t is very often said that humor does not travel well. This approach seems to have inspired many scholars to research on the (un)translability of humor. Nevertheless, it is enough to turn on the TV or to go to the cinema to realize that, regardless of any possible travel inconveniences or even a possible motion sickness, humor does travel across linguistic and cultural barriers. Having assumed that humor travels, the objective of the study was to establish how it travels across different languages, cultures, and translation methods.

There were two basic objectives of the study which was carried out in order to write a master thesis. The first one was to revise the validity of the translation tendencies identified and described by Juan José Martínez-Sierra, who analyzed the transfer of humor in The Simpsons (English being the source and Spanish dubbing the target language). In order to achieve that objective, the method designed by Martínez-Sierra was used in a different corpus and included two target languages (Polish and Spanish).

The second objective was to identify the tendencies in humor transfer with respect to the target language and different types of audiovisual translation (dubbing and subtitling). It is this second objective that is presented in this article.

For the purpose of this research, I have selected Shrek, an animated American film, which was successfully broadcast worldwide. Although there are multiple reasons to justify this choice, initially there were just two requirements which the potential research material had to meet: it needed to have dubbed and subtitled versions in Spanish as well as in Polish and it had to be a comedy. Since in Spain all films broadcast on the TV and in the cinemas are dubbed and DVD releases also have subtitles, finding an appropriate material in Spanish did not present any problems. Poland, on the other hand, is a different case. With the exception of products intended for young audiences. foreign films are presented to the audience with voice-over (television) or with subtitles (cinema). With these restrictions. choosing an animated film was an obvious choice.

The second reason is linked directly with the phenomenon of the film's Polish dubbed version, which marked the "pre-" and "post-" Shrek era. Poles, as all other societies, are very attached to the traditional audiovisual translation methods used in their country (voice-over in the television and subtitles in the cinema), and when asked about their translational preferences they list dubbing as the least desirable one, pointing out its lack of naturalness and poor quality. That is not the case of Shrek. Many people, including the ones who claimed to prefer watching films in the original version, describe its dubbing as excellent and even "funnier" than the English version. The success of the Polish dubbing was so great that its translator, Bartosz Wierzbięta, became not only recognizable, but even popular (there were many press interviews with him and the films which he translated later on were advertised as "translated by Bartosz Wierzbięta").

The study is based on the taxonomy of humorous elements formulated by Juan José Martínez-Sierra (2005), who used it to analyze audiovisual jokes in the popular American animated series The Simpsons. Basically, this taxonomy follows the trail blazed by Patrick Zabalbeascoa (1993, 1996, 2005) who talks about different joke types. Assuming that the category of joke types might sometimes be too narrow to apply (it is hard if not impossible to classify a given joke exclusively as national or language-dependent) Martínez-Sierra suggests using the category of humorous elements, which in different combinations can be combined to result in a joke. According to Martínez-Sierra (2005: 290-291), there are eight different humorous elements: Community-and-Institution (elements, which make a reference to cultural or inter-textual features bound to a concrete culture such as politicians, celebrities, organizations, newspapers, books, films, etc.), Community-Sense-of-Humor (topics which appear to be more popular in certain communities than in others by preference, rather than cultural specificity), Linguistic (elements based on linguistic features), Visual (humor produced by what can be seen on screen, not what can be read), Graphic (humor derived from a written message inserted in a screen picture), Paralinguistic (non-verbal qualities of voice associated with expressions of emotions as well as narrative silence), Sound (sounds found in the soundtrack and the special effects which by themselves or in combination with others may be humorous), and Non-Marked (miscellaneous instances that are not easily categorized but are, nevertheless, humorous).

The analysis was performed in three stages. The first one consisted in detecting source jokes and their four translations (Polish and Spanish dubbing and subtitling) and tagging the humorous elements as their constituents. Forty-two jokes were detected in the source text, which gives a total of two hundred and ten jokes analyzed. In order to carry out this part of the study, I resorted to a card designed by José Martínez-Sierra (2005: 293), modifying it to suit my research. The card used in the study is presented below.

Afterwards I proceeded to the "global quantity analysis" which consisted of counting all the humorous elements previously detected in the source and target texts. This part of the study allowed me to formulate conclusions regarding some general tendencies in humor translation in dubbing and subtitling.

Unfortunately, the quantitative analysis alone does not present a full picture of the differences in humor transfer in dubbing and subtitling. As Sierra-Martínez (2005: 292) aptly noticed, jokes after their translation can suffer not only from a quantitative but also qualitative losses or changes. This basically means that not all the elements are rendered on the one to one basis. The fact that the target text contains i.e. twelve linguistic elements and the source text sixteen, does not mean that twelve elements were translated and four lost. According to what Martinez noticed (2005: 292), humorous elements can of course be lost in translation, but they can also change type or even be added. To give an example: in the source text of Shrek a total of sixteen linguistic humorous elements were detected and in the Polish dubbing text only twelve. The detailed analysis carried out in the third part of the study enabled me to observe that out of the sixteen linguistic humorous elements from the source text eleven were rendered without the change of type (i.e. English word-play was substituted by Polish word-play), three were lost, two changed type (given joke is still laughable but not because of linguistic reasons) and one element was added (a linguistic element appeared in target text even though it was absent in the source text). The results of both parts of the analysis as well as the conclusions are presented in the following chapters.

Judging from the global quantitative analysis, we can say that, regardless of the language and translation method, the target texts tend to contain less humorous elements that the source texts. The proportional loss of the humorous elements is almost 5% in dubbing (2,5% in Polish dubbing and of 7% in the Spanish dubbing) and 18% in subtitling (a 18% both in Polish and Spanish subtitling).

It is also clear that the translations had the weakest performance in the rendition of the linguistic and community-and-institutions elements. Only 70% of the original linguistic humorous load and 55% of the community-and-institutions humorous load was rendered in translation. Also in the case of these two particular elements, the divergence between both translation methods is most visible. In the case of the linguistic elements, dubbing with 84% of the original quantity of elements rendered (75% in Polish and almost 94% in Spanish) performed far better than subtitling with 56% of the original elements rendered (62,5% in Polish and 50% in Spanish). Dubbing in general managed to render 75% of the original quantitative load of the community-and-institutions elements (93% in Polish dubbing and 57% in Spanish dubbing), which is over 40% more that subtitling with a 35% of the original amount rendered (43% in Polish subtitling and 28,5% in Spanish subtitling).

It should also be pointed out that as much as 46% of the humorous load in the original text belongs to the category of visual or non-marked elements. Precisely these elements are the ones which were rendered in 100% (visual elements) or even in 126% (non-marked) in the translated texts. This basically means that the translations contain more non-marked elements that the original. It is also noticeable that the amount of the non-marked elements is higher in subtitling (32% more than the original amount) than in dubbing (20% more than the original amount).

Before talking about the elements preserved which did not suffer any qualitative change during the translation process we have to point out that according to Martínez-Sierra a humorous element is rendered without changing its type as long as a substitute in the target language or culture is found (i.e. Poles would consider a given element as a Polish word-play or a reference to Polish culture).

Regardless of the target language, dubbing preserved on the average 85% and subtitling 72,5% of the original humorous elements. The most significant divergence between dubbing and subtitling can be noticed in the case of the community-and-institutions and the linguistic elements. Dubbing managed to render almost 71,5% (86% in Polish and 57% in Spanish) of the original community-and-institutions elements without changing their type, while the subtitled versions in only 35% (43% in Polish and 28% in Spanish) of the cases maintained the type of the elements unaltered.

As far as the linguistic elements are concerned in the dubbed versions as much as 75% (69% in Polish and 81% in Spanish) of the elements were rendered without a type change, while only 56% (56% both in Polish and Spanish) of the elements remained unchanged in the subtitled versions. Differences between dubbing and subtitling are slightly smaller in the case of the non-marked and paralinguistic elements. The dubbed versions preserved the type of the original elements in 95% and 70% of the cases respectively, while the subtitled ones in 87,5% and 60%. The visual elements were rendered in 100% in both translation methods.

As far as elements which suffered from a qualitative change during translation are concerned, in the study only three types of elements were detected as type-changing. It is the case of the community-and-institutions (65% of all changes), linguistic (27% of all the type changes), and community-sense-of-humor (7,5% of all the changes) elements. The change of type in all the cases happened always towards a non-marked element and was more common in subtitling (70% of the type changes) than in dubbing (30% of the type changes).

The study also showed that some humorous elements are lost in the translation process. Surprisingly it is the linguistic elements which disappear most frequently in both languages and in both translation methods, as they constitute a 48,5% of the entire loss. They are followed by the paralinguistic (21% of the entire loss) and community-and-Institutions elements (18% of the entire loss). The loss of the community-sense-of-humor elements amounts to 12% of the entire loss. Looking at the results of the study it is obvious that the loss of the humorous elements is more significant in subtitling (64% of all the total loss) than in dubbing (36%). Subtitling lost 86% of all the paralinguistic, 69% of all the linguistic, 66% of all the community-and-institutions and 50% of all the community-sense-of-humor elements lost.

One of the most interesting features of this study was the fact that jokes were found in which the humorous load of the target version increased as compared with the original version. As can be seen in the table presented below, the augmentation of the humorous load occurred mostly in dubbing, where 91,5% of all added elements were detected. If we compare the elements added in dubbing we can clearly see that this practice was more common in Polish, where almost 80% of all elements were added, than in Spanish.

The results obtained from the global and the detailed analysis allow us to draw several conclusions regarding the translation of humor in dubbing and subtitling. First of all, it should be pointed out that regardless of the language or the method of translation, the target texts tend to contain less humorous elements that the source texts. The quantitative loss of the humorous load in the translated texts in general is 10%.

It is also clear that compared to dubbing, the subtitled versions preserve less of the original humorous load (the percentage of humorous elements preserved without changing the source elements' type is significantly lower in subtitling) and contain less humorous elements in general. This conclusion is valid for subtitling in general, irrespective of the language of the target text. In other words, it means that dubbing in general is closer to the source text as far as the quantity and the quality of the humorous elements is concerned. The closeness of quality of the humorous load in dubbing is the expression of the domesticating tendencies of this translation method. Thank to the latter, dubbed versions of humorous films might appear more amusing to the target audiences.

There is one more observation that we can draw from this study, which concerns Shrek in particular. It is quite probable that the fact that almost half of the humorous elements in this animated film belong to the category of visual and non-marked is not accidental. Since these elements present less problems in transfer than other humorous elements, we could risk the statement that their use was intentional, with the aim to make the humor of the film internationally available. This leaves us just one step away from saying that part of the translation, understood as facilitating the understanding of a given product, begins at the time of creating the script.

Martínez-Sierra, José (2005) "Translating Audiovisual Humor. A Case Study", in Henrik Gottlieb (ed.) Perspectives: Studies in Translatology, Copenhagen: Routledge, 289-296.

Zabalbeascoa Terrán, Patrick (2000) "Contenidos para adultos en el género infantil: el caso del doblaje de Walt Disney", in Veljka Ruzicka, Celia Vázquez and Lourdes Lorenzo (eds.) Literatura infantil y juvenil: tendencias actuales en investigación, Vigo: Universidade de Vigo, 19-30.