Translation of Word- and Language-Play

Definition

Translation primarily refers to interlingual translation here, i.e., the transfer of texts (broadly understood) between distinct languages. Wordplay is used synonymously with puns, while language-play is a broader concept referring to the intentional manipulation of linguistic elements for special effects. Language-play thus includes, but is not limited to, wordplay. This paper provides an overview of the issues involved in the translation of language-play, including wordplay. To this end, definitions, categories, and arguments related to translation on the one hand, and to word- as well as language-play on the other, are presented. This includes an introduction to different modes, purposes, and challenges of translation, as well as a discussion of the possible forms, functions, and effects of language-play, partly in relation to humour. Based on this groundwork, claims and findings related to the combination of the fields, i.e., the translation of word- and language-play, will be summarized. The paper concludes with an attempt at placing word- and language-play translation into a broader context, including societal and educational concerns, as well as the latest changes affecting the translation profession.

1. Outline of the Paper and Frequently Used Abbreviations

The topic of this entry paper features three constituent concepts—translation, as well as word- and language-play—and these first need to be introduced and discussed in their own right, which will happen in Section 2 and Section 3, respectively, before they can be brought together in Section 4. Section 5 will provide concluding remarks and an outlook.

The following abbreviations are established in the academic field of Translation Studies and will also be used in the present context:

ST:

the source text, i.e., the original text in the original or source language, that serves as the basis for a translation

TT:

the target text, i.e., the resulting product of the translation process in the target language

SL:

the source language, i.e., the language of the ST

TL:

the target language, i.e., the language of the TT

2. Translation

This section is intended to set down what is generally meant by translation in broad terms, including some of the professional and academic distinctions made in the area (Section 2.1), as well as a number of common issues and challenges encountered by translators and students of translation (Section 2.2). (For a more comprehensive overview of translation, its history, and its study, one of the classic introductory textbooks, such as ref. [1], ref. [2], or ref. [3], can be recommended. Collections of influential and/or thought-provoking essays and studies, such as ref. [4], as well as encyclopedias and handbooks, such as ref. [5], or ref. [6], will provide additional insights.)

2.1. The Concept of Translation

The English word translation is a borrowing from French and Latin, as explained by the Oxford English Dictionary [7]. According to the same source, the word’s etymological origin in classical Latin generally referred to the “action of moving a thing from one place to another”, as well as to various more specific changes and movements, both concrete and abstract, including the “action of translating from one language to another”. The latter is also the predominant meaning of translation in Modern English, together with the outcome of such an action. (As to words for translation in languages other than English, their underlying, etymological roots generally seem to be quite in line, in terms of meaning, with those of the English term [8,9].) Both of the modern senses of translation, i.e., ‘process’ as well as ‘product’ related to the rendering of linguistic units (from words to texts) between different languages, will be in focus here.

That said, even within Translation Studies, as the academic field dealing with the phenomenon came to be called after a proposal by James Holmes [10], various types of translation have been defined, as in the classical distinction by Roman Jakobson between interlingual, intralingual, and intersemiotic translation [11]. Interlingual translation refers to the prototypical scenario of making linguistic material in one language available in a distinct other language, and it is this type that Translation Studies is almost exclusively concerned with. By contrast, intralingual translation takes place within a language, between two dia- or synchronic varieties, to render a text more accessible to an intended audience that would be struggling with the original version (as in the case of ‘modernized’ versions of the classics, or of news items ‘translated’ from a formal, standard variety into a simplified one). Intersemiotic translation, finally, overlaps with the more general idea of translation as a movement from one state to another, as it encompasses the transformation (which would be a possible alternative term) from one sign system, e.g., a written text, into another, e.g., a picture or film [11] p. 233.

Interlingual translation, i.e., the type of interest in the present context, can itself be divided into various categories, each with its own traditions, challenges, and norms when it comes to translational behaviours and outcomes. Translation norms can differ between geographically or socially defined cultures and depend on the type of material translated. (For some classic publications in this regard, see ref. [12] on the possible influences exerted on and through (literary) translation in different cultural systems, as well as ref. [13] and ref. [14] explicitly on norms in translation.) As far as fields and genres are concerned, commonly accepted categories of translation and its study include literary, technical, legal, business, and audiovisual/screen translation. The translation of the Bible (and other religious texts) constitutes its own subfield(s); the same can be said for theatre, opera, and song translation. Notably, although interpreting is often treated as distinct from written translation, the rendering of speech into another language is a form of translation as well—it also can take various forms, e.g., synchronous vs. consecutive interpreting, depending on the context (see, e.g., ref. [15] for an introduction and overview).

In reality, translation types and categories are often fuzzy and overlap. For example, the terms audiovisual and screen translation as such refer to a medium (or a limited range of media), but the actual content of what is being translated (e.g. a romantic comedy vs. a documentary) can be more or less literary vs. technical in nature, and the translation can involve different semiotic channels (see ref. [16] for an attempt at classifying subtitling, in contrast to e.g., dubbing, in this regard). In fact, the multimodality of much screen material (including computer games, websites and other digital content), which tends to combine writing with speech, sound effects and/or images, complicates the matter of translation further, and so do some specific textual features that may occur in almost any kind of source material (such as proper names [17], metaphors [18] or, as in the present paper, word- and language-play). Such features may constitute their own foci in studies on translation.

Additional distinctions in the field are based on the insight that texts in general and translations in particular can have different forms and functions, and that the purposes of a translation need not be the same as those of the original [19,20]. This aspect touches on the question of where translation proper ends and re-writing and re-creation begin. What freedoms can translators take without their work ceasing to be translation? One approach in view of the virtual impossibility of distinguishing in objective terms between, on the one hand, supposedly accurate translation and, on the other hand, freer, more creative or—to critically put it—less faithful transformations, such as ‘adaptations’ or ‘re-creations’, is to leave it to the stakeholders more immediately involved, notably the translators themselves and their clients, to decide whether what they are conducting or dealing with is translation or rather something else. (This would be in line with a common practice in the study of language(s), namely to let speakers of a variety decide for themselves whether what they are communicating in is a language in its own right or a variant (dialect) of some broader variety (language) [21] p. 64.)

The question of how ‘faithful’ different approaches and solutions are or can be, in translation contexts, constitutes a fundamental issue in both practical terms and in the academic subject, and it is relevant not least when it comes to the translation of word- and language-play specifically, as will be seen.

2.2. Issues in Translation

2.2.1. General Constraints, Expectations, and Possibilities

Some of the issues related to translation, including those touched upon at the end of the preceding section, have been addressed since antiquity (Cicero and St. Augustine are among the chronologically more distant figures frequently referred to in historical overviews of Western translation traditions, such as [22]), while other concerns and challenges are more recent, reflecting theoretical as well as technical developments in modern times.

In what used to be a prototypical translation scenario, where a human translator is exclusively or primarily responsible for the transfer of material—or a text, broadly understood—from a source language (SL) into a target language (TL), a fundamental decision to be taken by that agent is how close the resulting target text (TT) should remain to the source text (ST), in terms of form, contents, as well as function or effect. In other words, what liberties, if any, can, should, or must the translator take with, for example, word choices and grammatical structures (form), informational content, and the overall intended experience by and effect on the TT recipients, compared to what can be assumed to apply to the consumers of the ST? Frequently, a distinction has been made between a) staying close to the ST in the translation, with all the challenges that this may entail for the TL audience (‘moving the reader to the writer’, in German philosopher Friedrich Schleiermacher’s metaphor), and making adaptations that render the text more accessible or palatable to the TL audience (‘moving the writer to the reader’; see, e.g., ref. [2] for an introduction).

Of course, an extreme word-by-word translation, with strict adherence to original word order, would be very hard to process and understand by target readers, at least if they are unfamiliar with the SL, which can be assumed to be the normal case and the reason why translations are required in the first place. A more commonly adopted approach is therefore to translate sense-for-sense, making the necessary sacrifices to the differing linguistic systems of SL and TL, but conveying the original ideas as well as possible. ‘Change as little as possible, but as much as necessary’ is a general guiding principle likely to be followed by many translators, including the author of the present paper, though it obviously does not clarify in itself where to draw the line for what is or can be considered ‘necessary’, and this is precisely what much of Translation Studies is about.

For example, the inescapable constraints imposed on the translator by the SL and TL linguistic systems in terms of grammar and lexis, which may differ significantly in various respects even between languages within the same language family, will still leave room for TT formulations that seem vaguely non-standard or exotic in the TL (intentionally marking the translation as a translation, if that should be part of the desired effect)—or else for a range of perfectly standard TL solutions that nevertheless will carry subtly distinct nuances in meaning or style. In other words, translation always entails choices among different approaches, strategies, or procedures in the process, and among various conceivable solutions in the product.

2.2.2. More Specific Constraints and Challenges

Different languages also reflect the cultural differences between their speakers and the geographical areas where they are spoken, and an obvious type of translation challenge in this regard is how to deal with references to phenomena specific to the source culture but unknown in the target culture, such as the names of certain institutions, food, or items of clothing. Direct loans, explicitations and omissions are among the possible translation procedures that might be adopted (see, e.g., ref. [23] pp. 201–207 or ref. [24] pp. 74–76 for an overview), and as will be seen in Section 4.2.3, these procedures overlap with what can be conducted with word- and language-play as well. Relatively complex vs. relatively simple systems of address terms, as in Japanese vs. Swedish, represent another well-known, culture-bound challenge in translation. See ref. [25] for issues in this regard—even between rather closely related English and Swedish, both of which have comparatively simple rules for addressing others—and for how these issues can be and have been dealt with in translation contexts.

The communicative channel is another factor to consider in translation practice and analysis, and one that comes with its own restrictions and requirements. Not long ago, the typical translation product—or at least what the average person would think of spontaneously when confronted with the concept—might have been a text on a printed page, say a novel for humanists and/or free-time readers, or a manual for engineers and/or (other) users of technical devices. However, screen translation, notably subtitling, dubbing, and voice-over, has been a mainstream phenomenon for a century now and, in fact, a more commonly consumed form of translation in many places than printed material. (Gottlieb confirmed this to be the case in Denmark already in 1993 [26] pp. 151–153). The point here is that translators of traditional novels enjoy much greater freedom from medium-based constraints than do dubbers or subtitlers, who, among other things, need to take into account the connection of the spoken word to what is shown on the screen (including lip movements in the case of dubbing, if the target version is supposed to feel natural, or the physical space available for subtitles). Subtitles are also particularly open to criticism, in that any viewer could, at least in theory, compare them to the original soundtrack [27]. Translation of other multimodal material, such as drama, comics, adverts, webpages, or video games, is associated with partly similar, partly additional constraints, as hinted at already in Section 2.1. In opera and song translation, the TL choices need to fit the melody [28]. In poetry translation, the preservation of the metre or other formal aspects, or of cultural connotations, may be a major concern or challenge [29].

Yet another notable type of ‘issue’ in translation, as briefly mentioned above, is the intended function or effect of the TT. The default situation may be that the purpose of the translated text is essentially the same as that of the ST (e.g., entertainment—and ultimately sales of copies—in the case of novels, or safe handling of devices or machinery—and ultimately the avoidance of lawsuits—in the case of manuals), but the aim might also differ, as in the case of a ‘translation’ in the form of a simplified version of a short story for language learning [30] pp. 223–226, 247–250. Conversely, where the channel and overarching aim are the same for both ST and TT, as in the case of commercials and advertisements intended to increase sales of a product or service in different language communities, it is this sameness that may constitute a constraint if one way of approaching potential customers (e.g., through surprise effects, humour, or familiar forms of address) works in one community but not the other. (See ref. [31] on advertising slogans translated from English into Vietnamese.)

A reminder is in order, furthermore, that it is far from always the case that an individual translator is solely responsible for the final product. After all, there is usually a client who desires a translation to begin with and who thus initiates the process, and who may well want to have a say in how it is to be carried out and what the outcome is supposed to look like. There may also be guidelines to be followed, e.g., in a publishing house, and editors ‘correcting’ the translator’s work. In addition, the author(s) of the original material and/or the readers of the translated text may have opinions on the overall quality of the translation or on individual choices, which may lead to revisions in subsequent editions, if any. In other words, many stakeholders can influence translation both as a process and as a product—and in terms of what gets translated to begin with [32].

Finally, the skills and working conditions of the translators themselves must not be forgotten as a potentially decisive factor when it comes to translation quality [33] pp. 362–363. Translators can be assumed to be very proficient in both the SL and the TL, and while this may generally be the case, it is still a relative matter. Any translator could be somewhat better in one of the languages than the other, leading either to misunderstandings of the source material (and thus translation errors) or unintended so-called Translationese in the TT. In fact, mere proficiency in two languages does not guarantee good skills in translating between the two, even though people who are not involved in the field may think otherwise.

As to the circumstances under which translators (have to) perform their job, they, too, will play a role [34]. These include overall workload, time frames and deadlines, payment, and access to aids and resources. Even the physical workplace itself (e.g., a noisy and cramped apartment with a laptop literally on the lap, compared to a quiet office with large screens and competent colleagues next door to turn to for help) will make a difference, either directly or because it affects the motivation to do a good job. In times of artificial intelligence (AI) having entered both private and vocational contexts for real, the translator’s working conditions are quickly and drastically changing, and customers’ willingness to pay for their services is likely to be diminishing, ultimately threatening the profession as such (see also Section 5 below).

3. Word- and Language-Play

As a second prerequisite for discussing the translation of word- and language-play, after introducing translation as such, it needs to be established what kinds of phenomena the labels wordplay and language-play are taken to refer to. However, the discussion also requires an introductory note on a basic aspect of these labels, which take the form of compound nouns, namely their spelling: While wordplay (usually written in this concatenated form) is an established term, with an entry in most dictionaries of the English language, including [7], language-play (however spelled) is not. While there is a certain tendency in English for short, frequent and/or well-established compounds to be written in a “solid” form, i.e., in one word, and long, rare and/or more recently created compounds in an “open” form, i.e., with a space ([35] [36] (p. 1569)), the not-uncommon intermediate alternative with a hyphen, in this case language-play, is adopted here to reflect the relative infrequency of the term at large at the same time as its central role in the present context, as well as its parallel structure to wordplay.

With this decision on spelling having been clarified, the following subsections introduce the difference as well as overlap between the two key concepts (Section 3.1), provide overviews of basic distinctions regarding wordplay (Section 3.2) and language-play other than wordplay (Section 3.3), and elucidate how these concepts relate to humour (Section 3.4). For a more in-depth treatment, see, e.g., refs. [33,37].

3.1. Language-Play and Wordplay vs. Non-Playful Language Use

To start with, it should be acknowledged that word- and language-play, just like translation, can be both a process and a product, or an action as well as a result [38] (p. 47). That said, the focus in the present treatise will mostly be on linguistic play as an outcome, as this is what can or will normally be translated, in contrast to processes.

In either case, however, the view adopted here is that wordplay is a subcategory of language-play, meaning that the latter is the broader concept. In [33], the discussion of how language-play (as a product) might be defined culminated in the following proposal:

Language-play, contrary to normal, or non-playful, fragments of conversation or writing, is marked in the sense that the linguistic building blocks involved draw attention to themselves and their form, in addition to functioning as transmitters of content. In other words, language-play is present where the peculiarities of a linguistic system (or linguistic systems) have been exploited in such a way that an aural and/or visual (and by extension: cognitive) effect is achieved that would not be present, and perhaps consciously avoided, in language used with a focus on propositional content.

pp. 78–79.

As acknowledged already there and then, this definition is rather broad and does not permit an easy distinction between language-play and non-playful use of language. Borderline cases are inescapable, part of the reason being that it cannot always be determined with a satisfactory degree of certainty whether a seemingly playful instance was intended as such, i.e., whether it was indeed the result of play, or perhaps just of a blunder or coincidence. For example, rhymes draw attention to themselves and to the fact that the latter parts of two (usually unrelated) lexical items are phonologically—and often orthographically—identical. Where a rhyme was devised on purpose, as in many slogans, poems, and song lyrics, it represents a playful manipulation of linguistic building blocks for special effect, as in the line from an ABBA song quoted in Example 1:

(1)

Money, money, money /ˈmʌni/

Always sunny /ˈsʌni/

In the rich man’s world

However, a rhyme may on occasion be produced unintentionally in conversation, perhaps being noticed afterwards, perhaps not. Unintended juxtapositions of rhyming items, as in the made-up utterance in Example 2, are not playful per se:

(2)

Have you got some money /ˈmʌni/ so that I can buy more honey /ˈhʌni/? Oh, that actually rhymed!

According to Lecolle, “wordplay is not just a fact; it has to be recognized” [39] (p. 63), and the same can be said about language-play more generally. This is in line with the insight that (what looks like) language-play, including wordplay, can have been produced unintentionally on one level but be or become intentional on another. If someone inadvertently utters something that could be construed as, say, a double entendre by listeners so inclined, it is not language-play, but if it is then reacted to, repeated, discussed, and/or expanded on as a play with meaning and/or form, by others and/or the original producer, it becomes language-play. Consider Example 3 (this one also having been invented for illustrative purposes), where speaker B employs a formula, with an emphasized she, that is commonly used to turn an innocuous comment, like the one made by speaker A, into a sexual allusion:

(3)

A: This is bigger and harder than I expected.

     (commenting on the unexpected scope of a challenging task)

B: That’s what *she* said.

In a related vein, a character in a novel, play, or film may be made to utter malapropisms, such as Mrs Malaprop in R.B. Sheridan’s play The Rivals, who regularly produces word choice errors such as the one in Example 4 (where the spelling pine-apple, with a hyphen, has been copied from the script):

(4)

He is the very pine-apple of politeness. [instead of pinnacle]

The Oxford English Dictionary [7] defines malapropisms as “The ludicrous misuse of words, esp. in mistaking a word for another resembling it; an instance of this.” The very label for the phenomenon has been derived from the name of the character mentioned. In the case of malapropisms in The Rivals, they are errors—and thus not language-play—on the level of the theatrical role or plot, as in most ‘real-life’ cases. Yet the playwright intentionally had his character produce the malapropisms to make her appear pretentious and uneducated, and ultimately to achieve a comic effect on the audience, which makes these instances playful on a higher level.

Note that the difference between characters’ behaviour and authorial intention when it comes to language-play in scripted material is related to, but not the same as, the distinction between covert and overt word- and language-play. In some cases, this may play a role in the choice of translation procedure and in whether the instance is recognized/recognizable as playful by the translator to begin with (see Section 4.2.2 below).

The above examples of rhymes and malapropisms, as long as they have been produced on purpose, can also serve to illustrate the difference between language-play that is not wordplay, and language-play that is: Rhymes like the one quoted in Example 1 only or predominantly rely on form, creating a presumably pleasing sound effect, and should the lyrics cause listeners to become conscious of the fact that money and sunny actually rhyme (at least in some varieties of English, including the one employed by ABBA), there is not much more to it than that. By contrast, wordplay relies on form as well as meaning. Therefore, in the given context, if a certain word that would be expected is intentionally replaced by a more- or less-similar-sounding one that causes listeners or readers to reassess the described situation (in Example 4 from the Sheridan play, perhaps visualizing the person talked about as a pineapple), the result is a wordplay or pun. Note that wordplay and pun(s) are used as synonyms in the present context, in line with Delabastita’s standard-setting publication on wordplay translation [40].

Here is a definition of wordplay proposed by Delabastita:

[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 structure of the language used (level of competence or langue) are exploited in such a way as to establish a communicatively significant, (near)-simultaneous confrontation of at least two linguistic structures with more or less dissimilar meanings (signifieds) and more or less similar forms (signifiers).

[40] p. 57.

In this definition, the crucial elements distinguishing wordplay from the broader concept of language-play emerge in the last two lines, notably the “confrontation of … dissimilar meanings”, which goes beyond the mere play with sound. This will be explored in more detail in the next section.

3.2. Wordplay

As an illustration of what forms wordplay can take and how it works, consider Example 5, featuring a classic joke:

(5)

I was wondering why the baseball was getting bigger and bigger. Then it hit me.

Apart from the ambiguous referent of it, the pun—and the effect it may trigger (see 3.4 below)—depends on two meanings of hit that are invoked simultaneously by how the context has been set up: on the one hand, a baseball is mentioned, triggering in the listener’s or reader’s mind an image schema of the item and of the activity with which it is associated; on the other hand, the realm of cognition and (lack of) understanding are evoked through the word wondering. These two contexts are confounded by the punchline, where two relevant meanings of hit come into play at once: a concrete, literal meaning (as in ‘being hit by a flying object’), and an abstract, metaphorical one (as in ‘suddenly realizing something’). A crucial aspect of this particular joke is that being physically hit will lead to understanding, though, one imagines, at the cost of sudden and severe pain.

Just as puns constitute a subcategory of language-play, they can themselves be classified according to different criteria that have been shown to have some impact on how puns are dealt with in translation. One distinction regards the question of whether the key item or items feature twice or only once in the playful sequence. In the baseball joke above, the pivotal hit only occurs once, making the pun a so-called vertical one. It could be turned into a so-called horizontal pun if hit were mentioned twice, with the two contrasting meanings dominating one occurrence each. Example 6 constitutes an alternative formulation of Example 5 along these lines:

(6)

I was wondering why the baseball was getting bigger and bigger. It hit [‘became clear to’] me when it hit [‘physically struck’] me.

While the effectiveness or ‘punch’ of the revised punchline could be debated, it ought to work in principle, in the sense that the two meanings of hit (and the two referents of it) are triggered “(near)-simultaneously” (see Delabastita’s wordplay definition above). In fact, the unusual and seemingly redundant or tautological construction "It hit me when it hit me," which turns the pun into a horizontal one, may make it more likely to be recognized as a pun to begin with. For an extended discussion of vertical vs. horizontal puns, including some special cases, see ref. [33] pp. 160–163.

A second set of wordplay categories can be based on the nature of the relationship between the forms and meanings of the pivotal items. In Examples 5 and 6 (keeping the discussion simple by disregarding the ambiguous it), we are dealing with one form, i.e., hit, that carries—and evokes—two related meanings. Puns based on this particular pivot can therefore be labelled polysemic.

Polysemic means approximately ‘with many signs’ or ‘with many meanings’, in terms of its Greek-based constituent elements, but is used here to refer to puns based on polysemous items, i.e., forms that have two or more related meanings. The sense of hit as suddenly being struck by an insight is almost certainly derived from—and so clearly related to—the physical sense, in what might have been a creative metaphor originally, though it is probably not perceived as such in current everyday language use. Competent modern speakers of English would, in any case, see a connection between the two meanings of hit at play here, if made aware of the issue.

Polysemic puns can be contrasted with homonymic ones, where two unrelated meanings are triggered by one and the same form. Homonymic puns are much rarer than polysemic puns, at least in the data used in ref. [33], presumably because homonymy, thus defined, is itself much rarer in language than ubiquitous polysemy [41] (p. 245). Consider Example 7, which has been devised with a view to expanding on the baseball theme:

(7)

Why did the animal rights activists sue the baseball player?—Because he hit the ball with a bat.

This pun depends on the juxtaposition of animal (rights activists) and baseball (player), as well as additional items such as sue and hit, which together create overlapping contexts and thus simultaneously trigger the double interpretation (‘nocturnal flying mammal’ vs. ‘striking implement’) of the pivotal bat, which is a homonym in the sense adopted here, namely as having the same spelling and pronunciation for two seemingly or factually unrelated meanings.

It should be acknowledged that other definitions of homonymy, whose Greek-based constituents approximately mean ‘same name’, do exist. These definitions may, for example, rely on pronunciation only and not take into account the spelling and/or whether the meanings are related or not. Here, I follow the criteria employed, e.g., in ref. [42] (p. 206) and ref. [33] (p. 164). For a more complete analysis of Example 7, note also that sue is homonymous with the name Sue, as both spelling and pronunciation are identical (disregarding the capitalization of S in the name) while the meanings are entirely unrelated. However, neither the polysemy of hit nor the homonymy of sue/Sue has been exploited in this particular pun, which centres on the (vertical) use of bat and its two perceptually and etymologically unrelated meanings.

Apart from polysemic and homonymic puns, additional distinctions based on form can be made. In particular, one might wish to distinguish puns based on homophony and homography, respectively, i.e., on key items that are identical in pronunciation but not spelling (homophony) or vice versa (homography). While homographic puns thus defined appear to be extremely rare and therefore perhaps negligible from a translation perspective, homophonic ones are somewhat more frequent, at least in English (again, see ref. [33] for some numbers, including a couple of homographic puns on (p. 166)). Example 8 features a homophonic pun, adapted from an AI-generated suggestion:

(8)

When the police came to the daycare centre, several three-year-olds resisted a rest.

The pivotal item in Example 8 is a rest, which is homophonous with arrest, and due to the difference in orthography, this pun might work best in writing. If it were spoken, the potential double meaning of /əˈrɛst/ might well be missed. In fact, the pun would also be more likely to be missed, even in print, if it were written … resisted arrest, since the signals triggering the unusual alternative interpretation ‘resisted a rest’ might not be strong enough. This is a reminder that not all puns are equally noticeable, even when intended as such, which will affect their treatment in translation (see Section 4).

Note, furthermore, that the AI suggested a pun based on a rest/arrest in response to a query asking for a homonymic pun! This thus confirms that different definitions of homonymy may be applied by different people and, apparently, non-human intelligences.

A final form-based category of wordplay ought to be mentioned, namely paronymic puns, where the key items played with are identical neither phonetically nor orthographically, but merely similar. Example 4 above, based on pine-apple vs. pinnacle, is an example of that.

3.3. Language-Play Other than Wordplay

As mentioned, language-play comprises wordplay, but also other playful manipulations of linguistic elements and features. In listeners or readers, provided they notice that something in what they are exposed to goes beyond run-of-the-mill language use and that a playful intention appears to be behind it, this will cause a certain effect that involves increased attention to the elements and features that have been manipulated.

In contrast to wordplay, however, other types of language-play do not necessarily rely on play with both form and meaning, though in some cases they do. Mostly, non-punning language-play relies on form only, notably sound, but even where play with meaning is involved, differences from prototypical puns will be in evidence (see below).

Starting with exclusively sound-based language-play, rhymes constitute a common and comparatively well-defined category. As illustrated by the pair, money–sunny, in Example 1, rhyming relies on the identical pronunciation of two items from the last stressed vowel onwards, i.e., the /ʌni/ part in the example pair. In languages with consistent relationships between pronunciation and orthography, the phonologically identical parts of rhyming items will normally be spelled the same as well, but in languages like English, where many sounds can be spelled in different ways and where many letters and letter combinations can represent different pronunciations, this is far from always the case, as confirmed by Example 1.

Rhymes are very common in many song lyrics as well as poetry, where they may even follow consistent rhyme schemes, such as ABBA. (The pun on the name of the pop group behind Example 1 was intended in this case. For a practical introduction to rhyming and rhyme schemes in rap songs, see ref. [43].) Rhymes can also render catchphrases more memorable, as in Example 9, which quotes a sign of the US Occupational Safety and Health Administration, featuring rhyming items based on /зrt/ (in Standard American pronunciation) at the end of the second and third line, respectively:

(9)

  SAFETY FIRST

    STAY ALERT

DON’T GET HURT

To be clear, there is a meaningful relationship between staying alert and not getting hurt, but such a relationship prevails in all language use that is not pure nonsense, and the meanings of alert and hurt are not played with as such in the quoted slogan, only their forms, by including them both in prominent places in a short text, whereby their similarity is highlighted.

Other plays on form include half-rhymes, as in one of the verses of Dido’s song Mary’s in India, quoted in Example 10:

(10)

Danny came over last night and I cooked for him

He talked about you, Mary, and how much we loved you still

He told me he’s packed up your books and your letters and things

As the sun sets on Mary, it’s rising on him

In contrast to rhymes, so-called half-rhymes are not very clearly defined (they are not literally ‘half’ a rhyme, whatever that might be), but they fulfil essentially the same function as rhymes, particularly in song texts, by linking lines and items within lines through sound similarity. In Example 10, it is the items him, still, and things (and then him once more) that tie the verse together by being vaguely similar in sound. In fact, apart from the fact that they occupy a prominent place at the end of their respective line (which half-rhymes, like rhymes, tend to do in poetry and song lyrics), the overlap consists only in the vowel sound /ɪ/ and the fact that the words consist of just one syllable. Still, it seems there is no coincidence that each of the four lines ends with such a word, i.e., that they have been intentionally placed there to achieve a certain aesthetic effect. (Arguably, a half-rhyme is also present in Example 9, between first on the one hand and alert as well as hurt on the other. The true rhyme in lines 2 and 3 could be taken as evidence both for and against the idea that a half-rhyme might have been intended as well.)

One can also play with sounds in other ways, one of which consists of intentionally juxtaposing words that begin with the same sound or sounds—this is one possible, if rather informal, way of defining so-called alliteration, which in Example 11 is illustrated with a line from Frank Sinatra’s song That’s Life:

(11)

I’ve been a puppet, a pauper, a pirate, a poet, a pawn and a king

Note that it is quite unusual to have five words starting with the same sound in one line (and the same letter, too, though that is strictly speaking not a requirement for alliteration defined in phonological terms) and that placing only two such words close to each other can be enough to achieve a playful effect – just as only two rhyming or half-rhyming items can be enough. On the other hand, alphabetical ordering in dictionaries, indexes, and other lists is just a practical convention and quite the opposite of playful.

There are many additional types of language-play that can be distinguished, but presenting and discussing each of them in detail would go beyond what is required in the present context. Suffice it to say that one can play with language by, for example, repeating items in ways that draw attention to themselves (money money money at the beginning of Example 1 could qualify as such); one can purposefully break the rules of grammar for special effect; or one can create playful neologisms or nonce words that may or may not be given clear definitions. The first two lines of Lewis Carroll’s famous poem Jabberwocky, quoted in Example 12, feature no fewer than six such nonce creations (with the name of the poem constituting a seventh):

(12)

‘twas brillig, and the slithy toves

Did gyre and gimble in the wabe

Also, one can mix, in one and the same conversation, utterance, or even phrase, different languages or varieties of languages, e.g., English and French, or standard and non-standard (perhaps regional) dialects. Needless to say, as with other types of linguistic manipulation, the dividing line can be blurry between playful “stylisation” or “performance” on the one hand and, on the other, non-playful language use such as habitual and perhaps even subconscious code-switching or code-mixing. (For introductions to these concepts, see, e.g., [44] pp. 42–54, 351–353.)

As a final concrete instance of language-play, consider Example 13, created specifically to make a point here:

(13)

A: Who do you think you are!?

     (reacting to subordinate person B being out of line, in A’s opinion)

B: I’m Sam Smith, second assistant to Kim Jones. I thought we had met.

In modern standard usage, the fixed formulation Who do you think you are, especially if said in an aggressive tone and possibly a loud voice, is special in so far as it is the rhetorical or metaphorical rather than the literal interpretation that would first come to mind among proficient speakers of English without a language disorder. It has simply become a formula to express displeasure with somebody’s behaviour, and especially the perception that the addressee overstepped the boundaries of what is appropriate for their social standing. If, contrary to expectations, the addressee interprets such a confrontational move as if it were a genuine request for information, it becomes odd (and might infuriate the challenger even more). However, a genuine misinterpretation, perhaps due to difficulties in understanding non-literal language use, is not playful. (This is common among young children and some groups of people with neuropsychiatric disorders [45], p. 346). But if such a case has been created on purpose, for example in a script for a situation comedy or, as in the present case, to function as an illustrative example, it does constitute language-play–of the type that is not purely based on sound, but on a play with meaning.

Does Example 13 not represent a pun, then? If it is, it is certainly not a prototypical one, for where is the pivotal item with contrasting meanings triggered simultaneously by the linguistic co-text? The entire expression, Who do you think you are, would have to be considered polysemous, and the context causing the double interpretation would be the unexpected response. In [33], cases like these were subsumed in an unwieldy and ultimately perhaps unsatisfactory category entitled “Play with metaphors, similes, idioms, and related figures of speech”. An alternative label might be Play with pragmatics. However, pragmatics seems to be involved in all language-play—and in fact all language use —by default; thus, such a label would also be open to criticism. In any case, the point here, apart from the repeated warning that categories are fuzzy, is that some language-play can involve play with meaning without being an instance of more narrowly conceptualized wordplay.

3.4. Wordplay, Language-Play and Humour

An introduction to language-play, including wordplay, may suitably be concluded by addressing the question of how these phenomena relate to humour. The short answer, supported by Examples 1–13, is that they are not the same, but that there tends to be a connection. Simply put, there can be language-play without humour, and there can be humour without language-play. However, humour can and often does take the form of language-play, and language-play can and often does function as humour.

A couple of further clarifications can be useful at this point: The discussion here regards humour in the sense of ‘that which is intended to be humorous or funny’, meaning that humour is not synonymous with laughter (either), as the latter is a possible, but by no means given, physical reaction or effect, yet not humour in itself (unless it is somehow made the object of humour). In some languages, e.g., French, German, and Swedish, humour is (also) something an individual can have, as a mental disposition to appreciate that which is amusing, perhaps including the ability to purposefully cause amusement oneself. In English, such a disposition is normally referred to as a sense of humour, and this is not what is in focus here. For a fuller discussion, see [33].

One obvious difference between language-play and humour is that the former, by its definition and very label, is based on language use, while humour can be purely visual or situational (just consider, e.g., silent slapstick films, performances by mimes, or the picture of a person or scene that simply looks funny). Humour can also be based on or supported by music and other non-linguistic sounds, e.g., a car horn or a dog’s bark frightening a pedestrian. But even where humour is based on language, as in a traditional joke being told verbally, it need not involve language-play. Example 14 is a case in point:

(14)

Why are elephants wrinkly?—Because you can’t iron them.

Taken from an online collection of short jokes, almost all of which were ‘punny’, Example 14 represents an exception in that it does not rely on one or two identical or similar forms triggering more than one meaning. Nor does it involve other types of language-play. The humour derives, instead, from an unexpected answer to what appears to be a question inviting a scientific explanation. After all, a reference to the impracticality of applying a handheld domestic tool for smoothing clothes and other textiles to a large and usually wild animal’s skin is incongruous, and surprising incongruity is precisely what tends to characterize humour. While the joke may trigger amusing visualizations in the receiver’s mind (as may Examples 4–7 above, which were based on a person being described as a pineapple, or on conflicting meanings of hit and bat in a baseball context), there is no play here with specific linguistic forms as such.

Conversely, as we have seen in Examples 9–11 of purely sound-based linguistic manipulations, the intended effect of language-play, let alone the effect actually achieved, need not be that of humour. Rhymes and alliteration as such are not normally funny, even though in some quantities and/or in some contexts they may become so (if the amount and/or the occasion are experienced as surprising or incongruous).

While an in-depth discussion of the properties and possible purposes and effects of humour is beyond the scope of this paper, suffice it to emphasize that much language-play, especially but not only wordplay, is meant to function as humour. Usually, this will be for amusement, though as has been pointed out in the literature (e.g., [46]), some humour, notably if it is directed against others, can have other primary functions such as self-aggrandizement, subversion, mockery, or social control. There is also grim or black humour, where amusement, if it is involved at all, may be subservient to garnering strength or rallying forces in adverse circumstances.

The point here is the following: Humour has potentially significant functions in the contexts in which it occurs, and the same can be said for language-play. If humour takes the form of language-play, or if language-play contributes to the humour, the functions overlap. However, even if they do not, as in non-linguistic humour and non-humorous language-play, the intentions and effects may be significant and cannot simply be dismissed in, for example, a translation context.

4. The Translation of Word- and Language-Play

This section finally brings together language-play with its translation. It begins with additional clarifications, in Section 4.1, why the translation of language-play generally matters, though perhaps to different degrees in different contexts, and then proceeds with a consideration of how language-play can be and has been dealt with in practice (Section 4.2).

4.1. Why, Where, and How the Translation of Word- and Language-Play Can Matter

It could be argued that, ideally, a translated text should feature the same amount and types of language-play, in the same places, as the corresponding source text, provided there are no reasons to the contrary. In fact, since language-play (always including wordplay) is generally understood to come with an underlying intention or purpose—with chance outcomes of unreflected language use that look like possible language-play not counting as such—and since TTs are generally expected to remain as close as possible to their corresponding STs in both form and function, this would imply that not only the quantity, distribution and types of the ST language-play should be reflected in the language-play incorporated in the TT, but its roles and effects as well. For if the original author(s) felt the need or desire to incorporate language-play in the ST, and assuming that the actual effects on the SL audience match the authorial intentions, the TL audience should, in a standard translation scenario, be given the chance to experience the same effects, which will normally be achieved by the same or comparable means.

Of course, there may be various reasons not to accept this principle for a concrete translation assignment, e.g., because the function or effect of the TT is actually meant to differ from that of the ST, or because the same effect would have to be achieved by other means in the target culture than in the source culture. However, again, a 1:1 correspondence with regard to language-play (and, indeed, any aspect of the text) should be the aim when there are no such reasons to depart from it. That said, it is another matter, to be addressed in Section 4.2, whether such an ideal is possible to attain in practice.

At this point, a reminder is in order as to what the possible functions and effects of language-play may be. In the case of puns or wordplay, as well as some other forms of language-play, especially if they contain an element of surprise or incongruity, humour is likely to be the most common reason why instances were created and included in a text. Humour, in turn, can have various primary and secondary functions, including, of course, amusement, but also the release of tension, social management, criticism of people, concepts, or circumstances, etc. (see Section 3.4 above). Language-play can also be ironic (sometimes without being humorous at the same time), cognitively relaxing or challenging (as in language-based puzzles), attention-getting and/or retention-supporting (e.g., in adverts, slogans and mnemonic devices), distinguishing (e.g., for its creator, or a character, or a text as a whole), concealing (of ‘hidden truths’ in the face of power or censorship), etc. For a list of some of the possible functions and effects of language-play, see [47] pp. 140–141. For a similar list of wordplay functions, including the promotion of a philosophical or religious point, see [48] pp. 13–15.

None of these functions is trivial and easily dismissible a priori. Therefore, if a text to be translated contains language-play, a conscious decision as to its treatment—especially if this treatment should result in non-language-play—can be expected. Nor should quantity be a decisive factor: If there are only one or two instances in an otherwise non-playful text, maybe these instances are particularly significant and effective precisely because they stand out. Conversely, if there are dozens or hundreds of instances in the ST, each helps characterize it and create an overarching effect. As Landers claims with regard to “untranslatable” puns in the context of literary translation, “Maintaining the tone is the main concern” [49] p. 110. If this is the case, and if the primary function of ST language-play, and perhaps of the text as a whole, is to entertain and amuse, as in the case of a comedy, one could even argue that, ideally, there should be more (and perhaps even ‘funnier’) language-play in the TT than in the original. In general, the way language-play has been treated in translation, in terms of both quantity and quality, may function as an indicator of the overall quality of the target text in relation to the source text [47] pp. 149–150.

One way of assessing the quality of TT language-play (or of TT humour, for that matter, to the extent that the two overlap) is to try measuring its actual effect on a TL audience, preferably in comparison to the effect that the ST language-play has on an SL audience. While such reception studies require considerable time and effort compared to the more traditional approach of individual researchers analyzing and perhaps even evaluating TL solutions on their own, these studies can yield more robust results. At any rate, the actual reaction of consumers to translated products is worth knowing more about than is generally the case, as argued in [50], p. 122.

For example, one result that has emerged from a reception study [27] is that TL audiences seem to appreciate attempts at creating TL wordplay even in subtitling and even when they have a good understanding of the SL dialogue, which, after all, is still accessible in subtitled screen productions. This would thus disconfirm a speculation in [51] that the very accessibility of the original dialogue could be a reason not to create TL wordplay that is formally very different from the corresponding SL instance, due to the risk of confusion and cognitive overload. In other words, even in subtitling, which is a rather special and challenging translation context in terms of modalities and constraints, the (re-)creation of TL language-play may be the ideal solution to aim for.

Or consider another special case, namely that of translating written text into sign language. Incidentally, punning is productive and widespread not just in spoken and written discourse but also in sign languages. This means that written puns can be translated into sign language [52]; thus, there would be no reason in principle (though possibly in practice) not to attempt this in such a context either.

However, before addressing in some more detail how language-play can be and has been treated in translation, which will be the topic of the next section, a few more underlying reasons might be added for engaging in the activity to begin with and/or for making an effort when faced with language-play in a translation context. Regarding the question of effort, it should perhaps not be underestimated what satisfaction a translator can derive from hitting upon a TL equivalent that ‘works’ in the TT and for the target audience, especially when a more or less direct transfer from the SL was not possible (without loss of the playful aspect), and a solution required considerable ingenuity. A personal success story along these lines, in the context of literary translation, is told by Landers [49] p. 110, including the pride felt by him after having come up with an adequate TL counterpart to a crucial pun. Similarly, contentment with having ‘solved’ an, at first sight, seemingly impossible task (translating concrete poetry from Chinese into English) shines through in [53]. Many seasoned translators of texts that are not purely factual or technical may have had comparable experiences, and while hardly of concern for clients and consumers, the translator’s own sense of achievement should not be disregarded as a reason to treat language-play with respect in translation.

At least two more profound types of reason can be adduced, however: political and educational ones. Regarding the former, already hinted at above, it may be the case that a certain element in a text, such as non-conventional language use, including language-play, carries or supports a message that deserves to be preserved in translation. This could be criticism of perceived societal shortcomings, calls for change, or the like. A case along those lines is described by Bouton-Kelly, who addresses the potential political implications of poetry translation, in a (post-)colonial context, and then proceeds to give an account of her own translation into French of a Jamaican poem from colonial times, written by a woman in the local Patois, that requires various ‘playful’ solutions in an attempt to carry over some of the non-standard features in the original and, by extension, some of its political import [54]. Arguably, any language-play that makes recipients think about and perhaps reconsider issues beyond language as such carries a weight that may deserve extra-careful treatment in translation. (See also [55] p. 181 for a brief discussion of the possible functions of puns beyond purely meta-linguistic effects, e.g., “socially transgressive” ones, which in turn may affect their treatment in translation.)

As to educational reasons for taking the translation of language-play seriously, and perhaps letting students make their own practical and self-critical attempts in this regard, it could of course be claimed that any way of raising awareness, for example, in the kinds of political contexts just mentioned, can be considered ‘educational’ in a broad sense. However, a more prototypically educational motive to analyze and possibly actively engage in the translation of language-play is to boost linguistic and (inter)cultural understanding and proficiency. While the traditional grammar-translation method in foreign-language teaching and learning came to be viewed with suspicion and even disdain, due to its limited potential for generating motivation, enjoyment, and communicative competence in learners [56] pp. 16–17, translation activities of the more communicative, engaging, and relevant types have recently been promoted again as a rewarding strategy in language education. Laviosa does this specifically for language-teacher and translator training at the university level, with an emphasis on a holistic perspective [57]. Pavan also let (university) students do and discuss a complex translation task with various challenges in the form of multimodality (interrelated text and image), culture-specific references, stereotyping, dialectal variation, as well as word- and other language-play. While acknowledging linguistic skills and competence as a necessary basis in the endeavour, she particularly emphasizes the cultural aspects of translation and communication across language barriers and the value of raising awareness in this regard [58], p. 143. In fact, there seems to be no obstacle in principle why these aims and approaches could not be extended and adapted to, e.g., upper-secondary language classrooms as well.

Of course, analyzing language-play and jokes in general, without an explicit focus on translation, can serve to introduce, raise awareness of, and practice the application of linguistic concepts and processes. Aarons shows how this can be done for anything from grammar to semantics and pragmatics, mostly in monolingual contexts [59]. Interestingly, however, some of her example jokes are in more than one language, and some are explicitly about language. From analyzing these, it is only a small step to considering language-based jokes from a cross-linguistic perspective, including translatability (which Aarons also does on occasion).

As a final remark regarding the question of “why, where and how the translation of word- and language-play can matter”, it can be pointed out that, occasionally, the translation is the language-play, as when false friends between languages, i.e., words that are similar or even identical in form but with diverging meanings, are purposefully employed for special, usually humorous, effect (see some of the contributions in [37])—an effect that, like the translation of language-play, can contribute to increased linguistic awareness and proficiency (see specifically [60]).

4.2. The (Un)translatability of Word- and Language-Play—In Theory and in Practice

4.2.1. Preliminary Considerations

Consider the following claim: “Most, if not all, linguistic jokes are untranslatable” [59] p. 12. Being highly relevant in the present context, this claim requires a number of clarifications (some of which are provided by the author herself) that will help put it into perspective and, in fact, cast some doubt on its veracity. These clarifications will be provided in the next few paragraphs.

Firstly, it needs to be emphasized that what Aarons calls linguistic jokes only partially overlaps with the concepts in focus here. To begin with, a joke is by definition intended to be funny, while this is not necessarily the case for wordplay, let alone language-play more generally, as discussed in Section 3.4 and Section 4.1. Also, while the author appears to understand linguistic jokes as being based on the playful manipulations of one or more linguistic systems, i.e., on word- or language-play, a joke may require more than the crucial, playful items at its core, e.g., a story leading up to a punchline. In any case, jokes that are not based on language-play, of whatever kind, like Example 14 above, are of no interest in the present context and not covered by the quoted claim either. In principle, provided that no cultural differences stand in the way, jokes relying on content rather than linguistic form should be readily transferable between languages. (That said, the culture-specific aspects of many jokes should not be underestimated, as emphasized, e.g., in the chapter on wordplay translation in [61], or in [62] with regard to the dubbing of jokes in situation comedies.)

Secondly, as Aarons herself points out, she is referring to “faithful translation, not translation equivalence” [59] p. 19 when she asserts the untranslatability of linguistic jokes. This means, presumably, that a direct transfer of the SL elements to the closest TL equivalents (whatever they may be) is not possible if, at the same time, the joke as such (the humour, the originally intended effect) is to be preserved. Since SL and TL systems are by definition different (or else we would not be dealing with different languages), and since these systems are arbitrary when it comes to the relationship between form and meaning (as established in [63] p. 90), play relying on form and meaning, as in wordplay/puns, is not translatable in the narrow sense. This is a standard line of argumentation leading to—or supporting—the kind of claim quoted above.

Thirdly, however, Aarons opens up for another understanding of translatable, namely the possibility of creating a TL counterpart that is somehow ‘equivalent’ to the original item (in her case, a joke; in the present context, an instance of word- or language-play). This makes a big difference, for with such an interpretation, there is a shift from a claimed or suspected zero percent translatability to a potential one hundred percent.

Fourthly, credit needs to be given to Aarons for softening her claim “that the fundamental arbitrariness of language ensures that a faithful translation cannot replicate the joke” by demoting it to the status of a hypothesis that is “open to empirical enquiry” [59] pp. 12–13. In other words, the theoretical possibility appears to be acknowledged that even a direct, uncreative transfer (i.e., “faithful translation”) may occasionally lead to an equivalent effect, i.e., a preservation of the joke as such, in the target language. That is, in fact, precisely what has been shown in some studies investigating the matter (see Section 4.2.3 below).

4.2.2. Factors Influencing the Treatment of Language-Play in Translation

It is difficult to estimate how language-play, including wordplay, has been treated in translation in general, as there are simply many factors involved. Arguably, the exact percentages when it comes to playful vs. non-playful TL solutions, including the various subcategories that can be distinguished, do not even matter much in themselves, except for the insight that they could indeed range from very little to very much TL play in relation to the play present in the corresponding ST. In theory, as pointed out in Section 4.1, a TT could even feature more language-play than its ST, meaning that some concrete translation choices are playful where the SL counterparts are not, because the translators, or their clients, aimed at a TT characterized by increased overall playfulness, e.g., to boost the chances of commercial success in the case of comedic genres.

Below is a list, likely to be non-exhaustive, of factors that can be assumed to influence how language-play has been—and is being—treated in translation. Note, however, that not all of these factors need to play a role in every concrete scenario, and that even those that do have an impact in a given case may exert their influence to different degrees:

• the type and genre of the ST (e.g., TV sitcom vs. business report)
• the part(s) of the text where the language-play occurs (e.g., title/heading/summary vs. body text; introduction, main part or conclusion; dialogue vs. description; special features, such as diagrams, images or songs)
• the mode or channel of the translation (e.g., dubbing vs. subtitling, with its differing affordances and constraints)
• the status of the ST (e.g., recognized/influential classic vs. niche product)
• the (assumed or explicitly intended) status of the TT (e.g., standard-setting and/or commercially successful vs. rough aid to understanding for a limited audience)
• (other) details of the task description (what, if anything, the translator has been asked to achieve by the client, if any, in terms of outcome and effect)
• the role and impact of various other players who may influence the final product (ST authors, editors, users of the TT adapting it for their ulterior purposes, etc.)
• general translation norms applying to the task at hand (in the cultural context in which it is carried out, including norms on language-play translation specifically, if any)
• (other) external factors pertaining to the translator (working conditions, including access to relevant resources, time frames, material incentives, etc.)
• the translator’s personal aims and ambitions (as a professional in general and regarding the concrete task in particular)
• other internal factors pertaining to the translator (experience, skills, etc.)
• the type of language-play (e.g., wordplay/pun vs. rhyme or alliteration) and how amenable it is to direct transfer vs. in need of a more creative approach
• the prominence of the language-play (how easily it is recognizable as such, which can depend, for example, on whether it is overt or covert, i.e., openly reacted to by (other) characters within the story of a literary text, including, e.g., screen productions, or not)
• the (perceived) role of the language-play (e.g., a seemingly minor stylistic accessory or a major attention-getter or plot device)
• the reliance on or link to non-linguistic elements (such as images or melodies)
• whether the translation was partly or entirely made by a machine translation programme or artificial intelligence

For a succinct summary of many of the factors involved in the (un)translatability of challenging features in a text, see also [62] pp. 236–237. The last point in the above list was included as a reminder that the prototypical scenario of a human translator being solely or primarily responsible for (published) translations and their quality may soon be a thing of the past (see Section 5 below). That said, most of the factors that may have an impact on the treatment of language-play in translation do presuppose the agency of a human translator. Many of them are interrelated and overlapping, and they can be grouped into sets of overarching categories, for example, the following:

• factors related to the general nature of the ST and TT, respectively, and the norms and expectations associated with different types of text;
• the circumstances surrounding the concrete translation task, including the client’s instructions and expectations, the internal and external preconditions for the translator to do an adequate job, and the possible influence of other stakeholders;
• the details regarding the concrete instances of language-play, including their type, distribution, distinctiveness, and relation to other textual elements.

Considering the complexity of the situation, it may come as no surprise that a general claim regarding the actual treatment of language-play in translation is barely possible.

4.2.3. Procedures for Dealing with Concrete Instances of Language-Play

Homing in on the treatment of individual instances of word- and language-play in translation, a basic distinction needs to be made when it comes to the concrete options available to the translator: on one hand, the translation procedures that can be employed in theory; on the other hand, what has actually been done in a given translation context. Regarding the former, it is common to consider a range of approximately half a dozen procedures, typically including the following (with more fine-grained distinctions being possible, as hinted by the various alternatives included in the first point):

• direct transfer of the SL language-play into the TT (of the actual SL wording (borrowing) or as a direct or ‘faithful’ translation (calque), with or without preservation of the playful aspect)
• creation of the same type of language-play in the TT
• creation of a different type of language-play in the TT
• complete omission of the playful sequence

To illustrate these options with a hypothetical yet concrete scenario, it could be considered what a translator might do if faced with the task of translating the safety sign in Example 9 above (Section 3.3.) from English into some other language (as will be remembered, the text of the sign features an obvious instance of language-play in the form of a rhyme between alert and hurt, in addition to a just possibly intended half-rhyme):

• The sign might be kept as it is, in English, for example because most of the TL audience can be assumed to understand it anyway. If the sign is part of a larger text, e.g., a novel, it could be left unchanged to retain a source-culture element in the TT, perhaps to avoid hiding the fact that the text is a translation, and/or to challenge the TT readers and lift them out of their linguistic comfort zone (i.e., ‘moving the reader to the writer’; see Section 2.2.1). Using SL elements in a ‘translation’ can also simply be a way of giving up or taking the easy way out.
• The sign could be translated more or less verbatim into the TL (though, normally, by respecting the grammatical rules of the TL). As a consequence, the general sense of the message is likely to be preserved in the TT, while the rhyme is likely to be lost.
• The translator could make an effort not only to preserve the general message of the sign but also its playful aspect. This would normally mean recreating a TL rhyme, presumably, but not necessarily, in the same place as in the original. As a variant of this, the translator could include another kind of language-play, e.g., an alliteration or even a pun, be it because such an alternative comes more readily to the translator’s mind, or be it because it is believed more likely to fulfil the desired effect.
• The sign as such could simply be omitted from the TT, e.g., because it is deemed to play a negligible role in its context, such as the plot of a novel or film. (This is thus an option only if the sign is in fact part of a larger text to be translated, not if the task is to translate the sign specifically.) Alternatively, only the non-playful part of the sign could be kept and translated, but the lines with the rhymes could be omitted (either, again, because the omitted lines are not deemed important enough, or as a way to avoid the perceived difficulty of translating them).
• Occasionally, the omission of certain elements in the TT, such as language-play, can be compensated for by inserting similar elements in another place in the TT, where the ST does not have them, to preserve the overall effect of the text, rather than that of a specific passage in the text. In other words, the translator may deem it wise not to be creative in one place (e.g., to save time and effort, and/or to avoid an inadequate solution), but recognizes the value of the aspect in question (i.e., language-play, in the present context), thereby inserting an instance where it can be done with less effort and/or more satisfaction. In the case of the safety sign, this may mean a non-playful translation of its contents, but adding playfulness some lines before or after.

Of course, alternatives 3 and 5 only apply if the translator a) recognizes the SL language-play to begin with, b) considers it important enough to preserve or re-create it in one form or another, and c) has the motivation, time, and skills to achieve a playful TL solution. If the SL language-play is not even understood as such by the translator, the likelihood of a non-playful TL solution increases.

Studies on the actual treatment of language-play in one or several target texts generally find that a range of translation procedures has been employed, meaning that there will usually be less TL than SL language-play, but rarely if ever none (see, e.g., the collection of studies in [64]). The nature of the text seems to play a considerable role in this regard, such that, for example, a substantial reduction (or loss, to use a more traditional and evaluative term) was found in many of the translations of Shakespeare’s play Hamlet, in various target languages [40]. In fact, that seems to be a common finding for translations of literary classics [51] p. 166. On the other hand, there can also be considerable degrees of preservation or re-creation of language-play, as shown by studies focusing on movies [33,51,65] or TV shows [66]. González Davies, providing an example from literary translation but no numbers, claims that compensation, i.e., a particularly creative approach, since it lacks a direct ST counterpart, “is also a common strategy in dubbing sitcoms” [67] p. 207.

Importantly, and in contrast to general beliefs (see Section 4.2.1.), some TL language-play, even including wordplay, can be and will be preserved in translation without any major effort or creativity required by the translator, because even a direct transfer or “faithful translation”, to re-use Aaron’s formulation [59] p. 19, can sometimes preserve the playful aspect of a sequence. This is due to crucial similarities between source and target languages being more common than generally assumed, e.g., in the case of polysemic and paronymic puns (see Section 3.2), which will frequently rely on similar sound-meaning relationships across languages [33]. This is confirmed by, e.g., Ermakova et al., based on their experiments with the machine translation of wordplay from English into French: “We observe that in many cases the successful translations are due to the existence of the same lexical ambiguity … in both languages” [68] p. 463.

All other factors being equal (such as those relating to the task, culture or translator in general), language-play in the ST is more likely to survive or be re-created as such in the TT if the following conditions apply: (a) the instances are easily noticeable, (b) they have an important function and/or effect, and (c) they are of a type that lends itself to a direct transfer where the playful aspect is kept intact.

5. Conclusions, Additional Remarks, and Outlook

The discussion thus far can be summarized as follows: While the rendering of ST language-play into ‘equivalent’ TT language-play, in terms of both quantity and quality, can be considered the theoretical default aim in translation contexts, there are numerous factors that will affect the actual treatment of texts and their constituent elements, including language-play, leading to widely diverging approaches and outcomes. While this may seem like a trivial and—to some—even disappointing conclusion, it is what might be expected in the humanities and social sciences, whose objects of study are multifaceted and subject to all kinds of interrelated conditions and influences, including social, cognitive and emotional ones, most of which are hard to ‘control’. On the other hand, it confirms the creative component—and thus the prevailing relative freedom—in quintessentially human endeavours, such as communicating and playing with language, and in translating between different languages.

That said, technical development has reached a point where many such endeavours are taken over by non-human agents. All kinds of texts, even novels, poems, songs, and screenplays, not to mention non-fiction, can be composed by artificial intelligences (AIs), and this includes their translation. It also includes language-play: Example 8 above represents a modified suggestion by an AI, though it is unclear whether the suggestion was based on a human-produced original used as training data. Be that as it may, the translation of language-play seems to be one of the few remaining challenges in the development of effective and truly reliable machine translation, but there are those who are actively working on overcoming it, such as the Joker project [69]. While human translators still clearly outperformed an AI employed in the project a few years ago [68] p. 464, it only appears to be a matter of time until human agency and intervention have become largely redundant in translation contexts, with, perhaps, just a few experts being employed to double-check and adjust machine-generated output where considered particularly desirable—where the AIs may need a few more years until they perform on a par with or better than humans, as in the case of language-play translation.

In the meantime, some writers and translators may, if they are inclined, engage in experimental and unconventional forms of playing with language and dealing with language-play in translation, in ways that may depart from what is generally understood to fall within ‘proper’ translation [53,70] and that may illustrate the human capacity for true creativity not currently attained by the machines. In more mainstream contexts, it still makes sense to increase current and future translators’ awareness of the issues involved in language-play translation and to provide training in that regard (or in the partly overlapping realm of humour translation [67]).

After all, language-play matters. It helps people relax and enjoy themselves. It is a core element of much humour, and humour is obviously important in its own right, for various psychological and social reasons. Yet language-play can also help raise awareness: of language itself and how it functions (and sometimes malfunctions), and of the various aims and topics that language can be used for and about, be they interpersonal, societal, ethical, or commercial. Language-play is part of learning and mastering a language. Children acquiring their first language engage in it, and proficiency in a second or foreign language also comprises the ability to play with its elements for special effect—and to recognize when others do the same. This, in turn, is linked to broader cultural, including inter-cultural, awareness and competence.

As argued in ref. [71], the study of wordplay translation (though the case can easily be extended to the translation of language-play more generally) may appear to be a rather inconsequential undertaking, with limited relevance to the major challenges the world is facing today. Yet if we realize the potential of language-play in entertainment, societal discourse, and education, we immediately establish a link to some of the core concerns affecting our lives, well-being, and future. It then becomes meaningful to explore how language-play works on the linguistic, cognitive, and social levels—and, by extension, how it is treated in translation, i.e., in cross-linguistic and cross-cultural communication. Customers of translations are entitled to good-quality services and products, and that entails a respectful and well-founded treatment of the various constituent elements of translated texts, in terms of both content and form, including word- and language-play. Whether human translators continue to have a role in this regard in the foreseeable future or whether AI will take over virtually all such tasks soon, the adequate choice of translation strategies or procedures for challenging elements in a text, such as language-play, including creative solutions where appropriate, is part of the respect that texts and their translation deserve.