–165, 186–187, 196, 200, 223, 225, 281–282, 288, 294, 299, 306–307, 313–314 embarrassment 109, 176, 178 emotion recognition 269 ethnic humor 4, 112, 153 exposure 139, 141–144, 148, 153, 156, 158–162, 243 expressivity 45, 47 failed humor 93–94, 97–98, 111, 187, 191–195, 197, 199, 201–209, 211, 213, 215, 217, 219–221, 223–225 felicity condition 293–294, 302 figure of speech 49 French 44, 99, 151, 191, 193, 195, 197, 199, 201, 203, 205, 207, 209, 211, 213, 215, 217, 219, 221, 223–225, 314 fuzzy logic 257–259, 261, 263, 265–267

322

Subject Index

gelotophobia 115, 126, 134, 136 grammatical category 43 Greek financial crisis 139, 142–143 Gricean maxim 281, 288–291, 298 hearer’s feedback 172, 176, 182, 186 humor 1–7, 26, 28, 35, 37, 40, 77, 91, 93–113, 115–137, 139, 141, 143–144, 149, 151, 153, 155–156, 162–165, 167, 186–188, 191–197, 199–209, 211–213, 215–225, 229–233, 235–237, 239, 241, 243, 245, 247, 249, 251, 253, 255–256, 269–270, 272–277, 281–283, 286–288, 290, 293–306, 309–311, 313–314 humor recognition 5, 269, 272 idiom 93, 298–299, 302 illocutionary force 291 image schema 181, 184, 186 imagery 36, 167, 171–173, 180–181, 185–187 implicature 6, 45, 49, 73, 169, 281–282, 288–291 implicit information 47 incongruity 37, 96–97, 112, 116, 118–119, 121–122, 128, 131, 133, 136, 201, 234, 236, 255–256, 297 indirect speech act 291–293, 298, 304 informational status 43, 53 internal scope 77, 81 irony 126, 129, 201, 223, 225 isotopy 312 jab line 167, 171, 176, 180–181, 184–186 joke 3–4, 6–7, 78, 87, 89, 94–104, 106–113, 115–120, 123–124, 126–128, 130–131, 133–134, 137, 139–165, 167–190, 192–196, 224, 229–232, 234, 236–253, 255–256, 274–276, 281–283, 285–286, 288–289, 291–292, 294, 296–297, 299–304 joke construction 4, 167–169, 171, 173, 175, 177, 179, 181, 183, 185–187, 189 joke experiment 171 joke performance 98, 167, 169, 171–172, 177, 180, 182–183, 185–186, 189

joke telling 130, 169, 171, 174–175, 179, 188–189, 192–196 knowledge about humor 3, 93–95, 97, 99, 101–105, 107–111, 113 knowledge representation 16, 38, 40, 47 laughing voice 177, 183–185 laughter 94, 99–101, 111–112, 125–131, 135, 164–165, 169–170, 176–179, 181–185, 195, 203, 215, 222, 304 level of conceptual granularity 46 logical mechanism 37, 96–97, 111–112, 119, 123, 229, 255, 275 machine learning 13, 269, 274 macroscript 167–168, 186 main clause 44, 51, 53 Maxim of Iconicity 45 Maxims of Manner 45 meaning representation 1, 46, 243–245, 247–248, 250 media 103, 134, 149–150, 152, 160–161, 163, 165, 273, 276, 305, 311, 313 metaphor 3, 23, 31, 34–36, 55–57, 59–63, 65, 67–76, 153 metapragmatic 109–110, 112, 182, 187 multi-word verb 298–299, 302 name 13, 23, 47, 51, 148, 177, 180, 190, 239, 241–247, 298, 300–301 non bona fide 98 non-bona-fide mode 4, 28, 169, 171–172, 179, 182–183, 185, 187, 224 object-oriented 2, 11, 17 Ontological Semantic Theory of Humor 96, 187, 229–233, 235, 237, 239, 241, 243, 245, 247, 249, 251, 253, 255 Ontological semantics 2–3, 5, 7, 31, 37, 40, 47, 54, 57, 76, 227, 229–233, 239, 247–248, 253, 255, 257, 259–261, 263, 265, 267 ontology 2, 5, 7, 31, 47, 55–56, 59, 63–65, 68, 70, 75, 111, 230–231, 233–234, 236–237, 244–247, 250–253, 255, 267

Subject Index

operation 176, 257, 259–267 paradigm 56, 168, 281–283, 285–287, 302, 304 paratactic construction 44 phrasal stress 77–79, 81–89 phraseme 43–44 play frame 171–172 political humor 4, 139, 144, 156, 164–165 political jokes 4, 139–147, 149, 151, 153, 155–157, 159–165 practical knowledge 93, 105, 108, 110 presupposition 6, 19, 109, 281, 297–298, 302–304 presupposition trigger 298 Principle of Expressivity 45 punch line 116, 148, 167–168, 171, 176, 178–179, 181, 184, 186 redundant structures 47 rhetorical relation 44, 52 Rhetorical Structure Theory 44, 50, 54 ridicule 111, 126, 147, 152–153, 164, 218, 308 Russian National Corpus 45, 53, 89 script 1–7, 9, 11–17, 19–21, 23–31, 33–35, 37, 39–41, 43, 45–55, 59, 63, 71, 73–75, 78, 93–97, 99, 101, 109, 111, 119, 123, 134, 139, 141, 146, 149–151, 153, 156, 158, 160, 162–164, 167–169, 185–186, 192, 195–197, 229–243, 245–255, 269, 275–276, 281–282, 284, 288, 290, 293, 299, 303, 310 script opposition 6, 28, 37, 96–97, 101, 119, 123, 139, 141, 146, 151, 158, 163, 192, 195–196, 229, 243, 255, 275, 281–282, 284, 288, 293, 310 Script-Based Semantic theory 43, 46, 54, 167, 229 Script-Based Semantic Theory of Humor 229 Semantic Script Theory of Humor 4, 139, 167, 192, 281 semantic shift 77, 82, 87, 90

323

semantic valency 18, 77, 81, 90 sense of humor 4, 97–98, 101–102, 105, 107–108, 113, 115, 124–125, 127–136, 224 seriousness 99, 110, 115, 124, 126, 128, 130, 133, 137 simultaneity 43–44 situation 2, 15–16, 19, 21, 25, 27–28, 36, 43, 48, 51, 78, 80, 82, 86, 95–97, 104, 106, 108–109, 119, 127, 132, 137, 142–143, 156, 158–162, 168–170, 172, 174–179, 184–186, 189, 194, 210, 212, 215, 218, 229, 233–236, 241, 251, 253, 259, 273, 287–288, 290–291, 295, 299–300, 306 social behavioral biometrics 275, 277 social biometrics 269, 271–274 social network 269, 271–273, 275, 277 sociology 93, 95, 112, 164, 168, 187 speech act 54, 191, 195–197, 281–282, 291–294, 298, 301–305 storytelling 191–192, 197–200, 202, 204, 207–209, 211–216, 218–225 structural ambiguity 7, 281–287, 304 subordinate clause 43–44, 50–52 taxis 43–45, 47–49, 51–53, 89, 143, 151, 153, 308 taxonomy 41, 118–120, 135, 281–282, 284, 287–288, 295, 298, 302 temporal localization 50–52 temporal relation 43, 45, 47 text generation 46–47, 53 text understanding 46, 53 time interval 43–49, 53 tourism 6, 305–314 Tunisia 305–314 user identification 5, 269 well-formedness 53 wordplay 281–283, 285–287, 302 yes/no question 292, 298, 301