An analysis of Terry Pratchett’s wordplay and its
translation based on Delabastita’s There’s a Double Tongue
Bryn van Helden
student no. 1457675
Supervisors: J. P. M. Jansen and W. M. Lowie
’In truth, the turtle doesn’t have anything to do with most of the stories except, as it were, to carry the plot’ — Terry Pratchett —
Masterscriptie Engelse Taal en Cultuur
List of abbreviations . . . 2
Preface . . . 3
Introduction . . . 4
Chapter 1: Wordplay . . . 7
1.1 The rite circumstances. An almost definite definition of wordplay . . . 7
1.2 It takes all sorts . . . 11
1.3 Whose punning? . . . 15
Chapter 2: Translating Wordplay . . . 18
2.1 The rocky road to wordplay translation . . . 18
2.2 The typical ways of translating wordplay . . . 21
2.3 Producing a congenial TT pun: The x-factors . . . 26
Chapter 3: Analysis . . . 31
3.1 Ambiguity squared . . . 31
3.2 The truth revealed . . . 35
3.3 Translating the truth . . . 36
3.4 Simply the best . . . 41
3.5 Dismantling Delabastita . . . 45
Conclusion . . . 48
Bibliography . . . 52
Appendix A: It takes all sorts, the basics . . . 56
Appendix B: Wordplay overview . . . 61
Appendix C: Data of wordplay analysis The Truth . . . 89
Appendix D: German, Dutch and English through the looking glass and under the microscope . . . 91
CED Collins English Dictionary
Del. Delabastita’s There’s a Double Tongue
EWN Etymologisch Woordenboek van het Nederlands OALD Oxford Advanced Learner’s Dictionary
OED Oxford English Dictionary
VDN Van Dale Groot Woordenboek van de Nederlandse Taal VDND Van Dale Groot Woordenboek Nederlands-Duits
Most people in Lancre, as the saying goes, went to bed with the chickens and got up with the cows.
[footnote] Er. That is to say, they went to bed at the same time as the chickens went to bed, and got up at the same time as the cows got up. Loosely worded sayings can really cause misunderstandings. (Maskerade 91)
This quote from a Discworld novel by Terry Pratchett addresses an important as-pect of this thesis; loosely worded sayings can indeed cause misunderstandings, for many words and phrases have more than one meaning. The phrase ‘to go to bed with the chickens’ can mean both ‘to go to bed early’1 as well as ‘to have sex with chickens’.2 Without the footnote, the second reading would probably not have crossed the reader’s mind, and that was exactly why Pratchett added it: the comical ambiguity is intended. This is called ‘wordplay’, a literary technique which appears frequently in Pratchett’s work. Pratchett’s wordplay, together with its translation, will be the focus of this study.
Pratchett is one of the most successful British authors of the late twentieth and early twenty-first century and is best known for the Discworld series. The first Disc-world novel, The Colour of Magic, was published in 1983, and Pratchett has written 36 since, plus 9 non-Discworld novels. In 2007 he was diagnosed with a rare form of early-onset Alzheimer’s disease, but that has not stopped him yet from continuing his writing career.3 The genre of the Discworld novels has been described as, among other things, ‘humorous science fiction writing’ and ‘humorous fantasy’,4and Prat-chett himself as ‘the soul of wit’ (Miller 2004) and a man who ‘plays with language the way a juggler tosses burning torches from hand to hand’.5 Humour, in short, is a decisive factor regarding the success of his work. The Discworld novels have been translated into 37 different languages, and as much of the humour contained in his work can be ascribed to the use of wordplay, I became interested in the trans-lational problems posed by this literary technique. After all, as wordplay exists by
1‘to go to bed with the chickens’ (Palmatier 173). 2‘to go to bed with someone’ (OED).
3Terry Pratchett Biography, by Colin Smythe (his agent).
which does not possess these particular qualities poses a number of problems. These problems have been discussed by Dirk Delabastita in There’s a Double
Tongue: An investigation into the translation of Shakespeare’s wordplay with spe-cial reference to Hamlet (1993).6 He brought together what little had been written at the time with his own ideas and produced the most comprehensive study on the translation of wordplay existent today. While he wrote his book with a focus on Shakespeare, he devoted a large part of it to the discussion of general theories of wordplay and theories on translation, because, as he writes in his preface: ‘study-ing the translation of wordplay makes little sense without a more general theory of
translation and a theory of wordplay’ (xi). On the basis of these theories he
com-posed a range of translational methods.
At first sight Delabastita’s book seems to be a useful tool for analysing Prat-chett’s wordplay and its translation, but only a critical investigation can show if it really is suitable for a work other than that of Shakespeare. This thesis therefore examines the practicability of Delabastita’s theories and methods by using There’s
a Double Tongue as a basis for analysing Pratchett’s The Truth (2000) and its Dutch
and German translations.7 The choice for this particular Discworld novel is based on its relatively higher wordplay density compared to the average Discworld novel;8 it provides more data than, for example, The Light Fantastic. Puns from the other novels are, however, used abundantly to illustrate the terminology and theories dis-cussed in this thesis. I chose to base the analysis of wordplay translation on two translations because it enables me to compare the translators’ methods, providing a more thorough analysis of Delabastita than with just one translation.
The first chapter is devoted to establishing what is understood by wordplay, thereby substantiating why Delabastita’s definition is most accurate. In addition, it distinguishes between different types of wordplay and explains wordplay intention-ality and understanding, thus providing a basis for wordplay analysis in Chapter 3.
6 He also edited two volumes about the translation of wordplay: Wordplay and Translation (1996) and Traductio. Essays on Punning and Translation (1997).
7 Dutch translation: De Waarheid (2002), by Venugopalan Ittekot (real name: Ruurd Groot). He translated all Discworld novels. (Interesting Times and Maskerade had initially been translated by Joke de Vries and Kiki Zondervan, respectively, but were retranslated by Ittekot in 2009.) German translation: Die volle Wahrheit (2001), by Andreas Brandhorst. He translated all Discworld novels bar one (Making Money, translated by Bernhard Kempen).
8 I have drawn this conclusion after counting the number of puns in the two earliest novels (The
its alleged (un)translatability. It also introduces Delabastita’s classification of dif-ferent techniques available to the translator of wordplay which are relevant to this thesis. Furthermore, this chapter presents Delabastita’s ‘statistical laws’, which are assumptions on the chances of a congenial pun translation based on the extent to which the source language and target language are related. I have added these laws in order to investigate if they can help to explain the outcomes of my wordplay translation analysis.
Chapter 3 sheds a critical light on There’s a Double Tongue. It tests the knowl-edge assembled in Chapters 1 and 2 by discussing the workability and trustwor-thiness of an analysis of wordplay and wordplay translation such as suggested by Delabastita. These discussions are based on a wordplay analysis of the (English) original The Truth and on a wordplay translation analysis of the Dutch and Ger-man translations. The wordplay translation analysis is also used to discuss which translator did best. The last section criticizes Delabastita’s laws.
W
ORDPLAY
1.1
The rite circumstances. An almost definite definition of
wordplay.
Unfortunately, there seems to be no consensus when it comes to defining word-play. In its broadest sense, wordplay simply constitutes playing with words, such as is found in insult-competitions, bragging-matches (Redfern 13), crossword puzzles and punning. This definition is of little use here, because it is too broad a concept and does not distinguish between wordplay based on language use and wordplay based on language structure. Only the latter is of interest here, which significantly narrows down that which is understood by wordplay.
Two consulted dictionaries of the English language come up with a stricter sense of wordplay. According to the Oxford English Dictionary (OED), wordplay is ‘the action of playing with words; witty use of words, esp. of verbal ambiguities’. It is the second part of this entry that comes closer to the focus of this study, but it still leaves much to be desired; that is, the term ‘verbal ambiguity’ is too vague and undefined. The Oxford Advanced Learner’s Dictionary (OALD) did make an attempt to describe the verbal ambiguity of wordplay and is therefore one step closer to the definition of wordplay that will be used here: ‘making jokes by using words in a clever or amusing way, especially by using a word that has two meanings, or different words that sound the same’.
This definition emphasizes the structure of the language where wordplay takes place, that is, wordplay that is based on those aspects that are inherent in the language structure: form and meaning. However, by using the word ‘especially’ this definition does not rule out other instances of playing with words (i.e. generally ‘making jokes by using words in a clever or amusing way’). If this study were based on the interpretation of the OALD, instances of ‘wordplay’ such as the following from Men at Arms would have to be taken into account:
Here, the joke is based on language use, not structure. That is, Pratchett formed this witticism by using logic, a concept which lies outside language and is therefore not inherent in language itself. The OALD definition is too broad in this respect, because it is wordplay based on language structure that only concerns this study.
While the first part of the OALD definition is too all-embracing and should be dismissed out of hand, the second part is too restricted. It says that wordplay can come about by ‘using a word that has two meanings, or different words that sound the same’. In other words, a signifier (form of a word) has two signifieds (meanings), or different signifiers are pronounced the same. Of course, going with this definition one can come up with clear illustrations of wordplay. The following fragments from
The Light Fantastic and The Colour of Magic give examples of wordplay based on
a word that has two meanings (polysemy) and wordplay based on different words that sound the same (homophony), respectively:
‘[ . . . ] These are suits of space armour.’
‘They don’t look very roomy to me,’ said Rincewind [ . . . ] (Colour of Magic 267)
‘Rincewind, all the shops have been smashed open. There was a whole bunch of people across the street helping themselves to musical instruments, can you believe that?’
‘Yeah,’ said Rincewind [ . . . ]. ‘Luters, I expect.’ (The Light Fantastic 238)
In the first example, the pun is based on two meanings of the word ‘space’ which will be indicated with s1 and s2 (s stands for sense). The first meaning (s1) is ‘the area outside the earth’s1atmosphere where all the other planets and stars are’ (OALD). This is what the first speaker meant by ‘space armour’, a collocation unknown to Rincewind, who adopts the second meaning: ‘an amount of an area or of a place that is empty or that is available for use. Syn.: room’ (OALD). The second example makes use of the phonetic similarity of looters who ‘steal things from shops/stores or buildings after a riot, fire, etc.’ and luters who play ‘an early type of musical instrument with strings’ (OALD).
While there are many instances of wordplay that work this way, there are even more that do not entirely satisfy the standards of this clear-cut definition; that is, wordplay does not only occur on word level, as the OALD definition suggests, but grammar may also play a role (Del. 57). Take the following example from The Truth, which represents the title of a newspaper article:
Patrician Attacks Clerk With Knife (He had the knife, not the clerk) (179)
The first sentence is ambiguous because it yields two possible readings: s1 the Patrician has a knife with which he attacks the clerk, and s2 the clerk has a knife and is attacked by the Patrician. The author of the article realized that the phrase might be misinterpreted and therefore added a subtitle in brackets.2 This is a pun on the level of grammar (or, more precisely, syntax), because it is based on word order.
Another reason why the OALD definition is not clearly defined enough concerns the phonetic aspect of wordplay. According to the OALD, wordplay based on sound requires homophones, thus ruling out the numerous instances of wordplay where different words may sound almost the same, such as the following example from
The Light Fantastic:
‘[ . . . ] Mind you, there is a saving on the rates.’ ‘Rates?’
‘Yes, they’re-’ the shopkeeper paused [ . . . ]. ‘I can’t quite remember, it was such a long time ago. Rates, rates-’
‘Very large mice?’
‘That’s probably it.’ (226)
‘Rates’ and ‘rats’ do not exactly sound the same but they do sound alike, which makes it possible for the reader to make the connection between the two when faced with the question ‘very large mice?’ As the joke is based on ambiguity on the level of language structure (phonetic resemblance) it will be considered as an instance of wordplay or pun.3 As a consequence, the OALD definition — which would discard this example as being a non-pun because of the near-similarity in sound — is too much at odds with this study’s interpretation of wordplay so that another, more complete definition is in order.
A definition that allows for near-similarities of sound and meaning and which also takes into account other levels than word level on which punning can take place can be found in Dirk Delabastita’s There’s A Double Tongue (1993). Delabastita’s
2The subtitle might also be read in two different ways if one would go that far: s1 the Patrician was the one with the knife and the clerk was not; s2 the Patrician was holding the knife and he was not holding the clerk.
definition is most complete and will be used here for analyzing Pratchett’s wordplay. It is formulated as follows:
[W]ordplay is the general name indicating the various textual phenomena (i.e. on the level of performance or parole) in which certain features inherent in the struc-ture of the language used (level of competence or langue) are exploited in such a way as to establish a communicatively significant, (near)-simultaneous confronta-tion of at least two linguistic structures with more or less dissimilar meanings (signifieds) and more or less similar forms (signifiers). (57)
The terms langue and parole are derived from linguist Ferdinand de Saussure’s
Course of General Linguistics and refer to the ‘self-contained system of signs’ and
‘actual individual acts of speech’, respectively (Amsterdamska 231). That is, langue corresponds with language as a whole (e.g. the words and expressions that are part of a language). Parole, on the other hand, is concerned with speech (i.e. the actual
use of language). This distinction is of great importance in Delabastita’s
defini-tion above because it emphasizes a feature that is characteristic of wordplay: the interaction of parole and langue; individual acts of speech that make use of a self-contained system of signs.
Apart from including near-similarities of sound and meaning, wordplay on the level of grammar and the distinction between langue versus parole, Delabastita’s definition differs from that of the OALD in another respect: the OALD considers the function of punning to be ‘making jokes’, while Delabastita’s ‘communicatively significant’ is much broader than that and ‘merely insists on the need for some communicative significance’ (58). According to him, wordplay does not necessarily have to be funny but may also serve other goals, for example persuasion or irony (145–146).
As Delabastita’s definition is based on the study of wordplay in Shakespeare’s plays it is possible that Pratchett’s wordplay requires a slightly different approach because Pratchett’s work is of a different genre; it seems, for example, that Prat-chett’s wordplay is almost if not completely hinged on extracting a laugh from the reader. Nevertheless, ‘extracting a laugh’ is also ‘communicatively significant’, so while the focus of Pratchett’s wordplay may be somewhat different than Shake-speare’s, Delabastita’s definition provides a convenient, satisfactory basis for this study.4 I will therefore adapt it while remaining critical of any incongruities that may be encountered.
1.2
It takes all sorts
In The Humor of Humor (1954), E. Esar wrote:
The variety of puns must be infinite. Years ago I began to write a book on the subject and had little difficulty with its history, literature and other phases. But when I came to record the different types of puns, I gave up after identifying dozens of different species, for it seemed an endless task. (qtd. in Redfern 5)
Forty years later, Delabastita’s restrictive definition eliminated a great number of puns, significantly facilitating the classification of wordplay. He distinguishes be-tween vertical and horizontal forms of wordplay, for example. These forms are concerned with the arrangement of the pun components.5 When the components are positioned in one and the same portion of text, the pun is considered to be a vertical form of wordplay (Del. 78). Take the following example:
‘Therefore I will have dinner sent in,’ said the priest. ‘It will be roast chicken.’ - ‘I hate chicken.’
Dios smiled. ‘No, sire. On Wednesdays the king always enjoys chicken, sire.’ (Pyramids 157)
To enjoy s1: [formal] to have something good that is an advantage to you
s2: to get pleasure from something (OALD)
This is an example of vertical wordplay because the two components of the pun (the two meanings of ‘to enjoy’) are located in one and the same word and this illustrates what is meant by ‘simultaneous confrontation’ in Delabastita’s definition of wordplay. In order to clarify the distinction, here is an example of horizontal wordplay:
Thunder rolled. . .
It rolled a six. (Guards! Guards! 24)
To roll s1: to make a long continuous sound
s2: an act of rolling a dice (OALD)
have to be modified if languages other than the Standard Average European languages referred to are to be taken into account’ (57). As the languages referred to in his study are English, Dutch, German and French, it may be assumed that his definition of wordplay can also be applied to the analysis of the Dutch and German wordplay translations in Chapter 3.
Here, the two meanings of ‘roll’ is distributed over two portions of text; the two components occur one after another, thus resulting in a near-simultaneous con-frontation.
The distinction between vertical and horizontal wordplay is based on form and Delabastita calls it the ‘first axis’ (78). The second axis of form is concerned with the kinds of relationship of form between the two pun components (i.e. the ways in which the two components look or sound alike). These relationships of form may be based on homophony, homography, homonomy, and paronymy. Appendix A discusses these forms of punning as well as other types of wordplay that have been analyzed by Delabastita and are relevant to this study.6 The rest of this section will be devoted to idiom transformation, a punning technique that is only touched upon briefly by Delabastita but which appears abundantly and in many different forms in The Truth. The main source for the following detailed discussion is therefore not There’s a Double Tongue, but Andrejs Veisberghs’ essay ‘The Contextual Use of Idioms, Wordplay, and Translation’.7
The contextual transformation of idioms is a punning technique where idiom encompasses all ‘stable word combination[s] with a fully or partially transferred meaning’ (Veisberghs 155). ‘Kicking the bucket’, for example, is an idiom with a fully transferred meaning because its original literal meaning is now lost, while ‘be careful what you wish for’ also has a stable word combination but has an original meaning which is much clearer. Sometimes an idiom is transformed in such a way that it creates a contrast with its ‘normal’ reading, such as ‘He didn’t kick the bucket — he pushed it mildly’. Here, the reader’s expectations are defeated because the fossilized idiom has undergone an unexpected change. There are several ways in which to achieve this effect and they provide ample opportunity for wordplay.
The types of idiom transformation fall into two basic groups: structural and
se-mantic transformation (Veisberghs 157). Structural idiom transformations affect
both the idiom’s structure and meaning, while semantic transformation only al-ters the meaning and leaves the structure intact. Both kinds can be considered as wordplay because they both involve a confrontation of two linguistic structures with more or less similar forms and more or less different meanings (i.e. the idiom in its usual form and meaning and the changed form and meaning (158)). In
addi-6 Apart from homophony, homography, homonymy, and paronymy, these are: polysemy, malapropism, bilingual wordplay, and the garden path sentence.
tion, idiom transformations are contextually significant; in the Discworld novels as in most other instances, they ‘create a comic effect and attract the attention of the reader [ . . . ] to a specific point in the text’ (id.).
There are several subtypes of structural transformation, such as addition, inser-tion, and ellipsis, but the subtype that is important to this thesis because it occurs frequently in The Truth is substitution. Here, one or two words of the original idiom are substituted by a synonym, antonym or paronym while — if it has been done well — the original idiom is still ‘present in the minds of the author and the reader or listener’ (Veisberghs 158). In the following two examples, only one word has been substituted and the original idiom is still clearly visible:
Pointy shoes, even if they have a very long point, cannot absorb the impact of a human foot kicking what is, when all is said and screamed, a piece of wood with a thin cloth and leather wrapping. (Unseen Academicals 151)
‘This is what it comes to!’ muttered Colon. ‘Decent women can’t walk down the street without being eaten! Right, you bastards, you’re. . . you’re geography —’ (Guards! Guards! 271)
In the first fragment, ‘done’ is substituted by ‘screamed’, thus adjusting a perfectly normal idiom to the context, making it more lively and funny at the same time. In the second fragment, Colon confuses ‘history’ with ‘geography’, thus unwittingly altering the idiom ‘you’re history’.8 The same idiom transformation also appears earlier on in Guards! Guards! where ‘geography’ is used intentionally to empha-size its literal meaning: ‘[Dragons] lived on a chemical knife-edge the whole time. One misplaced hiccup and they were geography’ (132). Pratchett probably felt that ‘geography’ more graphically described what would happen to exploding dragons than could the standard idiom.
Semantic transformation, as opposed to structural transformation, does not
al-ter an idiom’s structure in order to change its meaning — it is the context which ‘activates the semantic change’ (Veisberghs 158). The most common form of se-mantic transformation is dual actualization.9 This term refers to those cases that trigger both the ‘idiomatic reading and non-idiomatic (literal, compositional) read-ing’ (158). An example of dual actualization from Unseen Academicals is ‘Wizards
always had a place in their hearts for cooks, because it was near their stomach’ (66). The idiomatic reading of ‘to have a place in someone’s heart’ (i.e. ‘to be important to someone’) applies as much to the context as does its idiomatic reading (i.e. the heart is literally, anatomically, close to the stomach).
Semantic idiom transformation may also take place in the form of a sustained or
extended metaphor, that is, a piece of language that ‘uses several vehicles10from the same area of thought’ (Montgomery 154). Extended metaphors contain ‘linked ve-hicles, grounds and tenors,11which provides aid in interpreting each new metaphor in the sequence as it arrives’ (154). The following example of a sustained/extended metaphor in an idiom transformation from The Light Fantastic helps to explain the workings of this phenomenon:
It looked the sort of book described in library catalogues as ‘slightly foxed’, al-though it would be more honest to admit that it looked as al-though it had been badgered, wolved and possibly beared as well. (11)
This fragment contains the idiom ‘slightly foxed’, which is often used to describe the papers of old books and writings, with ‘foxed’ meaning ‘stained with brownish-yellow spots’12 (OALD, OED). ‘Badgered’, ‘wolved’, and ‘beared’ are all verbs de-rived from forest animals, just like ‘foxed’. The tenors of the additional metaphors are unclear, extinct or non-existent (except for ‘badgered’13), but the ground of the primary metaphor provides a context which suggests that a book that has under-gone all those things (or animals) has suffered very much indeed. Pratchett thus extended the metaphor by adding vehicles from the same area of thought, making this an extended or sustained metaphor.
The third and last type of semantic transformation identified by Veisberghs is the zeugma, which involves ‘a single word, commonly a verb and used only once, [which] goes with two or more other words, in different senses’ (Clifton 263). In the context of an idiom transformation, this verb is part of an idiom, as in the following example:
Sometimes people left in disgrace, in a box or, in a few cases, in bits, but there was no tradition of resigning at all. (Unseen Academicals 23)
10A vehicle in this case indicating the ‘literal meaning of a metaphor’ (Montgomery 153).
11 The tenor is the non-literal meaning of a metaphor and the ground the relation between the vehicle and the tenor (id.).
12According to John Ayto, the stain ‘is so called because it is of the colour of a fox’ (Brewer’s
Dictio-nary of Phrase and Fable).
‘To leave in disgrace’ is an idiom of an abstract notion, from which the reader leaps to two different, material meanings without there being a change in verb. While there is a likeness in form (they all share the same verb), their meanings are dis-similar (abstract and material). The grammar is correct but the left-out verb leaves an unexpected pattern for the reader. That is, the idiom is left intact but the lit-eral meaning of ‘to leave in’ is given renewed attention caused by the subsequent phrases in which the shared verb is used in the material sense.
The wide variety of wordplay types may suggest that there is a clear-cut bound-ary between different kinds of wordplay, but one text fragment or even one or two words may contain more than one pun. The phrase ‘you’re geography-’, for example, is both a malapropism and an instance of structural idiom transformation. The fact that several types of wordplay may be contained in one wordplay fragment presents an additional problem for the translator who, as Chapter 2 will show, faces enough problems as it is.
1.3
Whose punning?
After establishing a definition of wordplay and distinguishing between different types of puns, there is another matter that needs to be discussed in preparation for the analysis of wordplay translation in Chapter 3. This matter is concerned with wordplay understanding and wordplay intention (Del. 125), both of which consist of at least two levels and which show the difference between, for example, puns that are recognized by the reader and not by the characters themselves and puns made consciously by a character without them being recognized as such by the other characters. These distinctions are of importance to the translator of wordplay, as will be explained below.
There are two levels of wordplay understanding and they distinguish between
overt and covert wordplay. An overt pun is a pun that is recognized as such by
the characters themselves, while a covert pun is understood mainly by the reader and may be made consciously by the character, but is never recognized as a pun by the other characters.14 The distinction made in wordplay intentionality is that be-tween character puns (or personal puns) and author puns (or auctorial puns), where character puns are intended by both the author and the character and author puns are only intended by the author alone and unperceived by the character (Del. 125). Character puns and overt wordplay do not always coincide and nor do author puns
and covert wordplay; a pun not intended by a particular character (author pun) may still be recognized by other characters (overt wordplay) just as a pun not recognized by the other characters (covert wordplay) may well be intended by the character who made the pun (character pun) (126).
All puns in the Discworld novels that do not appear in dialogue or thought are obviously covert and auctorial; the characters are not aware of the text because it is only visible to the reader. There are also covert auctorial puns that do appear in dialogue, however, as is the case in Unseen Academicals when the ticket collector interrupts Glenda:
‘Just speak with a little more class, eh? You don’t have to sound like-’ ‘My fare, lady?’ (49)
The ticket collector nor Glenda nor anybody else in the novel would have heard of the musical, so only the reader is aware of the pun on My Fair Lady. There is quite an amount of covert auctorial wordplay to be found in Pratchett’s work, which is both a blessing and a curse for the translator. As Chapter 2 will show, the translation of wordplay poses a number of problems; however, covert author puns may in some cases be omitted without seriously damaging the text.
Overt puns, on the other hand, are much harder to manage, because these puns are not only recognized by the reader but by the characters as well. It is sometimes hard to tell if other characters recognize a pun because they may recognize it but ‘choose’ to ignore it. For the sake of clarity all puns appearing in dialogue but that are not responded to by other characters will be classified as covert puns. What this means is that with overt character puns the characters react to the pun and add to it, as in the following example from The Truth, where the editorial staff (Sacharissa and the dwarf Goodmountain) have just hired a vampire who is feeling uneasy:
‘It’s all right, we won’t bite you,’ said Sacharissa.
‘And one good turn deserves another, eh?’ said Goodmountain. ‘That was a bit tasteless, Mr Goodmountain,’ said Sacharissa. ‘So am I,’ said the dwarf [ . . . ]. (137)
again, makes this kind of wordplay much harder to omit15 than an overt auctorial pun. The idiomatic quality of ‘[w]e won’t bite you’ and ‘That was a bit tasteless’ is not present in every language16 so that in many languages another solution will have to be found in order to reproduce both the idiomatic and literal meaning.
15Chapter 2 discusses this and several other techniques that a translator can apply if direct trans-lation seems impossible.
T
RANSLATING WORDPLAY
2.1
The rocky road to wordplay translation
Delabastita wrote: ‘if [a pun] is intentional, it has to be retained but will resist any attempt to that end. Whichever way one looks at it, translation and wordplay form an impossible match’ (153). This discouraging statement probably explains why so little has been written on the subject. It is indeed not a hopeful thought, especially not when considering that the first problem a translator encounters does not even have to do with the actual translating but with the first step in the process of translating wordplay: detecting the pun. As the following constraints will show, it is usually not enough for a translator to just follow his intuition and gear his attention to identifying puns.
Translators fixed on detecting wordplay are at risk of over-reading, for exam-ple (Del. 160). In his analysis of ambiguity, William Empson warned for this over-reading, with which he meant the ‘ingenious, over-drawn, and sometimes contradic-tory explications of a literary word or passage’ (Abrams 12). This way, an ambiguity may be considered as an instance of wordplay while the context does not support this view. This is the case, for example, when an ambiguity is marked as wordplay while the second meaning of this suggested pun does not make sense in the piece of text in which it occurs.
been written at some point in the past three decades. They may, however, suffer from another type of ‘under-reading’: missing out on puns because they are not expecting to find any or are not reading closely enough.
There is, however, no ‘safe middle course’ between the extremes of over-reading and under-reading, because things are more complex than that. Delabastita states:
Besides the puns about which one can be positive, and next to the numberless potential puns that are not significant in the given contexts, one can construct many interpretations of a text that do, or do not, accept the relevance of dou-ble meanings but which cannot be so easily dismissed as either over-readings or under-readings. (161)
The relevance of wordplay in a piece of text is often indicated with the terms
significant versus non-significant wordplay, that is ‘a more or less subtle
manipula-tion and a more or less negligent transgression of speech convenmanipula-tions’, respectively (Del. 123). Broadly speaking, the distinction is one of wordplay intended by the au-thor and wordplay which is unintended. Instances of non-significant wordplay are called punoids or non-puns and these may come in the forms of Freudian slips, unin-tentional ambiguities, potential ambiguities that lack sufficient contextual support, etc. Most of these can be disambiguated by the context (166–167), but is not always easy to distinguish between intended and unintended wordplay.
Apart from the problems of distinguishing between significant and non-signifi-cant wordplay there is a serious constraint in the next step of the translation pro-cess: the translation itself or, rather, the untranslatability of wordplay. The tradi-tional view has been very well expressed by Joseph Addison in The Spectator:
The only way [ . . . ] to try a piece of wit, is to translate it into a different language. If it bears the test, you may pronounce it true; but if it vanishes in the experiment, you may conclude it to have been a Pun. (81)
They consider it ‘(relatively) untranslatable’ (Catford qtd. in ibid.) and think that it represents ‘special challenges to both translators and translation studies as a scientific discipline’ (Toury qtd. in ibid).
Puns are considered (relatively) untranslatable because they owe their exis-tence to the structure of the language used. Wordplay ‘exploits the discrepancies of linguistic forms and meanings’ (Del. 154) and it is therefore in the majority of cases impossible to find an equivalent in another language because it has different linguistic qualities. In other words, the double meanings of wordplay exist because the language they appear in has the semantic, phonetic and/or morphological qual-ities that make those double meanings possible. This is opposed to other instances of wit which are easier to translate because they are based on notions that exist
outside the boundaries of language. The following excerpt from Sourcery,1 for ex-ample, contains no wordplay whatsoever and can be translated easily: ‘It was said that everything in Ankh-Morpork was for sale except for the beer and the women, both of which one merely hired’ (47). In contrast, another excerpt from Sourcery shows an instance of wordplay which poses a translational problem:
‘[ . . . ] Death walks abroad,’ said Nijel.
‘Abroad I don’t mind,’ said Rincewind. ‘They’re all foreigners. It’s Death walking around here I’m not looking forward to.’ (170)
abroad s1: at large (as in: free, without constraint) (OED)
s2: in [ . . . ] a foreign country (OALD)
The double sense of ‘abroad’ is an example of a linguistic quality of the English lan-guage that does not occur in, for instance, Dutch or German, thus posing a problem for the (Dutch and German) translators of this fragment.
While there are several objections against the translatability of wordplay, there are also objections against the untranslatability of wordplay. Delabastita gives three, the first one stating that ‘the doctrine of the untranslatability of wordplay [ . . . ] overestimates the degree of anisomorphism2between natural languages’ (182).
While languages are to a large extent idiosyncratic they do sometimes show simi-larities which may ease the transfer of a pun into another language.
The second objection to the doctrine of untranslatability focuses on the disre-gard of the textual perspective (183). As puns are textual elements that are
de-1Found in The Wit and Wisdom of Discworld.
pendent on contexts and signals, their translation should not be considered from an autofunctional viewpoint only. That is, the translatability of a pun should not be theorized about without taking into account its textual setting and functions. Delabastita means to say that if a pun plays a ‘primary role in the total cluster of textual features that make up the ST’ — that is, its ‘peculiar forms and meanings [ . . . ] perform an essential functional role’ — it will be harder to reproduce in the T.ling.code3than when it plays a subordinate role, because in the first case the pun has to be reproduced as literally as possible so as not to lose its functional signifi-cance in the TT, while in the latter the translator has more translational freedom because ‘little or no synfunctional loss will be sustained if [the pun] undergoes cer-tain shifts in the translation’ (184).
Delabastita’s third objection argues against the doctrine of untranslatability’s ‘restricted notion of translation’. This notion of translation is restricted in the sense that translation is perceived ‘as a transfer process that operates with substitutions only, and that can guarantee full equivalence or even identity of meaning by virtue of the fact that meanings can be dissociated from their verbal formulation’. Trans-lation in that sense therefore does not include the various techniques used in word-play translation, such as
low-equivalence substitution (the pun is adapted), deletion (the pun is dropped in the TT), addition (new puns are introduced), permutation (the TT contains word-play but in different textual positions) or repetition (the ST pun is simply quoted in a footnote).
There is no reason to ban these techniques from a complete definition of translation, because they are applied to ‘normal’ translation as well — the only difference being that they occur more often in wordplay translation. (186–188)
2.2
The typical ways of translating wordplay
The next step in the process of translating concerns the range of translational tech-niques available to a translator of wordplay. The following presentation of said techniques provides a basis for analysis in Chapter 3 in which the translational solutions for the Dutch and German translations of The Truth will shed light on It-tekot and Brandhorst’s translational methods. The list below gives brief, structural
descriptions taken from There’s a Double Tongue. The English examples added to illustrate the techniques have been taken from the Discworld novels and the Dutch translations are mine. Before turning to the list of techniques it is important to note that it is possible to combine two or more techniques. Also, not all techniques are applicable in all cases (Del. 191).
Pun > Pun. This first technique is described as follows:
The TT contains a pun that can be identified as a translational solution to the ST pun in question through the similar positions of the ST pun and the TT pun within the ST and the TT respectively, and/or through their comparable charac-teristics. (192)
There are several ways in which an ST pun can be transformed into a TT pun. For instance, the type of TT pun may or may not correspond to that of the ST pun.4 In the same way, the ST and TT semantics may or may not coincide. In the following fragment from Jingo, for example, where Vimes addresses his servant who plans to go off to war, s1 (ST) and s1 (TT) are in a relationship of relative equivalence, and so are s2 (ST) and s2 (TT):
‘Well, we shall miss you, Willikins.’ Others may not, he thought. Especially if they have time for a second shot. (86)
to miss s1: to feel sad because you can no longer see somebody
s2: to fail to hit (OALD)
‘We zullen je missen, Van Wichelen.’5 Anderen misschien niet, dacht hij. En al helemaal niet als ze tijd hebben voor een tweede schot.
missen s1: de afwezigheid (van iets of iem.) [ . . . ] pijnlijk voelen
s2: niet treffen, niet raken (VDN)
It may also be possible for the TT pun to be in relative equivalence where only one sense is concerned; the fragment below puns on the likeness of sound and spelling (paronymy) of ‘feud’ and ‘feudal’. In the translation, the relative equivalence be-tween s1 (ST) and s1 (TT) has been retained while s2 (TT) is not equivalent with s2 (ST):
4Both the ST and TT pun could be substitutions, for example, or the ST pun is polysemic while the TT pun is based on syntax.
Lancre operated on the feudal system, which was to say, everyone feuded all the time and handed on the fight to their descendants. (Carpe Jugulum 42)
s1 feudal system: feudalism
s2 feud: verb. to have an angry and bitter argument with somebody over a long period of time. (OALD)
Lankhr opereerde vanuit het leenstelsel, wat zoveel betekende als: iedereen had de hele tijd ruzie over het niet teruggeven van geleende goederen en deze ruzies werden dan doorgegeven aan de nakomelingen.
s1 leenstelsel: feodalisme
s2 lenen: te leen geven; tijdelijk (m.n.) kosteloos ten gebruike aan iemand afs-taan. (VDN)
Lastly, the two senses of the TT pun may be in no relationship of relative equiva-lence with the ST senses at all. An example of this kind of translation is that of the ‘death walks abroad’ fragment in section 2.1. There, ‘abroad’ signifies s1 ‘at large (as in: free, without constraint)’ and s2 ‘in [ . . . ] a foreign country’. A Dutch translation could be:
‘De Dood zit in de lucht,’ zei Niezel.
‘In de lucht is niet erg,’ zei Rinzwind. ‘Daar kom ik toch niet. Ik maak me pas zorgen als de Dood hier op de grond zit.’
lucht s1: aanstaande, ophanden zijn
s2: uitspansel, hemel; atmosfeer (VDO; VDN)
The two senses of ‘abroad’ have not been retained because Dutch does not have a word signifying both meanings. Instead, ‘lucht’ has been used which has two different meanings but which captures the general spirit of the ST pun.
Pun > Non-Pun. ‘The TT fragment which can be identified as the T.ling.code substitute for a particular punning ST fragment does not contain any wordplay’ (Del. 202). One of the possibilities is that only one of the two meanings is translated. This is the case, for example, in the translation of the following paronymic pun from Unseen Academicals where a blacksmith tries to explain to the decent Miss Healstether what ‘mucky’ philosophers are and what they do:
‘They are the ones who go on about what happens if ladies don’t get enough mut-ton, and they say cigars are-’
‘That is a fallacy!’ [said Miss Healstether.] ‘That’s right, that’s what I read.’ (53)
s1 fallacy:/fæl@si/a false way of thinking about something
s2 phallus: /fæl@s/(technical) the male sexual organ, especially when it is erect (OALD)
‘Zij zijn degenen die door blijven zagen over wat er gebeurt als dames niet genoeg schapenvlees eten, en ze zeggen dat sigaren-’
‘Dat is een misvatting!’ ‘Als jij het zegt. . . ’
misvatting: een verkeerd oordeel over iets (VDO)
The paronymic ambiguity of fallacy/phallus has gone lost in the translation; only one meaning has been retained and the result is a non-pun.
Pun > Punoid. In this case, the translator has tried to ‘recreate [a ST pun’s] textual effect by using some other, wordplay-related rhetorical device’ (Del. 207). The translator may have used rhyme, for example, or irony. In the ‘fallacy’ example above, the pun is triggered by the blacksmith’s remark: ‘That’s right, that’s what I read.’ The translation partly makes up for its non-pun solution by means of the blacksmith’s remark in Dutch, which shows that the blacksmith is making fun of Miss Healstether by suggesting — because of her fierce reaction — she must know all about cigars.
Direct Copy. Instead of translating the pun (pun > pun) or using another rhetor-ical device (pun > punoid), the translator copies the ST pun straight into the TT without translating it. An example from the Discworld novels where direct copy is possible can be found in The Truth: some of the dwarf names, like Boddony and Caslong, are paronymies which refer to names of typefaces; these names of type-faces are not restricted to the English language or culture but are known by that name in, for instance, the Netherlands and Germany as well.6 These dwarf names can therefore be copied directly into German and Dutch.7
Transference. Both direct copy and transference introduce ‘linguistic material not belonging to normal repertoires of the T.ling.code’, but whereas direct copy transfers S.ling.code signifiers into the TT, transference brings S.ling.code
signi-fieds into the TT. That is, the TL text or parts of the text has SL meanings and
they come in the form of neologisms or word-signifiers with a prior existence in the T.ling.code. Transference is a technique used when the TL does not have a word that sufficiently covers a particular SL meaning. Most English words have one or more Dutch counterparts, but this is usually not the case where neologisms are concerned. In The Truth, for example, Pratchett introduces the ‘Big Wahoonie’ (83) as a nickname for Ankh-Morpork, thus punning on its American equivalent the ‘Big Apple’. ‘Wahoonie’ is a word made up by the author so a Dutch equivalent did not exist before Ittekot translated The Truth. The translator therefore had to think of a Dutch-sounding equivalent and came up with the ‘Grote Wahoenie’ (De
Waarheid 57), thus making use of transference.
Addition: Non-Pun > Pun. The TT contains wordplay in a passage where the ST does not. This is a technique usually applied to compensate for the fact that the translator failed to translate a ST pun into a TT pun elsewhere in the text (Del. 215–216).
Addition (new textual material): Zero > Pun. This is a more dramatic form of compensation because the addition of a TT pun takes place in a sentence or a group of sentences for which it is impossible to find a counterpart fragment in the ST (217). For instance, the translator may have added a piece of dialogue containing wordplay; the addition does not add anything to the plot but it is put there merely
6Boddony: Bodoni; Caslong: Caslon.
7 It is, of course, up to the translator to opt for direct copy (as did the German translator of The
Truth) or to alter the names slightly so as to adapt them to the TL (as did the Dutch translator who
to compensate for a pun > non-pun translation earlier on in the text.
2.3
Producing a congenial TT pun: The x-factors
According to Delabastita, the success of turning a ST pun into a congenial TT pun largely depends on the interdependence of the type of ST wordplay and the relations (historical, typological, etc.) between the S.ling.code and the T.ling.code. The as-sumption that follows from this consideration is that the more alike the languages are culturally, genetically, or historically, the more likely the success of turning a ST pun into a congenial TT pun. Delabastita discusses this assumption by means of four hypothetical laws,8 each of which is given below with an explanation, ex-ample(s), and possible restrictions. One must keep in mind, though, that punsters may use more than one linguistic feature for one single pun, which means that ‘its reproducability may be subject to more than one of the [hypothetical] laws’ (247). Also, the laws must be seen as a set of basic principles whose relevance ‘can only be assessed in detailed historical analysis of each particular case’9 (id.).
Delabastita’s first law states that ‘[i]ntralingual similarities of form have a greater chance of being reproducible in another language to the degree that the languages involved are historically related’ (234). What this means is that word-play based on phonetics is more easily reproducible if the TL is historically related to the source language. A pun with as its components bread and dead, for example, is easier to reproduce in historically related languages like Dutch and German10 (brood/dood and Brot/Tod respectively) than in a language further removed from English, like Spanish (pan/muerte). However, two factors (semantic and formal shifts), may reduce the chances on a pun > congenial pun translation. Firstly, his-torically related words may have acquired a totally different meaning while their formal similarity has been retained (236). English stool and Dutch stoel, for ex-ample, are historically related and used to have the same meaning, but chair took over the meaning of stool and the latter can now only be translated to kruk in Dutch. Secondly, ‘sound laws do not operate in a mechanical manner, since we have to allow for the effects of analogy, borrowing, interactions between different laws,
8 Delabastita calls them ‘statistical laws’ because they refer to the chances that ‘a given relation-ship will occur or that two (or more) events will be associated with each other in a specified way’ (Theodorson 227). However, these laws have not been tested statistically (that is, there is no data available to support them). The term ‘hypothetical laws’ is therefore deemed more appropriate here. 9This remark strengthens the argument in favour of the term ‘hypothetical laws’.
etc.’ (236) This explains why many pairs of words within a language can show for-mal similarities without being repeatable in the historically corresponding pairs in other languages.
Delabastita’s second law is concerned with polysemic wordplay; he writes that ‘the chances of finding a congenial TT pun are relatively independent of the his-torical relatedness of the S.ling.code and the T.ling.code’ (236). While hishis-torical relatedness does increase the chances of a polysemic pun > congenial pun, there is another factor that influences the reproducibility of polysemy and it does not lie in language itself (237).11 This factor occurs in metaphors — the motivation of the relationship between s1 and s2 in this case lies beyond language, namely in reality. A discussion on the following metaphor will clarify this statement. Consider the following pun from Pratchett’s Reaper Man, where Death says, ‘I EXPECT [ . . . ] THAT YOU COULD MURDER A PIECE OF CHEESE?’12 (135) The relation be-tween the two meanings of ‘murder’ intersecting in this pun is metaphorical; s1 (to kill) and s2 (to eat) are not linguistically related but their relation is grounded on the assumption that one can be so hungry that when finally being able to eat something, one will be eating it with so much speed and vigour that it may bear the resemblance of ravishing or killing it.
In his book Metaphor and Thought, Andrew Ortony writes: ‘[i]n some cases the metaphor is so obvious that it shows up in numerous languages and therefore might be misdescribed as a universal tendency of language rather than a natural tendency of thought’ (45). The polysemy in the example above is the result of the ‘natu-ral tendency of thought’; this instance of metaphor belongs to the extra-linguistic world and this explains why some instances of polysemy recur among unrelated languages. The comparison between eating and killing is not just a British experi-ence but a human experiexperi-ence and may therefore be universal or near-universal. In Dutch, for example, the equivalent of the metaphorical pun above might be: ‘Vol-gens mij zou jij een moord doen voor een stukje kaas’. While it is not a congenial translation, it does combine the notion of eating and killing. This translation is not linguistically motivated and it is very likely that equivalents exist in languages that are not related to English.
Metaphors do of course not always yield a foreign language equivalent. They are the result of a ‘cognitive act and are not given beforehand in the nature of
things themselves’ (Del. 129). This means that many more relations could have been drawn in forming a metaphor and so the choice made might very well not be the choice made in other languages because human experience is not necessarily universal. It may be universal but can just as well be ‘widespread, culture-bound, or even highly individual’ (239). There is a factor which may increase the chances on a congenial pun translation: if the metaphorical pun is culture-bound, the polysemy has a greater chance of being reproduced if the target culture is related to the source culture (240).
The third hypothetical law that Delabastita introduces is the following: ‘When the ST pun is syntactic or morphological, the chances of finding a congenial TT pun are relatively higher to the extent that the S.ling.code and the T.ling.code are ty-pologically related’ (242). Two languages are tyty-pologically related when they show feature similarities, e.g. they have the same word order or they both use preposi-tions instead of postposipreposi-tions (Cysouw 554; Singh 412.1.1). Syntactic ambiguities in English such as ‘the ugly doctor’s wife’ and ‘the cute girl’s hat’, for example, can-not be reproduced in French because in French, the ‘adjective must agree in gender and number with the word which it qualifies’ (‘la grosse femme du major’ or ‘la femme du gros major’) (Kees & Hoppe qtd. in Del. 243). Dutch cannot show the am-biguity of the examples in English because of another language feature: its word order. While the ‘s composition does occur in Dutch (as in ‘mama’s hoed’), an adjec-tive always comes right before the noun it describes, so only ‘mama’s lelijke hoed’ is correct, or ‘de lelijke hoed van mama’. In that respect Dutch and English are not typologically related while they do show feature similarities in other ways, such as the use of prepositions instead of postpositions.
Delabastita admits that there has not been any extensive comparative research in order to formulate specific hypotheses (243) and this becomes most obvious when analysing this third hypothetical rule. While it does to a certain extent explain why languages which are not genetically related can render a pun > congenial pun trans-lation because of typological similarities,13 it does not take into account syntactic puns like the following, found in Pratchett’s Moving Pictures:
‘I don’t understand her,’ he said. ‘Yesterday she was quite normal, today it’s all gone to her head.’
‘Bitches!’ said Gaspode, sympathetically.
‘O, I wouldn’t go that far,’ said Victor. ‘She’s just aloof.’ ‘Loofs!’ said Gaspode. (Briggs 84)
This is an example of a syntactic pun through and through. The pun lies in Gas-pode’s mistaking ‘a’ in ‘aloof’ for an article.14 In order to find a congenial TT pun, however, it is not enough for the TL to share the use of articles before nouns. The TL should also have a word signifying ‘aloof’ or anything like it and this word should start with a morpheme that could also be mistaken for an article. Dutch, for example, does share the syntactic use of articles before nouns (een boot) and it has words starting or sounding like een (eentonig, interessant), but there is no word equivalent to ‘aloof’ which does both. Eenzelvig might be a solution, but as the first syllable is pronounced differently than the article een, it would only work when seeing it in writing and not on the level of pronunciation. In addition, ‘zelvigen!’ does not have the same punch as ‘loofs!’. It is therefore not a congenial pun.
Delabastita’s fourth and last hypothetical law has to do with phonetic and pol-ysemic puns. He says that ‘the chances of finding a congenial TT pun are rela-tively higher to the extent that there exists interlingual borrowing between the S.ling.code and T.ling.code’ (244). Instances of interlingual borrowing usually have to do with words from a language that are added to the lexicon of another, but there are also other kinds of lexical borrowing, namely loan translations and semantic borrowings.
A loan translation — also called a calque — is ‘a special kind of borrowing whereby a language borrows an expression form of another, but then translates literally15 each of its elements’ (Vinay 32). An example of a calque taken from
There’s a Double Tongue is ‘skyscraper’. By borrowing the expression from English
but translating its elements, the French now use gratte-ciel, the Italians grattacielo, the Dutch wolkenkrabber, and the Germans Wolkenkratzer (245). It is evident that puns which make use of these kind of metaphors lend themselves to a congenial pun translation if the word in the TL is the result of a loan translation — after all, the pun relies on the independent elements of the ST word and these elements are the same in the TL.
Semantic borrowing, just like loan translations, does not involve the introduc-tion of new morphemes. It ‘covers the diverse ways in which one language can influence another by changing the meaning of an existing word’ (Del. 245). The subtype which Delabastita states is most concerned with the fourth hypothetical law is that in which a ‘target’ word with one meaning becomes polysemic ‘under the influence of a comparable polysemy in the “source” language’ (id.). An example is the polysemy of ‘lord’ (s1 master - s2 god). Jews were not allowed to utter the name of God so they used the word for ‘master’ instead. When the Bible was translated into Greek the translators used ‘master’ as well, also giving it the additional mean-ing of ‘God’. This spread to the modern European languages and that explains the polysemy of English ‘Lord’, French Seigneur, German Herr, Dutch Heer, etc. (246) Semantic borrowing facilitates the translation of puns in languages that are in-volved in the same instance of semantic borrowing because it creates ‘more or less equivalent instances of polysemy in different languages’ (id.). In other words, a pun based on the verbal ambiguity of the word ‘lord’ can easily render a congenial pun in Dutch because heer has inherited the same additional meaning.
The form of borrowing which gives the translation of puns the greatest chance of success is lexical borrowing, which involves the importation of the entire set of form and meaning (signifier and signified) into the borrowing language (246). There are several ways in which this borrowing can take place. For instance, ‘head’ and many of its European counterparts is polysemic and has been borrowed as such from the Latin polysemic word caput. It is also possible that a polysemic word has been borrowed monosemically twice at two different moments in history, as is the case with the German words Dose (box) and Dosis which both derived at different moments in history from Greek/Latin dosis. The chances of producing a congenial pun in the TL is greater when the two languages involved are European languages that share a large number of Greek/Latin words. (246–247)
A
NALYSIS
3.1
Ambiguity squared
If two researchers with the same knowledge of wordplay made an analysis such as I have done of a novel like that of Terry Pratchett, they would without a doubt yield different results. There are several problems that prevent wordplay analysis from being an exact science, and they are concerned with the detection of puns and defining different types of wordplay.
The first problem is the high number of ‘ambiguous cases’, which float some-where between being a pun and a non-pun or punoid. It is exactly because there is no clear border between that which is defined as a pun and that which is most defi-nitely not a pun, that these ambiguities arise. The following fragment, for example, did appear in my first, raw analysis of wordplay in The Truth, but was removed from it after closer inspection because it failed to meet the conditions for it to be called wordplay:
‘I have presented a writ of exeo carco cum nihil pretii on the basis of olfacere violarum and sini plenis piscis. Tomorrow I shall move that you are ab hamo [ . . . ]’ (416)
The example discussed above shows that there is no clear-cut boundary between puns and non-puns or punoids. ‘olfacere violarum’ may be considered bilingual wordplay by someone else; it all depends on where one draws the line, and these lines have to be drawn separately for each ambiguous fragment on the basis of the extent to which they meet the requirements for wordplay.
The second problem is concerned with the limits of a researcher’s knowledge of concepts that lie outside language and in, for example, popular culture or jargon. That is to say, some puns may escape the researcher’s attention if they refer to mat-ters he has no knowledge of. Puns based on popular culture will not go unnoticed that easily for researchers because they refer to knowledge that is known to basi-cally everyone, such as the name ‘Rocky’ for a troll who has had a history as a boxer. There may be younger readers, however, who have not heard of the film and will miss out on the pun. Other puns are harder to spot and require more than general knowledge, for example because they refer to a specialized subject such as printing. Take the following example:
[I]f you took the leaden letters that had previously been used to set the words of a god, and then used them to set a cookery book, what did that do to the holy wisdom? For that matter, what would it do to the pie? (The Truth 55)
I would not have detected a pun in this fragment if I had not consulted ‘The Anno-tated Pratchett File’ (The L-Space Web), which states that ‘pie’ does not only refer to cooking, but also to ‘Printer Pie’. For confirmation I consulted Industrial
Edit-ing: A Handbook on House Journals, which says: ‘When type-matter is accidentally
upset, the resulting jumble is said to be pied. Hence the expression “printer’s pie” ’ (Smith 214). There are many puns in Pratchett’s novels that refer to concepts out-side language (e.g. films, books, and printing), and they may be passed unnoticed by a researcher with no knowledge of the subject or subjects concerned.
The third problem concerns the act of defining the different kinds of wordplay. Once a list has been drawn up with all instances of wordplay that one can be fairly certain about, these must be divided into groups in order to draw conclusions about the author’s use of wordplay and to make a statement about the possible transla-tional issues and solutions.
Dwarf expressions are more concentrated. (23)
to concentrate s1: to give all your attention to sth and not think about anything else
s2: to bring sth together in one place (OALD)
There can be no doubt about the type of wordplay and the word that is punned upon has even been signalled (i.e. printed in italics to make sure the reader does not miss it); it does not get any clearer than that.
Unfortunately, in practice the greatest number of puns cannot be classified in a wink of an eye. It is, for example, at times hard to decide whether a certain pun can be called a substitution, because it is not always easy to distinguish between a stable word combination — an idiom — and a word combination that cannot be classified as such. The following pun will illustrate this:
‘You’re saying that if this goes bad you’ve never seen us in your life-’ ‘Ahem. . . ’ Mr Slant [the zombie] coughed.
‘Your afterlife,’ Mr Pin corrected himself. (79)
I opted to call ‘seen [ . . . ] in your life’ an idiom because it is a word combination that appears frequently1but above all because the substitution ‘afterlife’ creates a contrast with the original and makes the reader aware of the literal meaning of ‘in your life’ in addition to its figurative, idiomized meaning ‘ever’. Someone else might not define this word combination as an idiom. In effect, they would not classify this pun as a substitution; perhaps they would not even identify the fragment as a pun. Another problem encountered when defining types of wordplay is concerned with distinguishing homonymies from polysemies, because it involves consulting dictionaries as well as etymological dictionaries and the distinction is not always absolute. This may not always be a time-consuming activity, but sometimes the word that is punned upon is homonymic but has been polysemic in the past,2which makes it hard to give a clear definition. In effect, one analyst may draw the line at the entry of a word in a modern dictionary, while the other analyst digs into its past and comes up with a different definition.
1225 million hits on Google (14 March 2010).
The fourth aspect that causes inaccuracy in wordplay analysis has to do with wordplay intentionality. In the section ‘Whose punning?’ in Chapter 1, it was al-ready mentioned that it is sometimes hard to determine whether a character pun is recognized by other characters. I therefore proposed to call every pun that appears in dialogue but that is not responded to by other characters a ‘covert pun’. While this facilitates distinguishing between overt and covert wordplay (wordplay under-standing), the problem of wordplay intentionality is much harder to handle. When a character utters a pun it is often impossible to know if he/she is aware of the pun, as in the remark made by Otto the vampire when his head has just been cut off:
‘Everyvun all right?’ said the vampire. ‘Zat vas a close shave, yes?’ (282) [dual actualization]
While at first sight it looks like he is making a joke, nothing in that bit of text or anything that follows it demonstrates that he is aware of the pun he made. It therefore cannot be said with absolute certainty that this is a character pun and not an author pun.
Wordplay analysis, in short, is not an exact science.3 The total number of puns, the number of, for instance, substitutions, and the quantity of character puns in a Discworld novel are not fixed; the ambiguity of pun-defining techniques and the ambiguity of puns themselves require an approach that is based on estimates in which each case is analysed individually. In addition, an analysis requires some personal assessments on behalf of the analyst him/herself. Fortunately, the large number of puns in The Truth enabled me to make some generalizations that are still true if the percentages given are slightly off. Yet it is important to keep in mind that it is nothing more than an approach, for there is no such thing as a one true ‘perfect’ analysis.
The above-mentioned evaluation does not only provide the analyst with the amount of reservation needed when taking on the task of analysing wordplay in a Discworld novel; it also gives information about how to interpret such an analysis to those who read it or want to use it. With this in mind one is properly prepared for the wordplay analysis of The Truth in section 3.2 and the wordplay translation analysis in section 3.3.
3.2
The truth revealed
The estimates in my analysis of wordplay in The Truth are based on a total amount of puns of 183. With the above-mentioned reservations in mind, this number could be slightly bigger or smaller, but the percentages following from it provide a suffi-ciently accurate average. A wordplay fragment that can be analysed in two different ways (it is both a substitution and a morphological pun, for example) counts as two puns, not as one.
Almost half of these 183 puns are polysemic or based on dual actualization (24% and 22%, respectively). They constitute the two biggest wordplay groups, followed at some distance by the third group, substitution (16%). All the other types of wordplay are under 10%, with — for instance — morphological and paronymic puns both making up 7% of the total amount of wordplay, homophony 5%, and nomen est omen 3%. Zeugma, homography, and insertion appear only once throughout the whole novel.4
It is interesting to note that there are 35 cases where more than one pun applies to the same fragment, where by far most combinations only appear once (dual ac-tualization + homonymy and morphology + polysemy, for example). There are two cases where the combination appears twice (substitution + paronymy and nomen est omen + paronymy), and substitution + morphology appears three times. The combination dual actualization + polysemy, however, appears 14 times, making up 40% of the total amount of combined puns. This may not seem that surprising, see-ing that these two types form the two biggest groups of wordplay; but it does show that Pratchett is very fond of this technique. He may as well have combined dual actualization with paronymy, for example, but all dual actualization combinations bar one (dual actualization + homonymy) are complemented by a polysemous word. Of all the puns in the novel, 87% is vertical and 13% horizontal. This can be ex-plained by the fact that the vertical axis dominates in nearly all types of wordplay; horizontal wordplay in the two big groups dual actualization and substitution, for example, makes up only 8% and 3%, respectively. Zeugma is the odd one out, for it exists by virtue of its horizontal nature; yet there is only one zeugma in the whole novel, so it does not really make a difference. The reason why there is still 13% hor-izontal wordplay is because of the large number of polysemic puns, of which 25% is horizontal.
