2. PURISM

Preservation of a language’s purity is, according to Coulmas, part of this ‘prescriptivist agenda’ (Reference Coulmas, Tieken-Boon van Ostade and Percy2017: 49). Purism, defined by Walsh (Reference Walsh2016: 1) as ‘a belief that the language of a particular group of speakers is in decline or threatened in some way by foreign borrowings or by internal changes, and that the language therefore needs protection from these undesirable changes’, is often considered as ‘un trait typiquement français’ (Paveau and Rosier, Reference Paveau and Rosier2008: 37). This is likely due to the work of the Académie française and the country’s history of linguistic intervention (Walsh, Reference Walsh2016: 2). Moreover, Paveau and Rosier (Reference Paveau and Rosier2008: 73) assert that purist commentary on language is natural and normal: ‘chacun de nous est un peu, beaucoup, à la folie, passionnément puriste, particulièrement en ce qui concerne la langue’. If we are all to some extent purist, there may be a wealth of lay purist attitudes displayed on Twitter.

Thomas’ (Reference Thomas1991) theoretical framework for describing purism, which is the most substantial to date, suggests that purism is motivated largely by non-rational factors, such as aesthetics or nationhood, but that the distinction between rational and non-rational is not a clear-cut binary (Reference Thomas1991: 35–36). Rather, motivations are more complex and may be considered both rational and non-rational. Thomas also distinguishes between two types of purism: internal and external. External purism is ‘xenophobic’ (connotations of racism from xenophobia not present), targeting foreign elements in language (Reference Thomas1991: 81), such as Anglicisms in France (Ricento, Reference Ricento2009: 120). Internal purism is a belief that linguistic change and/or error within French is contaminating the purity of the language, e.g. non-standard spellings, or in this case, the linguistic intervention in the orthographic system. This study focuses on internal, rather than external, purism.

Purist discourse draws upon a number of images to express opinions on language use (Thomas, Reference Thomas1991: 19), which have been widely studied, particularly in the French and German contexts. In Paveau and Rosier’s (Reference Paveau and Rosier2008) in-depth study of primarily internal purism in France, it is suggested that French purist discourse draws on the following arguments: aesthetic (beau/laid); political (langue de la liberté); pseudo-linguistic (clarté de la langue); and metaphorical (langue en bonne santé) (Reference Paveau and Rosier2008: 57). Thomas’ model also includes aesthetics, but in a broader sense, encompassing values such as ‘correctness’, ‘wholeness’ and ‘pristineness’ (Reference Thomas1991: 39). It is of note that these arguments focus on the language itself, and not the user, as a point of criticism. Thus far, the purist imagery and arguments used online have not been analysed in detail in the French context.

Walsh (Reference Walsh2016), in her comparative study of internal and external purism in France and Quebec, explores purist attitudes at the official, group and individual level in both places. The metalanguage used by language societies and their attitudes is considered but is not examined at the level of individual speakers. The current study focuses on the opinions of individuals in reaction to direct government intervention into the language in a computer-mediated communication (CMC) context.

3. THE SPELLING REFORM OF 1990

Attempts to maintain the primacy of Standard French have punctuated the language’s history (Bourhis, Reference Bourhis, Ryan and Giles1982: 34) and the spelling system has been reformed numerous times (Groupe RO, 2012: 130). In the years leading up to the Rectifications orthographiques in 1990, there was a general feeling of acceptance for a moderate reform ‘to modernise the spelling of French, without threatening the foundation upon which it is built’ (Catach, Reference Catach and Sanders1993: 141–142). Both linguists and laypeople were discussing the possibility of a potential reform (Arrivé, Reference Arrivé1994: 3); a study published in L’école libératrice, claimed that 90 per cent of French teachers backed a simplification of French spelling (Goosse, Reference Goosse1991: 120).

In reaction to the apparent desire for reform, Prime Minister Michel Rocard founded the Conseil supérieur de la langue française (CSLF) in 1989, an institution headed by Rocard himself, which united journalists, linguists and writers, amongst others, to work together on a number of rectifications to the spelling system (Arrivé, Reference Arrivé1994: 5–7). During their first meeting, on 24 October 1989, Rocard presented five areas to be targeted: the use of hyphens, the pluralization of compound words, the circumflex accent, the past participles of reflexive verbs and a number of ‘anomalies’ (Arrivé, Reference Arrivé1994: 5–7). Once achieved, Rocard and Maurice Druon, as Secrétaire perpétuel of the Académie française from 1985–1999, approved the corrections (Catach, Reference Catach and Sanders1993: 143). The official report was subsequently published in the Journal officiel de la République française in December 1990 (Ousselin, Reference Ousselin2004: 491).

Despite the positive sentiment demonstrated from the late 1980s (Dupriez, Reference Dupriez2009: 20), the rectifications were not well received by the media (Goosse, Reference Goosse1991: 118) or the speech community (Reyes, Reference Reyes2013: 337) and were quickly relabelled as a réforme de l’orthographe (Arrivé, Reference Arrivé1994: 7). This was despite clear attempts in the official documentation to highlight that the changes do not constitute a reform (Académie française, 1990: 3), which may be seen as an upheaval of the system. Such resistance is not uncommon (Reyes, Reference Reyes2013: 337) or surprising given the strong cultural tie that the layperson creates between spelling and culture (Goosse, Reference Goosse1991: 119). As Watts explains: ‘Les Français plus que tout autre nation associent langue et culture’ (Reference Watts1991: 88). Rocard continued to defend the proposed changes, despite the negative reaction, while the Académie distanced itself and did not recommend the usage of the new spellings (Arrivé, Reference Arrivé1994: 7). Dictionaries began to include the reformed spellings alongside the old spellings and, in April 2007, the Ministry of National Education announced the official recognition of the new spellings in the school system from 2008 (Dupriez, Reference Dupriez2009: 23).

In February 2016, discussion of the reform resurfaced when it was announced that from September 2016 the corrected spellings would be used in educational textbooks (Le Hanaff, Reference Le Hanaff2016), leading to widespread discussion on social media. This prompted the Académie to publish an online statement, which, despite acknowledging Druon’s official approval of the report, declared: ‘L’Académie française tient tout d’abord à rappeler qu’elle n’est pas à l’origine de ce qui est désigné sous le nom de « réforme de l’orthographe », dont la presse se fait l’écho depuis quelques jours’ (Académie française, 2016). On 4 February 2016, the Education Minister, Najat Vallaud-Belkacem (Reference Vallaud-Belkacem2016) also reacted to the media backlash via her official website, stating that the proposal of changes was outside her remit, but rather that ‘ce travail revient à l’Académie française’. This was followed by an interview, published on Le Figaro’s website, with the current Secrétaire perpétuel of the Académie française, Hélène Carrère d’Encausse, who claimed that the Académie had in fact not approved the reform itself, but rather a limited number of changes to anomalies (Aïssaoui and Corty, Reference Aïssaoui and Corty2016). In the face of social media backlash, neither the Académie nor the French government were willing to take responsibility for the Rectifications orthographiques.

5. REACTIONS

#JeSuisCirconflexe tweets fell into four categories of reaction towards the spelling reform (created post hoc following analysis of the data): negative; informational; hashtag seen as an overreaction; and positive.Footnote ⁶ Neutral tweets which shared a link or information about the reform were classified as informational (79 tweets – 6.6%): e.g. ‘La réforme de l’ #orthographe, résumée en deux points’ (JE-0027). The hashtag seen as an overreaction category includes tweets which criticise other Twitter users for using the hashtag #JeSuisCirconflexe (41 tweets – 3.4%): e.g. ‘#JeSuisCirconflexe #JeSuisCharlie vous associez de l’orthographe au drame de Charlie hebdo c’est du grand n’importe quoi’ (JE-0159). The sample by reaction types is shown in Table 2.

Table 2. Composition of reactions by type (#JeSuisCirconflexe)

The majority of tweets in the sample, 998 (83.7%), display a negative reaction to the reform: e.g. ‘Cest vraiment trop trop nul cette réforme de l’orthographe #JeSuisCirconflexe’ (JE-0282). Within the negative reactions, an additional emotional response was found in 104 tweets (10.4%) of which shock, anger and sadness were the most prevalent: ‘M…mais je pensais que c’était un hoax!’ (JE-0193); ‘Je suis véritablement INDIGNÉ’ (JE-4476); ‘Ils ont brisé mon coeur’ (JE-1500). These emotional responses may suggest a non-rational response to the reform.

The #JeSuisCirconflexe hashtag was critiqued in 22 tweets (1.9%): ‘LE HASHTAG LE PLUS DEBILE DE L’HISTOIRE DE TWITTER !’ (JE-0738); ‘Tous les fdp [fils de putes] qui tweet #JeSuisCirconflexe et qui ont tweeter #JeSuisCharlie l’année dernière nvm [niquez vos mères], j’vous jure vous respecte pas.’ (JE-1080); ‘Y a qd même des gens qui osent utiliser #JeSuisCirconflexe en référence à#JeSuisCharlie’ (JE-4261). The language used in these examples – débile, fdp, and osent – is more aggressive than in other categories of the sample, and this is typical of tweets criticizing the hashtag. The choice of vocabulary expresses the anger felt towards the appropriation of the #JeSuisCharlie hashtag for a very different occurrence.

Of the three tweets (0.3%), which reacted positively,Footnote ⁷ two mention language evolution: ‘Une langue qui n’évolue pas est une langue morte…’ (JE-3510); ‘notre langue évolue !’ (JE-4313). Both users acknowledge that languages change over time, and by using the hashtag #JeSuisCirconflexe, link such natural evolution to the reform. The final positive tweet states ‘Je comprends parfaitement l’utilité de cette réforme de l’orthographe. Il suffit de lire les tweets de votre TLFootnote ⁸ pour comprendre pourquoi’ (JE-1991). This user suggests that non-standard language is not acceptable, even online, and that a reform is both needed and likely to improve the situation. All three positive tweets acknowledge a need for a language to change over time, but provide no explicit support for the changes proposed by the reform.

Table 3 displays the frequency of each reaction type for the two-hour samples of tweets from search terms ognon, #JeSuisCirconflexe and reforme orthographe. Thirty-three tweets (19.5%) using the neutral term reforme orthographe were informational, in comparison to seven ognon tweets (8.6%) and five #JeSuisCirconflexe tweets (5.6%). This is unsurprising given the neutrality of the search term, reforme orthographe, and the neutrality strived for in official information-sharing. Two ognon tweets (2.5%) and seven reforme orthographe tweets (4.1%) displayed positive reactions to the reform; if we look at the #JeSuisCirconflexe tweets collected during these two hours, there were no positive reactions. Whether the search term was negatively marked or neutral, negative reactions formed the vast majority of tweets. It is not possible to say, at least with any certainty, whether this seemingly widespread negativity is representative of popular general opinion; first, as there is no guarantee that this sample is representative of the population; and second, it may simply be that those who agree with the reform, or are ambivalent to it, did not tweet on the subject.

Table 3. Types of reaction towards the reform (all samples, 2 hours)

6. THE OBJECT OF CRITICISM

#JeSuisCirconflexe tweets present differing opinions about where responsibility lies for the resurfacing of the reform. The then French President, François Hollande, and current Minister of Education, Najat Vallaud-Belkacem, were the two most blamed figures in the sample. Neither of these politicians was directly involved in the spelling reform, first published in 1990, but, arguably, their positions of power in 2016 mean they are to some extent responsible for its revival and introduction into textbooks from September 2016. The Académie française is mentioned much less than the political figures, despite the significant role that the Académie played in the spelling reform (Catach, Reference Catach and Sanders1993: 143).

Chômage was mentioned by 27 of the tweets which referenced Hollande (64.3%): ‘Hollande a voulu alléger les chômage, il l’a fais! Sans accent circonflexe…’ (JE-1573); ‘Fin du chômage et des chômeurs Bravo #Hollande’ (JE-1985); ‘#Hollande a enfin vaincu le #chômage’ (JE-0204). A cartoon by Hervé Baudry, which shows Hollande atop of the word chômage with its circumflex in his hands saying: ‘Je vous avais promis, d’alléger le chômage: c’est fait !’, is shared in 18 tweets.Footnote ⁹ The subtext in many of these tweets being that Hollande has bigger issues to resolve than the spelling system, such as unemployment rates: ‘Dites Me @fhollande si au lieu de supprimer accents circonflexes vous pouviez carrément supprimer “chômage” et “impôts”‘ (JE-4000). To these users, orthography is a triviality in comparison to other societal problems.

Of the tweets which mention Najat Vallaud-Belkacem, seven directly address her Twitter account, showing a desire to engage the minister in online discussion. One such tweet, JE-0008, shares a Le Monde article entitled ‘Non, l’accent circonflexe ne va pas disparaître’;Footnote ¹⁰   not necessarily indicating support of the reform but at least a wish to share reliable information. Shortly before my collection period, Vallaud-Belkacem labelled the general reaction to the reform as a ‘polémique absurde’ in a tweet that included a link to a L’Express article,Footnote ¹¹ which similarly called the backlash a ‘manipulation politique’. While this tweet does address general criticism of the reform, the minister avoids direct engagement with Twitter users.

Governmental priorities were again questioned in tweets directed at Vallaud-Belkacem: ‘Que le gouvernement s’occupe des chômeurs avec accent même sans accent ils seront toujours dans le même cas’ (JE-0722); ‘Simplifier l’orthographe tire la langue vers le bas. Vous n’avez rien d’autre à faire? ’ (JE-4212). Unlike in Hollande references, five tweets directly insult Vallaud-Belkacem: ‘NUL NAJAT’ (JE-0722); ‘débarrassons-nous de Najat VB’ (JE-1214); ‘elle est bête la ministre’ (JE-2590). Research commissioned by British newspaper the Guardian found that articles written by female contributors receive ‘more abuse and dismissive trolling’ via online comments than articles written by men (Gardiner et al., Reference Gardiner, Mansfield, Anderson, Holder, Louter and Ulmanu2016). The difference in reactions to Hollande and Vallaud-Belkacem may suggest that this phenomenon is present in a wider CMC context.

No tweets which reference the Académie are insulting or aggressive, as is the case in tweets about Vallaud-Belkacem. It seems that the institution managed to escape the wrath of the Twitter backlash, despite having, arguably, the greatest involvement. This may be because the Académie denounced their involvement in February 2016 via a newspaper interview (Aïssaoui and Corty, Reference Aïssaoui and Corty2016) or because French-speakers are used to the prescriptive nature of the Académie, whereas linguistic intervention by the government is less usual.

7. RECURRING THEMES

This section will analyse the recurring themes in the 998 #JeSuisCirconflexe tweets classified as displaying a negative reaction towards the spelling reform of 1990. Where relevant, retweets and tweets from the other two search terms are also discussed. All tweets are reproduced exactly as collected, including non-standard orthography and punctuation.

The circumflex accent is the subject of 128 tweets (12.8%). In 104 tweets, instances where a sentence’s meaning would be affected by the inclusion/exclusion of the accent were quoted: ‘« Ⓔtre un homme mûr » et « être un homme mur », pas si sûr que ce soit la même chose…!’ (JE-0947); ‘Avant la réforme: “je suis sûr ta mère elle va crier” après “je suis sur ta mère elle va crier”‘ (JE-0668). The second, third and fourth most retweeted tweets use examples in the same way: sur/sûr (JE-1018 – 162 retweets and JE-0451 – 128 retweets); mur/mûr (JE-0042 – 132 retweets). The ninth tweet, JE-0020 (67 retweets), mentions first names which are spelt with circumflex accents and states: ‘C’est dur de perdre une partie de soi’. Highfield (2015: 2715) states that humour is a key factor for the ‘retweetability’ of a tweet; the popularity of these tweets seems to uphold this claim. Table 4 displays all examples cited more than once (excluding retweets).

Table 4. Circumflex examples cited in #JeSuisCirconflexe tweets

Of the 20 examples, only four are proposed to change by the spelling reform, while 16 are unaffected. In fact, jeûne, mûr and sûr are explicitly mentioned in the official documentation as words which will not change (1990: 5). Despite this, these three words are among the most commented upon, which again may be because of the comedic value of confusing the homophones: ‘Je suis sûr ta soeur elle va bien. Je suis sur ta soeur elle va bien’ (JE-1452); ‘Je vais me faire un petit jeune Je vais me faire un petit jeûne’ (JE-0377); ‘Ne jamais confondre un homme mûr avec un homme mur’ (JE-2725). It is unclear whether Twitter users chose these examples for their humour or because they believed that the circumflex was being removed from these words. Langer, in analysing reactions to the German spelling reform, found that the facts were unimportant. What was important was that language was in danger (Reference Langer2001: 34). This seems to be reflected in the sample. Furthermore, the tweets focus on singular examples, rather than on the principles behind the changes; a tendency also shown in reactions to the German spelling reform (Hohenhaus, Reference Hohenhaus, Langer and Davies2005: 206).

The future decline of the French language is alluded to in 80 tweets (8.0%): ‘Ça commence par supprimer les accents et ça finira par inscrire “sa va” dans le dico…’ (JE-0142); ‘C’est choquant, on finira à utiliser les émoticônes à la place des lettres’ (JE-1226); ‘2015: Oignon. 2016: Ognon. 2023: Truc qui fait pleurer’ (JE-0578). This theme was also shown in the tweets containing search terms reforme orthographe and ognon (6 and 12 tweets respectively). Langer and Davies (Reference Langer and Davies2011: 1) suggest that this slippery slope metaphor, of progressive changes leading to eventual decline, is a ‘recurrent phenomenon’ amongst non-linguists, with ‘modernising and foreign influences’ often cited as the threat. The mention of emoticons (JE-1226) is a clear example of such modernizing influences. The tweets present hypothetical future changes that are often hyperbolic such as un wazo for un oiseau, Ojd for aujourd’hui and an extension of the change proposed to nenuphar (ph to an f) to the words pharmacie (farmacie) and orthographe (orthografe). Reyes similarly found that Spanish speakers were likely to associate spelling reform with language decline (Reference Reyes2013: 354). It is unclear whether this negative attitude, although seemingly cross-cultural, is specific to this instance of government intervention or to language change more broadly.

In 76 tweets (7.6%), users suggest that the reform represents an undesirable simplification of the language or its decline into mediocrity: ‘Plutôt que de prôner l’excellence, ce gouvernement préfère le nivellement par le bas’ (JE-0030); ‘Simplification – Glorification de la médiocrité – Nivellement par le bas…’ (JE-3455); ‘“Simplification” de orthographe: RAS LE BOL du culte de la médiocrité et du nivellement par le bas’ (JE-1640). This is further shown in the retweets: ‘Apparement en France on aime accompagner les gens dans la médiocrité’ (JE-0003 – 199 retweets); ‘#ReformeOrthographe : simplification, glorification de la médiocrité, nivellement par le bas…’ (JE-0006 – 90 retweets). The use of simplification mirrors the official report that uses the verb simplifier twice (1990: 11, 13) and states that ‘une langue simple ou simplifiée à l’extrême est une langue pauvre’ (1990: 5). For those who have mastered the complicated spelling system, the social status associated with ‘correct’ usage of the standard language (Lodge, Reference Lodge1993: 12) and ‘correct’ spelling (Paveau and Rosier, Reference Paveau and Rosier2008: 143) is at stake. As Johnson (Reference Johnson2005: 129), when looking at the German spelling reform, suggested it is not the lowering of the standard but rather the lowering of its value which concerns commentators.

Seven tweets (0.7%) include variations of the verb appauvrir, which echoes Druon’s use of pauvre in the official documentation outlining the reformFootnote ¹² : ‘un appauvrissement de notre langue’ (JE-3668); ‘Appauvrir la langue pour que les mauvais ne fassent pas l’effort de s’améliorer, à l’image du pays: médiocre’ (JE-1013); ‘Comment justifier cette volonté de tout appauvrir? ’ (JE-3417). Both Druon and the Twitter users use the same rhetoric but from opposing angles. It seems hugely unlikely that this is a deliberate mirroring of official rhetoric as tweets which gave examples of words which will not be changing suggested that Twitter users are not informed on the intricacies of the reform and its official documentation. Rather, concepts such as simplification and appauvrissement have become classic arguments within metalinguistic discussion; the metaphor of an impoverished language has been used in German purism since at least the seventeenth century (Jones, Reference Jones1999: 59–60).

The perceived inherent qualities of the French language are praised in 45 tweets (4.5%): ‘En gommant l’accent circonflexe, on gomme aussi l’histoire et l’étymologie. La logique de la langue!’ (JE-1096); ‘défendons la richesse de notre langue’ (JE-2930); ‘La langue française, chef d’oeuvre de l’humanité, élégante et subtile, fille héritière du latin… Ne la bradons pas !’ (JE-0227).  A number of recurring purist images are displayed; a link to Latin, for instance, has been salient since the Middle Ages (Catach, Reference Catach and Sanders1993: 139) and qualities of logic and richness resemble Paveau and Rosier’s pseudo-linguistic arguments (Reference Paveau and Rosier2008: 57). In linguistic terms, no language is inherently more logical or richer than any other language; all such judgements are arbitrary (Joseph, Reference Joseph1987: 4), yet metalinguistic commentary has long included such judgements (Jones, Reference Jones1999: 121).

Beauty is mentioned in nine tweets: ‘ⓞ beauté de la langue française’ (JE-2400); ‘il faut préserver notre si belle langue française’ (JE-0403). Comments on the beauty of French resemble Thomas’ aesthetic purism, a non-rational argument for the purity of a language (Reference Thomas1991: 39) and the aesthetic argument which Paveau and Rosier suggest is characteristic of a purist discourse (Reference Paveau and Rosier2008: 57). Beauty is presented as reason enough to protect a language; whether this makes an efficient or usable language is insignificant.

Nostalgia is apparent in 21 tweets (2.1%), which refer to the circumflex accent as le petit chapeau or le chapeau chinois, echoing presumably how users referred to the accent as children. The French have a special relationship with the circumflex accent which ‘symbolise l’orthographe française’ (Cerquiglini, Reference Cerquiglini, Alonso Montilla, Bruña Cuevas and Muñoz Romero1996: 65) and a similar fondness for the accent was observed in the years following the initial publication of the Rectifications: ‘Le fait de pouvoir écrire plus facilement et simplement ne justifiait pas l’abandon des signes diacritiques auxquels ils trouvaient un certain charme’ (Ousselin, Reference Ousselin2004: 492). Four tweets looked back to the seventeenth century and la langue de Molière – the French playwright who has become synonymous with the bon usage of French: ‘laissez nous la belle langue de Molière..!’ (JE-0390); ‘Une réforme inutile qui fait de cette belle langue française une langue qui perd son symbole: la langue de Molière’ (JE-4284); ‘Molière doit se retourner dans sa tombe’ (JE-4379). Twitter users hark back to a ‘golden age’ of French language, as is common in prescriptivism (Watts, Reference Watts and Wright2000: 35) and purism (Thomas, Reference Thomas1991: 76). The golden age construct implies that change following a golden age can only be negative (Watts, Reference Watts and Wright2000: 35) and that ‘something needs to be done to restore it [the language] to its former glory’ (Langer and Davies, Reference Langer and Davies2011: 1). Furthermore, references to Molière, in particular, link the French language to culture and, in turn, the spelling reform becomes an attack on this culture.

Moving from themes to a recurring feature, humour is present throughout this corpus of tweets, particularly in the form of, what could be interpreted as, hyperbolic reactions: ‘On s’en ballecouilles de vos oignons de merde la ! C’est mon prénom qu’on va peut être tuer… #JeSuisCirconflexe’ (JE-2026); ‘si mon gosse plus tard il écrit ognon à la place de oignon il se prend une tarte dans la gueule’ (OG-0081); ‘2000: bon les jeunes ne respectent plus l’orthographe on va réformer l’enseignement. 2016: ok on laisse tomber on réforme l’orthographe’ (RO-0369). This form of humour is a dimension which seems to have been less prominent in the reactions to the reform in 1990. This may be symptomatic of the medium; humour seems to be an important feature on Twitter and the restricted character limit of tweets lends itself effectively to punch lines (Molyneux, Reference Molyneux2015: 929), or may reflect a changing perspective on the reactions.

8. CONCLUSION

The Rectifications orthographiques, published in 1990, were presented officially as moderate and acceptable to all. However, the media, followed by Twitter users, quickly dubbed the rectifications a reform,Footnote ¹³ immediately radicalizing the proposal, and negative reactions to the reform certainly prevailed in the corpora analysed here. The data seem to support Langer’s (Reference Langer2001: 34) claim that laypeople object to spelling reform outright as they believe change is bad, but rarely reflect on the principles of the reform; as evidenced, for instance, by users’ outrage at words losing their circumflex accents even when no change was planned. Given that language planning and intervention relies on the support of users of the language (Ricento, Reference Ricento2009: 9), this seems to have negative implications for the uptake of the new orthographies.

Tropes seen in metalanguage since the sixteenth century, such as the beauty, clarity and logic of the French language, (Jones, Reference Jones1999: 121) still appear pervasive, particularly when the language is felt to be under threat. The more extreme imagery and language present in tweets are seemingly a more recent addition to written discussions of language; they are evidence perhaps for suggestions of increased online disinhibition (Casale et al., Reference Casale, Fiovaranti and Caplan2015: 170). Measured against Thomas’ categories of rational and non-rational purist arguments (Reference Thomas1991: 37), the tweets analysed generally made non-rational assertions against the reform, with aesthetic arguments amongst the most used. Johnson argues that complaints against a reform cannot be disregarded simply as a ‘case of linguistic purism’, but that they are a form of bourgeois public debate against ‘absolutist forms of authority exercised on the part of the state’ (Reference Johnson2005: 130). Moreover, in the context of a spelling reform, a desire to maintain the status quo, that is to say the current orthography, is not irrational, even if aesthetically motivated. French speakers adopting the reformed spelling system will have to learn the new orthography, which takes time and effort; time and effort which could be saved by keeping the existing system. The reform will benefit children learning the French orthography for the first time, but for the Twitter users passing comment in this sample, it represents an overhaul of existing knowledge. It would seem rather reductive, then, to classify the reactions as wholly non-rational for their use of an aesthetic argument.

The samples used in this study are proportionally small in comparison to the amount of Twitter data sent daily. In future work, a larger corpus would make it possible also to incorporate geographic and demographic variables, potentially using the geographic information to construct a comparative study between reactions in, for instance, France and Quebec. Equally, a study which compared more comprehensively these online reactions to those displayed in 1990, when the reform was first published, could prove interesting; how much, if at all, have opinions towards the reform changed? Despite the limitations of the current study, it demonstrates the value of using Twitter to study the reactions of users to the re-emergence of a spelling reform, showing that sociolinguistically interesting discussions of governmental language intervention are happening online.