Linguistic Theories of Humor

1

Of course, the idea of linear organisation, both of the disambiguation process and (presumably) the process of text generation, is not new, it was part of the "functional sentence perspective" and the concept of "communicative dynamics" as developed by the Prague school of linguistics.

The author finally touches on the problem of the quality of jokes and puns (two kinds of humorous language use which the author does not always succeed to keep apart): "the best puns are those in which either the two senses coexist in a difficult balance, or in which the connotating sense brings a meaningful contribution to the global senses of the text", i.e. puns should possibly transport a "hidden meaning".

Chapter 3 is an attempt at an analysis of puns. First of all it is important to keep in mind that not any ambiguous word or utterance can be regarded as a pun. Furthermore, puns seem to be or incite errors, but their effects are not erroneous but humorous.

The author also makes a point of the fact that puns could also occur on semiotic levels different from linguistic ones, e.g. visually, esp. if different letters are in use, not to mention jocular punning by means of Chinese characters. The central and fundamental phenomenon upon which puns operate is similarity, 2 more often than not phonemic similarity is concerned. Here, it is important to observe that certain features, or rather positions are essential and crucial (bathtub effect), i.e. initial and final positions are prominent in the triggering of certain effects (this gives a plausible answer to the question why there are such stylistic devices as rhyme and alliteration. 3M oreover, puns tend to be more similar to their targets than "normal" malaproprisms! The rule seems to be: marked structures oust unmarked ones. Errors in puns are fake (deliberate) errors, this is manifest from the fact that the confusion of certain phonemically different strings (connection vs. collection) would never occur in normal communication.

Because punning speech is not casual and eroneous, but planned and deliberate, it is an example of exceptional language, concentrating on marked structures. One of its favourite schémas is antithetic/chiastic (so is alliteration and rhyming, by the way):

(28) Phonological syntagmatic reversal: You've had tee many martoonis. (Milner 1972:18) (31) Lexical syntagmatic reversal:

(Definition of'hangover'): The wrath offrapes. (Ibid.)

All this might be regarded as proving that the speaker is aware of his language system, including insight into such processes as syllabification, "even into the nature of reasoning itself (p. 141). Thus it appears that puns articulate a certain metalinguistic goal, though this Coates (1987) proposes "nearness" instead of "similarity". Rhyme insinuates -by way of phonemic similarity -semantic similarity and distance simultaneously (p. 160). The "startling" effect of rhyme originates in the realisation that similar sounds do not correspond to similar senses ("note that a word cannot rhyme with itself!", p. 161), the rhyme effect is stronger the more semantically distant the two rhyming words are. So-called clang responses prove the "existence of soundbased associations in the mental lexicon" (p. 162); the same holds true for word retrieval strategies. may be implicit and follow a special kind of logic, since the rules of the normal world and communication are suspended, a non-bona-fide (NBF) attitude holds instead. Consequently, the rules of a playful paleo-logic apply. There is really nothing extraordinary or absurd about this exceptional use of language: Something very similar occurs in poetry for the sake of aesthetic effects. Humor causes a defunctionalization of language, i.e. for play instead of communication. The philosophy that lurks behind the playful exploitation of similarities in semiotic devices is a Cratylistic language theory, i.e. the speakers "do indeed have a motivated, non-arbitrary view of the linguistic sign, (in which homonyms could be treated as synonyms)" (p. 149).

Chapter 4: Resolutions in puns.

Puns and jokes have to be understood in order to produce their humorous effect. This involves a heuristic process, or rather some decoding effort on the part of the recipient/listener/reader, in fact, because of the co-presence of two senses in opposition, the text would have to be reconsidered, re-passed, i.e. the recipient is confronted with the task of back-tracking, he has to search for inferential clues in order "to get the joke". 4 For this job of disambiguation certain disjunctive elements should prove useful, although there is no general structural feature that would help to identify these elements.

Chapter 5 on Semiotic and text theories discusses Koestler's bisociation idea and Schmidt's so-called "Text-Theorie", relying on a complex of presuppositions in which texts like jokes figure as examples of the well-known incongruity theory in pragmatic terms: the joke teller is contradicted by the deictic field (Eco would call this the violation of a frame). Wenzel's concept of frame change is mentioned which is the cause for the listener's decoding effort (see above) and reminds one of Barthe's understanding of the narrative as a puzzle. Other authors, e.g. Nash and Redfern, are mentioned.

Chapter 6 deals with a neighbouring aspect, viz. script-based theories. Raskin has struck upon a widely known but extremely underestimated fact: the (native) speaker's competence for humorous texts: he/she normally knows whether a text is meant to be humorous, i.e. the recipient is usually aware of its perlocutionary goal and effect (notabene: this applies to the competence, not the performance!).

Raskin's concept of the script-switch trigger is introduced along with the all-important idea of -beyond-Gricean -non-bona-fide maxims which are a peculiarity of humorous, (possibly also of ficticious) texts. These concepts form the central parts and ideas of Raskin's Semantic Script Theory of Humor (SSTH). An attempt at an expansion of SSTH is GTVH= General Theory of Verbal Humor.

A well-known joke-tellers' experience is the mis-or non-understanding of a joke or pun because of various reasons, although the narrative structure presented may have been impeccable.

Such an extended theory needs various KRs (Knowledge Resources); a gamut of abbreviations for well-known ingredients of definitions for situational approaches to discourse analysis follows (cp. continental text theories like Schmidt's or Steger's):

LA • language; NS = narrative strategy; SI = situation; LM • logical mechanism; SO = script opposition; TA = target. According to Attardo the GTVH is a 6-ruple (there might easily and obviously be more than 6), these are not binary values although they seem to operate in a hierarchical order: LA, SI, NS, TA, SO, LM.

Chapter 7 deals with register-based humor.

Since jokes and puns (the main material of the present study) are texts they as a matter of course are guided in their generation by "a set of choices" (this is how Halliday describes the term "register"). 5 The guiding principle of register choice is stylistic, the primary function of stylistic decisions in the formulating of a text is the activation of affects in the listener/reader (the author mentions Bally who distinguishes two kinds of affects: natural and evocative. Obviously, textually activated affects are evocative). Much more productive in this connection is the term and idea of formality, i.e. the recipient-oriented choice of the communicative means in a given interaction. We should not be surprised to find that humorous texts and humor in general are associated with informality.

Still, an overall theory of register beyond impressionistic accounts seems impossible, even more so since essentialistic approaches yield just yes-no decisions and solutions.

Attardo decides to try a different, i.e. a polythetic approach, grounded on family resemblances (cp. Wittgenstein and Lakoff 1987) and inspired by a semantic theory according to which prototypes form the centre for elements arranged around it in various combinations: there is no longer an essential crucial feature which would define a class by necessary and sufficient conditions. Instead, the distance in which these clusters of semantically related elements around a prototype are located, the "length" of the "links" which relate the various "family"-members, becomes a crucial value.

Thus, the sum of the connotations (script elements) activated and the relationship between them will represent a certain register.

Chapter 8 is an attempt to apply the findings and assumptions of the GTVH to humorous texts longer than the normal length of jokes (there presumably are other differences than just length!). The author tries to show why such stories provide humorous effects. Here are some of his observations: E. A.Poe's short story "The System of Dr. Tarr and Professor Fethers" derives its humorous effect from the fact that essential parts of information are being withheld: "it is impe-Others call it by different names: Crystal/Davy (1969) use "province"; Fishman sees it as "domain" etc.

rative that something in the text never be mentioned" (p.260). We all know that jokes are marred if the teller tries to explain the punch line.

A passage from L.T.Peacock's "Headlong Hall" (1815) is noteworthy for its incongruity between the trivial events and the linguistic register (latinate, flowery, formal = high brow) in which they are related.

Still another passage, this time from Voltaire's "Candide" (note the differences in the times of publication!), uses the register of highly philosophical discourse to describe sexual intercourse (besides satirizing Leibnizian philosophy). One should keep in mind that for the description of sexual subject matter normally only two registers are available: medical, or obscene.

The last item is an example of the so-called menu-propos, a medieval and Renaissance kind of nonsense texts that follow only one rule: the first line must rhyme with the last line of the previous character's utterance: the incoherence of such nonsense texfc" violate both semantic rules and the principles of SSTH" (p.269). The scripts do not really overlap, they are practically non-contiguous. The author is here confronted with the question how to explain the obvious opposition between two different scripts (e.g. MAD vs. SANE in Poe's story) and the disambiguation of their discrepancies without the use of a Raskian switch trigger. He presumes it could be the cumulation of allusions which actualize the second script.

The really -pragmalinguistically -interesting issues are discussed at the end: in Chapter 9 which deals with the Cooperative Nature of Humor and in Chapter 10.

The fact that joking is much used in everyday communicative interaction, despite its seeming non-cooperativeness (one should remember that Grice himself considered irony as an example of implicature that one could not possibly neglect when dealing with normal discourse). Moreover, jocular texts are not perceived as lies or ill-formed, cryptical speech, i.e. they are a socially accepted linguistic activity. This is why humorous texts must be approached from a NBF (non-bona-fide) mode of mind as Raskin proposes.

Moreover, Attardo assumes a hierarchy of maxims, since all jokes violate the maxim of relevance, apart from violating others as well.

Chapter 10: Humor in Context. There are two kinds of jokes that are structurally not different but vary in their situational functions: canned jokes, and conversational jokes. The latter are not easily transferred onto another situation ("you had to be there"), whereas canned jokes can be completely decontextualized, they frequently link on to previous jokes. Besides the element of surprise essential for the joking effect, the teller (rule: just one speaker!) frequently uses forewarning (irony is similar as to the NBF-mode but has no forewarning) and disclaimers, i.e. the teller appeals to the NBF-modes of his listeners: they are supposed to switch to a paratelic (playful) turn of interest, laughter being implemented as the sign of having understood (note, that laughter as a consequence of joke-telling is an exception to the non-overlapping rule). Laughter can be spontaneous, delayed or even absent (the teller has to remain silent, pretending he believes in what he has told).

Because of its "denying any harmful intention for an action" (Kane 1977: 14f.) the different kinds of humorous communication facilitate in-group interaction and strengthen ingroup bondage as well as out-group rejection.

Humor is similar to irony not only in that it is deniable, but also by the prominent feature of exaggeration, which usually serves as an irony signal.

In social gatherings humorous contributions serve to make oneself present and memorable, moreover humor connotes and presupposes familiarity and sociability by negotiating the emotional/affective aspects of a given situation.

There are variations of what counts as humorous: e.g. there is ritual joking and there is-, of course teasing with its element of criticism (Bergson was one of the authors who advocated humor as a means of social correction). Still, even aggressive teasing is retractable to some extent, provided it is exaggerated.

All this works because humor is not bound to the maxim of quality. Generally speaking, humor does not involve as much status investment as other kinds of utterances, this is because of its paratelic orientation similar to children's playing with words, in other words: it is a pleasure rather than a commitment.

There is a last chapter (11) which tries to give an outlook on humor research in the future.

The idea that the investigation into paronymie puns could provide an insight into "phonemic distance" seems intriguing enough, it should have been under way, since immemorial times.

One other issue, viz. how jokes are "born" and popularized and recycled is a textlinguistic problem of long standing, that has never really been tackled. A list of unresolved questions remains -no wonder, since we are concerned with something that is present in everyday communication of all of us (except Sir Popper (cp. p.319!)).

The author's adherence to his master's (Raskin's) credos is obvious if not virulent. Printing errors are not easily avoided, some of them can propel a text passage into unintelligibility: p. 184 talks of "the 'use' of humor by the speaker either for social critjgjsm or for the (!) of taboo instincts ... " (cp. p.253, 203, 268).

Abbreviations can be a serious obstacle to smooth reading.