Metaphors and Jokes in the Fragments of Cratinus
What’s the difference between a metaphor and a joke? At first this might seem a strange question. Indeed, the differences are so obvious, and the similarities so much less so, that the question itself sounds like the opening to a joke in the “What’s the difference between an elephant and a blueberry?” style.1 However, in this article, I am going to suggest that the line between these two categories is far from clear-cut. Using examples from the comic poet Cratinus, I will show firstly that the same fundamental mechanism underlies jokes and metaphors, and secondly suggest that Cratinus’s metaphorical language in fact actively exploits the slippage between these two modes of speech to create humour and expose the inherent absurdities of literary language and convention.
An understanding of the potentially ambiguous line between jokes and metaphors, and the extent to which these two modes of speech may not be entirely distinct categories, provides us with a particularly useful insight into the metaphorical language of Old Comedy. Comedy is hugely rich in metaphorical language,2 and its metaphors range from the large-scale [End Page 231] and structural (for example, the polis-as-oikos metaphor which underpins much of the plot of Aristophanes’ Knights),3 to the frequent short metaphors which colour its poetry. Given that the genre is also, needless to say, rich in humour, a model of metaphor which shows its close likeness to jokes and puns may go some way towards explaining the almost excessive fondness which comedy has for metaphorical language.
Old Comedy also seems to be a good place to start when looking to understand how literary (as opposed to everyday) language works. Comedy often operates at the fault lines of language, exposing the impossibilities which, despite their centrality to the function of language (and especially literary language), might otherwise go undetected. Comedy in general, and (as I argue) Cratinus in particular, also has a tendency towards amplification. We therefore find in comedy literary phenomena such as metaphors in a somewhat exaggerated form, and this very lack of subtlety makes for useful case studies. I hope therefore to show not only that theoretical models of metaphor (and especially the model based on joke theory which I shall propose) have much to offer in our understanding of Old Comedy, but also that comic language in its very comicness has much to offer to our theoretical understanding of literary metaphor.
METAPHORS AND JOKE THEORY
The contention that metaphors and jokes are underpinned by the same fundamental mechanism rests on the observation that, at their core, both involve taking two apparently disparate ideas and mapping them onto one another. In metaphors, this mapping takes the form of the comparison drawn between the vehicle and the tenor (or source and target), whereby one is understood in terms of the other. From the 1990s onward, there has been interest in understanding this relationship between vehicle and tenor from the perspective of cognitive linguistics. This cognitive approach takes as its starting point the idea that metaphorical comparisons involve the mapping not only of linguistic categories, but also of underlying conceptual [End Page 232] domains. Cognitive approaches stress the embeddedness of metaphorical language in everyday speech. For example, common idiomatic expressions such as “the foundations of a theory” or “constructing an argument” are understood to be the linguistic manifestations of an underlying conceptual metaphor THEORIES ARE BUILDINGS.4
Of these cognitive models, the most dominant has been Gilles Fauconnier and Mark Turner’s theory of “conceptual blending,”5 which they set out in their 2002 book: The Way We Think: Conceptual Blending and the Mind’s Hidden Complexities. According to Fauconnier and Turner, metaphorical language not only involves the two inputs of vehicle and tenor, it also creates a third mental space in which these two inputs are blended together in an impossible way. The most famous and often cited of Fauconnier and Turner’s examples involves an effort in 1993 to break a longstanding sailing record set in 1853. Newspaper reportage described the 1993 ship as being “4 days ahead” of its 1853 predecessor. According to Fauconnier and Turner (1998), the use of this language summons up an impossible blended conceptual domain in which the two ships are imagined as co-existing in the same time and space, although the 1993 ship simply followed the same course as the first and could not in any real or literal sense be “ahead” of a ship which set sail some 140 years before it. Fauconnier and Turner are keen to stress the ease with which these impossible blended spaces are summoned up, and their characterisation of metaphor as a process whereby apparently incompatible domains may be seamlessly blended together is consistent with the majority of cognitive approaches to metaphor in presenting the cognition of metaphorical language as a fundamentally frictionless process, essential to ordinary language. This seamlessness is arguably a function of an approach which takes everyday metaphorical language (sometimes termed “dead metaphors”) as its starting point, with literary metaphor something of an afterthought.
Within the field of cognitive linguistics (and indeed beyond), the work of George Lakoff has been very influential; the title of Lakoff and Mark Johnson’s seminal 1980 work, Metaphors We Live By, is indicative [End Page 233] of the degree to which ordinary language use forms the basis of Lakoff’s studies. Although Lakoff did go on to write about literary metaphor (Lakoff and Turner 1989), the results of this work are mixed. Throughout their study, Lakoff and Turner argue that metaphor takes place on two levels, the conceptual and the linguistic, and that before there can be any discussion of language, the underlying conceptual structures of a text must first be understood. This analysis significantly underplays the interaction between the linguistic and the conceptual, and the degree to which what Lakoff and Turner are keen to designate as “mere language” in fact shapes thought.6 Moreover, there is a tendency towards a rather domesticating approach to poetic language which emphasises shared underlying structures at the expense of differences in poetic language and style.7 Lakoff and Turner often de-emphasise the sheer strangeness of the poetic language under discussion and conceive of difficulties of language or expression as “questions to be answered” or problems to be solved.8
In contrast to cognitive theories of metaphor that argue that the impossibilities of metaphorical language are de-emphasised in the process of cognition, joke theorists are specifically interested in the impossibilities and disjunctures generated when two domains are mapped onto one another. The idea that jokes, like metaphors, are structured by mapping two different and apparently incompatible frames of reference onto one another is central to the work of many joke theorists, most notably Victor Raskin, whose work applies script-based semantic theories to verbal humour. Raskin’s core argument is that (1985.99): “A text can be characterized as a single-joke-carrying text if both of the conditions are satisfied: i) The text is compatible, fully or in part, with two different scripts; ii) The two scripts with which the text is compatible are opposite,” where a [End Page 234] “script” (essentially analogous to the cognitive terminology of a “domain”) is defined as a “large chunk of semantic information surrounding the word or evoked by it. The script is 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).
While it remains influential, Raskin’s theory of opposition has been challenged by later scholarship, including a piece by Raskin himself (Attardo and Raskin 1991), in which his earlier model is critiqued as excessively binary, and it is suggested that jokes rely not on a single opposition between scripts, but a hierarchy of interrelated inputs of script, logic, and language. Studies of absurd humour by Jerry Palmer (1987)9 and Alexander Brock (2004) further challenge theories of the joke which rely on models of binary opposition, and suggest that we should instead see “constellations” (Brock 2004.359) of multiple oppositions working simultaneously alongside one another.10 Finally, Raskin’s language of opposition has itself been questioned, most notably by Eliot Oring, on the grounds that the divergent scripts within a joke are not so much opposites in any real sense but can be better understood as simply “incongruous”;11 and Oring’s theory of “appropriate incongruity,” which holds that, in a joke, two incongruent ideas are mapped onto each other through some (often spurious) linking word or [End Page 235] idea, such as a double meaning that has significance in both of the joke’s two scripts, accordingly refines Raskin’s earlier ideas. In Oring’s model, the joke: “Q: Why should you wear a watch in the desert? A: Because a watch has springs,” is based on an incongruity between the scripts DESERT and WATCH, with the double meaning of “spring” as both a body of water and a mechanical device providing the linking mechanism through which the two scripts are made temporarily “appropriate” to one another.12
Oring is one of a number of theorists, ancient and modern, who recognise the central role of incongruity in both jokes and metaphors. Aristotle in Rhetoric 3.11 designates jokes (τὰ ἀστεῖα) as a kind of subset of metaphor, and suggests that both modes of speech rely on an element of surprise whereby the listener’s expectations are subverted. Aristotle’s remarks (as well as Cicero’s discussion of jokes in de Oratore 2.52.240)13 appear to have influenced Freud, who likewise comments that metaphors and jokes are both forms of indirect representation (Freud 2002.76).14 In Oring’s concise formulation, both metaphors and jokes involve mapping two “clashing conceptual categories,”15 and metaphors, like jokes, involve a degree of absurdity due to their suggestion that one thing is not only like another, but actually is something else—“a logically absurd proposition,” in Oring’s terms.16
However, while the observation that these two categories of speech share the same fundamental structure is surely correct, Oring’s claim that, [End Page 236] in a metaphor, the connections between conceptual categories are not fundamentally spurious, as they are in jokes, is overly schematic.17 In particular, in common with “blending” theory, Oring’s discussion does not sufficiently recognise the defamiliarising aspects of literary metaphor, whereby poetic language makes us see connections between categories which are unusual and allows us to see each category in a new and unfamiliar way.18 To take an example in English, Auden’s line: “I’ll love you till the ocean is folded and hung up to dry”19 makes a comparison between the ocean and laundry on the basis of their shared quality of wetness. This link is surely no less spurious or ridiculous than those found in jokes, and one could easily imagine this metaphor reformulated as a joke (“Q: How do you turn an ocean into a desert? A: Hang it out to dry.” I appreciate that my effort is not a particularly good joke, but it is recognisably a joke nonetheless). This suggests that the line between a joke and a metaphor cannot be exclusively based on the level of spuriousness of the comparison between the two fields of reference (for in both Auden’s metaphor and the joke above, the comparison between oceans and laundry is the same), but must take into account the manner of presentation. Specifically, the difference between metaphors and jokes seems to me to lie not in the spuriousness of the connection per se, but rather the extent to which the formulation foregrounds this spuriousness and the resulting absurdity and impossibility.
It is worth noting before proceeding that joke theorists such as Raskin and Oring tend to use examples from fairly limited subsets of humorous language, showing a preference for examples taken not from literature or film or even everyday conversation, but instead from what we might [End Page 237] term “stand-alone” jokes which can be told independently of any specific context. In the modern tradition of joke telling, such stand-alone jokes often take the form of a question and answer (“Why did the chicken cross the road?” “What’s the difference between . . . ,” etc.), or a short story (“So a man goes to the doctor suffering from insomnia . . .”), and it is jokes such as these which form the basis of, e.g., theories of appropriate incongruity. This kind of stand-alone joke is not really a feature of Old Comedy, and this inevitably raises the question of whether the joke, in the sense meant by modern joke theory, is a meaningful category with regard to Old Comedy.20
I would suggest, however, that the absence of these kinds of stand-alone jokes in Old Comedy is not due to the lack of any concept of the joke as a discrete entity or category of speech, but rather because, in the context of what is a fundamentally narrative genre, any jokes are likely to be embedded in the structure of the narrative.21 To take only one of the almost limitless examples in Aristophanes, the Acharnians contains an extended riff on the double meaning of χοῖρος as both “piglet” and a slang term for [End Page 238] the female genitalia (particularly of young girls).22 This double meaning is used as a linking device between two incongruous scripts, and the scene reads simultaneously as marketplace transaction and sexual procurement.23 The χοῖρος pun culminates in a highly obscene punchline: Dicaeopolis, not convinced by the girls’ disguise despite the Megarian’s insistence, jokes that “they might seem like piglets now, but they’ll grow up into right cunts” (Ach. 781–82: νῦν γε χοῖρος φαίνεται. / ἀτὰρ ἐκτραφείς γε κύσθος ἔσται). To repackage this joke as a stand-alone one-liner is certainly possible. However it would be considerably less funny than it is in situ, since the χοῖρος pun is only one strand of a whole nexus of jokes which make up the humour in this scene (including a further series of obscene puns concerning what the “piglets” might like to eat, the stereotype of the tricky Megarian, etc.); there is also an extended set-up that includes an element of costuming which adds another level of visual humour.24 It is precisely this punchline’s embeddedness in the overall narrative and its enmeshment in a whole series of interrelated and interdependent jokes which give it its comic force.25
While, in practice, jokes in Old Comedy are not discrete entities, this does not mean that there is no concept of the joke as a distinct category of speech. This notion of the joke as a stand-alone entity which can be wheeled out over and over again is strongly suggested by the opening of Aristophanes’ Frogs. The slave Xanthias begins the play by asking Dionysus whether he should say “one of the usual things which make the audience laugh” (Εἴπω τι τῶν εἰωθότων ὦ δέσποτα, / ἐφ᾽ οἷς ἀεὶ γελῶσιν οἱ θεώμενοι;). When Dionysus insists that he should avoid any vulgar puns on πιέζομαι or θλίβομαι,26 Xanthias asks if he should tell ἕτερον ἀστεῖόν τι [End Page 239] (“some other witty thing,” Ran. 5) or τὸ πάνυ γέλοιον (“something really funny,” Ran. 6), with ἀστεῖόν and γέλοιον seeming to correspond closely to a joke in the modern sense. (It is worth noting that τὰ ἀστεῖα is used by Aristotle in Rhetoric 3.11 specifically to designate puns.) Even in the absence of any kind of set joke-telling idiom (most notably, the question-answer format) which dominates the modern tradition, this sort of meta-joke—a joke about the conventions of joke telling27—suggests a genre not only alert to the joke as a distinct and established category of speech, but one which treats this category as a topic of interest.
Cratinus’s oeuvre, or what little is left of it at least, suggests a poet similarly interested in the conventions of comic poetics; and his Pytine, which seems to have dramatised the playwright’s own struggle with writer’s block, is arguably among the most overtly metapoetic works of antiquity. The poet’s interest in, and exposure of, poetic convention is exemplified in his approach to metaphors and by the way he exploits the ambiguous space between the metaphor and the joke. Cratinus’s metaphors—particularly those metaphors which describe poetic speech—frequently utilise stock comparisons in such a way as to amplify their inherent impossibilities and absurdities, generating humour which often comes at the expense of the conventions of serious poetry. With this mockery, Cratinus provides us with a how-to guide, laying bare exactly how such serious poetic language works.
METAPHOR JOKES IN CRATINUS
One of the stock metaphors which makes relatively frequent appearances in the fragments of Cratinus involves the comparison between speech, particularly poetic speech or song, and liquids. Such a comparison is well embedded in Greek poetic idiom, in which words are commonly described as “flowing,” as, for example, in this passage from the Iliad (1.247–49):
τοῖσι δὲ Νέστωρἡδυεπὴς ἀνόρουσε λιγὺς Πυλίων ἀγορητής,τοῦ καὶ ἀπὸ γλώσσης μέλιτος γλυκίων ῥέεν αὐδή. [End Page 240]
And among them NestorSweet of speech leapt up, the clear-voiced speaker of the Pylians,And from his tongue honey-sweet speech flowed.28
This conventional metaphor of speech (or poetry) as liquid is especially natural in drama, given Dionysus’s role as the god of both wine and the theatre. The relationship between poetry and wine is, in fact, the subject of Cratinus’s Pytine, in which the poet presented himself on stage as a character torn between the charms of Drunkenness and Comedy (personified as his mistress and wife, respectively),29 and which seems to have ultimately asserted the essential connection between wine and poetic production.30 Given the plot of the Pytine, it is unsurprising that one substantial fragment from the play features an extended metaphor in which speech is described as having the qualities of a liquid (Pytine frag. 198):
ἄναξ Ἄπολλον, τῶν ἐπῶν τῶν ῥευμάτων.καναχοῦσι πηγαί· δωδεκάκρουνον <τὸ> στόμα,Ἰλισὸς ἐν τῇ φάρυγι· τί ἂν εἴποιμ᾿ <ἔτι>;εἰ μὴ γὰρ ἐπιβύσει τις αὐτοῦ τὸ στόμα,ἅπαντα ταῦτα κατακλύσει ποιήμασιν.
Lord Apollo, what floods of words!Splashing streams, a twelve-springed mouth,An Ilissus in his gullet. What can I say?If someone won’t stop up his gob,He’ll flood everything here with his poetry.
The first thing to note is that, in contrast to the Homeric passage, the metaphor of speech-as-liquid is far more sustained and not simply contained in a passing word or phrase (cf. ῥέεν αὐδή). Descriptions are piled up on top of one another, with the two fields of reference repeatedly [End Page 241] juxtaposed (ἐπῶν . . . ῥευμάτων, δωδεκάκρουνον . . . στόμα, ἐπιβύσει . . . στόμα, and κατακλύσει ποιήμασιν), so that this excess of description itself takes on an absurd quality.31 The words-as-liquid metaphor here turns on the double meaning of στόμα in line 2, which (like the English “mouth”) may refer to either the mouth of a person or animal, or metaphorically to the mouth of a river or stream. As in a joke, this double meaning is used to bridge the gap between two disparate fields of reference.32
Importantly, the passage makes no attempt to smooth over this disparity, but instead amplifies the collision between the abstract (speech) and the concrete (water). Speech is not just described verbally or adjectivally as “flowing,” but is embodied in a series of nouns as “floods” (1: ῥευμάτων), “streams” (2: πηγαί), and even as one particular river, the Ilissus in Attica (3: Ἰλισός).33 This final proper noun caps the previous two plurals with its greater exactness, increasing yet further the gap which has opened up between the abstract (speech) and concrete (a specific river which exists in a specific geographical space). This embodiment of speech as a physical body of water builds up to a final reductio ad absurdum in the last line of the fragment, where speech takes on the destructive powers of water, flooding everything around it (ἅπαντα ταῦτα κατακλύσει). This image of something so abstract as speech spilling into, and filling up, the physical space around the speaker pushes the impossible and absurd qualities of the speech-as-liquid metaphor to the forefront, and the two fields of reference collide much as they would in a joke.
An additional element of humour is perhaps opened up by the prayer to Apollo which begins this fragment. In the Homeric Hymn to Apollo, the god is warned that the spring at Telphusa may make the location [End Page 242] unsuitable for a temple due to the noise of passing horses and mules being watered (Hom. Hymn Ap. 261–63). The god responds by blocking up the spring with stones (Hom. Hymn Ap. 379–87) and building an altar there anyway. The prayer to Apollo therefore adds further emphasis to the collision between the concrete and abstract in this passage, as the speaker asks a god known in the mythic tradition for stopping up a literal noisy spring to here stop up a metaphorical one.
The potential absurdities of the speech-as-liquid metaphor are also highlighted in a short fragment from the Didaskaliai (frag. 38): ὅτε σὺ τοὺς καλοὺς θριάμβους ἀναρύτουσ᾿ ἀπηχθάνου, “When you were hated for drawing up from a well beautiful thriamboi” (i.e., songs for Dionysus). The metaphorical language is in some ways appropriate and conventional, with the poetry-as-liquid metaphor again fitting the apparently Dionysiac context.34 However, as in the previous fragment, the clash between the abstract and concrete is increased by an almost ridiculous level of specificity, here created through the rare verb ἀναρύτω.35 The more common form would be ἀρύω, meaning “to draw water” (itself an oddly specific action in this abstract context), and the addition of the prefix ἀν- (“up”) locates the action even more firmly in physical space, thereby increasing the clash between the concreteness of this action and the abstract notion of “gaining inspiration for song” for which it seems to stand. Further, by insisting that the water is “drawn up,” the verb inserts the rather everyday image of a well (since it is only from a well that water can be drawn specifically upwards), whose insistent physicality as an object sits rather uncomfortably in this otherwise abstract phrase. This discomfort further emphasises the contrast between the “high” act of poetic composition, and the “low,” everyday, action of drawing water to which it is compared. The multiple collisions in this line between abstract and concrete, spatial and non-spatial, banal and poetic, seem to flaunt the ultimate spuriousness of the conventional poetry-as-liquid metaphor and the result is playful and silly—funny, even.
The inherent peculiarities of the standard metaphorical language used to discuss poetry are similarly foregrounded in a fragment of the Archilochoi, which strays even further into the territory of jokes. The [End Page 243] fragment appears to come from the play’s agon, in which the iambic poet Archilochus faced off against Homer, or perhaps Homer and Hesiod, and in which Archilochus was most probably declared the victor.36 The fragment begins with the speaker(s)—most likely the chorus37—apparently passing judgement on Archilochus’s performance in this agon: εἶδες τὴν Θασίαν ἅλμην οἷ᾿ ἄττα βαΰζει; “Did you see this Thasian pickle-juice, what sort of thing he’s barking?” (Archilochoi frag. 6.1). Hiding in this rather odd line is a fairly conventional metaphor whereby invective or unpleasant language is referred to as “salty”38 or “vinegary.”39 However, once again it is not an adjective that is used, but rather a substantive with a highly specific semantic range, namely ἅλμη, meaning “brine,” and used most frequently in comedy in relation to food preparation and pickling.40 To add to the obscurity, it is not the poetry, but the poet himself who is referred to as being ἅλμη, an attribution which is jarring given the noun’s feminine gender (τὴν . . . ἅλμην).
The obscurity is further compounded by the fact that this metaphor of Archilochus-as-pickle-juice is followed almost instantly by a second metaphor in which the poet is described as “barking” (βαΰζει; cf. the noise that dogs make in Greek, βαὺ βαύ). Again, this metaphor of human-speech-as-dog-bark is fairly conventional (and a representation of human speech as dog barks is central to the trial scene in Aristophanes’ Wasps);41 and a description of Archilochus as barking would not be unusual. However, the combination of these two metaphors results in the extremely peculiar image of barking pickle juice, as Archilochus becomes both brine and dog simultaneously. The accumulation of incompatible images results in an exaggerated sense of impossibility, as two altogether different fields of reference are mapped onto one another. The link is entirely spurious: the fragment does not seem to propose any real shared characteristics or [End Page 244] structural similarities between dogs and pickle juice themselves; rather, the mapping is created on the level of their linguistic usage, since both are implicated in conventional metaphors concerning harsh speech. This double metaphor is, in its structure, almost indistinguishable from word-play jokes which use abstract coincidences of language (such as double meanings)42 to map ideas whose similarities exist only on the level of their signifiers, but which, at the level of the signified, are entirely unconnected.
Such double metaphors, in fact, seem to be characteristic of Cratinus. Two further fragments also demonstrate how the poet magnifies the incongruities inherent in metaphorical language by collapsing multiple fields of reference within a single short phrase. For example, a fragment from the Drapetides uses this device in the description of libation cups: δέχεσθε φιάλας τάσδε βαλανειομφάλους, “Receive these libation cups with acorn/bath-shaped belly buttons” (frag. 54).
This compound word is rather ambiguous, since its first section could derive from either βαλανεῖον (“bath”; LSJ ad loc.) or βάλανος (“acorn”).43 Acorn might perhaps make more sense in the compound word itself, since it is easier to see a comparison (in shape at least) between a navel and an acorn—as opposed to navel and bath. However, in the context of libation cups, baths are not wholly irrelevant, since both baths and cups are containers of liquid (and the resultant implied comparison between libation-wine and bathwater is rather gloriously grotesque). This ambiguous word and its jumble of potential acorn/bath/navel imagery add to the absurd impossibility of the metaphor, as the cups take on multiple characteristics: human bodies with navels, whilst also sharing features with acorns (in shape) and/or baths (in function). This impossible, almost nonsensical,44 accumulation of metaphorical imagery is further augmented by the possible sexual meaning of “acorn,” which was an apparently standard metaphorical term for the glans penis as used not only by comedians (e.g., Ar. Lys. 413), but also by technical writers.45 [End Page 245]
Fragment 314 similarly accumulates multiple metaphorical images with a few short words: ἔχων τὸ πρόσωπον καρίδος μασθλητίνης, “He has the face of a shrimp like a leather thong.”46 Again we have a double metaphor: someone is described as having a face like a shrimp which is then likened to a leather thong in its appearance. The collision between three separate frames of reference, and the resultant excess of imagery, makes the language difficult to process, and this difficulty is compounded by the presence of an apparently new coinage, unattested elsewhere, μασθλήτινος, which itself uses a relatively obscure word for leather (μάσθλης, as opposed to more commonly attested words such as, e.g., σκῦτος and its many compounds) as its basis.
In common with fragment 54, there may also be a fourth, sexual field of reference adding to the multiplicities of the imagery. A recent article by Carl Shaw persuasively argues that the sexual overtones of fish imagery in Old Comedy have been under appreciated, and that crustaceans in particular, on account of their shape and colour, were a rich source of phallic euphemisms.47 While “shrimp” by itself is not enough to indicate the presence of a sexual euphemism here, its combination with the adjective μασθλήτινος does perhaps suggest a reference to the comic phallus, which was stitched from red leather and which is even referred to by Aristophanes as “the leather thing” (Ar. Nub. 538–39: σκυτίον καθειμένον ἐρυθρὸν ἐξ ἄκρου παχύ). If καρίδος μασθλητίνης can be read as a reference to the comic phallus, then this fragment would be another example of joke-like double-meanings being used in tandem with metaphor to create an excess of imagery. As in fragment 54, this excess compounds the inherent absurdities of metaphorical language by insisting not only that “one thing . . . is something other than itself,”48 but that it is, in fact, three or four different things all at once.
At times, the double meanings in Cratinus’s metaphors stray even further into the territory of jokes, as in a fragment from the Nomoi. This play seems to have contained some elements of the golden-age theme which makes [End Page 246] frequent appearances in Old Comedy,49 and fragment 136’s combination of food and sex would be consistent with this: τυρῷ καὶ μίνθῃ παραλεξάμενος καὶ ἐλαίῳ, “After lying with cheese and mint and oil” (Nomoi frag. 136). The metaphorical convergence between sex (παραλεξάμενος) and eating (τυρῷ καὶ μίνθῃ . . . καὶ ἐλαίῳ) rests on a pun, since Tyro and Minthe were also the names of heroines who slept with gods.50 Indeed, this line could easily be reformulated as an (albeit fairly terrible), joke (Q: What do Greek gods do when they’re hungry? etc.).51 The pun is made funnier still by the inclusion of ἐλαίῳ (“oil”) right at the very end of the line, because this word, unlike τυρῷ and μίνθῃ, does not have a double meaning as a name, and therefore it confronts us more fully with the impossibility of having sex with foodstuffs.52 There may, of course, be a second kind of double meaning at play in this absurd turn; while the datives τυρῷ and μίνθῃ must clearly be governed by παραλεξάμενος, ἐλαίῳ might, given the association between oils and sex,53 be read as a dative of instrument. Either way, the placement of ἐλαίῳ in many ways acts like a punchline, and it is difficult to know whether to classify the line as a joke or a metaphor (although the use of παραλεξάμενος is certainly metaphorical—or at least euphemistic), since it rather straddles the divide between the two.
METAPHORS AND JOKES IN COMEDY AND BEYOND
As I hope these examples have shown, Cratinus’s metaphors often seem not only to sit in the ambiguous space between metaphor and joke, but to actively exploit the potential slippage between the two. Far from minimising [End Page 247] the frictions which arise when mapping two fields of reference onto one another, Cratinus foregrounds the impossibilities inherent in metaphorical language by emphasising the tension between abstract and concrete, piling up multiple fields of reference, and employing double meanings. These devices, especially when they are used in combination with unusual compound words or apparent neologisms, have the effect of making the imagery difficult to process at times and creating overt, joke-like absurdities. Such absurdities are not only a feature of Cratinus’s more unusual metaphorical comparisons (e.g., the face-shrimp-leather compound), but are also evident in his deployment of highly conventional metaphors (e.g., words-as-liquid). It would therefore seem that the level of impossibility and spuriousness in any given metaphor lies not in the aptness of the comparison per se (as Oring suggests), but rather in the way in which the comparison is presented and particularly the degree of emphasis given to the inevitable impossibilities which arise in claiming one thing is actually another. To put it simply, when Homer describes words as “flowing,” this is not a particularly funny formulation, and the impossibility of the comparison of words and liquids passes by all but the most observant listener. When Cratinus takes this same comparison and describes words not just as flowing, but as uncontrollably flooding the physical space around the speaker and drowning everyone in the process, the absurdity of the comparison becomes clear, and the metaphor becomes joke like. The absurdity was always there in Homer, but Cratinus asks us to notice it and to take our newfound appreciation of the ridiculousness of the metaphor back to our reading of more “serious” poetry.
Unlike theories of metaphor based on blending, which insist that the impossibilities and frictions of metaphorical language are always mini-mised in the process of cognition, the model of metaphor suggested in this paper foregrounds the impossibilities generated through metaphorical comparisons. Instead of treating these impossibilities as a cognitive stumbling block to be overcome, a model which approaches metaphor through the lens of joke theory allows us to appreciate impossibility as a point of literary interest in itself. Further, this approach recognises that metaphorical language sits on a spectrum from the conventional metaphors embedded in everyday speech, to the highly defamiliarising metaphorical language which we find in literary texts, and through to jokes that exploit the absurdity created when divergent fields of reference collide. A single comparison may sit at different places along this spectrum depending on how it is framed; we have seen that many serious metaphors can quite easily be reformulated as jokes in the modern, groan-inducing, question-answer format. Indeed, [End Page 248] if jokes are, in fact, only a more extreme manifestation of metaphor, this provides some explanation for the pervasiveness of metaphorical language in Old Comedy, which, as a genre, sometimes appears to employ this literary device almost to the point of excess.
Cratinus’s exploitation of the inherent impossibilities of metaphor—and the ways in which metaphorical language can slip into the funny and absurd—are not only sources of humour, they are arguably also another manifestation of comedy’s impulse towards exposure and revelation. Just as the genre exposes and ridicules the body and its functions, so, too, does it reveal the absurdities of dramatic convention through metatheatre; and by flaunting the impossibilities inherent in metaphor, the absurdities of linguistic and literary convention are similarly exposed. In understanding the close structural relationship between metaphors and jokes, we also come to appreciate yet another way in which Old Comedy lays itself bare.54
1. For a discussion of elephant jokes, see Oring 1992 chap. 2. The observation that jokes and metaphors share some similarities of structure is, of course, made by Aristotle in Rhetoric 3.11. However, while this observation is frequently discussed by joke theorists (e.g., Attardo 1994.20 and Perks 2012; Oring’s remarks on jokes and metaphors will be discussed below), neither jokes nor Aristotle’s remarks are regularly discussed at any length by theorists of metaphor. For example, in Kövecses’ 2010 introduction to metaphors with its overview of the major theoretical approaches to the topic, jokes and puns do not make a single appearance, and the humorous potential of metaphor is not discussed.
2. Newiger 1957 remains the benchmark for studies of metaphor in Old Comedy, although he largely limits himself to the role of metaphor in comic personifications. For more recent studies, see, especially, Silk 2000a chap. 3 and Ruffell 2011.60–85.
3. For an extended discussion of metaphors in the Knights, see Newiger 1957 chap. 1. Ruffell 2011 chap. 3 also analyses the play’s metaphors in detail and the relationship between these metaphors and jokes. Ruffell’s observations are discussed below. The overall plot of the Knights is often referred to as allegorical (e.g., Hubbard 1991.66: “The whole play operates as an allegory”; Slater 2002.70 describes the play’s opening as preparing the audience for an “exposition of the political allegory”; Dover 2004.239: “At least one play of Aristophanes is allegorical: Knights”); for a discussion of why this term is problematic with reference to Old Comedy, see Kidd 2014.69–71.
5. This theory of “conceptual blending” in many ways emerged from the earlier “interaction theory” of metaphor, for which see, e.g., Black 1993 and Glicksohn and Goodblatt 1993. The term “interaction” is also used by Silk 1974 in his study of metaphor in early Greek poetry. For an example of how blending theory may be successfully applied to Greek poetry, see Budelmann and LeVen 2014.
6. See, e.g., Lakoff and Turner 1989.107: “Linguistic expressions—mere sequences of words—are not metaphors in themselves. Metaphors are conceptual mappings. They are a matter of thought, not merely language” (emphasis added).
7. See, for example, chap. 1, which discusses metaphors for death in a wide range of poets including Dickinson, Tennyson, Horace, Auden, and Shakespeare. Lakoff and Turner 1989 find that these poets employ similar cognitive structures, however their analysis is often insufficiently alert to how poetic language interacts with and even shapes these structures, and the differences between these very varied poets are underplayed. For a detailed criticism of Lakoff and Turner’s study and the insufficient attention paid to language and aesthetics in their analysis, see Jackendoff and Aaron 1991.
9. Note that while many of Palmer’s observations are useful, he at times displays a disappointing blindspot where misogynistic humour is concerned. See, for example, 1987.45–46, where in his discussion of comic violence against authority figures in silent comedy, he categorises “women, and especially well dressed and beautiful women” as figures of authority over the implied male audience, thereby downplaying the way in which women both in life and in art are frequently disempowered.
10. Palmer 1987.52–53 discusses a graffito where to the words JESUS SAVES, a second author, using a different hand, added the words “and Kegan scores on the rebound!” Palmer recognises multiple levels of opposition here, not only between the two scripts (Christ vs. goalkeeper), but also between a religious tone followed by iconoclasticism, earnestness followed by triviality, etc. Brock’s work focuses on the absurd humour of Monty Python and suggests that sketches produce “unexpected script constellations” (2004.359) where multiple jokes operate simultaneously.
11. Oring 1992 chap.1 and 2003 chap. 1. Oring suggests that Raskin’s claim that, for example, the scripts LOVE and GOING TO THE DOCTOR are “opposite” is not true in any meaningful sense. While some jokes clearly do rely on an opposition between, e.g., the abstract and concrete, it is reductive to claim that only opposition may underpin jokes, and Oring is surely correct to suggest that LOVE and GOING TO THE DOCTOR are incongruous rather than opposite. Oring’s theory of incongruity is accordingly far more flexible—and can be successfully applied to a wider variety of humour—than Raskin’s narrower opposition. Raskin has, however, distanced himself from incongruity theories, see, e.g., Attardo and Raskin 1991.33.
12. Oring 2003.2–3 argues that the failure to make the link, however spurious, “appropriate” in some way leads to nonsense and not an actual joke. So, for example, if for the above joke, the answer given was “because camels live in the desert,” it would not be funny, since there has been no attempt to force a mapping of scripts through some appropriate linking device. Oring notes that small children who have developed an awareness of the question-answer joke format, but have not yet fully understood the structural rules which govern such jokes, often make unsuccessful jokes in this nonsense-riddle manner. Children’s jokes are also discussed by Freud 2002.122–23.
13. Cicero suggests that jokes can be split into two categories: those which play on words (dictu) and those which play on ideas (re). For a more detailed discussion of Cicero’s remarks and their influence on Freud, cf. Attardo 1994.26–29 and 54–55.
14. Note also that Freud’s argument that “in a joke when the use of the same or a similar word takes us from one sphere of ideas to another, remote, one . . . The re-discovery of what is familiar . . . is pleasurable” (2002.118) overlaps in interesting ways with Ricoeur’s theory of metaphor’s “power to ‘redescribe’ reality” (2003.5).
15. Oring 2003.5. Oring’s use of the word “clashing” is, I think, key, as it suggests an understanding of metaphor which diverges from those cognitive models (such as blending) which seek to de-emphasise the “clash” between the two fields of input.
17. For the opposite end of this spectrum, see Ruffell’s discussion, which although it contains many extremely useful observations on the similarities and differences between jokes and metaphors, ultimately concludes wittily in aporia: “If jokes are structurally related to metaphor, is there a clear, tenable difference between a comic and a non-comic metaphor? Is there, indeed, anything formally and structurally distinctive about comic utterances, or is humour entirely a matter of context? The answer, perhaps appropriately comic, is probably ‘both’” (2011.101).
18. Cf. n. 14 above. Compare Palmer 1987.69–70, who suggests (in terms very similar to Oring’s) that the difference between a metaphor and a joke is that, in a joke, implausibility dominates over plausibility, while in the metaphor, plausibility dominates over implausibility. As with Oring, Palmer’s analysis suggests that the balance of plausibility or implausibility rests in the comparison per se; however, his analysis is considerably more alert to the strange and defamiliarising qualities of literary metaphor.
19. This example is taken from W. H. Auden’s poem As I Walked Out One Evening, lines 13–14.
20. There is evidence for the stand-alone joke in antiquity in the form of the so-called Philogelos, a joke collection which most likely dates to the third or fourth century ad. For a discussion of the Philogelos, see Beard 2014 chap. 8. The jokes mostly take the form of a short (between two and five lines) story, and are often built around a stock character (the skolastikos, the miser, the simpleton; national stereotypes about the stupidity of Sidonians, Abderites, and men from Kyme also abound). Puns and double meanings occur frequently. For example, number 172 in the collection relies on the double meaning of myrmex as “ant” and “boxing glove”: Κυμαῖος πύκτην ἰδὼν πολλὰ τραύματα ἔχοντα ἡρώτα πόθεν ἔχει ταῦτα. τοῦ δὲ εἰπόντος Ἐκ τοῦ μύρμηκος, ἔφη· Διὰ τί γὰρ χαμαὶ κοιμᾶι; (“A man from Kyme, seeing a boxer with many injuries, asked where he got them from. ‘From the myrmex,’ he replied. ‘Well what did you sleep on the ground for?’”). A similarly terrible pun on the word μέχρι(ς), which can be used in reference to both time and space, forms the basis of joke number 99 in the collection: Σκολαστικῶι τις λέγει· Χρῆσόν μοι βίρρον μέχρις ἀγροῦ. ὁ δὲ Μέχρι σφυροῦ, εἶπεν, ἔχω· μέχρι δὲ ἀγροῦ οὐκ ἔχω (“Someone asks a skolastikos, ‘Lend me a cloak as far as the country.’ ‘Well I’ve got one that goes as far as my ankles,’ he says, ‘but not one that goes as far as the country.’”) The jokes which make up the Philogelos must surely predate their collection quite considerably, however its substantially later date in comparison to Old Comedy means that it cannot be used as evidence for the existence of such stand-alone jokes in the 5th century bc.
21. As Palmer 1987 chap. 7 shows, this embeddedness is common in many forms of performed comic narrative. Palmer argues that jokes do not interrupt narrative, nor do they operate independently from it; rather “the enunciative mechanism [of the joke] emerges within the flow of the narrative” (1987.147), and the narrative world itself is central to establishing the rules of plausibility which the joke may then abuse or contradict. Palmer’s examples are largely drawn from film and television, which (despite the obvious differences in technology!) are in many ways more similar to Old Comedy than the stand-alone joke, since they are narrative media which employ a combination of visual and verbal effects. For an example of how film and television comedy may be deployed as a comparator to useful effect in the analysis of Old Comedy, cf. Ruffell 2011 passim.
23. Compare Robson 2006.16–17, who suggests that rather than simple “opposition” or “incongruity,” narrative comedy (and especially Aristophanic comedy) often operates according to a model of “frame abuse.” Robson uses the example of “buying a train ticket,” and suggests that there are certain actions (asking the price, finding out which platform the train will leave from, etc.) which are associated with this frame. Humour can then be created by abusing this frame and introducing non-associated elements. The example above from the Acharnians doesn’t so much abuse the frame of “shopping in the agora” as splice together two incompatible frames, but this can still be seen as a phenomenon closely related to Robson’s frame abuse.
26. See Henderson 1991.188 for a discussion of the scatological puns in this opening exchange. The jokes rely on the double meanings of these verbs as both “I am pressed” (i.e., by the weight of the baggage which Xanthias is carrying) and “I need to shit.”
27. On the phenomenon of metajokes, see Oring 2003.18, who argues that such jokes expose the artificiality of the conventions of joke telling. Xanthias’s comments in lines 12–13, where he asks what the point of his carrying luggage was if he is not allowed to make the standard jokes about its weight, seems to me to expose the artificiality of this apparently standard joke set-up in much the same way as the metajokes discussed by Oring.
28. Translations are my own throughout.
30. On the plot of Pytine, see Rosen 2000, Biles 2002, Ruffell 2002, and Bakola 2010.60–64. Even if the play did end with the poet “reformed” to a degree (note that Biles refutes the idea that this is necessary), Pytine clearly plays on the conventional connection between wine and poetry.
31. On this phenomenon of accumulation and lists in Aristophanes, see Spyropoulos 1974, and Silk 2000a chap. 3, esp. 144–59, on accumulations which also contain disjunctions (including collisions between the abstract and concrete). It is probably unsurprising that frag. 198 has many similarities with Aristophanic language, given its apparently close relationship with Eq. 526–28 (for which see Σ Ar. Eq. 526a). For a more extended comparison of Aristophanes’ and Cratinus’s styles, see Silk 2000b. There is, however, a tendency in much scholarship towards Aristophanic exceptionalism, and Silk is no exception in this regard; see also Taillardat 1965.1, who suggests that “L’abondance des images est peut-être le trait qui marque le plus vivement le style d’Aristophane.” As this article’s analysis should make clear, I do not consider Cratinus’s style to be any less virtuosic than that of Aristophanes.
32. Cf., e.g., above for the joke: “Q: Why should you wear a watch in the desert? A: Because a watch has springs,” in which the double meaning of spring is used to map the two domains of desert and watch onto one another.
34. Very little is known about this play, since only its title and this single fragment survive; however, the reference to specifically Dionysiac songs (i.e., thriamboi) would seem to fit comfortably within a play whose title apparently refers to dramatic productions/lists of dramatic productions/dramatic rehearsals.
35. This compound is otherwise only attested in Plutarch de Primo Frigido 949f.
36. A full discussion of this play, and particularly its agon, can be found in Bakola 2010.70–79. Bakola suggests that the reference in frag. 6 to a blind man (6.3: ὁ τυφλός) may indicate that Archilochus’s opponent was Homer, but there is also some evidence (cf. Diogenes Laertius 1.12) for Hesiod being present on stage.
38. Cf., e.g., Ath. 3.121e: ἁλμυροὺς λόγους (“salty words”). Compare the Latin use of sal to mean witty language and (potentially invective, if only jestingly so) banter in, e.g., Catull. 13.5: “et sale et omnibus cachinnis” (“both salt and all the laughs”).
40. E.g., Ar. Vesp. 1515, Ar. frag. 432, and Antiph. frag. 221.
41. See Vesp. 903. For βαύζω used of human speech, see, e.g., Ar. Thesm. 173 and Aesch. Ag. 449; of dogs barking, Heracl. frag. 97.
42. See again the discussion above of the joke “Q: Why should you wear a watch in the desert? A: Because a watch has springs.” Springs (i.e., bodies of water) and springs (mechanical devices) have no concrete qualities in common, but are only linked by an abstract and entirely coincidental double meaning.
44. On the phenomenon of excesses of meaning leading to nonsense, see Kidd 2014.6–7, who notes that often, “Utterances which are generally called ‘nonsense’ do not ‘mean nothing’ but rather mean too much.”
46. This fragment is unassigned to any play, and in Ath. 106b is attributed to Eupolis and not Cratinus. However, since the fragment displays this multiple metaphor device in a manner which is consistent with Cratinus’s style elsewhere, this lends support to the idea that the fragment belongs to Cratinus and not Eupolis, who does not utilise this multiple metaphor device to the same degree.
50. Storey 2011.333. The similarity between the dative form of “cheese” and “Tyro” is purely coincidental; however Minthe was turned into a mint plant by a jealous Persephone in an aetiological story explaining the plant’s name. For this story, see Strab. 8.3.14. On female personifications in Old Comedy, see Hall 2000 and Kidd 2014.77–83.
51. I use the modern question-answer joke format here for the sake of illustration; in the context of Old Comedy, a gag such as this would likely be distributed between two speakers, one giving the set-up and the other the punchline, as in the examples from Acharnians and Frogs discussed above pp. 238–40.
52. Cf. Kidd 2014.77–83 on “Having sex with abstractions.” Kidd discusses the inherent impossibilities created by the embodied abstractions which are a common feature of the comic stage (e.g., Theoria and Opora in Aristophanes’ Peace).
53. See, for example, Acharnians 1064–66, where Dicaeopolis instructs a bridesmaid to instruct the bride to anoint her husband’s penis with oil in preparation for sex.
54. For anyone still wondering (cf. p. 231), blueberries always forget.