2. PICARD AS A ‘COLLATERAL’ LANGUAGE

France’s Oïl varieties provided the inspiration for a conference and collection of papers (Eloy, Reference Eloy2004a) devoted to langues collatérales (‘collateral languages’), whose indeterminate status stems from their linguistic similarity to and shared history with a dominant language, falling somewhere between ‘languages’ and ‘dialects’ in public perception (Eloy, Reference Eloy and Eloy2004b: 6):

Writing in the same volume, Trudgill views ‘collateral’ languages as a subset of what Kloss (Reference Kloss1967) has labelled ausbau varieties. Recognized primarily on socio-political or cultural grounds (Reference Trudgill and Eloy2004: 70), the status of the latter can alter at any time: from ‘dialect’ to ‘language’ in the case of Letzeburgesch, for example, or from ‘language’ to ‘dialect’ in the case of Scots, Provençal or Low German. By contrast abstand languages, in Kloss’ terminology, are separated by a linguistic distance great enough for them to be universally considered different from their dominant competitor – Basque, for example, cannot be grouped together with its Romance neighbours French and Spanish. Authorities may seek to downplay differences between ausbau varieties and the dominant language, casting the former (as for example with Catalan in Franco’s Spain) as a dialect of the latter. In abstand situations, on the other hand, they may seek to accentuate differences between a minority variety and a dominant language used in a neighbouring state, thereby discouraging contact between speakers of closely related varieties across a national border.

While linguists, Trudgill argues, are no better placed than others to decide questions of minority language rights, or to dictate language policy, they do have the expertise to distinguish linguistic from non-linguistic arguments, and to identify claims falsely made on the basis of political, social or cultural motives, and bolstered by pseudo-linguistic evidence. Picard would seem to be an archetypical collateral language in Eloy’s terms, and indeed both a section of the editor’s introduction (pp. 20–21), and four other papers in the 2004 collection are devoted to it. Although officially recognized since 1990 (alongside Wallon, Champenois, and Lorrain) as a regional language of Belgium, Picard enjoys no comparable status in France. It was not mentioned in the loi Deixonne of 1951,Footnote ² nor in a succession of goverment reports on regional language diversity published since the Revolution; Abbé Grégoire’s 1794 report to the National Convention Rapport sur la nécessité et les moyens d’anéantir les patois et d’universaliser l’usage de la langue française (see De Certeau, Julia and Revel, Reference De Certeau, Julia and Revel2002: 331–51) had barely referenced the Oïl varieties at all, viewing them apparently as little more than ‘bad French’. Not until Cerquiglini (Reference Cerquiglini1999) is Picard mentioned directly as one of 75 ‘langues de France’, in a report received with some suspicion by language activists, who saw in belated official recognition of linguistic diversity little more than an attempt to stymie ratification of the Council of Europe’s (1992) Charter for Regional or Minority Languages (CRML), by over-emphasizing the impracticalities of societal multilingualism.Footnote ³

Structural similarity to another dominant variety need not, however, be an impediment to language status, as we saw above, and would not in itself justify the downgrading of the langue d’oîl varieties, such as Picard, in comparison with France’s other regional languages. Eloy (Reference Eloy1997: 210), in this vein, draws parallels with Corsican, Scandinavian and creole varieties to offer an unequivocal answer to his own question, ‘Qu’est-ce que le picard?’:

Eloy is at pains to emphasize the word ‘potentiellement’, and does not pretend that conditions are currently propitious. Even for its own speakers to recognize Picard as a language alongside French would require what he terms a ‘démarche identitaire forte’ (ibid.). Though the precise nature of this démarche is unspecified, it is likely to face some very significant obstacles. Historical evidence for an identifiable Picard language rests on shaky ground (see section 4), and the label ‘Picard’ itself has little resonance in much of the picardophone area identified by linguists (section 5). Perhaps more significantly, a binary French-Picard model, which is not, and probably never has been, reflective of actual usage may even have been detrimental to maintenance of Picard varieties.

3. MODELLING PICARD USAGE

Evidence suggests that speakers within the picardophone area (see section 5) see language variation in terms of a diglossic model which contrasts français and either picard or patois. This was evident from metalinguistic comments by my own informants in Avion (see Hornsby, Reference Hornsby2006a), who used the term ‘patois’ to denote a set of forms (which might vary between localities) with local rather than national currency; the diglossic model also underpins Carton’s (Reference Carton1981) typology, discussed below. Support for this model comes from compelling evidence of discrete differences between Picard and French at the grammatical level (see Auger, Reference Auger2010; Auger and Villeneuve, Reference Auger, Villeneuve, Meyerhoff and Nagy2008; Villeneuve and Auger, Reference Villeneuve and Auger2013;Footnote ⁴ Auger and Villeneuve, this volume), and in phonology (Auger and Villeneuve, Reference Auger, Villeneuve, Barra-Jover, Brun-Trigaud, Dalbera, Sauzet and Scheer2014; Hendrickson, Reference Hendrickson, Kingston, Moore-Cantwell, Pater and Staubs2014). But the perception of discrete varieties is often belied by the reality of highly heterogeneous outputs. In the example below, recorded in Avion, Pas-de-Calais (see Hornsby, Reference Hornsby, Ayres-Bennett and Jones2007: 77), French elements (underlined) are mixed in the same turn as corresponding Picard forms (bold), which are geographically restricted and trigger identification of speech as ‘patois’ among Avionnais themselves (e.g. tu/té; une/enne; dans/dins). The bulk of the lexicon can be taken to be either French, or common to both French and Picard:

Mixing of Picard, French and shared elements is evident in the speech of Eloy’s Amiénois picardophone informant Maurice Boucher (Reference Eloy1997: 145–174), and is also reported by Villeneuve and Auger (Reference Villeneuve and Auger2013) in Vimeu, where language mixing has acquired the term ‘dravie’. The latter distinguish regional French (2) from Franco-Picard (3) on the basis of a greater proportion of French forms for which Picard equivalents are available in (2) than in (3), where the matrix language appears to be Picard (Villeneuve and Auger, Reference Villeneuve and Auger2013: 115). Vimeu speakers perceive (2) to be French and (3) Picard, but nonetheless, as the authors point out, marked Picard forms (bold) are embedded in (2) and French ones (underlined) in (3):

Such heterogeneity at the level of the speech turn might, of course, be a reflection of conversational code-switching and belie greater homogeneity over shorter sequences of talk. To investigate this possibility, clause and turn-level co-occurrence for a number of binary Picard/French morphological variables in the Avion data were examined by Author (2006: 58–64). For each variable pairing, ‘cohesion scores’, based on the percentage co-occurrence of ‘like’ variants (French+French or Picard+Picard) were indeed found to increase significantly as the unit of analysis was reduced, but were noticeably lower for Picard+Picard than for corresponding French+French pairings, and fell away sharply at turn level. Picard speakers in Avion, in other words, sustained fairly homogeneous output in what they would term patois only over short stretches of talk; over longer sequences mixed output was normal. Switches to patois were signalled primarily through (inconsistent) use of Picard morphology and some high-frequency Picard lexical items. Such behaviour recalls young Londoners’ targetting of ‘London Jamaican’, as reported by Le Page and Tabouret-Keller (Reference Le Page and Tabouret-Keller1985: 180):

Carton’s (Reference Carton1981) model of variation in the Nord-Pas-de-Calais represents an attempt to resolve the contradiction between speakers’ perception of a two variety (français-patois) model on the one hand, and the reality of highly mixed usage on the other, through appeal to what Carton calls dialectalité, i.e. the variety targetted by the speaker him/herself, irrespective of actual output. The four-term typology he offers bears some similarities with post-Creole continua described by Bailey (Reference Bailey1973), Bickerton (Reference Bickerton1975) and others, in so far as an ‘acrolectal’ Variety 1 (français général) betrays no regional marking at all, while at the other extreme ‘basilectal’ Variety 4 (patois) represents a pure Picard variety showing no influence at all from French. Here and elsewhere Carton is clear that Variety 4 is a notional construct, which no longer corresponds to any speaker’s actual output, if indeed it ever did: ‘Le “pur picard”, s’il a existé, n’existe plus. Mais la “picardité” est bien vivante’ (Carton and Lebègue, Reference Carton and Lebègue1989: Introduction).Footnote ⁵ The ‘mesolectal’ Varieties 2 and 3 represent intended French (français d’intention) and intended patois (patois d’intention) on the part of the speaker. While both mix regional and national forms, the former are more localized and greater in number in Variety 3 (see (2) and (3) above):

While the model is not unproblematic (see Hornsby, Reference Hornsby2006b), Carton’s insistence (Reference Carton1981: 17) that his Varieties should be seen as a reference points in a continuum (‘il n’y a aucune solution de continuité entre ces variétés’) frees us from the reductive fiction of a binary model and allows for a wide range of behaviours in which local, supralocal and national features are mixed.

It was notable that Picard forms were not used at all in Avion by informants under 30 years of age, and further evidence of declining competence in Picard is presented in Pooley’s (Reference Pooley2004) research among younger speakers in the Lille conurbation. Pooley’s informants showed low levels of lexical recognition in Picard and even lower levels of active competence in a translation test, where less than 10 per cent of informants on average could correctly render French items in Picard. Pooley concluded that there was ‘very little latent competence in Picard’ (Reference Pooley2004: 487) among young Lillois. Low rates of intergenerational transmission would seem to be a classic symptom of language death, and Picard would appear, on Fishman’s (Reference Fishman1991) GIDS (Graded Intergenerational Disruption Scale), to be close to the highest level of attrition, level 8 (‘most vestigial users of Xish are socially isolated old folks’). Certainly, as Pooley (Reference Pooley2004: 576) points out, the view that contemporary northern patois are varieties which emerged from a respectable medieval language called Picard but have ‘fallen from grace’ through contact with French is well established in the education system. A case could certainly therefore be made that Picard is an obsolescent heritage regional language worthy (and in urgent need) of at least the support afforded to, for example, Breton or Corsican, a view arguably bolstered by belated recognition of language status in the Cerquiglini Report. Seductive though this narrative might appear, it is not supported by the historical evidence, which suggests that parallels with France’s other regional languages are of limited value.

4. PICARD AND FRENCH IN HISTORICAL PERSPECTIVE

There is a long-standing divide in French dialectology between ‘separatists’ such as Remacle, Gossen and Chaurand, who argue for discrete dialects in France’s Gallo-Roman patchwork – Gaston Paris’s ‘immense bigarrure’ – and ‘continuators’ such as Paris himself, Meyer, Gilliéron, Brunot and Tuaillon who view such divisions of the Romance continuum as arbitrary, and potentially a distortion of the linguistic facts. Divisions of the langue d’Oïl area are particularly controversial given the prolonged contact, and extensive similarities, between these varieties and French. What is not disputed is that there is little evidence of differentiation within the Oïl zone before the ninth century (see Delbouille, Reference Delbouille1970; Eloy, Reference Eloy1997: 54–55), though it should be emphasized that there was very little vernacular writing at this stage, and written documents are in any case a notoriously unreliable indicator of spoken norms, to which we have no access.

While local differences do become evident from the ninth century, medieval texts such as the Serments de Strasbourg and the Séquence de Sainte-Eulalie (see Ayres-Bennett, Reference Ayres-Bennett1996) are highly mixed, as scribes often selected forms of wide currency at the expense of less familiar local ones. The consensus view among French historical linguists (see e.g. Delbouille, Reference Delbouille1970; Gossen, Reference Gossen1970: 30–31; Carton, Reference Carton1992: 33) is that early Oïl texts are written in what amounts to a pan-regional koiné (‘On a le droit de parler de la littérature d’oui en ancien français comme d’un tout.’: Cohen, Reference Cohen1967: 84), leavened with a proportion of local forms according to the origin of the writer. However, Picard forms associated with prosperous northern mercantile centres such as Arras figure strongly enough alongside francien items for some to posit a medieval scripta franco-picarde, i.e. a de facto Oïl written standard in which French and Picard predominate (a claim rejected by Dees Reference Dees1980, Reference Dees1985), or even to suggest that Picard was a serious medieval rival to French in terms of prestige – see for example Wartburg (Reference von Wartburg1946: 87):

A strong Picard visibility in medieval texts, however, does not amount to a well-defined and clearly delimited norm, and in terms of prestige, Picard soon starts to lose out to its Parisian rival. The presence of Picard forms declines markedly in the thirteenth century with economic decline in the mercantile centres of the north, and French forms soon start to predominate (see Cohen, Reference Cohen1967: 84): Picard forms had all but disappeared in texts by around 1400. Brunot (Reference Brunot1966: I, 328–331); Picoche and Marchello-Nizia (Reference Picoche and Marchello-Nizia1989: 21–22); Lodge (Reference Lodge1993: 98–100) and others, moreover, cite evidence that, long before this point, Picard norms were already being compared unfavourably to those of Île-de-France. Some four centuries before a stylized Picard would be used to comic effect by Molière (for example through the character of Nérine in Monsieur de Pourceaugnac), a mocking pastiche of Picard speech appeared in a play dating from 1285, cited by Picoche (Reference Picoche, Darras, Picoche, Debrie and Ivar1985: 57), who also quotes a mid-thirteenth-century assertion by Barthélémy l’Anglais that Picard ‘est idiomatis magis grossi aliarum Gallie nationum’ (‘is the ugliest tongue of all the peoples of France’). The most celebrated illustration of Picard social inferiority, however, comes earlier still. A number of commentators reference twelfth-century poet Conon de Béthune’s annoyance at being mocked, by Queen Alix of Champagne and her son Philippe-Auguste, for his Artesian speech:

While for Picoche (Reference Picoche, Darras, Picoche, Debrie and Ivar1985), Conon has been labouring under the misapprehension that Picard and French are social equals, one might equally suggest that his frustration stems rather from the fact that he was already well aware of the gulf in prestige, and had been doing his best to accommodate to his socially elevated guests by attempting to speak French, only to fall back occasionally on Picard forms (‘se j’ai dit mots d’Artois’). What does seem clear is that, at least at Conon’s social level, convergence with French at this stage was already significant enough to allow mutual comprehensibility (‘Si la puet on bien conprendre en franchois’), and that Picard forms are not deemed appropriate in high-status company. It is perhaps also noteworthy that Conon refers not to Picard, but to ‘mots d’Artois’, suggesting that the former may not yet have any resonance as a glottonym. Gossen (Reference Gossen1970: 27) and others cite two references to langage pickart in the 1283 Livre Roisin, a legal document written in Lille, in a section prescribing use of the local Romance vernacular in preference to Latin (and, almost certainly, to Flemish) for the swearing of oaths. But as Boisier-Michaud (Reference Boisier-Michaud2011: 74) points out, these are the exception rather than the rule, and in so far as direct reference to the vernacular is made at all, the term roman is usually preferred in legal documents drafted in the region.

In spite of Holmes and Schutz’ (Reference Holmes and Schutz1935: 43) claims for an early thirteenth century ‘standard’ Picard, which fell from acceptability by the turn of the fourteenth century, standardization in the case of Picard can be described as abortive at best. While it is possible to argue that it was briefly selected to perform some of the H functions previously reserved for Latin, elaboration was minimal and Picard has been neither codifiedFootnote ⁶ nor accepted in Haugen’s (Reference Haugen, Pride and Holmes1966) terms.

Gossen stresses the heterogeneous character of medieval Picard texts (‘Ce que nous exposerons au cours des chapitres suivants sera donc l’élément picard de la scripta franco-picarde et non « le dialecte picard du moyen âge »’: Reference Gossen1970: 45), contrasting 18 Picard features with 32 other regularly occurring forms which are shared with other dialects (pp. 153–155), while Picoche (Reference Picoche, Darras, Picoche, Debrie and Ivar1985: 59) argues that even in the most dialectal texts the proportion of unambiguously Picard forms never exceeds 30 per cent. The fourteenth-century historian Froissart, for Pooley (Reference Pooley2004: 175) ‘the last author of any importance in the French canon to manifest any Picard features’, uses the latter to a much greater degree in quoting direct speech than in the main body of his text (Lodge, Reference Lodge1993: 132), again suggesting that Picard forms were already deemed less appropriate for ‘serious’ writing.

In the absence of a codified, pan-regional standard, later Picard writers and performers made their work accessible over a broader area of what is now northern France in the same manner as medieval scribes had done, by substituting localized forms for those of wider currency, which were often French. In his analysis of seven Picard poets writing between 1710 and 1892, Carton (Reference Carton1992; quoted by Pooley, Reference Pooley2004: 219) finds the proportion of identifiably Picard forms to range between 22.2 and 48 per cent.Footnote ⁷ Mixing was seen as the normal mode of expression, while attempts to render an overly ‘pure’ Picard were seen as inauthentic, striking a false note. Referring to the Lillois minstrel François Cottignies (‘Brßle-Maison’; 1678–1742), Carton observes (Reference Carton1965: 59): ‘une trop grande densité de dialectalismes fait suspecter une chanson de n’être pas vraiment populaire’. A further study by Pierrard of 960 Picard songs from the modern period finds evidence of conscious francofication by Desrousseaux, Watteuw and others. The nineteenth-century chansonnier Louis Debuire Du Buc is quite candid about the reasons for this (Pierrard, Reference Pierrard1966: 36):

We are thus left with something of a paradox: speakers’ perceptions of a two-variety français/patois model seem entirely at odds with an apparently unbroken tradition of mixing Picard and French elements, which extends all the way back to the medieval period.

In the light of historical and contemporary evidence, Eloy concludes that Picard is ‘un idiome [my emphasis] dont l’autonomie par rapport à un autre est demeurée partielle’ (Reference Eloy1997: 209), having never enjoyed a concerted enhancement initiative which might have cemented its status as a language in its own right (p.110):

One might suggest that the success of such a démarche identitaire forte would depend on at least three elements: codification, acceptance, and territorial unity. All of these present sizeable, and probably insurmountable, obstacles in the case of Picard.

5. STATUS ENHANCEMENT FOR PICARD

Recognition of ‘language’ status normally implies use of a codified norm in some H functions, the range of which, in the case of a regional or minority language, may in practice be limited. Certainly it is difficult to envisage the ‘effet structurant sur l’identité’ envisaged by Eloy where fragmentation of varieties hampers any genuine sense of ‘speaking the same language’ among speakers themselves, and indeed compromises mutual comprehensibility in some cases for speakers of different varieties. For national languages, the norm selected is what Trudeau (Reference Trudeau1992: 16 fn.4) has called a norme spontanée, which emerges through association with a social elite, often based around a national capital or centre of power. Regional languages are generally denied a norme spontanée, not only because by definition they are generally spoken at some distance from centres of power, but also because the very elites whose norms might have been selected are generally the first to abandon the language for a more prestigious rival (in this case French). In the case of ausbau languages such as Picard, the problem is further compounded by difficulties in determining the boundaries of the language with respect to the standard or dominant variety, and with respect to other, neighbouring varieties which form part of the same linguistic continuum.

Using a model developed by Junkovic and Nicolai (Reference Junkovic and Nicolai1987), Eloy (Reference Eloy1997) posits two sets of ‘Réferentiels normatifs’ (RN), i.e. Picard and French, from which speakers may diverge to a greater or lesser degree. The RNs are defined primarily on the basis of morphology, for which distinct Picard and French forms are available; items for which no such formal contrast exists are deemed to be common to both RNs. Thus the usage of Eloy’s own informant Maurice Boucher, an Amiénois local radio presenter, is classified as ‘Picard’ on the basis of fairly consistent use of marked morphological items which contrast with equivalent French forms, even though much of his lexicon is shared with French. In fact, as Pooley (Reference Pooley2004: 605–607) points out, much of the picardité of Boucher’s output can be attributed to frequent use of the imperfect tense, which has the distinct Amiénois Picard ending -[wε] (sg.)/-[wεt] (plu.), in his capacity as host of a nostalgic radio programme which invites Picard speakers to reminisce. Even were this not the case, morphology alone seems a poor basis on which to base a claim for languagehood and, in Pooley’s words (Reference Pooley2004: 608): ‘there would be no difficulty in finding examples of related varieties where more sharply contrasting sets of paradigms over a wider area of the grammar are not deemed sufficient for speakers to perceive the two varieties as distinct languages but rather related dialects of the same language’. Unfortunately, the alternative approach adopted by Pooley himself in an earlier work (Pooley, Reference Pooley1996), which requires item-by-item identification of forms as either ‘Picard’ or ‘French’, proves no less problematic. The complexities of such an exercise can be seen from Pooley’s analysis of French-Picard /ɑ~/-/ɛ~/ alternation. The specific phono-lexical set affected is defined as follows (Pooley, Reference Pooley1996: 99; after Viez, Reference Viez1978 [1910]: 82):

There is an obvious contradiction here between (ii), which suggests that all –ment items are subject to /ɑ~/-/ɛ̃/ variation and (iii), which implies that only some are: category (iii) in fact amounts to a catch-all grouping allowing potentially any SF /ɑ~/ to be realized /ɛ̃/ by some speakers, and raises the problem of what do and do not constitute ‘loan words from French’. Pooley suggests that membership of the –ment set should be determined on the basis of citations in early Picard glossaries, providing a checklist of patois forms (p.100), which includes for example rudemint but not mélancoliqu’mint, the latter having been rejected by Landrecies (Reference Landrecies1992: 72) as a French form ‘[habillé] régionalement à la hâte’. Such decisions appear to turn ultimately on considerations of formality, with higher register terms generally presumed to be unavailable in the Picard form,Footnote ⁸ but register in French is generally viewed as a continuum (see Offord, Reference Offord1990: 121) and it is unclear where the threshold for acceptability in Picard might be set. And, of course, it hardly needs saying that the exclusion of ‘learned’ lexical items only weakens the case for Picard as potentially a fully-fledged language fit for a range of H and L functions.

In the absence of identification with an autochtonous social elite, acceptance in Haugen’s (Reference Haugen, Pride and Holmes1966) terms depends on association with a clearly bounded territory with which speakers identify.Footnote ⁹ This again proves problematic for Picard, given that the Picard area as defined by linguists (see Figure 1), largely but not exclusively on the basis of isoglosses (see Dubois, Reference Dubois1957; Loriot, Reference Loriot1967), is an abstraction which has never corresponded to any historical province or administrative area.

Figure 1. The Picard linguistic area (from Dawson, Reference Dawson, Dotte, Muni-Toke and Sibille2012: 49, Carte 1).

The zone identified in Figure 2 sits rather awkwardly on a contemporary political map, encompassing the entirety of the modern Somme and Pas-de-Calais departments, covering all but the traditionally neerlandophone Westhoek area of Nord, and extending north-west into Belgium to include Tournai and Mons. The southern boundary cuts through the mostly Francien-speaking Aisne and Oise departments. Not surprisingly, this area has no resonance among non-linguists, and is notable for its lack of internal unity in both cultural and linguistic terms. It lacks any ethnographic political or independence movement comparable even with Le Parti pour la Normandie indépendante in Normandy. A further complicating factor is the creation in 1982 of an administrative Picardie region encompassing Somme, Aisne and Oise, but not the Nord or Pas-de-Calais; this was replaced in 2016 by the larger ‘Hauts de France’ administrative region, which includes all five departments but is still not coterminous with lingistic Picardie. The absence of any genuine pan-Picard identity stems in large part from a cultural divide between ‘francs picards’ (Carton and Poulet, Reference Carton and Poulet2006: 114) in the predominantly rural areas south of the Somme and ‘Ch’ti(-mi)s’ in the traditionally industrial urban north. This divide is significant enough to merit publication of two separate popular works on Picard (albeit with a significant overlap in content) by Alain Dawson: Le “Chtimi” de Poche (Reference Dawson2002) and Le Picard de Poche (Reference Dawson2003).

Figure 2. Typology of Nord-Picardie Varieties (after Carton, Reference Carton1981: 17).

Working-class inhabitants of the Nord and Pas-de-Calais identify as Nordistes or Ch’ti(-mi)s rather than as Picards, and while Pooley’s Lillois informants (Reference Pooley2004: 665) distinguished ‘patois paysans’ and ‘patois ouvriers’, the term ‘picard’ was hardly ever used, even by patois associations, and Pooley soon dropped it during fieldwork to avoid confusion. The socio-cultural division within the Picard area between ‘deux régions aux mentalités bien distinctes’ (Carton and Poulet, Reference Carton and Poulet2006: 114) owes much to an accident of geology: industrialization of the Nord-Pas-de-Calais was triggered by the discovery of rich and exploitable coal reserves there in the eighteenth century. Internal variation is not in itself, of course, a barrier to languagehood: major differences between varieties have hardly impeded the success of English, and all of France’s langues régionales are highly fragmented. Nonetheless, for most of the latter – a notable exception being Corsican (see below) – recognition of langue régionale status has gone hand in hand with a measure of codification for a limited range of H functions. Any attempt to codify Picard, however, is likely to be severely hampered by the peculiar demographics of the area in which Picard varieties are spoken. The problem is neatly illustrated by the imperfect ending which was so totemic for Eloy.

Evidence from the Atlas Linguistique de la France suggests that a wide range of imperfect tense forms was available within the Picard zone at the turn of the last century (see Author, 2006: 92). These reduced over time to two principal variants, one (-[o] singular/-[ɔt] plural) associated very broadly with the Nord and Pas-de-Calais, the other (-[wɛ]/-[wɛt] used mostly south of the Somme (see Eloy, Reference Eloy1997: 156–157; Flutre, Reference Flutre1977: 89–90). Of these, the northern variant is likely to be the majority form, given that the population of the Nord and Pas-de-Calais is a little more than twice that of Somme, Aisne and Oise combined.Footnote ¹⁰ Furthermore, given what is known about geographical diffusion, we would expect the form associated with densely populated industrial agglomerations in the north to be more widely adopted than the form used in the less densely populated south. This appears to be borne out by Eloy’s informant Maurice Boucher, who mostly uses the -[wɛ] form of his native Amiens, but also has some occurrences of –[o] (Reference Eloy1997: 157); by contrast neither Author (2006) nor Pooley (Reference Pooley1996; Reference Pooley2004) report any occurrence at all of -[wɛ] in the north. The form which appears to recommend itself as a standard on grounds both of majority usage and diffusion, then, is the one used by speakers who generally do not identify their own patois as Picard. Selection of the minority form -[wɛ], on the other hand, would be unlikely to command widespread acceptance given that it is almost unknown in the Nord-Pas-de-Calais.

It is not hard to imagine a multitude of similar dilemmas, particularly if, as Dawson (Reference Dawson, Dotte, Muni-Toke and Sibille2012: 49) suggests, east-west differences within the Picard area are at least as significant as those separating the north and south. For orthography, Carton (Reference Carton and Eloy2004: 185) neatly summarizes the tension between a desire to respect local differences on the one hand, and the need on the other for an accessible, standardized writing system in the absence of a codified norm: ‘Un picard “unifié” est une utopie, mais on a besoin de cette utopie pour progresser.” Abandoning this utopie altogether in favour of polynomic model, as proposed by Dawson (Reference Dawson, Dotte, Muni-Toke and Sibille2012: 49–50), is unlikely to offer a way out of this impasse. While polynomic communities such as Corsica have no single codified norm, but show wide mutual comprehension between, and tolerance of, different varieties, the model as described for example by Marcellesi et al. (Reference Marcellesi, Bulot and Blanchet2003: 199–306) crucially requires a strong sense of common identity within a clearly bounded territory, both of which are notably absent in the case of Picard.Footnote ¹¹ Without these, there is no consensus among speakers on what does or does not constitute ‘Picard’, nor on where the boundaries between it and French or, say, Norman lie (on this point see Pooley, Reference Pooley2004: 17–19).

Standardized varieties which have to be constructed from the top down often involve uneasy compromises between dialects, and rarely command widespread acceptance in Haugen’s (Reference Haugen, Pride and Holmes1966) sense. France’s regional languages provide ample evidence of artificial standard languages which command little acceptance among native speakers. In the case of Breton, for example, Jones (Reference Jones1996: 321) highlights the chasm which separates generally middle-class néo-bretonnants, who promote a standardized Breton, often learned as a second language, from the mostly poorer and less-educated users of traditional, more localized varieties (see also Kuter, Reference Kuter and Dorian1989: 85): ‘For all their endeavours to arrive at a more “grass roots” type of Breton, the variety spoken by many néo-bretonnants is still identifiable as such and remains worlds apart from that of traditional native speakers’.

In such cases, those who see standardization as a route to status enhancement or revitalization are all too often driven by a purist desire to establish a standard variety which is true to its roots, and which maximizes differences from the dominant language, even where this runs counter to the established usage of most speakers. Hornsby and Quentel (Reference Hornsby and Quentel2013: 74) for example report the coining of Celtic neologisms in neo-Breton ‘by a very small, literate elite whose authority is questioned or simply not recognized by a majority of speakers’. They cite as examples the Welsh calque poellgor ‘committee’, which is preferred to the long-established French loan komite, the coining of kontelezh ‘county’ where kontad already exists, and the redefinition of marc’h houarn (literally ‘iron horse’), already used by some speakers to mean ‘locomotive’, as an official term for ‘bicycle’.

While language-based activism in the picardophone area has a lower profile than in Brittany, similar tendencies can certainly be observed. Pooley notes that patoisant gatherings tend to attract people ‘who have achieved a degree of upward mobility and a confidence to express themselves in French’, and exclude those who feel comfortable in neither French nor Picard.Footnote ¹² Some of these endeavour to promote a notionally pure but historically inauthentic variety of Picard (Reference Pooley2004: 679):

Here, as in Brittany, we see evidence of a social divide between intellectuals promoting a pan-regional norm and generally less educated speakers who use their patois on a daily or sometime basis, for whom it expresses a more localized identity. At worst, the purist tendencies of the former can exclude the very native speakers in whose name they campaign. The Comité Régional Picard for example, cited by Pooley (Reference Pooley2004: 642), lauds the traditional varieties of the Nord-Pas-de-Calais for their fidelity to a probably mythical Picard norm:

only to dismiss the mixed ‘Ch’ti-mi’ varieties of the industrial north in highly charged terms (p.644):

Ch’ti-mi, then, is not ‘pure’ Picard, and therefore it is not Picard at all. Such reasoning ignores the fact that, as we have seen, ‘pure’ Picard in any historical era proves remarkably elusive, and is better viewed, as Carton suggests, as an abstraction which corresponds to no-one’s actual usage. Indeed if, as Eloy (Reference Eloy1997: 84) contends, ‘l’étonnant est que le picard reste aujourd’hui une réalité’, this is due in very large measure to those very Franco-Picard Ch’ti-mi koinés which emerged with industrialization through contact in the burgeoning towns of the north, and the close-knit social networks which maintained them, from the corons of towns such as Avion in the Pas-de-Calais (Author, 2006: 12) to the courées of textile workers in Roubaix (see Pooley, Reference Pooley1996: 30–34). To sacrifice such varieties on the altar of linguistic purity is both misguided and perverse.