Internet Jokes: The Secret Agents of Globalization?

Authors


  • Accepted by previous editor Maria Bakardjieva

Abstract

In this article, we use the somewhat unusual lens of joke translation to examine the process of "user-generated globalization" – cross-national diffusion of content by Internet users. We tracked the translations of 100 popular jokes in English into 9 languages and analyzed them quantitatively and qualitatively. Our findings indicate that (1) web-based diffusion of translated jokes is common but varies greatly, both between languages and between jokes; (2) "global hits" differ from "translation-resistant" jokes in their themes and incorporation of American markers; and (3) translated joke versions tend to include only minor cultural alternations (such as name shifting), thereby preserving the original messages. Overall, these findings suggest that Internet jokes serve as powerful (albeit often invisible) agents of globalization and Americanization.

The Internet, more than any previous medium, provides technological facilities for global exchanges of content. The realization of this potential, however, is not determined merely by technology; it is the complex nexus of practices and choices made by Internet users that generates forms of globalization. But these users often do not speak the same language. Whereas in its early days the Internet's lingua franca has been English, most Internet users nowadays are not native speakers of English, and the medium's accelerating growth in countries such as China is expected to strengthen this trend (Danet & Herring, 2007; Hale, 2012). This linguistic variation invokes a series of fundamental questions about cultural globalization and the Internet: To what extent does the Internet facilitate global flows in a multilingual setting? What is the role of user-generated translations in this environment? Do some types of content spread globally better than others? And finally, how are processes of localization, globalization, and Americanization interwoven in these Internet-based flows?

In what follows, we address some of these broad questions through the ostensibly narrow prism of translated Internet jokes. While jokes may be perceived as light-hearted entertainment, they are often important arenas in which troubling issues are processed and negotiated. Humorous texts relate to the symbols and stereotypes specific to the place and time of their creation, reflecting norms and aspirations, power structures, and collective fears (Billig, 2005; Meyer, 2000). Since jokes are ambiguous modes of communication, they often enable people to invoke controversial issues without the sanctions accompanying "serious" communication on such topics. In the past, when jokes were mostly told orally, it was difficult to trace their evolution and variations. In the digital era, however, people post jokes in forums, blogs, and other social media – leaving traces that are ripe for scholarly harvesting. Moreover, the Internet makes the international spread of humor easier than ever and, technically at least, may also facilitate the spread of translated joke versions. Jokes are often difficult to translate (Laurian, 1992), particularly when the basic ambiguity that invokes humor in anchored in language-specific puns and word-plays. However, initial evidence suggests that popular Internet jokes do cross linguistic and cultural borders, in various ways (Kuipers, 2002; Shifman &Thelwall, 2009).

The first part of the paper sets the analytical framework for our investigation. We start by conceptualizing Internet jokes as facilitating user-generated globalization – a process of cross-national diffusion of content by ordinary Internet users. We then discuss the unique role that the United States plays in such processes of cultural globalization. The concepts of banal nationalism, banal globalism, and banal Americanism help us articulate the duality of the US as a nation state and a global brand. We further illustrate how these negotiations over identity permeate translation processes, surveying literature about format TV. Building on these theoretical foundations, we then present our research questions, which relate to (1) the scope of user-generated globalization; (2) its directionality (that is, testing the assumption that Web-based diffusion originated in English versions); (3) the themes (and American orientation) of jokes that succeed (and fail) in spreading globally; and (4) the nature of changes introduced to translated jokes. In the third section, we describe our research design and the main methods of investigation: Combining quantitative and qualitative analyses, we traced the translations of 100 jokes that are popular in English into nine other languages. The results are followed by a discussion that ties the findings back to the fundamental debates about new media, language, and cultural uniqueness in an era marked by globalization and Americanization.

Internet Jokes as Modes of User-Generated Globalization

In many ways, research into globalization and the Internet echoes a larger debate that is polarized between those who see globalization as a site of Americanization and homogenization, and those who portray it as a more heterogeneous, bottom-up process. While some studies have pointed to American or Western biases in aspects such as language, commercialization, and technological infrastructures (e.g., John, 2013), others have provided evidence for glocalization or localization, both with regards to commercial websites and bottom-up use (e.g., Liu et. al, 2002). Many of these studies looked into globalization as a process based on conscious actions of highly motivated actors such as advertisers or political activists. However, we claim that globalization, Americanization and localization may also be shaped by small, invisible steps.

In the process that we refer to as "user-generated globalization," ordinary Internet users translate, customize, and distribute content across the globe. The literature has overlooked this concept for three reasons. Firstly, it is not driven by high-profile commercial agencies, such as major production companies, but by the daily practice of many ordinary users. Secondly, unlike the transnational flow of mass media content, this arena of globalization does not explicitly involve economics. Finally, once a text has been translated and local markers such as names and currencies have been added, its foreign origins may be less evident to readers and posters.

Our exploratory study of user-generated globalization examined the process through the case study of one Internet joke (Shifman & Thelwall, 2009). We found that the joke was widely translated to most languages explored, and that translations were localized to a limited degree. The main shortcomings of our previous work lie in its focus on a single case and its narrow theoretical scope. Therefore, the current study evaluates user-generated globalization through a much larger corpus and using a more nuanced theoretical framework. In particular, we use sociological and literary theories to look into the ways global, local, and American cultural markers are intertwined in translated jokes.

Internet Jokes as Banal Nationalism, Globalism, and Americanism

Conceptualized by Michael Billig (1995), the term "banal nationalism" refers to the mundane, often unnoticed symbols through which a sense of national belonging is constructed in people's lives. Nationhood is articulated through artifacts such as coins, flags, and maps; events such as the Eurovision Song contest or the Olympics; and a wide array of linguistic uses. This daily ‘flagging’ of a nation constructs a strong and uncontested feeling of ‘we.’ Ironically, assumptions about national uniqueness, as well as the means through which nationality is constructed, are international (Meyer & Strang, 1993). Moreover, whereas nations are becoming increasingly similar to each other in many organizational fields, some institutions apply subtle processes of erasure to give national meanings to global structures (Kaplan, 2012).

Alongside this prevalent national construction, people around the world are also exposed to mounting manifestations of globalization. Drawing on Billig's work, Beck (2002) coined the term "banal cosmopolitanism" to refer to the experience of globality as grounded in people's everyday lives. Taking a slightly different angle, Szerszynski, Urry, and Myers (2000) referred to "banal globalization" as the constant flagging of global images in mass media, which may remind people of their roles and responsibilities as global citizens. They found the following three types of images that may promote a sense of global action: the planet as a whole, emblematic environments, and people who are associated with actions of care and helping.

Within the surge of transnational symbols, one nation occupies a unique position. The scale and scope of American cultural transmissions are so vast that the terms globalization and Americanization are often used interchangeably. Indeed, Americanization is a process related to Globalization but is not identical with it (Ritzer, 2011). Globalization and Americanization are not two equal and parallel processes, but rather there are "deviations, lines of flight and counter movements" (Sznaider & Winter, 2003, p. 6). Nevertheless, even in an era marked by mounting claims about the decline in America's power, analyses of global processes of production, circulation and reception of cultural commodities reveal the American dominance in fields such as film, music, fashion (Sznaider & Winter, 2003), and news (Segev & Blondheim, 2013). In this process of cultural diffusion, American symbols are often presented as global; Hollywood stars, Cola-Cola, and Levi's are not regarded as American products but as universal icons.

Whereas the studies cited above looked into various aspects related to the flows of American products and symbols, in this study we focus on the ways in which America is signified, following Michael Billig's (1995) approach. According to Billig, understanding banal Americanism requires what he terms "a taxonomy of flagging"; that is, a nuanced depiction of the various ways in which America is marked in popular texts. Americanization may be incarnated in concrete symbols such as the Stars and Stripes; in more subtle markers such as "sky-scraped skylines" or "the cowboy gulch" (p. 150); and in abstract values such as individualism and consumerism. These abstract values are the most difficult to analyze because they assume numerous incarnations, are shared across many Western cultures, and cannot be regarded as distinctively "American." Yet the challenge that Billig's conceptualization raises is even deeper. When rethinking America as both a nation and a global brand, a question that arises (and, as far as we can ascertain, remains unaddressed in the literature) is whether we can identify distinct "global" and "local" aspects of American culture. In other words, the "taxonomy of flagging" we wish to develop aims to differentiate between "national" American markers that do not transfer to other languages (America as a nation state), and American markers that have become global standards (America as a global "brand").

While banal nationalism, banal globalism, and banal Americanism have been discussed here separately, they are actually tangled in multifaceted ways. Below, we will explore the manifestation of such interactions in contemporary jokes, both in English and in their translated versions. As our analysis will demonstrate, when a joke that originated in America is translated into a different language, such as French, it may have some markers of French banal nationalism introduced to it, while still preserving some of its original American markers. Therefore, as elaborated in the next section, translations of texts from one language to another serve as important sites of identity formation in an era marked by cultural hybridity.

Translation as Cultural Negotiation

Translated texts are important junctions of cultural flows as they inevitably involve at least two languages and cultural traditions. Translation studies have investigated this unique symbolic meeting place, examining the paths that translators take when negotiating the texts' original linguistic and cultural codes with those prevalent in the so-called target culture (Heylen, 1993; Palm Åsman). Two strategies discussed in this respect are "domestication" versus "foreignization." In a domesticating translation, the foreign text is strongly modified to fit the receiving culture's values. Conversely, the foreignizing translation stresses the linguistic and cultural difference of the foreign text, thereby "sending the reader abroad" (Venuti, 1995, p.20).

The domestication of texts is a complex task that is carried out through various practices, including one that Pedersen (2007) called "cultural substitution"; that is, the replacement of extralinguistic cultural references (ECRs) with items that are well known in the target culture. Extralinguistic cultural references are expressions that refer to entities outside language that a person who is not part of a specific culture may not be familiar with. When encountering such a reference, the translator may choose to "domesticate" it through three main strategies: generalization (replacing a phrase with something more general); substitution (replacing a phrase with something that denotes another type of cultural context); and omission (eliminating the foreign element). Using this framework, Palm Åsman and Pedersen (2011) outlined the following ECR types: names of people and places, food and beverages, sports and games, units of measurement, and literature and music. To a large extent, the changes introduced to these references in the course of translation dictate whether the text is domesticated or foreignized.

Translation studies such as those discussed above tend to focus on dyads, measuring two texts (source and target) against each other. In recent years, however, questions of translation have become increasingly intertwined with questions about globalization and Americanization, as similar media products are diffused worldwide. One of the most interesting areas of translation in contemporary culture relates to the flow of television formats. Formats are defined as program templates that are reproduced in different countries in domesticated versions (Moran, 2009). A key concept that underscores recent studies of formats' translational flow is "glocalization" (Robertson, 1995). The term challenges the clear distinction between cultural homogenization and heterogenization by depicting contemporary culture as a synthesis of internal and external influences. Thus, rather than simply accepting or rejecting global models, local actors amalgamate the foreign and the familiar to create multifaceted, hybrid cultures (Kaplan, 2012; Kraidy, 2005). Glocalization is strongly manifest in TV formats, in which global flows of money and ideas are often wrapped in national dress. Such shows are modified in ways that make them seem local: Their participants represent familiar ethnicities, are pictured in local settings, and highlight local customs (Moran, 2009). The use of vernacular is another key factor in this process of domestication. As a form of banal nationalism (Billig, 1995) – a transparent marker of national identity – vernaculars add a strong local flavor to the global recipes embedded in the television format.

Although formats are domesticated in numerous ways, they still embed generic cultural values that are shared across many programs (Waisbord, 2004). Thus, whereas the manifest content of such shows is local, their deep structure – the rules on which they are based – is often charged with Western values, particularly American ones. For instance, Galander's (2008) analysis of the Malaysian versions of Wheel of Fortune and Who Wants to be a Millionaire found frequent references to local historical sites, religious occasions, and local political figures. At the same time, however, Galander analyzed the overall philosophy of these shows as utterly Western, as they present "consumerism as a value, gambling and risk-taking as fun, entertainment as an end, and material wealth as an ultimate social value" (p. 14). In a similar vein, Murray (2001) analyzed Survivor (which was actually conceived in the UK and first aired in Sweden) as being "infused with American cultural myths of success and social mobility" (pp. 43–44), as well as and with self-commodification and social Darwinism that valorizes inequality.

Domesticated TV formats and literary translations are institutional sites of translation. Both the translator and the producer are employed by a media organization, which is often market-driven. By looking into translated jokes, we expand the scope of analysis to a new site of cultural negotiation. This arena is often detached from any formal organization and is energized by individuals who may (or may not) catalyze processes of "user-generated globalization."

Research Questions

In this study, we addressed four questions related to user-generated globalization through joke translation. The questions concerned the process's scope, its direction, the main themes underpinning it, and patterns of localization.

RQ1: To what extent do popular jokes in English diffuse cross-linguistically on the web?

By addressing this question, we hope to provide initial evidence about the scale of user-generated globalization, as facilitated by joke diffusion. We differentiate between: (a) diffusion of texts in English on websites in other languages; and (b) circulation of translated versions.

Beyond the question of scope, we need to explicitly assess our assumption about the dominance of English (and the US) in the global diffusion of jokes. We do this by answering the second and third research questions.

RQ2: Do popular Internet jokes in English that are also found in other languages appear first in English and only then in these languages?

While RQ2 relates to English in general, addressing RQ3 will enable us to assess the role of American culture in this process more specifically.

RQ3: Do some comic themes spread globally more successfully than others and, conversely, are there some topics that do not translate widely?

We hypothesize that some themes – such as computers and gender differences (Shifman & Blondheim, 2010; Shifman & Lemish, 2011) – will travel across national borders better than others. Thus, an analysis of jokes that succeeded in spreading globally may reveal a "core" of internationally-shared themes. While using English as the point of departure dictates that the initial set of themes/values embedded in the jokes will be "Western" in orientation, the question which of these is globalized is fundamental. Conversely, an analysis of jokes that did not translate well may shed light on the "local" dimensions of specific English-speaking cultures.

RQ4: To what extent do translated versions of the jokes introduce culture-specific elements, and which practices of localization do they apply?

We aim at understanding the practices that Internet users apply as translators who may choose to "domesticate" or "foreignize" texts, emphasizing national, American and/or global markers in the process. We wish to find out not only whether translated jokes tend to copy the English message without change, but also which strategies of domestication they apply.

Methods

Our research design included five successive phases: (1) sampling 100 popular Internet jokes in English; (2) tracking the Web presence of the jokes in the top 10 Internet languages; (3) tracing the timeline of the appearance of the first versions; (4) probing the differences between "global" and "local" jokes; and (5) assessing the nature of changes introduced into the translated jokes.

Stage 1. Sampling 100 Popular Internet Jokes in English

Our main aim was to compile a list of comic texts that are widely circulated on the Internet. This was a major challenge, as there are no popularity measurements for written texts equivalent to those found for videos. Therefore, we constructed the following sampling protocol.

  1. Based on the method developed in our previous studies (Shifman, 2007), we identified 10 popular humor websites (in English). This was achieved using a combination of Google Page Ranks and multiple search engines. The websites selected scored at least 5 on the Google Page Rank index and appeared in the first 30 results of at least two out of the three major search engines (Google, Yahoo!, and Bing) when the word "jokes" was used as a search string.
  2. We then automatically retrieved all of the jokes from these 10 websites (n = 29,211).
  3. To provide an initial estimation of the Web presence of this vast number of jokes, we used the first 10 words of each text as a search string (for example, "Sherlock Holmes and Dr. Watson went on a camping trip") in Bing's automated search (The Bing API version 1). This phase ranked the entire corpus according to one unified popularity indicator.
  4. The top 5 percent of the results (n = 1500) was further screened by 10 coders who filtered out nonjoke content that appeared on the websites, such as long stories and short citations.
  5. Finally, two coders looked for additional copies of each of the top jokes, including versions that did not appear in the original computer-based search. For instance, instead of using the string "Sherlock Holmes and Dr. Watson went on a camping trip," they used "Sherlock Holmes" "camping trip." This procedure resulted in a more accurate estimate of the jokes' popularity, enabling a more reliable selection of 100 jokes as the final sample.

Stage 2. Assessing the Web Presence of the Jokes in the Top 10 Languages

The main aim in this stage was to assess the presence of translated versions of the 100 jokes in the 10 languages most commonly spoken by Internet users: English, Chinese, Spanish, Japanese, French, German, Portuguese, Arabic, Korean, and Russian (a breakdown can be found at http://www.Internetworldstats.com/stats7.htm, accessed 08.10.2011). Tracking the translated versions required the ongoing engagement of native speakers of the investigated languages. Eighteen coders, all of whom were students at the institute at which this research was carried out, were involved in two rounds of joke-searching. We used the following protocol to determine the jokes' presence in the investigated languages.

  1. Coders were asked to translate a set of key phrases that had proven to be useful for detecting many joke copies in English and to use them as search strings in Google (google.com was used for all searches). They were asked to report both the number of obtained URLs and their accuracy. To measure accuracy, they scanned the first and last 20 hits of each search. The set of key phrases was only regarded as sufficiently accurate when more than 15 of 20 results (in both the first and last pages) were indeed the joke or a version of it.
  2. Coders were instructed to find another set of keywords in order to accurately detect many copies of the texts. As in the first stage, the string was only used when more than 15 out of 20 results in both the first and last pages were indeed the joke or a version of it; if the count was lower, the coder worked to modify the key phrases.
  3. Since there is a subjective component in the process of finding keywords, we validated the results compiled by the first group of nine coders. A second group of nine coders independently repeated the same process. The main aim was to ensure that the coders did not miss joke copies due to the use of unsuitable key words. When correlating the average number of jokes found by the two independent coders across the nine examined languages, the result was very high (0.84).
  4. In addition to finding jokes in translated versions, we looked for joke copies in English that were published in non-English websites (for example, in a Chinese blog). This was done by using the same set of English key phrases as search strings in Google's language-specific option.

Stage 3. Tracing the Timeline of the Appearance of the First Versions of the Jokes

The main purpose of this procedure was to assess whether the English versions predated versions in other languages (RQ2). While many jokes may have originated from outside the Internet and thus their "true" origin may me intractable, we can still trace their appearance on the Web. To this end, we used 21 jokes from our corpus that were defined as "global hits"; that is, jokes that spread across almost all of the languages we inspected. For each of these jokes (in all 10 languages) we used the keywords devised in stage 2 as strings in a manual Google custom range search for dates in order to detect the first joke copies in all languages. We visited each of the pages detected in this procedure to ensure that the joke was in the page at the apparent date.

Stage 4. Analyzing Global versus Local Jokes

In order to address RQ3, which refers to the global and local dimensions of contemporary Internet humor, we probed the differences between global hits and translation-resistant jokes. We applied the principles of thematic qualitative analysis (e.g., Lindlof & Taylor, 2002) in this analysis, focusing on two realms: (a) Global and local markers. Following Billig (1995), we examined the jokes to assess the extent to which they imbed markers of Western, American, or other identities; (b) Thematic analysis of the major topics and standpoints encoded in the jokes.

Stage 5. Assessing the Nature of Changes introduced into the Translated Jokes

The focus in this stage was on a group of 21 jokes identified as "global hits." In order to assess the nature of change in the translated jokes, random results were extracted from the results of the queries in the local versions of Google. For each joke, we sampled 20 translations in the nine languages. Some jokes appeared in fewer than 20 copies in some of the languages, so the total number of joke units examined was 3448. Nine native speakers assessing the extent to which the translated version differed from the English one. Using a detailed coding scheme based on previous translation studies, the coders tracked the changes in various extralinguistic cultural references, such as personal names, geographic locations, typical types of sport, religion markers, and cultural norms. In addition, the coders also looked into more general patterns underpinning localizing practices. In weekly team meetings, we formed a unified scheme of analysis and ensured that coders used similar criteria when forming coding decisions.

The main domains of analysis in phases 1-5, as well as some of the results, are summarized in Figure 1.

Figure 1.

Main domains of analysis

Results

User-Generated Globalization: Scale, Scope and Directionality

RQ1 asks about the extent to which popular jokes in English diffuse cross-linguistically on the Web. To address this question, we traced both translated jokes and English jokes on non-English web pages. We found that translated jokes are generally much more prevalent than English versions in non-English websites (see Table 1). In all the languages that we examined, there was a positive and significant correlation between the popularity of jokes in English (in non-English websites) and in translated versions.

Table 1. Joke diffusion by language
LanguageTranslations (mean)English in vernacular pages (mean)Translation/English in vernacular pages (1)Translation/ sum of all translations(2)
  1. Note:

  2. 1Spearman correlations between the number of jokes in the local language and the number of jokes in English in the local language page.

  3. 2Spearman correlations between the number of jokes in the local language and the total number of jokes in all non-English languages.

Portuguese355.3523.71.409**.788**
Russian324.2646.13.322**.621**
Spanish320.5926.54.496**.800**
French261.0227.13.377**.676**
German250.4653.1 .441**.713**
Chinese172.24122.71 .391**.531**
Arabic161.3760.71.524**.569**
Japanese 58.9318.44.577**.475**
Korean 44.7517.89.557**.524**

While the translation of jokes is common, it also varies greatly, both between languages and between jokes. There is a huge divergence between languages with high rates of translation and languages in which translating jokes is much less prevalent. When calculating the average number of URLs, including translated jokes, for the entire corpus of 100 jokes, Portuguese had the highest number (355.35), followed by Russian (331.64), Spanish (320.59), French (261.02), German (250.46), Chinese (172.24), Arabic (158.5), Japanese (58.93), and Korean (44.75). These results may be related to "cultural affinity" (Wu, 1998). The top languages in the list are spoken in so-called Western cultures that are mainly based in America or Europe, while the languages for which we found fewer translations are prevalent mainly in Asia and Africa.

Nevertheless, the gap between Chinese and Arabic (that are relatively open to joke translation) and Japanese and Korean (in which joke translation seems to be marginal) requires further attention. An explanation for the low number of joke translations in Japanese may relate to genre: The practice of telling short third-person stories with a punch line is not prevalent in Japan. The most common form of joking in Japan is a story about a personal experience, which is often exaggerated to create a humorous effect (Ōshima, 1999). Another explanation may be connected to linguistic distance. Hart-Gonzalez and Lindemann (as cited in Chiswick & Miller, 2005) studied the proximity of English to other languages through the measurement of foreign language training. They found that Japanese and Korean were the most distant from English, followed by Mandarin and Arabic; while French and Portuguese were found to be close to English, followed by Spanish, Russian, and German. The similarity between these rankings and our findings suggests that the volume of joke translation may be related to linguistic proximity.

Beyond this disparity by language, we also found great variation between jokes. While some jokes were extremely popular across almost all languages, a minority appeared only in the English sample; for these jokes, we could not find even a single translation on the Web. Moreover, as Table 1 indicates, the number of translations for each joke in each of the languages was clearly and significantly correlated with the overall number of times it was translated into other languages. This indicates that the global popularity of a joke is a very good indicator of its popularity in any one country, and vice versa. To reflect this diversity between jokes, we created a "globalization index" that represents the number of languages that have a large number of copies (at least 50 URLs) of a specific joke. The threshold of 50 was chosen due to its proximity to our median (52). The minimum value of the globalization index is zero and its maximum value is nine (when many translations are found in all the languages). Examining the jokes according to their globalization indexes yielded a threefold classification of the texts: (a) global hits: jokes circulated widely in at least eight of the nine languages (n = 21); (b) partially globalized: jokes that are popular in between three and seven of the languages (n = 55); and (c) translation-resistant: jokes that were found in substantial numbers only in English or in up to two other languages (n = 24).

As elaborated in the methods section, the group of global hits constituted the basis for testing the hypothesis about the direction of flows (embedded in RQ2). Our results clearly support this hypothesis: In 20 of the 21 global hits, the first date of appearance of the English versions predated the first appearance of the joke in all other languages. In most cases, the jokes in English were documented as appearing as early as the mid-to-late 1990s, whereas most versions in other languages appear in the 2000s. The only exception to this pattern was a joke about a mental asylum, which had its first documented appearance in Russian in 2000. Even in this case, however, the English version came second, several years before versions in any other language. In other words, the joke only spread to other languages after becoming prevalent in English. This resembles the process that Bloch and Lemish (2002) termed "the megaphone effect," by which non-American cultural products often penetrate the global market only after they have succeeded in the US. The next, qualitative phase of analysis showed that the jokes are not only in English, but that they are clearly marked as American.

Global and Local Dimensions of Joking

Our comparative analysis of global hits and translation-resistant jokes yielded a twofold differentiation that related both to the jokes' markers of locality and to the main themes underscoring them. The corpus of 24 globalization-resistant jokes was distinctively American. Twenty-two out of 24 texts had clear American markers, such as places and brands, phrases, and stereotypes. The jokes mention specific places such as Arkansas, Alaska, and Alabama, ascribing distinct characteristics to the people living in them. For instance, in a joke about three contractors from New Jersey, Tennessee, and Florida who are bidding to fix a fence at the White House, the contractor from New Jersey manages to get the contract by cunningly splitting the revenue between himself and the government official, manipulating the contractor from Tennessee to actually do the job. Another prominent American marker found in translation-resistant jokes is a focus on internal American politics (particularly the divide between Democrats and Republicans). In addition to geographic and political references, the texts are saturated with an array of American identity markers, including stereotypes (rednecks), companies and products (AT&T, Winchester Silver Tips), and national heroes (Amelia Earhart).

Similar to translation-resistant jokes, most global hits (16 out of 21) were apparently created in the US. However, the markers of American identity are much more subtle for this group. In three of these jokes, the only American signifiers were semiotic (for example, the use of "diapers" rather than "nappies"), whereas in other cases the references to America were mainly to well-known brands and figures such as Microsoft, Barbie, and Christopher Columbus. The texts included a small number of references to less internationally known brands, such as Victoria's Secret. No joke in this group referenced a town or state in America (in contrast to seven such references in the translation-resistant jokes), and no references were made to internal tensions underscoring American society. More frequent are general references to Western culture, such as consumption, Christmas, nuns and mental asylums.

These differences in relation to American markers might be somewhat expected; however, the thematic comparison between global hits and translation-resistant jokes yielded some surprising findings. Both groups deal with various themes related to capitalism, albeit with differing foci. Translation-resistant jokes tend to reveal negative capitalism's aspects. They deal with wages, unhappy employees, and exploitive corporations. For instance, a text titled "10 Best Out-Of-Office Email Replies" includes "Thank you for your email. Your credit card has been charged $5.99 for the first 10 words and $1.99 for each additional word in your message." This text and others embed a critical approach towards aspects of the capitalist system in America.

If translation-resistant jokes tend to focus on the agonies of capitalist production, global hits tend to be about consumption. Moreover, in 7 out of 21 jokes the exchange of money is tied to romantic relationships. In one text, a man enters a shop to buy a Barbie doll for his daughter and is told that Divorced Barbie is the most expensive one because it comes with "Ken's car, Ken's house, Ken's boat, Ken's furniture, Ken's computer, one of Ken's friends, and a keychain made with Ken's balls." The most common tie between romantic relationships and consumerism describes women as shoppers. This linkage between women and consumerism is part of a larger theme underpinning the global hits: a focus on gender differences. "Mars and Venus" humor, in which men and women are presented as different species, with contradictory emotional, social, and communicative needs (Shifman & Lemish, 2011), is manifest in five of the 21 global hits. While the theme of gender differences is prominent in American pop culture, it cannot be portrayed as distinctively American. Indeed, the roots of the emphasis on women's 'otherness' may be also ascribed to French scholars, such as Luce Irigaray and Hélène Cixous (Tong, 1998), and the early popularization of gender-difference-based representations has been seen in other countries as well.

The comparative analysis of the two joke groups led us to differentiate between "local" and "global" markers of Americanism. If in translation-resistant jokes America is marked by references to political issues, terrorism, rednecks, and internal cleavages between New Yorkers and Alabamians, globalized jokes incorporated internationally known icons, such as Bill Gates, alongside issues such as consumerism and gender differences, which may be associated more generally with contemporary Western societies.

What Happened to the Jokes in the Course of Their Translation?

Our fourth research question focuses on the extent to which translated versions introduce culture-specific elements, and on the practices of localization they apply. Out of 3448 translated jokes examined in this phase, 2415 did not incorporate any cultural changes in comparison to the English version; cultural changes were detected in about one-third of the translated texts (n = 1033). The ratio between culturally modified and nonmodified jokes was in the range of 1.85–3.52 for most languages: German (1.85 nonmodified jokes for each modified one), Portuguese (1.95), Arabic (1.95), Spanish (2.13), Korean (2.79), Russian (2.96), and Chinese (3.52). French had a stronger tendency towards domestication: Cultural modifications were detected in approximately half of the texts translated into French (at a ratio of 1.09). The high proportion of modified texts (in which the American origins are often concealed) is in line with the literature about well-entrenched anti-American sentiments in France, which has led the French to treat Americanization with a mixture of fear and loathing (Debouzy, 2006). Japanese is located at the opposite end of this axis: We found many more exact translations into Japanese than culturally modified translations – the ratio between the two groups was as high as 12.57. This may reflect a perception that these jokes were too apparently foreign to domesticate.

We further examined the group of 1033 jokes that introduced cultural modifications, probing the nature of localization processes. The qualitative analysis revealed three main patterns underpinning the cultural modifications of Internet jokes; we have termed these as "off-the-shelf localization," "custom-made localization," and "cutaway localization."

The most common pattern of domestication (found in about 55% of the modifications) was "off-the-shelf localization." This is a somewhat superficial form of local adaptation; when standard American/English markers such as names and currencies appeared in the texts, they were replaced almost automatically by their local equivalents. For example, the name Bob in the English version of one of the jokes became Nestor or Pedro in Portuguese; Roberto, Juan, or Ruben in Spanish; Kurt, Heiner, or Brend in German; and Yung Ju in Korean. Interestingly, the Chinese and Japanese versions maintained the English/American names in the translations. This "foreignizing" translation decision may stem from a sense of cultural gap that the jokes' translators felt they could not be seamlessly bridged through familiarization. Another example of off-the-shelf localization is currency replacement. In the abovementioned Divorced Barbie joke, for example, the cost in the American version is in U.S. dollars, whereas the translated versions of the joke use local currencies depending on the joke's language.

Contrasting with the instant off-the-shelf localization is what we have termed "custom-made localization." This more profound pattern of localization, found in approximately 30 percent of the modifications, involves the addition of culturally unique elements to the translated joke. Thus, if off-the-shelf localization is based on transformations in universal categories that exist in all modern societies, the custom-made localizations are often unique to a specific culture. Custom-made localization incorporates two intertwined dimensions: unique semantic expressions and references to culturally distinctive norms. For instance, some of the Spanish translations of the "Politics Made Simple" list (in which various political systems are described using a simple example of two cows) included the word narcofosa, which means graves of drug cartel workers. This unique word was used to describe the Mexican economy as corrupt and based on drug dealers. The Arabic version of "Husband 1.0" included the application Urfi 7.0 (urfi is a marriage without an official contract, customary in Egypt), and the equivalent text in German introduced the application Stammtisch 2.0 (a phrase that describes a scheduled informal meeting at a reserved table in a beer hall or some other establishment). An intriguing example was found in the Korean version of the "Men's Rules" list, in which the man protests against the local custom of keeping a couple's diary, in which the man and women reflect on their feelings. In some cases, a unique cultural element was added through charged expressions; for example, the pejorative slang for Americans in French (Ricains, Amerlocs) and in German (Ammis), in a joke describing the Americans' tremendous efforts to develop an antigravity space pen (ending with the punch line, "The Russians used a pencil"). As detailed below, such expressions may hint that alongside Americanization, the diffusion and interpretation of global jokes may also reinforce Anti American sentiments.

The third, and least common, pattern of localization, which was found only in about 15 percent of the modifications, involved the omission of American elements in the translation process; this is referred to here as "cutaway localization." Such omissions may occur due to either a lack of familiarity with specific American elements or due to cultural differences. For example, in the English version of the "Men's Rules," the male narrator regrets that his girlfriend doesn't look like a Victoria's Secret model, referring to the famous American retailer of women's lingerie. This brand was omitted in the Portuguese, Spanish, Russian, and Arabic versions of the list, since it is not familiar in many of the countries in which these languages are spoken. An example of an omission due to a cultural gap was identified in the Arabic translation of the humorous list entitled "Why Men Are Happier than Women." Many of the statements in the original English version included blunt references to sex, which were omitted from the Arabic joke versions, perhaps because they were considered too provocative for these cultures.

Conclusions

This study set out to examine globalization and Americanization through the somewhat unusual perspective of joke translation. The title of the article ends with a question mark rather than with an exclamation point. By way of conclusion, we wish to explain why.

In general, our findings indicated that user-generated globalization through joke translation is a common process: 76 of the 100 joke in English jokes appeared in three or more of the nine non-English languages that we studied. Our timeline analysis supported the hypotheses about the dominance of English in the directionality of this flow, as English versions preceded the appearance of jokes in other languages. However, the translation of the jokes varied significantly between languages. While some languages (particularly those prevalent in Europe and the Americas) incorporate many translations, the practice of joke translation in other languages (especially Korean and Japanese) was found to be marginal.

We also found significant differences between jokes with regards to global spread, leading to a threefold classification of global hits, partially globalized, and translation-resistant jokes. A comparative analysis of global hits and translation-resistant jokes revealed considerable differences between them. Whereas the former focus on consumerism and gender differences and include references to well-known American brands, the latter incorporate a different set of references to America, focusing on internal politics, regional stereotypes, and the agonies of the individual stuck in the capitalist production system.

With regards to the changes that were introduced to the translated texts, we found that in most cases, the translated versions did not incorporate any cultural changes compared to the English versions. When such cultural modifications were presented (in approximately one-third of the texts), they were often manifested in what we term "off-the-shelf-localization" – the simple replacement of American/English markers such as personal names, currencies, and sport-related brands with their local equivalents. Custom-made localization – the introduction of unique cultural elements to the texts – was found in smaller volumes. The three modes of localization identified in this study resemble processes described in the literature as "institutionalized erasures" (Kaplan, 2012), or as substitutions and omissions (Pedersen, 2007). While these processes were previously discussed mainly in relation to mass-mediated culture, we demonstrate that they are manifest also in bottom-up user-generated content. Beyond this, our work adds a distinction between deep/unique modifications and universal/superficial modes of domestication, on which we will elaborate shortly with regards to cultural uniqueness.

Overall, these findings suggest that popular Internet jokes may serve as powerful agents of globalization and Americanization. The ongoing process of joke translation formulates a global humorous sphere, even if its reach is often not evident to end users. In this new sphere, issues such as gender differences, consumerism and the agonies of computer users reflect and construct global mindsets in a reciprocal process. While this study indicates that America is dominant with regards to these values, it cannot be considered as their sole and uncontested source. Moreover, the diffusion of jokes featuring American symbols does not automatically equal Americanization: As discussed below, America may also serve as the butt of some of the jokes. Thus, the process is not negatively charged, monolithic, universal, and unconditioned– hence the question mark in the article title.

This study makes several contributions to the understanding of the intersection between new media, language, globalization, and cultural uniqueness. Since the use of language has been tied in many studies to cultural practices and local norms, linguistic diversity is often identified with cultural diversity. Without contesting the basic assumption about the intimate language-culture linkage, we claim that in the process of user-generated globalization through joke translation, local languages may contribute to homogenizing/Westernizing effects. Ironically, the widespread practice of off-the-shelf-localization has the potential to deepen globalization, as the combination of local markers with a familiar language makes the jokes seem local, camouflaging their Western/American origins and thus possibly undermining opposition to them.

Our analysis thus illustrates Roland Robertson's (1992) perception of globalization as "a twofold process involving the universalization of particularism and the particularization of universalism" (p. 102). On the one hand, jokes' domestication (or even merely their translation) can be read as "particularization of universalism" because universal structures are tailored and modified to better speak to local cultures. At the same time, this process also reflects the "universalization of particularism," since the same practices seem to take place in all languages: for instance, they all use local currencies as markers of uniqueness. Yet–corroborating Kaplan's (2012) conceptualization of "cultural erasures" through which national meanings are ascribed to global models which become invisible to end users–there seems to be a gap in the levels of signification characterizing the universal and the particular. Whereas the globally shared values are anchored in deep meaning structures which are often invisible, notions of locality and local uniqueness are invoked through explicit signs and symbols.

The only exception for this signifying system relates to the US. While existing research about banal globalism describes cosmopolitan markers that remind people of their standing as global citizens, in the jokes that we examined such markers appeared only with reference to American products (e.g., Barbie) or citizens (e.g., Bill Gates). At the same time, we showed that not all things American are globalized: a distinct set of signifiers (e.g., "Alabama", "Rednecks" and "Republicans"), was found only in translation-resistant jokes. In other words: While Americanization and localization have concrete markers, globalization often doesn't. Yet this does not mean that globalization is not happening – it is just more difficult to capture.

Alongside these contributions, the study has three main limitations: We focused solely on verbal humor, we examined only texts and not their reception, and the starting point of our cross-lingual expedition was English. We hope that these shortcomings will be addressed in follow-up studies. Whilst our focus on verbal humor enabled us to trace large numbers of texts through search engines, future research could benefit from looking into visually based humor types. Given that visual humor is becoming prevalent in online comedy (Shifman, 2007; Kuipers, 2002), and since it is much easier for images to travel internationally than it is for words, new sites and modes of visual glocalization are expected to emerge. Moreover, we ascribed the lack of translations of jokes into Japanese to the marginality of this genre in Japan. Looking into translation patterns of other genres may validate this explanation, or it may provide further evidence for a broader wall between these cultures and globalized American/Western culture.

While our study focused on texts that were stripped from the context of their production and consumption, future research should look into the ways in which different audiences interpret imported Internet jokes. This may be of particular interest since humorous texts are particularly polysemic forms of communication. The inherent association between humor and meaning multiplicity, or polysemy, has been documented in studies that showed how the same comic texts are interpreted differently by different groups (Park et al., 2006; Vidmar & Rokeach, 1974). This innate ambiguity has been associated with a broader argument about humor's societal implications: humor has been widely described as a double-edged sword that fulfills the contrasting societal functions of identification and differentiation, control or resistance. Thus, while a joke may serve dominant social forces through the mockery of deviants, it may also function as a liberating weapon through which the oppressed criticize those in power (Meyer, 2000; Lynch, 2002). Taking this knowledge into account, future studies may investigate to what extent jokes are used by Internet users across the globe as ways of laughing at America or the West. Thus, the texts that we describe as promoting Americanization may be used subversively to undermine it. For instance, jokes that mock Microsoft's belligerent business strategies may be easily interpreted as criticizing the nation with which the company is associated. Such further examinations of user interpretations may be enriched from taking into account the literature about cross-cultural values (e.g., Schwartz & Bilsky, 1987), which has been already applied to explore the consumption and reception of humorous content (Crawford, 2013).

Finally, we believe that future research should turn this study on its head by changing the direction of inspection. Not only was our initial set of 100 popular jokes in English, but we also found that it largely reflected an American perspective. While this may be a legitimate starting point considering the vast literature on globalization as Americanization (which was supported by our timeline findings), other directions of globalization should also be examined. Therefore, our dream study would start from the top 100 jokes in each of the nine languages examined in this project, tracing their translations to other languages. Such an analysis would shed new light on questions related to the substance of global flows, as well as on their trajectories.

Acknowledgments

We wish to thank the following individuals, who contributed significantly to various phases of this study: Rahel Blum, Lillian Boxman-Shabtai, Mila Chaseliov, Isabelle Cohn, Elena Epelman, Adam Fineberg, Allison Hantos, Tahani Hasssadiea, Jenny Hestermann, Gabi Holzhacker, Teruto Ishiida, Yejoong Jung, Stephanie Lee, Noé Michael, Suguru Nemoto, Asaf Nissenbaum, Amal Odeh, Frédérique Schillo, Jennifer Sebban, Sogoud Suleiman,Yujin Surh, Matias Szuchman, Lumin Tao, Juhae Yoon, Noa Zubarev. We are also imbedded to Nicholas John and two anonymous reviewers for their insightful comments on this manuscript. Finally, we are grateful to the German Israeli Foundation (GIF) for funding this project.

Biographies

  • Limor Shifman is a Senior Lecturer at the Department of Communication and Journalism, the Hebrew University of Jerusalem. Her main research interests are new media, popular culture, and the social construction of humor. Address: Department of Communication and Journalism, the Hebrew University of Jerusalem, Jerusalem, 91905, Israel

  • Hadar Levy is a PhD candidate at the Department of Communication and Journalism, the Hebrew University of Jerusalem. Her main research interests are sound studies and communication technologies. Address: Department of Communication and Journalism, the Hebrew University of Jerusalem, Jerusalem, 91905, Israel

  • Mike Thelwall is a Professor at the School of Mathematics and Computer Science, University of Wolverhampton, UK. His main research interests are the social web, sentiment analysis, and altmetrics. Address: School of Mathematics and Computer Science, University of Wolverhampton, Wulfruna Street, Wolverhampton WV1 1LY, UK

Ancillary