2.1 Ironman France 2010

Ironman France, held in Nice and its hinterlands, was until recently the only established competition in a Francophone context. In 2010, the seventh year of its running, approximately 2800 professionals and amateurs competed in the event. They were accompanied by several thousand additional supporters, fans and commercial participants.Footnote ² The linguistic background of athletes, determined by country of origin on official documents, reflects most sanctioned Ironman races held outside North America: the largest number of competitors came from the host country, although no nationality – and, presumably, language group – constituted an absolute majority. In order of importance, French was the most represented language among amateur athletes (with 1170 or 43% of enrolled participants), followed by English (548 or 20%), Spanish (326 or 12%), German (128 or 4.5%), and Italian (94 or 3.5%); the remaining participants hailed from more than 40 countries and presumably analogous linguistic backgrounds.Footnote ³

The de facto languages of Ironman France were French and English. Both were used for all communications during the months leading up to and during the days following the race, including athlete briefings, safety guides, tips for supporters and spectators, results and commercial marketing. Despite the parallel position of English in official print and online material, French overwhelmingly dominated both spontaneous and planned discourse during the 2010 event itself, as well as in related activities, including opening and closing ceremonies, informational and press briefings, expositions, race officiating and award ceremonies. All oral communications were made first in French and only subsequently, if at all, in English; the latter were without exception translated and truncated versions of their French equivalents. Thus, while English was present, it was not on equal footing with French.

2.2 English in Ironman France

One of the most prominent characteristics of the Ironman France community of practice is the appropriation and use of English words and expressions not seen outside of this context. In order to gain a clearer picture of these innovations, linguistic data were compiled from three sources: official race material; online discussion groups; and spontaneous communication. This selection covers a range of performances, facilitating investigation into varying communicative settings.

Official materials comprise examples of highly redacted linguistic performances, including athlete guides, briefings, race preparation advice, and other information made available via the race organizers’ website and distributed in print during the event.Footnote ⁴ These materials are notable for their simultaneous commercial and community-oriented characteristics: the organizers have the concomitant responsibility of promoting their event (a for-profit competition), maintaining safety, and complying with rules emanating from various governmental and sanctioning bodies, while also promoting a sense of mutuality among participants and their supporters. Official materials included 84 pages of printed material and the equivalent thereof in internet content, totaling approximately 3500 words.

Online discussion forums were selected to provide insight into more spontaneous but asynchronous output, where the careful redaction and self-censorship characteristic of print material is largely absent (for issues surrounding computer mediated language use, see e.g. Guédon, Reference Guédon1997, Herring, Reference Herring, Schiffrin, Tannen and Hamilton2001, Crystal, Reference Crystal2006). Four French language forums were surveyed on two popular websites dedicated to triathlon and related topics (www.xtriathlon.com and www.onlinetri.com).Footnote ⁵ The forums were chosen according to their original thread titles: preparation for Ironman France 2010; coverage of the Ironman World Championships; and subjects of interest to athletes engaged in Ironman competitions (training regimes, costs, equipment, injury, logistics and course evaluation). It should be noted that discussion frequently diverged to cover a wider range of distally related and wholly unrelated topics, notably doping, weather, professional French triathletes, family, travel and sex. The selected forums included a total of 1021 unique postings (81,946 words total, excluding signatures and footers) by 65 individuals, all of whom contributed under the guise of self-selected usernames. Although the linguistic background of contributors cannot be assured, it is reasonable to assume that all are native or highly proficient non-native French speakers.Footnote ⁶

Lastly, spontaneous discourse was solicited to provide examples of synchronous linguistic output. Five male, native French speakers (ages 31 to 48) were contacted via acquaintances of the author and agreed to be interviewed. Three were from the Parisian region (of whom one grew up in Normandy) and two were from Nice (of whom one grew up in Lyon). All interviewees had little English proficiency, measured as a self-reported ability to carry on informal conversation in English, although all studied English in secondary school, and none reported using English on a regular basis or having lived in or travelled to an Anglophone country for more than one month. Each interviewee participated in the 2010 event as his first Ironman competition, placing respondents in the camp of inexperienced athletes.

Interviews were conducted via Skype, a Voice and Text over Internet Protocol permitting two-way synchronous communication. This allows for users to type their responses such that the text is instantly transmitted in a manner more closely simulating distal oral communication. The use of text (rather than voice) communication facilitated subsequent data tagging and extraction, as what participants intended to communicate was clearly made available to the author with little risk of distortion in transmission. Text communications were saved for subsequent analysis without the need for impressionistic or phonetic analysis (phonological adaptation being a topic beyond the scope of the present work).

Participants were provided two prompts (Parlez de vos préparations pour Ironman France cette année, ‘Speak to me about your preparations for Ironman France this year’; and Racontez la compèt’, y compris vos emotions pendant les moments difficiles, ‘Speak about the competition, including your emotions during difficult moments’) and asked to respond for approximately five minutes. Additional probing questions were framed using the participants’ language, e.g. when one participant stated, ‘j'étais déçu par mon bike’ (‘I was disappointed by my bike leg’), the author prompted him by inquiring, ‘qu'est-ce qui s'est passé au bike?’ (‘what happened during the bike leg?’). All interviews involved questions of this nature, ranging in number from three to eleven per interview. Subjects were informed that their responses were part of an investigation into communication and interaction involving Ironman France athletes, but were not told until after the interviews were completed that the primary foci of the investigation were English words and expressions. The interviews yielded 5248 words (excluding hesitations, repetitions, and apparent phatics, as well as questions from the author).

All sources were examined for terms and expressions determined to be of English origin. A word or expression was included in the dataset by virtue of its etymological notation in the Grand Robert de la Langue Française (i.e. as having been borrowed from English) or by its lack of appearance in this reference. The issue of what constitutes a borrowing is, of course, far more complicated than the present work can fully address (see Ayres-Bennet & Carruthers, Reference Ayres-Bennett, Carruthers and Temple2001: 324–326). Several English borrowings were excluded from investigation, as they were felt to be widely used and relatively well-established, e.g. marketing. Proper nouns (e.g. Ironman); quasi-proper fixed expressions (e.g. Made in USA) were also excluded.

Tokens that appeared at least three times throughout the corpus as a whole were manually tagged, as were any pertinent native equivalents. These data were subsequently classified as one of two types: those lacking an accepted and widely-used French equivalent; and terms and expressions having an accepted and widely-used French equivalent. A distinct sub-category of tokens comprises morphosyntactic calques and blends, as well as slogans implying phonetic wordplay. The respective data are provided in Tables 1, 2, and 3.Footnote ⁷

Table 1.

Integrated tokens

Table 2.

Tokens with widely used or accepted native equivalent

Table 3.

Other innovations (morphosyntactic constructions, calques, ‘Frenglish’)

Frequency of occurrence for tokens of the second type, as well as their native equivalents (when used), were tabulated within each of the three source corpora, provided in Table 4. In the case of online forums, in which original posts are commonly copied verbatim into subsequent responses, frequency of occurrence tabulations required closer investigation after manual tagging. Calculations include the first occurrence of a token by a forum participant and those integrated in responses directly addressing an original posting by subsequent participants (e.g. when a forum respondent disagreed with part of an original post and specifically quoted it), but excluded tokens that appeared in automated copies of original postings (i.e. when a forum respondent did not directly address an original quote in his response and the presence of the token was deemed ancillary to a given response). As seen in Table 4, frequency of occurrence is relatively low (at or below 0.001) in the forum corpus. This is most likely due to numerous discursive excursions on the part of participants, who engaged in lengthy threads concerning topics having little a priori connection to triathlon, notably sex. These substantially increased the size of the corpus and reduced the overall frequency of occurrence of pertinent tokens. By contrast, the frequency of occurrence of English origin tokens in official media is relatively higher. This is putatively a byproduct of the highly focused, commercial communication engendered by this media and the close association of the sport and its promoters with English, a matter discussed further in §4.

Table 4.

Count, frequency of occurrence (rounded), and ratio of English to native equivalent (NE)

The overwhelming majority of tokens (all of Table 1; 14 of 16 of Table 2) are nouns, whereas only two adjectives (clean and oversize) function solely as adjectives. Several nouns variably occur as predicates, e.g. qualifier (‘tu es qualifer’; ‘you're a qualifier’), or adjectivally, e.g. dropoff (‘la zone dropoff run’; ‘the run dropoff zone’). Tokens given in Table 3 (adaptations and complex constructions) also show a clear tendency toward the appropriation of nominals, although there are a small number of verbal adaptations (see discussion below). These observations reflect Poplack, Sankoff and Miller's (Reference Poplack, Sankoff and Miller1988) scale of borrowability. All tokens can be classified as straight borrowings, appropriating both form and meaning (albeit at times with restriction) from the source language (Haugen, Reference Haugen1950; see also Ayres-Bennett and Carruthers, Reference Ayres-Bennett, Carruthers and Temple2001), as in Tables 1 and 2, as well as some examples from Table 3 (e.g. les big start, le long day), or as variably adapted, compounded or calqued borrowings, as in most data in Table 3. Because this study focuses only on written representations of language, it is not possible to determine the extent to which they are (or are not) otherwise adapted.

3.1 Language ecology and sociolinguistic niches

The notion of language ecology dates to Haugen (Reference Haugen1972), who employed this term as a means of capturing the nature of language within its social, environmental context. More recently, Mühlhausler has refined Haugen's original ideas, noting that a linguistic ecology is ‘a dynamic system consisting of a number of inhabitants and meaningful interconnections between them’ (Reference Mühlhäusler and Pütz1997: 5). Similar to biological and anthropological understandings, language ecology conceptualizes the relationships among inhabitants of a given space and at a given time, and may be defined widely or narrowly, depending upon a number of variables (geography, economy, demographics, etc.). Ecological relationships are sustained by a plurality of functional links which create mutuality and support; conversely, a lack of links or their erosion can create conflict and lead to ecological collapse (Reference Mühlhäusler and Pütz1997: 4–5). Importantly, ecologies are not advanced as substantive truths, but heuristics, i.e. as tools for conceiving of human performances and the competences putatively underlying these within a social framework. A language ecology is thus a metaphor within which a sociolinguistic environment can be understood and the interactions between its members be formalized. Ingredients contributing to an ecology include social networks, relationships between speakers and groups of speakers, and the functional distribution of languages or language forms (e.g. dialects, registers), as well as the formal statuses enjoyed by languages within the larger community and the effect of these statuses on function and individual performances (see also Mufwene, Reference Mufwene2001).

This paper focuses on a sociolinguistic context that is relatively restrictive in terms of the number of inhabitant-participants, the nature and scope of their connection, and the multiplicity and diversity of their interactions. The community of practice surrounding Ironman France more closely resembles that which Altmann, Pierrehumbert & Motter (Reference Altmann, Pierrehumbert and Motter2011) term a niche, i.e. as comprised of individuals pursuing a common or narrowly defined goal (see also Laland, Odling-Smee & Feldman, Reference Laland, Odling-Smee and Feldman2000, Beckner et al., Reference Beckner, Blythe, Bybee, Christianen, Croft, Ellis, Holland, Ke, Larsen-Freeman and Schoenemann2009). A sociolinguistic niche is defined by and does not supersede a singular or narrowly defined objective shared by its participants. Thus, a niche can be understood as a narrowing of the ecological criteria articulated in Mühlhausler (1997). In many regards, a niche is the complement to a dense social network (see e.g. Milroy, Reference Milroy1980): it is dynamic, multiplex, but shallow or narrowly defined. For these purposes, the ecology of a language niche can be understood as below:

1. (1) A niche includes a diversity of participants conjoined by a singular or highly restrictive associative link.

2. (2) The relationship between participants is dynamic, but is defined by their situation relative to the associative link, i.e. by an individual or individuals’ status as participant in the activity or enterprise constituting this link.

3. (3) The relationship between participants is founded upon and sustained by this link in an exclusive or quasi-exclusive manner.

4. (4) Without this link, the relationship between participants would devolve.

This is, of course, not the only sociolinguistic environment which could be understood with recourse to the above definition. Indeed, numerous studies have looked at similar, highly focused sociolinguistic environments and the linguistic performances of their respective participants. These include online communities, e.g. gaming societies (Soares Palmer, Reference Soares Palmer2010), and language use among marginalized sub-sections of larger cultures, e.g. drag queens (Barrett, Reference Barrett and Myers-Scotton1998). The particular niche under investigation is interesting as it emerges from a unique sporting event involving large numbers of amateur athlete-participants, themselves of varying socio-economic and cultural backgrounds; the linguistic performances seen among these participants are also particular to this context, reflecting the appropriation and use of English in a number of other niches, albeit with distinct tokens and putatively unique connotations and associations.

3.2 Language fetishization

Few terms used to describe human relations with or beliefs about the objects in their social context have enjoyed as checkered a history as fetishism and fetishization. Particularly relevant for the present discussion is Marx’ view of commodity fetishization, i.e. as the objectification of goods, services, and other traded articles, which are imbued with values removed from their physical substance (Reference Marx1867, 1 §4 = 1932: 78). Marx’ understanding, whereby an object's value is decoupled from its use and attached to its societal connotation, is called upon by Kelly-Holmes (Reference Kelly-Holmes1997, Reference Kelly-Holmes2000), who investigates borrowings in advertizing. She analyzes examples of language use in which form is divorced from (referential) function, i.e. where signifiers are chosen for the connotations and associations which they evoke. In her examples, words and expressions are not used to activate a referent on the part of interlocutors, but to activate connotations and associations: in essence, signifiers are used without the necessary implication of a signified.

Instances of language fetishization in advertising and product marketing abound. The slogan Fahrvergnügen, used by Volkswagen cars in the 1990s and cited by Kelly-Holmes, provides a particularly compelling example, especially as this neologism, purportedly the equivalent of ‘driving pleasure,’ was purposefully coined for the North American consumer. Other noteworthy examples include c'est si bon, used in television advertising for Yoplait yoghurts in the United States, and geen het jammeren (sic; literally ‘no the to moan,’ using a Northern Dutch verbal lexeme), the mistranslated slogan for Beljum Butter sports cream. In each of these, language forms become symbolic of stereotypical cultural attributes of speakers of that language: German conveys solid automotive engineering and construction; French, culinary sophistication and quality; and Flemish, a cycling history steeped in tradition and bravura. In a similar vein, according to Kelly-Holmes, English is used in advertising to convey openness and worldliness, where the language is symbolic of a global identity and not directly tied to a particular country or culture (2000: 76–77). This is echoed, albeit without use of this term, in Chesley (Reference Chesley2010), who presents evidence for English borrowing being a distinct phenomenon in contemporary French. Thus, whereas the connotations associated with German evoke technical prowess, French an ill-defined savoir vivre, and Flemish athletic tradition, the connotations evoked by English activate notions of openness and cosmopolitanism on the part of speakers, notably those living in Western market economies. It is argued below that the use of English evokes additional, community-specific connotations within the Ironman niche.

4.1 Straight borrowings

The tokens provided in Table 1 are, in many respects, unremarkable in Modern French, which has borrowed countless terms from English. Data such as hot spot (a viewing areas designated for supporters, fans, and the press) or gel (also spelled gelle, a high-calorie nutritional concentrate used during competition) function much like any recent other loan in French, e.g. airbag (see Martin, Reference Martin2007 and Reference Martin2008). These terms can be retraced to the language community in which referents originated and which provided the surface form. Whether or not a particular signifier has gained acceptance among users outside of this or similar contexts, and despite any acceptance or proscription by language authorities, dictionary editors or the like, such surface forms function mutatis mutandis like other borrowings and are used in a manner similar to nativized words. Additional terms noted in the corpora, such as leader and briefing, are included in the Grand Robert de la Langue Française, which dates them to 1829 and 1945, respectively. Data in Table 1 can thus be readily understood within any number of theoretical frameworks (see e.g. Winford, Reference Winford and Hickey2010 and Thomason and Kaufman, Reference Thomason and Kaufmann1988). By contrast, the terms given in Table 2 are not accepted by dictionaries and, more importantly, are not observed with any regularity outside of sports contexts or, even, outside of triathlon.

Many of the tokens in Table 2 are used more restrictively than the corresponding native item, as shown in Table 4. For example, finisher is used exclusively to refer to an individual who has completed an Ironman distance race and not another race format (e.g. an Olympic distance triathlon or a marathon), at least in the analyzed sources. Run is frequently used in reference to this component of Ironman, e.g. ‘j'ai eu du mal au run’ (‘I was sick during the run’) and only infrequently in discussion of training or other running events, for which course à pied or the acronym CAP/CàP are noted far more often. Likewise, clean is used only to refer to athletes who do not engage in doping and never in the wider sense of one who is morally or ethically upright, much less to describe the physical appearance or quality of an athlete or situation. Many such terms are reportedly in frequent use in other athletic contexts, e.g. coach which has seeming gained popularity in team sports such as football/soccer and rugby.

Many data were used interchangeably with an equivalent French form, as seen in Table 4. Frequency of use of native versus English forms was not straightforward, however. Several English-origin tokens are used more often than their French equivalents. This is the case for bike (referring to the cycling component of a race or training, e.g. ‘je l'ai croisé au bike’ (‘I came across him during the bike leg’), as well as in reference to equipment, e.g. ‘je voudrais bien son bike’ (‘I'd love [to have] his bike’)). This duality parallels vélo, which is seen less frequently in all source corpora. The use of run is similarly distributed across corpora, with restrictions as noted above. Other English-origin forms used more frequently than or to the exclusion of their French equivalents include finisher and la finish/finishline, qualifier, slot, half (referring to a half Ironman distance race), draft/drafting, dropoff, check-in and –out, and taper/tapering.

It is worth noting that certain native terms were used relatively more frequently than respective English origin equivalents in one or all corpora. This is particularly the case for swim (referring to the swim component of a race): in official media natation/nage was more frequently noted, whereas the reverse was true in online forums and discourse. Another example of this is seen for race, referring to the event itself (e.g. ‘RDV [rendez-vous] en trois mois pour le race’ (‘a date in three months for the race’)), for which the native equivalents compèt/compétition were far more frequently noted in all surveyed sources.

4.2 Morpho-syntactic innovations

In addition to the above, a number of morphological or syntactic constructions of putatively English origin were noted (Table 3). Several of these, especially those extracted from official media, recurred with high frequency and to the exclusion of an equivalent French expression or construction. This is notably the case for nouns directly modified by another noun, e.g. stand infos, produits running and bike service, which follow an established noun + noun pattern in modern French (viz. station service, point Internet). In no surveyed sources were the normative French equivalents observed.

Equally interesting is the use of multiple nominal modifiers in the case of la transition bike run or la transition swim bike, describing the timed sequencing of legs of a triathlon event. These constructions alternated with native equivalents la transition vélo CAP and la transition nage vélo; normative syntactic constructions (e.g. la transition de la nage au vélo) were not noted in any surveyed sources. Other innovations include apparent cases of expressions borrowed in toto, e.g. les big start (describing the favored or well-known professional racers),Footnote ⁹ and calques, e.g. le sur-entrainement (likely derived from the English overtraining). Morphological accommodation was also observed, e.g. adaptation of the English verb ‘to clip’ (as in, ‘to clip shoes to a pedal’) as clipser (e.g. ‘tu peux clipser tes chaussures sur ton vélo,’ (‘you can clip your shoes to the bike’), ‘[les] chaussures clipsés [sic] au bike’ (‘shoes clipped onto the bike’)). Also noted was an innovative compound or loan-blend, ultégracarbonné, derived from Ultegra (a bicycle component brand) and carbon (< ‘carbon fiber’, a lightweight material used in the construction of bicycle frames), e.g. ‘j'mouline sur vélo ultégracarbonné’ (‘I grind away on an ultegra-carbon bike’). Mention should also be made of use of the innovative use of trainer (< English to train), referring to the act of training. Perhaps because of homography and potential homophony with the normative verb traîner (‘to trail, linger’), there were relatively few occurrences of this that could be clearly viewed as the equivalent to s'entraîner, the latter being noted far more frequently.

A final series of tokens noted in the sources are more difficult to classify. These include two official slogans for the 2010 event, show devant and ça va être show, and three noun phrases, la Energy Party (an event held before the competition for participants, supporters, fans, and the press), la Heroes Night (an awards and recognition event held after the competition), and l'Expo Village (a large merchandizing area established during the days before and after the event for the purposes of promoting the race and its official sponsors). Far more restricted in their usage than the tokens discussed above, these are nevertheless seen with high frequency in official media and recur in both discourse and forums. In no surveyed sources were native equivalents for any of these tokens observed, although one token (l'Expo Village) alternated with l'expo/l'exposition and there were a plurality of approaches to the form of the definite determiner associated with Heroes Night: les Heroes Night, la Heroes Night and l'Heroes Night were alternatively observed.