The preserve of the rural elderly, or a language for modern life? Authenticity, anonymity and indexical ambiguity in Martinican Creole

. This paper investigates the effects of (ongoing) standardization on linguistic attitudes and representations in the French Caribbean island of Martinique, where traditionally stigmatized Martinican Creole (MC) boasts a quasi-official orthography and some representation in formal domains. We use socio-biographical, perceptual and attitudinal data from a questionnaire-based study to investigate the relation between respondents’ (i) exposure to ‘activist’ MC – as a proxy for standardization; (ii) attitudes to MC on the status dimension; (iii) purism and (iv) breaking away from traditional MC indexicalities. Two findings are particularly noteworthy. First, exposure to activist MC fails to predict purist attitudes towards MC, which are similarly high regardless of respondents’ degree of exposure. Secondly, we find a mismatch between highly positive status attitudes and the persistence of traditional low-status MC indexicalities. We argue that, while some traditional indexicalities may wane as the standardization process progresses, others are essential to MC’s enduring representation as an authentic language and, therefore, less likely to recede.

existence of positive status-related attitudes and traditional low-status indexicalities (Section 3).Such indexical ambiguity is accounted for by pointing to the enduring framing of MC as a language of 'authenticity' and traditional culture, even within the activist milieu.Section 4 takes a more speculative turn: by analyzing reported indexicalities against individuals' specific statusrelated attitudes and degree of exposure to activist MC, we make predictions as to which traditional indexicalities may wane over time, as standardization progresses.Finally, Section 5 provides a brief conclusion for the paper, foregrounding its contributions and limitations.
1.1.LANGUAGE STANDARDIZATION IN MARTINIQUE: FROM TRADITIONAL DIGLOSSIA TO MC'S GROWING REPRESENTATION.For the greater part of its history, Martinican Creole has entertained a diglossic relation with French, its lexifier (Bernabé 1983).With MC relegated to low-prestige environments, French had the monopole of formal domains such as schooling, the administration, and the media (Prudent 1980).The prestige differential between the two languages was also reflected in their respective indexicalities, with MC being associated with the underprivileged classes, and French with the elites.
In 1946, Martinique was elevated from the status of colony to that of French overseas department.Besides its political implications, this event had momentous consequences for Martinique's sociolinguistic setup.The combination of cultural assimilation and compulsory French-language education resulted in French becoming an L1 for most Martinicans (March 1996;Beck 2017).As a result, French has since gained ground in informal domains which had previously been the preserve of MC.
In addition to provoking a major departure from the canonical model of diglossia (Ferguson 1959), the shift to French had a profound and paradoxical impact on the prestige differential between the two languages.Having expanded beyond the elites and its traditional domains of usage, French can no longer be associated exclusively with formality and prestige.Moreover, the rapid advance of French was perceived as a threat by Creole activists, who reacted by promoting MC in spheres from which it had hitherto been banned, viz. the media, the arts, and school.Their endeavors also included corpus planning initiatives.Chief among those are the standardization efforts led by the GEREC-F research group (Groupe d'Études et de Recherches en Espaces Créolophones et Francophones), which included the development of a phonemic orthography and grammatical descriptions of MC (Schnepel 2009;Ardoino 2023).Previously stigmatized variants gained wider acceptance even among the middle class (Prudent 1980), not least because of their perceived authenticity and distance from French.
These various changes call for a reexamination of Martinican bilingualism.Is MC still viewed as a minoritized variety devoid of overt prestige, or does it display some of the indexical attributes of standard languages (e.g. an association with formality and modern life)?If the latter is true, can we really positin keeping with traditional understandings of the Martinican continuum (Bernabé 1983)that more Creole-sounding forms/varieties would be perceived as less prestigious than forms/varieties closer to French?

LANGUAGE STANDARDIZATION OUTSIDE MARTINIQUE: INDEXICAL CHANGE, PURISM AND
INSIGHTS FROM THE LITERATURE.The need to take stock of Martinican bilingualism is made even more glaring by crosslinguistic studies showing that standardization can have far-reaching consequences for minoritized languages and their speakers.These consequences are of two main kinds more positive attitudes (Vari 2021) and increased purismand can be attributed to the ideological shift brought about by the standardization process, which the linguistic anthropological literature has captured through the notions of 'authenticity' and 'anonymity'.
While international standard languages are anonymous in their seeming geographical/ethnical 'neutrality', minoritized languages are perceived as belonging to authentic speakers for whom they act as identity markers (Woolard 2016).When minoritized languages undergo standardization, however, this binary breaks downin ways that are yet to be fully understood.Although standardization has been linked to increased purism and the erasure of traditional indexicalities (Eckert 1983), several studies have cast a doubt on the strength of such a link (Jaffe 2003;Sallabank 2010;Urla et al. 2016), showing that purism can predate standardization and standardization needs not undermine the appreciation for traditional varieties.
In the Martinican context, the effects of (ongoing) standardization are even less clear.So far, the literature on purism has either foregrounded (Térosier, François-Haugrin & Duzerol 2022) or qualified (Ardoino 2023) the danger of MC standardization for Martinican speakers' linguistic security.Moreover, such studies only rely on public discourse and do not explore individuals' attitudes/perceptions and their causes, nor the wider indexical changes that might feed purist discourse.
By directly investigating Martinicans' attitudes towards Creole and the persistence/waning of its traditional indexicalities, this paper gauges both the effectiveness of Creole activist discourse in promoting a more prestigious image of MC, and the potential positive/negative impact of the standardization endeavor on Martinican speakers.Moreover, it also provides empirical evidence to inform crosslinguistic discussions about standardization, which are often based on purely qualitative data about language ideologies (for an exception, see e.g.Vari & Tamburelli 2020).
1.3.RESEARCH QUESTIONS.This paper asks whether Martinicans represent MC as a low-prestige, 'authentic' variety, or as an emerging standard language subjected to standard language ideology or, finally, as somewhere in between.This overarching question can be broken down into the following research questions (RQ): • RQ1.Is Creole(ness) still associated with informality and informal domains?
This question is about individual attitudes and shall be addressed by investigating both attitudes to MC on the status dimension and actual perceptions of MC speech.
• RQ2.Is there a MC purism and, if so, is it comparable to the degree of purism towards French?
• RQ3.Does MC enjoy 'authenticity' or 'anonymity' indexicalities?This question is about broader linguistic representations, which tend to be shared at the community level but may show some degree of variation across individualsespecially at times of indexical change.To address this question, one should investigate the extent to which Creole is (still) associated with people and places indexical of solidarity/authenticity (e.g.rural environments and the street), as opposed to status/anonymity (e.g.urban environments and university).1

Methodology and data collection.
To address these research questions, we used an online Qualtrics questionnaire that elicited the following attitudinal, perceptual and socio-biographical data.
2.1.ATTITUDES TO MC ON THE STATUS DIMENSION.Respondents were presented with two attitudinal statements (Creole should be an official language in Martinique, alongside French and Creole cannot ever become the language of trade and science) and asked to express their agreement on 5-point Likert scales ranging from 'fully disagree' to 'fully agree'.

PERCEPTIONS OF FORMALITY FOR DIFFERENT VARIETIES OF MC (MORE VS LESS FRENCHIFIED).
Respondents were presented with eight pairs of semantically identical oral stimuli that contrasted more Frenchified (traditionally less stigmatized) and less Frenchified (traditionally more stigmatized) varieties of MC.For each pair, participants were asked to choose the version they found more appropriate.Half or the pairs were presented in (fictitious) informal contexts such as gatherings with friends, and the other half in (fictitious) formal settings like conferences and newscasts.This perceptual task also included French stimuli pairs (contrasting more Creolized and less Creolized French) to be used as a term of comparison for perceptions of MC.The goal of this task is to find out (i) whether formal settings are associated with an increased preference for less Frenchified or more Frenchified MC, compared to the informal contexts (used as the 'baseline') and (ii) how formality 'norms' for MC compare to those applying to a fully standardized language like French.2Findings from this task provide further insights into MC's status.While a higher preference for Frenchified Creole would indicate that Creoleness still indexes low status, the opposite would point to its increased prestige and the emergence of purist norms wherebyin MC just like in Frenchformality is tantamount to the absence of language mixing.

PURISM (FOR MC AND FRENCH
). Respondents were presented with two attitudinal statements for each language (Martinican Creole/French is too Frenchified/Creolized and When speaking Creole/French, one should avoid using any expression that is clearly French/Creole) and asked to express their agreement on 5-point Likert scales ranging from 'fully disagree' to 'fully agree'.

SOCIAL INDEXICALITIES.
We tapped into respondents' stereotypical social representations of Creole by eliciting their association of 'good Creole' with more vs less traditional speakers and places (the elderly, men, the countryside, the street vs the youth, women, urban environments, university).
These contrasts were presented through bipolar scales, as in the example below: 'Creole is spoken…' (1) a lot better by men than women (2) slightly better by men than women (3) equally well by men and women (4) slightly better by women than men (5) a lot better by women than men While preferences close to the traditional pole would attest to the persistence of traditional indexicalities, ratings closer to the mid-point (i.e.no preference) or the less traditional pole would signify a change of indexicalities, towards more modern and/or high-status representations of MC.
2.5.EXPOSURE TO 'ACTIVIST' STANDARDIZED MC.Finally, we collected data about respondents' exposure to the unofficial 'standard' variety of MC associated with schools and the activist milieu.This measure is needed to estimate both Martinicans' overall familiarity with standardized MC andby comparing the attitudes of speakers reporting higher and lower degree of exposure the effect of standardization on the language attitudes/representations listed above.
Exposure to activist MC was estimated through a combination of questions.First, respondents were asked to express their (dis)agreement with regards to the following statement, using a 4-point Likert scale: I (have) regularly take(n) part in activities for the teaching/learning of Creole.Then, they were prompted to choose their 'favorite' orthography for four MC words, amongst a series of four options more/less aligned with the (quasi-)official MC orthography. 3 Answers were then rated depending on their closeness to such orthography.We subsequently obtained an individual measure of exposure to activist MC, by calculating the mean value between (i) reported participation in MC-language activities and (ii) the mean knowledge of MC orthography, across the four words.

Findings.
We administered our questionnaire to 123 Martinicans ranging from 15 to 80 years of age (M = 47.7,SD = 15.96).Most respondents were women (75%) and had received some form of post-secondary education (79%).This high level of educational attainment far exceeds official statistics for Martinicans holding university qualifications, which stand at approximately 23% (Insee 2023).The disproportionate representation of women and highly educated participants is, in fact, a common drawback of online questionnaires (Smith 2008;Bethlehem 2010).Although neither gender nor education was found to significantly influence respondents' attitudes to and perceptions of MC, 4 caution is warranted when generalizing the findings below to demographics not represented in this study.

ASSOCIATION BETWEEN CREOLENESS AND STATUS (RQ1
). Across our respondents, MC appears to enjoy relatively high status.As shown in the graph below (Figure 1), most of them believe that (i) Creole should be an official language in Martinique, 5 and that (ii) it can in fact become the language of trade and science. 6 3 We deliberately phrased this question in subjective terms ('favorite', instead of 'correct' orthography) to discourage respondents from looking up the words.We assumed that respondents who had been exposed to activist norms would still report the 'official' correct orthography, even if they were asked about their personal preferences. 4There are three partial exceptions.First, men display slightly less positive status-related attitudes than women and slightly higher purism.Secondly, education is correlated with associating 'good Creole' with rural vs urban environments.In all cases, however, the effect is small and slightly above significance.The effect of these and other sociobiographical factors on individual attitudes and perceptions will be explored in a future publication. 5The questionnaire was administered in February-March 2023.In May 2023, the Territorial Collectivity of Martinique voted to make MC a co-official language alongside French.We can thus conjecture that opinions in favor of officialization could be even more prevalent now, although the decision to co-officialize MC has not been recognized by the French State. 6Since the second statement features a negative polarity item, higher disagreement amounts to more positive attitudes.
Figure 1.Attitudes towards Creole on the status dimension.Reactions to the statements 'Creole should be an official language in Martinique, alongside French' (left) and 'Creole cannot ever become the language of trade and science' (right).
However, these are only self-reported attitudes about the desired/accepted position of MC in Martinican society.Being elicited explicitly and in the absence of actual Creole speech, they are both prone to desirability bias and rather abstract.
Complementary evidence of MC's increased status comes from the perceptual task described in section 2, which compares preferences for less/more mixed language (i) in formal and informal contexts and (ii) for Creole and French stimuli.The results show that presenting stimuli in formal settings leads to higher preferences for the less mixed versions.For French stimuli, this tendency is very pronounced, with higher formality associated with a 31% higher preference for less Creolized variants (82.11%, vs 62.80% for the informal settings).This is a fairly unsurprising finding, given the entrenchment of purist discourse surrounding the French language (Coppel 2007; for some nuanced and empirically based accounts of purism in France, see Oakes 2001, Boughton 2005 and Walsh 2016).What is more surprising, however, is that the same significant effect is also found for MC, albeit to a lesser degree.Although preferences for less mixed versions are overall lower for Creole than for French, for MC too formal settings are associated with a boost in preference/acceptance for the less mixed variants, i.e. less Frenchified MC (52.24%, vs 44.11% for the informal settings).7These trends are summarized in Table 1  One could hardly explain the relative preference for less Frenchified MC in formal settings, without positing that formerly stigmatized MC has developed some form of overt prestige.In this respect, the perceptual task confirms the positive attitudes explicitly elicited by the attitudinal items in Figure 1.
3.2.HIGHER PURISM FOR CREOLE THAN FOR FRENCH (RQ2).As seen in Section 1.2, the standardization process is thought to entail the emergence or exacerbation of linguistic purism, with variation and language mixing framed as threats to the newly developed 'standard' (Eckert 1983;Woolard 2016).The fact that our respondents have reported an increased preference for unmixed ('purer') MC in formal settings could be an indication that MC is not immune from such purism.Indeed, 75% of our respondents (strongly) agree that 'Creole is too Frenchified' and 65% of them (strongly) agree that 'when speaking Creole, one should avoid using expressions that are clearly French' (see Figure 2 below).These rates of reported purism are astonishingly high, for a variety traditionally referred to as patois, gibberish or bad French (Prudent 1980).Even more surprisingly, they are significantly higher than the corresponding rates for French (see Figure 3 below), where there is no consensus on whether Creole expressions should be avoided when speaking French, and just 24% of respondents (strongly) agree that 'French is too Creolized'.In the absence of qualitative data to complement the above picture, we cannot provide a conclusive interpretation for these findings.Instead, we propose a tentative account, to be tested in future research.We believe that MC might elicit more purist reactions than French because, being less standardized and less mastered/spoken (Beck 2017),8 it is also perceived as more threatened by language contact (Ardoino 2023;cf. Bernabé 1983).This perceived threat is likely sharpened by fears of language loss stoked by activist discourse, and the role of identity marker that MC plays in Martinican society (Pulvar 2004(Pulvar , 2005)).The fact that in our study reported purism is only weakly (or not at all) correlated with status-related attitudes and exposure to activist Creole supports an interpretation of purism as the response to fears of language and identity loss, more than a direct outcome of standardization.9If confirmed, this account would suggest that what stokes individual purismin Martinique and potentially elsewhereis not so much a standardization process that has reached completion but, perhaps, an incipient one that is just advanced enough to sharpen fears of language/identity loss, but not enough to quell them.

PERSISTENCE OF TRADITIONAL INDEXICALITIES.
So far, MC has displayed social attributes (relatively high status and purist norms) that set it apart from typical minoritized languages.This raises the question of whether its social indexicalities have changed, too.Is MC still associated, in folk beliefs, with its traditional speakers (men and the elderly) and low-status social places (the countryside and the street)?
In terms of speakers, we find a mixed picture comprising 'new' and traditional indexicalities.On the one hand, little distinction is made between genders, with 67% of participants reporting that men and women speak equally good Creole, vs only 27% who report a preference for men's Creole.On the other hand, 'good Creole' is still strongly associated with the elderly (73%), and only 2 out of the 123 respondents attribute equally good Creole to all age groups.
In terms of places, 85% of respondents (strongly) believe that Creole is better spoken in the countryside than in towns (Figure 4 below), and 67% (strongly) believe that it is better spoken in the street than at university (Figure 5   How can one account for these traditionaland seemingly low-statusindexicalities, given the findings described above?The preference for the MC of rural environments could potentially result from the word countryside evoking images of a pristine, idealized past.This, however, cannot explain the association of 'good Creole' with the 'street', which suggests that MC is still viewed as the language of informality. 10Together, these results point to the persistence of traditional (low-prestige) indexicalities in the face of standardization and growing prestigean indexical ambiguity that will be the focus of the next section.

Discussion.
Together, the findings presented in Section 3 raise important questions regarding MC's ongoing standardization and its implications for speakers' attitudes.On the one hand, MC has acquired a status that exceeds that of a minoritized language and appears to be subjected to purist normsboth explicitly (i.e.reporting of purist attitudes) and implicitly (i.e.increased preference for less mixed Creole in the perceptual task's formal settings).On the other hand, however, this increased status coexists with the upholding of traditional low-status indexicalities.
How to account for these conflicting representations?Is this indexical ambiguity the result of change in progress (i.e.incomplete standardization), or a more stable attribute of MC that could outlive the standardization process?While we do not have a definite answer to these questions, by examining how MC indexicalities are shaped by (i) status-related attitudes and (ii) exposure to activist Creole, we can tentatively distinguish between more and less likely indexical changes to come.

GOOD CREOLE IS IN THE STREET: INCOMPLETE STANDARDIZATION?
When we analyze MC indexicalities on the rural-urban axis against status-related attitudes and exposure to activist Creole, we can see that respondents who report a milder preference for street Creole (27%), no preference at all (22%) or, more rarely, a preference for urban Creole (11%) also tend to report higher status-related attitudes and exposure to activist Creole.This is illustrated in Table 2 below which shows, for each possible answer on the rating scale (from 'Creole is much better spoken in the street' at the top, to 'Creole is much better spoken at university' at the bottom), the proportion of respondents who chose that answer, their mean attitudes to Creole on the status dimension (out of 4) and their mean exposure to activist Creole (out of 4). 10 Admittedly, the overwhelming bias for the 'street' can also result from respondents' unfamiliarity with the Creole spoken and/or taught at university, or the university campus environment altogether.This, however, does not invalidate the interpretation offered in the text.Since our question taps into stereotypical representations rather than individual knowledge, more entrenched associations of MC with formality and teaching could have led to associating 'good Creole' with the university environment even in the absence of personal exposure to university MC.  showing, for each level of the response variable, the proportion of respondents who selected that level, their mean attitudes to MC on the status dimension and their mean exposure to activist MC As Table 2 shows, the transition from more traditional/lower-status indexicalities (the street) to less traditional/higher-status indexicalities (university) does not correspond to perfectly linear increases in status-related attitudes and exposure to activist Creole.For status attitudes, there is a gradual increase over the rating scale, but the strongest increaseand so, potentially, the strongest effect on the response variablecoincides with the reporting of 'no preference' vs a 'mild preference for street Creole'.For exposure to activist Creole, by contrast, the strongest increase (and, thus, the strongest effect on the response variable) corresponds to the choice of 'mild preference for street Creole' vs 'strong preference for street Creole'.11It is not clear why these variables may affect respondents' choices at some levels of the rating scale more than at others, or why they affect different levels from each other. 12hat is clear, however, is that the association of 'good Creole' with the street is overall weaker for respondents with higher status-related attitudes and exposure to activist Creole.We can thus hypothesize that this traditional low-status indexicality may wane over time, as MC continues to be standardized, taught and propelled into formal domains.This stands in contrast with the association of 'good Creole' with the countryside, which is further explored below.

GOOD CREOLE IS IN THE COUNTRYSIDE: A MORE STABLE INDEXICALITY?
Although similar at a first glance, the association of 'good Creole' with the countryside reveals meaningful differences from the case above.To start off, the preference for the traditional pole is even stronger, with almost no respondent reporting a preferenceeven a mild onefor urban Creole.Secondly, differences in the degree of preference for rural Creole cannot be predicted by either status-related attitudes or exposure to activist Creole, as shown in Table 3 below.While exposure to activist Creole shows a correlation with the choice of 'mild preference for rural Creole' over 'strong preference for rural Creole', the effect is smaller than for the street/university indexicality and fails to reach significance.Why is the association of 'good Creole' with rural environments so prevalent amongst our respondents, and why is it (seemingly) not eroded by MC's growing status and increasing degree of standardization? 13We argue that this could be due to the idealization of the 'countryside' in contemporary representations of MC, both inside and outside the activist milieu.Rural environments are widely regarded as a bastion of 'authentic', less Frenchified MChence a repository of potential lexicon to expand the emerging standardized MC (Bernabé 1983;Ardoino 2023).The privileged position that rural MC occupies in activist language/discourse might explain why neither higher status-related attitudes nor exposure to activist MC undermine its ideological leverage over urban varieties.As long as MC standardization is framed as the pursuit of (lost) 'authenticity', and authenticity is synonymous with rural life and the 'old times', the association of 'good Creole' with the countryside may remain a cornerstone of the MC standardization project.Therefore, 'rural' indexicalities may be less likely to wane over time than the association of 'good Creole' with the street which, despite apparent similarities, is better explained as a historical hangover from the times of strict diglossia and MC minoritization.

Answer
This ambiguous coexistence of modern and traditional indexicalities takes us back to our discussion of purism in 3.2.Although the literature often associates purism with language standardization, in this study neither exposure to activist Creole nor status-related attitudes show correlations with reported purism towards Creole.Where does the high Creole purism found in the study come from, if it is not a fallout of the standardization project?We conjecture that respondents' purism might result from a general longing for 'authentic' MC that can originate outside the activist milieu, although being potentially fed by it.This would tally with studies showing that purism can emerge even in the absence of fully organized standardization projects (Schieffelin, Woolard & Kroskrity 1998;Aikhenvald 2001;Sallabank 2017).

Conclusions.
This paper has drawn on different types of data to keep stock of social representations of MC at a time of sociolinguistic upheaval.What emerges is a picture of indexical ambiguity, with MC simultaneously displaying indexicalities of a 'standard' language (prestige, modernity) and a minoritized 'authentic' variety (low status, associations with old speakers and traditional places).
By looking at attitudes to MC on the status dimension and perceptions of formality for different MC varieties, we have shown that MC has acquired a status that exceeds that typical of minoritized languages: most respondents are in favor of the officialization of MC and believe that MC can become a language for high-status domains and functions.
At the same time, this increased social status coexists with traditional low-status indexicalities, such as the association of 'good Creole' with the street (vs university) and the countryside (vs the town).By investigating how these two indexicalities are shaped by attitudinal and sociobiographical factors, we have shown that, in fact, these associations likely follow different trajectories and require different explanations.While the association of 'good Creole' with the street appears less stable and may be wiped out by increased standardization, the idealization of countryside Creole is more integral to the standardization project and, thus, less likely to recede.
This paper has foregrounded the indexical complexity that can surround minoritized languages on the path towards standardization.Beyond its analysis of sociolinguistic dynamics in Martinique, it makes both theoretical and methodological contributions.From a theoretical viewpoint, this research draws attention to the presence of purist attitudes/ideology in understandardized languages (a purism of 'authenticity') and the persistence of traditional indexicalities in the face of waning diglossia.Both findings chime with studies showing that standardization does not always imply a departure from the 'authenticity' values associated with the minority language (Urla et al. 2016;cf. Ardoino 2023).
Methodologically, this paper illustrates a multifaceted approach to the study of minority language standardization, which combines the elicitation of linguistic perceptions, attitudes and representations to better uncover (and quantify) the multitude of stances lying behind language planning and publicly available discourse.This approach is not foolproof, though, as its quantitative focus can lead to glossing over important differences in speakers' attitudes and linguistic experience.The findings presented here should, therefore, be confirmed not only on a more representative sample of the Martinican population but, also, by combining the analysis of quantitative patterns with that of richer and more nuanced qualitative data.

Figure 2 .
Figure 2. Purism towards Creole.Reactions to the statements 'Creole is too Frenchified' (left) and 'When speaking Creole, one should avoid expressions that are clearly French' (right) below).

Table 1 .
Ratio of preference for the less mixed variants for French and Creole stimuli, in relation to formality condition

Table 2 .
Preference for university vs street Creole.Table

Table 3 .
Preference for urban vs rural Creole.Table showing, for each level of the response variable, the proportion of respondents who selected that level, their mean attitudes to MC on the status dimension and their mean exposure to activist MC.