Translating such a complex linguistic object begins with understanding the pun in its source language, deciphering where the pun is hidden, how it is uncovered by the English receptor (witness/spectator/reader) and what it may trigger to them. It is followed almost by a scientific process in which the pun is dissected and ripped apart word by word to track down the pun out of its hiding place. After this dissection comes the activation of the multilingual nature of the translator, a step in the process that will bring the translator to switch from the source to the target language in order to find THE perfect compromise. This translation phenomenon is all about making choices about translation strategies. The translator is constantly torn between favoring the lexicon and a stylistic-linguistic approach or favoring the message and a functionalist approach, between academic translation and professional translation, literary translation and pragmatic translation, faithful and unfaithful translation, a stylistic-linguistic approach and a functionalist approach.

In the end, translating a pun brings the translator back to the process of creation that is inherently linked to the process of translation. The translator is also a creator, a re-creator of a material that was created by someone else, the translator creates something both similar and different, he works on different levels and takes on many roles, as a French translator one acts both as an English receptor of the English pun, as a French re-creator of the English pun and as an artificial French receptor of the French pun. The task of the pun translator turns out not to be that easy. Translating, transcribing, transmitting and transcending a pun : why playing with words is far from being punless/pointless.

!Humor is the first gift to perish in a foreign language" Virginia Woolf

• The JokeR contest 2 Translating from English to French as many puns as possible from the list containing five hundred puns until February 24th. That is what the JokeR contest seemed to be at first sight for the both of us. The translations had to keep the semantic fields from the original text and had to contain a form of pun, these are the only limits this contest imposed on us. On the CLEF"s website the JokeR project is described as an automatic Wordplay and Humour Translation Workshop and its goal !is to bring together translators and computer scientists to work on an evaluation framework for creative language, including date and metric development, and to foster work on automatic methods for wordplay translation" . It is a workshop among others that make up the CLEF 2022 Conference and 3 Labs of the Evaluation Forum such as eRisk: Early risk prediction on the Internet for instance.

We spent a month translating as many puns as we could, everyday we would select a few of them and we would work at our own pace while still trying to keep an efficient rhythm. To start translating puns the first questions that were raised between us were : !What is humor ? What is funny ?" and then came another one, a much more complex one: !What is not ?". Defining humor and defining a frame around it was probably the first challenge we had to deal with. Many linguists worked on that very simple question such as Pierre Guiraud in Les jeux de mots , !Le premier problème est la définition même du concept de jeu de mots et de ses limites, interférences et imbrications avec des faits voisins [...]", which translated means that the first problem is the very definition of the concept of the wordplay and its limits, its interferences, its links with similar concepts. Before translating humor we had to understand it, we had to work on understanding why and how the puns worked in the original language, in what context they were rooted in, what did they trigger in the receiver, what did they rely on ? That was the deepest part of the iceberg.

EMAIL: email1@mail.com (A. 1); email2@mail.com (A. 2); email3@mail.com (A. 3) ORCID: XXXX-XXXX-XXXX-XXXX (A. 1); XXXX-XXXX-XXXX-XXXX (A. 2); XXXX-XXXX-XXXX-XXXX (A. 3) The top of it was the pun in its grammatical and written form. We undoubtedly had a two-part work to do, translating the pun"s essence and then we had to funnel it into a material shape, into letters, we had to multitask and coordinate these two aspects.

One thing is for sure, translating these puns was a way to get a better understanding of ourselves as translators, we were always in the process of testing the limits of translating puns. We even got to the point where making puns and placing them in the right spot in a sentence in an everyday context outside the context of the JokeR project was an automatism, by understanding puns and translating them it actually trained us to make others, in our mother tongue but also in English, which once again proves that translating is about much more than translating it is about creating, it is about endless creation.

• Our theoretical development Throughout this short essay we shall try to understand How translating wordplay goes beyond translation by crossing and blurring the lines with transcribing, transmitting and transcending the wordplay owing to the human aspect of the translator. Firstly we shall focus on the activation of mental frames through collocations, semantic fields and the multiplicity of semantic fields summoned at the same time in a pun which are three aspects that are inherently connected to the human nature and imagination, secondly we shall dwell on the subtleties that might elude to non-human translators such as the orality of a pun as well as the idea of common knowledge that surrounds and fills it, in a third part we will develop our essay around the translator and its satisfaction, the details of the presentation of a pun and finally we will end this essay with a reflection on machine translation. We think that our analysis aims at taking part in the many theoretical contributions that the literature surrounding translation and the translation of wordplays contains.

Collocations are one of the things that bring credibility to a pun, or to any translation for that matter, it immediately reassures the reader that the target language was respected. Take for instance the pun 57 : !He threw jello at his wife, who had him arrested for carrying a concealed weapon", any English-speaker immediately understands that it should be concealed instead of congealed because concealed weapon is a collocation, which Mona Baker defines by a !tendency of words to co-occur regularly in a given language" . That is the same thing in French, in our translation !Il a jetté son île 5 flottante à la figure de sa femme, qui l"a fait arrêter pour possession d"arme blanche" we used a collocation !possession d"arme blanche" a rather long one, and a collocation that activates the frame of the crime exactly as !concealed weapon" does. Not only does using a well-chosen collocation activate the right frame but it also helps the receiver to acknowledge and accept it faster and more efficiently.

When a pun was rooted in a specific context it could be difficult to choose whether to keep that same context and to look for words and sentences that could fit in this said context or to uproot the pun from its context to plant it in another one while still keeping its very structure. If we were to take the example of the pun number 28, the pun was set in a general school context and in a specific geometry class context hence the pun could rely on a joke about geometry and students by using the Baker, Mona. In Other Words: A Coursebook on Translation. Routledge. word !squares" which encapsulates both a way to define the students and a geometry term. We chose to keep the pun set in a school context but we decided to transpose it in another class, a carpentry class. We ended up conveying the pun through the use of the word " bûcheurs" which encapsulates both a way to define the students and also a word close enough to the word !bûche" which is a carpentry term. Transposing in a different setting is an efficient way of translating puns.

On some of those puns we had to be very careful of the multiplicity of semantic fields they were summoning, as it can be shown in the pun 105 : !The Net-surfing bandits always ended an operation by clearing out the cash", this pun relies on two semantic fields, the first being that of the internet, through the qualification of the bandits !net-surfing" and the reference to a computing action !clearing out the cache", and the second being that of the robbery through the reference to the !bandits" and to the !cash". Now this pun requires to be translated by taking into account those two fields and by intermingling them through a pun that is not necessarily the same as the English one, the original one relying on the similarity between the word !cash" and the other !cache" which allows a double game on the word. Here we translated it that way : !Les voleurs qui surfent sur le web ne comptent que sur les gros débits", indeed we first summoned the semantic field of the internet by inserting the computing action surfing on the web !surfent sur le web" and the reference to the broadband !débits". Then we activated the second field by qualifying the characters as thieves !voleurs" and by making a reference to cash withdrawal through the double meaning of the word !débit" which can be both used in these two contexts.

Whether the English or the French sentence, both rely on a pun that comes from the doubleness of a word, this word is where the two semantic fields touch, the contact area that allows the joke to work. Translating puns where a multiplicity of semantic fields touch requires to find that contact area as soon as possible and to exploit it until the balance between the two semantic fields is reached and brings enough satisfaction (a translated pun rarely brings a whole level of satisfaction to its creator) to the translator. This process of finding THE word that will act as a supreme master key required us to browse for hours on end on websites that list lexical fields and synonyms, this overflow of information was crucial in our way of working, maintaining this common flow of information was a way for us to constantly brainstorm and to spill all of the haystack on ourselves until the needle pricked us.

As Konrad Żyśko highlights in his book A Cognitive Linguistics Account of Wordplay , a good 6 way to define a wordplay would be to use Delabastita"s definition !the various textual phenomena in 7 which structural features of the language(s) used are exploited in order to bring about a communicatively significant confrontation of two (or more) linguistic structures with more or less similar forms and more or less different meanings" Throughout the extract we read he actually 8 explains how combinations of spelling, pronunciation can produce different effects on the words like homonymy homophony, homography and paronymy. The context can sometimes be stronger than the linguistic aspect because even if we just demonstrated that the presence of a textual phenomena or the combination of different linguistic structures is needed almost all the time to create a wordplay, Delabastita explains it through the example of the question : What do sea monsters eat ? Answer: Fish Żyśko Konrad. A Cognitive Linguistics Account of Wordplay. Cambridge Scholars Publishing, 2017, p.5. and ships. Not only does it work because fish and chips is a collocation but it works even more because chips is written ships and he explains that as ships are objects that are situated in the mental frame of the sea it triggers an immediate message of acceptation from the reader, it wouldn"t have worked the same with fish and chips or crisps for instance. Finally Żyśko Konrad concludes this part by saying that !it seems that the missing link for wordplay to emerge is an appropriate context", indeed the contextual aspect should never be underestimated in comparison to the linguistic aspect.

3. Thinking outside the box: aspects that might elude to a non-human translator

Even though one could consider at first sight that a pun is a written entity that solely relies on the process of writing, it turned out to be a multi-faceted challenge that worked through unexpected aspects. One of the most important aspects of our translating work was to be extremely aware of the mechanisms that were activated when we were translating, how our minds worked, how we got to where we wanted to go with the puns, what our process of translating requires (complete silence, interaction, music, darkness, etc...). We tried our best to feel what translation triggered in us and to exploit the information we gained through the whole process. One decisive point in our meta-fictional point of view about our work was to understand that we were definitely not only writing, never have we ever just sat in silence for hours and never communicated out loud or together in order to get to the best translation possible. Translation is not robotic (well that is up to debate) but translation is felt, translation is an art, it is creation and creation rarely occurs silently on the blankness of a page. We witnessed that the translation was oral. The translation was about exchanging points of view and speaking out loud and trying to see if the puns sounded good, if they could be said out loud without stumbling over a word. We often needed to check the orality of our puns, thus the oral aspect turned out to be at least as important as the written one, what makes a wordplay even better is to consider his hidden oral aspect.

The oral and visual aspects were pretty much unexpected but THE aspect that needs to be discussed is undoubtedly that of the common knowledge and how in every wordplay common knowledge is summoned in the reader, it obviously varies on a spectrum from small and almost insignificant references to the most complex ones. In Rhymes and PUNishment adventures in wordplay written by Brian Cleary there is a joke that plays on common knowledge but especially on a specific musical reference :!do, re, mi, fa, sol, la, and ti are notes on the major scale. To get the puns, 9 you need to know that these words are pronounced like !doh," "ray," "mee," "fah," "soh," "lah," and !tee."". The book is a children"s book that aims at explaining puns to children, it contains a pun or several puns per page and the bottom of the page contains an explanation, this system and this example perfectly emphasises and highlight the importance of the belonging to a certain social group in order to understand a specific group, for instance here musicians have a rather superior advantage on non-musicians. This is an aspect that definitely came into play when we translated, sometimes one of us felt more comfortable on some subjects while the other didn"t.

Cleary, Brian P. Rhymes and PUNishment Adventures in Wordplay. Millbrook Press, 2006, p.9.

As we translated we came to the realisation that sometimes some puns were so great in English that their greatness could not be replicated in the same way in French, hence some English puns are definitely better than the translations we attempted to give. However it also works the other way round, some of the French translations turned out to be of a better quality than the original English ones. If we were to take the example of the pun number 26 which is If you"re a gardener you might call yourself a #plant manager", the original pun contains the context of leadership and that of gardening but apart from the utilisation of these two contexts there is no striking use of an idiomatic or a contextual item to reinforce the pun. Nevertheless when we translated it we chose to keep the leadership and the gardening contexts while adding a new ingredient to the mix : a reference to the French culture that only French people could get (even if some could argue that this vegetable brand is not as trendy as it was 10 years earlier). This is a reference to a brand that was really popular a few years ago and that would often be in ads on tv, making it part of a French culture in a way. Sometimes by going further than the original joke we actually create a new pun, as a matter of fact if an English translator were to be given only the French joke as the original joke that they had to translate they would probably take into account the two well-established contexts and they would also notice and try to reproduce this reference to a pop culture element.

One could think that once the pun was translated then the work was done and that is partly true, the only thing we had left to do was to choose how we wanted it to look written. But we quickly came to the realisation that the presentation through stylistic effects would undoubtedly have consequences on the orality and the efficiency of the pun. If we take the puns 78, 83 and 127 we can identify different ways that we presented our puns to the reader.

In pun number 78 !When cannibals ate a missionary they got a taste of religion" which was translated in !Quand des cannibales mangeaient un missionnaire, ils avaient cet arrière-goût de foi(e)" we hesitated between writing foi (faith) or foie (liver) and we decided to propose both of these solutions by offering to the reader two simultaneous options in one.

In pun number 83 !The boating store had its best sail ever" which was translated by !Le magasin de bateaux fait des proue-sses" we decided that we needed to put the emphasis on the word that was a part of the boating industry lexical field !proue" which is the bow of the boat and hence we needed to unstick proue from !prouesses" which means achievements, as we could not use parentheses as we did in the example just above we decided to use another stylistic device; the hyphen in order to put the emphasis on the fact that there is a word within the bigger word while maintaining the unit of the word.

In the cases in which we could not use parentheses or hyphens we actually decided to commit to a bolder and riskier option : relying on the reader to fully understand the pun, by making this decision we made a one-sided contract by trusting them to engage in the understanding of the joke by saying it out loud or by peeling the sentence layer by layer to make sure that no pun is hidden in the cracks. For instance in number 127 !The sign on the nudist camp said, #Clothed"til May'" which we translated by !La pancarte sur la zone non-fumeur annonçait : !réouvre en joint", in this case we trust the reader to understand that even though we write joint (which is a joint in English) it is also pronounced like juin (June in English).

Other options also seemed interesting even if we didn"t choose to explore them like for instance putting the words or some letters in bold letters or by putting slashes in-between our options. The question is : Did we sometimes choose to present some of the word plays a certain way because we felt like it would be understood in a better way by another human ? Something that a machine perhaps could not have done.

In order to translate the puns from the JokeR contest, we put together the puns from the source language on a document created on the Google Doc platform. At that time, we were not yet aware of the help that the application would bring us. Indeed, we had chosen this application to be able to exchange rapidly and efficiently our translations, projects and ideas. Within a few hours after having translated the first few puns, we realised how crucial the computer would be for our work.

Using Google Doc, we realised that the "automatic correction" offered by the website would allow us to better translate our puns. Indeed, the Google Docs app has the particularity of correcting typos and grammatical errors in order to propose a more correct and a faster act of writing to the users. For example, Google Doc will correct conjugations, typing, punctuation, syntax or wording errors. This tool can work in many languages. That is to say, it is possible to change the language of the document in order to point out the mistakes made in the target language. This device is known to be useful when we are not fluent in the writing language.

While translating the word games for the JokeR contest, we had to play with the structure of the words. Furthermore, it happened that we used words that did not first make sense in the sentence where they were put because it was necessary to make these linguistic "mistakes" in order to create the joke. We can thus take the example of the sentence 61 : By underlining the word, the computer tells us that the words "chez amère" are an error. Thus Google Doc invites us to correct these words and replace them with "sa mère". Initially, we observe that in English the wordplay is based on an emotion: "mutter", as well as a word in the lexical field of mother and daughter: "Mother" ,clearly, in this case. In trying to translate this pun, we used a sound effect created and realised thanks to the connection between two words "chez" and "amère". We kept the category of bad emotions with a the vocabulary linked to the semantic field of pain and hurt. Also this same word encapsulates the aspect of the "mother" thanks to the pronunciation of the last two words put together. We knew that the word play was successful because the computer suggested that we correct the syntax of the sentence and cut the word into two other words first hidden in one. The game on the resemblance of a word with another word is understood and taken into account by the computer.

We observed this phenomenon in several sentences. For instance, we could see this phenomenon in the sentence number 64: In her dissertation on computer-assisted translation and machine translation, Sandrine Peraldi quotes her professor who describes computer-assisted translation as "the set of computer programs and tools that can help the translator and the writer in their research, in the constitution of resources and during the translation process". 10

To conclude we demonstrated and developed the idea that translating wordplay goes beyond translation by crossing and blurring the lines with transcribing, transmitting and transcending the wordplay owing to the human aspect of the translator, we understood that firstly human imagination is at the core of wordplay translation through the use of collocations, the human mastery of semantic fields that allows transposition and higher levels of complexity within a pun, through this part we understood that sometimes context can be more important than language regarding a wordplay. Then we focused on the aspects that are too subtle for a machine to grasp like the orality of a pun or its inherent link to a common knowledge shared by humans.

Then we ended our demonstration by focusing on the translator and on the choices they could make while translating a pun and how these choices impact the quality of the pun as well as its presentation and hence its reception. Consequently we also understood that these choices were made willingly, that translators not only tried their best to be understood by their readers but would also sometimes translate and present their translation out of pure human instinct. Finally we concluded on a reflection about how machines can help humans in their translation of wordplays through applications like Google Docs for instance. This last aspect brings us to a last question: Can computers translate wordplays as efficiently as human beings if they are able to understand them ? And if they do, does this mean that computers understand humour and hence reach a higher level of understanding of human emotions and humanity more generally as each day passes ?

In this part we would like to acknowledge the participation of Charlotte Daniel's mother who really enjoyed to help us in our translations, we would like to thanks the University of Western Brittany and its translation teachers for the courses on translation we got during both our three years of Bachelor's Degree and our first year of Master's degree. Moreover we would like to thank the Sandrine Peraldi, Muriel Bellamy, Marion Chesné, Camille Lehmann, Audrey Prost. Comparatif