Linguistic ideologies, personae and practices in seventeenth-century France

1 Introduction

Third-wave sociolinguistic work has profitably analysed the study of contemporary sociolinguistic variation through the lens of linguistic ideologies and personae. Indeed, Woolard (1998: 4) suggests that, beyond work which explicitly invokes the term “ideology”, there are ‘countless studies that address cultural conceptions of language, in the guise of metalinguistics, attitudes, prestige, standards, aesthetics, and so on’, and she contends that it would be valuable to rethink much of this material ‘within an explicitly social-theoretical frame of ideology analysis’. In this article I argue that ideology analysis and the exploration of personae can offer new insights on even relatively familiar material from the past,^[1] in this case the history of linguistic standardization in seventeenth-century France. In particular, they allow us to bring together concepts and materials which have to date been viewed atomistically and help explain how standardized registers and forms can spread from a relatively narrow sector of society to wider adoption.

In a foundational chapter, Silverstein (1979: 193) defines linguistic ideologies as ‘sets of beliefs about language articulated by users as a rationalization or justification of perceived language structure and use’. Irvine, conversely, places more emphasis on the social fact as Woolard (1998: 4) observes, defining language ideology as ‘the cultural (or subcultural) system of ideas about social and linguistic relationships, together with their loading of moral and political interests’ (Irvine 1989: 255). A third definition of linguistic ideology is Rumsey’s ‘shared bodies of commonsense notions about the nature of language in the world’ (cited in Irvine 2022: 225; cf. Woolard 2021: 964–965).^[2] Woolard (1998: 3) underlines the fact that language ideologies are not about language alone: rather, they envision and enact ties of language to identity, to aesthetics, to morality, and to epistemology.

Woolard (1998) identifies four recurring – although not universally accepted – strands to the conceptualization of ideology, which are useful for the discussion in this article: (i) ideology is understood as ‘ideational or conceptual, referring to mental phenomena; ideology has to do with consciousness, subjective representations, beliefs, ideas’ (1998: 5); (ii) ideology is conceived as ‘derived from, rooted in, reflective of, or responsive to the experience or interests of a particular social position, even though ideology so often (in some views, always) represents itself as universally true’ (1998: 6); (iii) ideology is seen as ‘ideas, discourse, or signifying practices in the service of the struggle to acquire or maintain power’ (1998: 7); (iv) ideology, understood as power-linked discourse, implies ‘distortion, error, mystification, or rationalization’ (ibid.).

Eckert (2008: 456) maintains that persona style is the best level for approaching the meaning of variation, for it is at this level that we connect linguistic styles with other stylistic systems such as clothing and other commoditized signs and with the kinds of ideological constructions that speakers share and interpret and which thereby populate the social imagination. A persona constitutes ‘a way of being and acting associated not just with a social identity in an abstract sense, but with its embodiment in a character, imagined or actually performed’ (Johnstone 2017: 285).^[3] As certain linguistic features cooccur, they begin to index the individuals or groups who use them, and through interpretation and reinterpretation they become linked with qualities ideologically associated with this group (D’Onofrio 2020).^[4] This occurs through enregisterment, ‘processes and practices whereby performable signs become recognized (and regrouped) as belonging to distinct, differentially valorized semiotic registers by a population’ (Agha 2007: 81). Personae are ‘holistic, ideological social types that are recognizably linked with ways of being and speaking’ (D’Onofrio 2020: 4). For Eckert (2008: 456), ideology is central to stylistic practice, that is ‘every stylistic move is the result of an interpretation of the social world and of the meaning of elements within it, as well as a positioning of the stylizer with respect to the world’. Linguistic variables constitute an indexical field, or ‘constellation of ideologically related meaning’ (Eckert 2008: 456). The indexical system consisting of variation at once embeds ideology in language and in turn constructs ideology.

Since styles and registers become associated with ‘typifiable social personae or practices’ (Agha 2005: 37), personae and their associated linguistic usages are an important way of embedding and spreading norms.^[5] In his study of the spread of Received Pronunciation (RP) of English, Agha (2003) presents the example of how the eighteenth-century court norm for pronunciation gradually transformed into the national standard, through the process of role alignment, whereby listeners sought to align their self-image with the ‘characterological figure’ or the persona depicted in the message, as it circulated across a number of genres. While notions of ‘good usage’ and the ideology of the standard similarly only gradually embedded themselves across different strata of French society, not least through universal education in the nineteenth century and the mass media from the twentieth century on, it is in the seventeenth century that the standard comes to be linked with social personae who serve as desirable models. Moreover, as in the case of English RP, the dominant linguistic ideologies spread across different metalinguistic genres: the deliberately non-technical genre of observations and remarks on the French language (Ayres-Bennett and Seijido 2011) – that may be considered linguistic etiquette books or courtesy manuals – play a major role in popularizing ideas, partly through their commercial success and partly through being cited in other metalinguistic genres including formal grammars, dictionaries, literary conversations, letter-writing manuals and works on rhetoric, elegance and style, as well as the associated personae appearing in essays, novels and even on the stage (Vaugelas 2018: 192–204).

Following the practice of third-wave sociolinguistics (Eckert 2012, 2021), I consider identity categories ‘not as static constructs, but instead as emergent from repeated practice, necessarily dynamic and informed by shifting sociohistorical and ideological contexts’ (D’Onofrio 2020: 2). I therefore begin by presenting some of the dominant ideologies in seventeenth-century France, situating them within their socio-cultural and political contexts (Section 2); this sociohistorical context at once shapes and reflects the intersecting ideologies and thus the discussion of language in seventeenth-century French metalinguistic texts. In Section 3, I outline how the ideologies of the standard, prescription and purism infuse the hierarchization of linguistic variants and the presentation of the qualities of what constitutes ‘good usage’, while in Section 4 I examine how the increasingly value-laden standard language creates linguistic insecurity in French speakers. In Section 5, I explore how in the sociohistorical and ideological context of seventeenth-century French society, a number of personae emerge who feature widely in courtesy books, literary discussions and metalinguistic texts, notably the honnête homme ^[6] and the honnête femme. As we will see, these personae are at once created by literary, discursive and metalinguistic texts, and recognized and labelled by them as existing personae. Another such persona is the précieuse, who can be viewed positively as a strong female role model for good linguistic usage or negatively as embodying what are considered to be the dangers of excess for women, just as another persona, that of the femme savante, displays what is deemed to be an excess of learning. Having demonstrated the value of bringing together ideological analysis and personae for understanding how certain linguistic ideas penetrated French society, taking them from a narrow circle to wider acceptance, in Section 6 I nuance the picture by considering briefly the realities of both linguistic practices and the way they are described and analysed by grammarians and authors of volumes of observations on the French language. As might be expected, the linguistic reality is more complex than the ideologies which overlay it. I conclude by underlining the benefits of applying the lens of ideologies and personae for sociohistorical linguistics.

2 Language ideologies in seventeenth-century France

In this section, I review some of the dominant ideologies of the period and how they are interrelated. As Woolard notes (Woolard 2021: 962), ideologies of language are morally and politically loaded ‘because implicitly or explicitly they represent not how language is, but how it ought to be’. Through what she calls ‘the alchemy of linguistic ideology’, the language usage of dominant groups is ‘endowed with distinction that seems inherent in the language’s essence rather than historical accident’. Moreover there is a circularity in the process: the linguistic forms construed in this way are then in turn taken as evidence of the same traits in the speakers (Woolard 2021: 962–963).

Linguistic ideological research, for many, has power and the creation of hierarchies at its centre, creating ‘a network of distinctions between “good” and “bad”, native and non-native, standard and non-standard, prestigious and non-prestigious and so on’ (Seargeant 2023: 160). The ideology of one language – one nation is inextricably linked to a belief in the separability of languages as distinct entities (critiqued, for instance, by Milroy 2001) and the ideology of monolingualism. This in turn generates the need for a standard language which, since Milroy and Milroy (1991: 22–23), linguists have rightly considered to be an ideology and a standard language ‘an idea in the mind rather than a reality’.

The equation of language and nation is a historical, ideological construct, frequently associated with the nationalist ideology of language and identity (Woolard 1998: 16–17). Anderson (1983) uses the concept of ‘imagination’ to help analyse the complexity of formation of the nation-state; if we follow him in treating nation as a ‘form of consciousness of (political) community’ rather than a ‘political creed’, then this form of ‘imagining’ is, Woolard (2004: 64) argues, well established in European circles by the sixteenth century. We certainly already find the intertwining of language and nation in sixteenth-century French texts.^[7] For instance in chapter 3 of the Deffence, et illustration de la langue françoyse, Du Bellay accepts that the French language is still relatively poor, but imagines a time when ‘this noble and powerful’ kingdom will achieve universal domination and the French language will be able to equal Greek and Latin and produce its own Homers, Demosthenes, Virgils and Ciceros (Du Bellay 2001 [1549]): 82). In the next chapter (2001 [1549]: 83), he contends that the French language is not, however, as poor or ‘infertile’ as some imagine, claiming that this is due, after to God, to the efforts of ‘nostre feu bon Roy, et Pere Françoys premier de ce nom, et de toutes vertuz’ (‘our late king and father François the first of this name and first in all virtues’).^[8] He continues:

Je dy premier, d’autant qu’il a en son noble Royaume premierement restitué tous les bons Ars, et Sciences en leur ancienne dignité: et si a nostre Langaige, au paravant scabreux, et mal poly, rendu elegant, et si non tant copieux, qu’il poura bien estre, pour le moins fidele Interprete de tous les autres. Et qu’ainsi soit, Philosophes, Historiens, Medicins, Poëtes, Orateurs Grecz, et Latins ont apris à parler François. (Du Bellay 2001 [1549]: 83–84)

I say first, since he was the first in our noble kingdom to restore all the good arts and sciences to their former dignity: and thus has made our language, which was formerly rough and unpolished, elegant and if not as rich as it could be, at least the faithful interpreter of all others. And the proof of this is that philosophers, historians, doctors, poets, Greek and Latin orators have learned to speak French [i.e. have been translated into French].^[9]

From the seventeenth century on, the quintessential symbol and organ of power in relation to policing the development and usage of the French language is the French Academy. As early as c. 1625 the idea of establishing an Academy able to rival the Accademia della Crusca in Florence had been mooted and had won the support of Cardinal Richelieu, chief minister of Louis XIII. Jean Chapelain wrote enthusiastically about the idea in a letter to Guez de Balzac in 1634, although the latter’s reply was more sceptical, asking ‘D’où vient le principe de l’autorité et la source de la mission?’ (‘where does the principle of the authority and the source of the mission come from?’; Pellisson and D’Olivet 1885: 390). This question of who has the authority to pronounce on and shape the French language is a recurring one, to which I will return. From the outset, external authority was imposed by Richelieu who pressed the Academy in turn for tangible signs of its authority and influence. In 1672 the Academy was given Royal protection and housed in the Louvre.^[10]

Why should the monarchy and its politicians be involved? The creation of the French Academy was seen as part of Richelieu’s mission to limit the power of the nobility and consolidate royal power, which had been severely shaken since the assassination of Henri IV, followed by a second Regency, and Huguenot rebellions in the 1620s, among other things. It is significant that the Parlement de Paris, the sovereign judicial authority, whose political role derived from their power to register royal edicts or remonstrate against them, took nearly three years to register the creation of the Academy. The reasons for this delay are unclear, but one probable explanation is that they saw this new institution as another means devised by the Cardinal of weakening their authority and strengthening that of the crown (Caron and Ayres-Bennett 2019).

In the letters patent founding the French Academy, signed by Louis XIII,^[11] the King notes that Richelieu has played a key role in some of the recent achievements in his realm, including the establishment of peace in his Kingdom, the relaunching of commerce, the observation of military discipline, and taking control of public finances; the king continues that he wishes to follow Richelieu’s advice on other things ‘pour la gloire et pour l’embellissement de la France’ (‘for the glory and the embellishment of France’). In response to this request, Richelieu, he continues, has advised that ‘une des plus glorieuses marques de la félicité d’un État étoit que les sciences et les arts y fleurissent et que les lettres y fussent en honneur aussi bien que les armes, puisqu’elles sont un des principaux instruments de la vertu’ (‘one of the most glorious indications of a State’s prosperity is that the sciences and the arts flourish in it and that learning should be as valued as arms, since it is one of the principal instruments of power’^[12]). The recommendation is to start with the noblest of the arts, eloquence, since the French language, which has to date been neglected, is capable of becoming the most perfect of modern languages and this justifies the founding of the French Academy. Thus from the outset the intwining of language and power is articulated, and the role the French language can play in nation building.

Gal and Woolard (2001: 1) discuss how named, distinct languages, dialects, standards, speech communities and genres are constructed out of ‘the messy variability of spoken interaction’; they are the product of the work of language experts and of widely shared linguistic ideologies. Whilst the separation of French and Latin long predates the Early Modern period, it is in the sixteenth and seventeenth centuries that the status of French as a language equal to, if not superior to, the ancient languages is asserted, and that expert knowledge is ‘enacted’ (Woolard and Genovese 2007: 488) in grammars, dictionaries, and other metalinguistic works, as well as in translations. Discrete national languages are part of the formation of nations and this therefore generates a monolingual ideology, whereby the rich diversity of the languages and dialects of France is erased and the French nation is equated with the single French language. For the language to have status and authority it has to be standardized, and so the ideology of the standard emerges strongly in seventeenth-century French texts.

In seventeenth-century France two questions were uppermost: which usage should form the basis of this ideal standard, and on what or whose authority. The question as to which regional variety should become the basis for the norm had long been settled. In the twelfth century Garnier de Pont-Saint-Maxence boasts that ‘Mes languages est buens, car en France [i.e. Île-de-France] fui nez’ (‘I speak well because I was born in the Ile-de-France’; Ayres-Bennett 1996: 10). By the early fourteenth-century, an anonymous author from Meung, writing c. 1325, apologizes for his language (Lodge 1993: 100) which he describes as ‘rude, malostru et sauvage’ (‘rough, uncouth and barbarous’) because he was not born in Paris and is therefore not as elegant as Paris (‘ne se cointes com fut Paris’). The prestige attributed to the Parisian variety does not depend on any intrinsic linguistic value of this variety but is based on external reasons: Paris was the home of the royal court, of the law courts, of schools (and later of the University of Paris) and of Saint Denis, an important spiritual focus. As Lodge (1993: 104) remarks, social, demographic and economic factors were decisive: with political and economic power coming to be concentrated in Paris, ‘so the language of the rich and powerful in Paris came to be accorded greater status and respect’. Certainly by the thirteenth century Francien had become the most highly valued of the northern dialects. Whether we adopt scripta theory, whereby the writing systems of thirteenth-century texts ‘attempt to conform, implicitly and with varying degrees of success, to a supraregional norm’ (Lodge 1993: 114–115) or follow Dees in thinking that this supraregional norm only emerged in the following century (Lodge 1993: 115–116), it seems that ‘the linguistic awareness’ of many Old French authors ‘and probably their linguistic ideal, transcended the boundaries of their own dialect’ (Rickard 1989: 44). Returning to Woolard’s recurring strands in the conceptualization of ideology, the references to Paris are thus subjective, reflecting a certain social positioning, and part of a discourse to acquire or maintain power, rather than a reality. The same strands recur as the definition of the basis of the standard evolves over the sixteenth and seventeenth centuries.

In the sixteenth century debate turned to which sociolect should form the basis of the norm (see Trudeau 1992). In the 1550s, the printer, grammarian and lexicographer Robert Estienne outlines the two competing sources, the royal court and the legal courts, and presents them as being in relative harmony (Caron and Kibbee 2016: 40):

Nous […] avons faict ung recueil, principalement de ce que nous avons veu accorder a ce que nous avions le tẽps passé apprins des plus scavãs en nostre langue, qui avoyent tout le temps de leur vie hanté es Cours de France, tant du Roy que de son Parlement a Paris, aussi sa Chancellerie et Chãbre des cõptes: esquels lieux le langage sescrit et se pronõce en plus grande pureté qu’en tous autres. (Estienne 1557: Au lecteur)^[13]

We have compiled a collection, principally of what we have seen to be in agreement with what in the past we learnt from the most knowledgeable in our language, who had all their lives frequented the courts of France, both of the king and of his Parlement in Paris, as well as his Chancellery and Chamber of Accounts: in these places the language is written and spoken much more purely than in any others.

His son, Henri Estienne in his Hypomneses de gallica lingua (‘Recommendations about the French language’, 1582), argues, on the contrary, that royal court usage has become too corrupted to serve as a linguistic model, concluding that the Parlement de Paris, the legal court, must be the focus of the linguistic norm because there ‘incorrect language is as rare as it is frequent at the royal court and, applauded in the latter, it is banished from the former’.^[14] Some sixty years later, the definition of good usage offered by Vaugelas in his Remarques sur la langue françoise (1647: Preface II.3) is very different, and the supremacy of the royal court over the legal courts for spoken usage is settled. Good usage is defined as:

la façon de parler de la plus saine partie de la Cour, conformément à la façon d’escrire de la plus saine partie des Autheurs du temps. Quand je dis la Cour, j’y comprens les femmes comme les hommes, et plusieurs personnes de la ville où le Prince reside, qui par la communication qu’elles ont avec les gens de la Cour participent à sa politesse.

the way the healthiest part^[15] of the Court speaks as it agrees with the written usage of the healthiest authors of the day. When I say the Court, I am including women as well as men, and several people of the town where the Prince lives, who through the communication they have with the people at Court share their politeness (politesse).

The usage of the legal courts, the Palais, is now criticized and relegated to a technical jargon with limited usage. This corresponds to a systematic erosion of the authority of the Parlement by the first three Bourbon Kings.^[16] In Woolard’s terms (e.g. Woolard 2021: 972–973), we have here the ideology of authenticity,^[17] which represents a language variety, ‘as the voice of particular speakers rooted in particular localities’, thereby emphasizing social indexicality (Woolard 2021: 972).^[18]

Turning to the second question, in his Institutio oratoria, Quintilian argues that language is based on four things: reason (supported chiefly by analogy, sometimes by etymology), antiquity, authority (in his case, of orators and historians), and usage (I, vi, 1). He concludes that ‘Usage however is the surest pilot in speaking, and we should treat language as currency minted with the public stamp’ (I, vi, 3).^[19] In seventeenth-century France, usage also dominates as the ‘Maistre et le Souverain des languages vivantes’ (‘master and sovereign of living languages’, Vaugelas 1647: Preface I). The use of metaphors of power, sovereignty and royalty to describe the role of good usage recur throughout Vaugelas’s text and relate back to our first ideology linking language and nation.^[20] In the very next section of the Preface, Vaugelas calls usage ‘le Roy, ou le Tyran, l’arbitre, ou le maistre des langues’ (‘the King, or the Tyrant, the arbiter or the master of languages’), terms which are particularly pertinent in the contemporary political climate. Aiming to describe usage ‘du temps’, Vaugelas often differentiates contemporary usage from that of his predecessors, notably of his sixteenth-century predecessors (Ayres-Bennett 2007). In other words, antiquity no longer plays a role in his work or that of most of the remarqueurs (authors of volumes of observations and remarks on French).^[21] As well as the two authorities cited above, namely the ‘healthiest’ part of the court for spoken usage and the ‘healthiest’ part of contemporary authors, he adds the authority of ‘gens sçavants en la langue’ (‘experts in the language’, Vaugelas 1647: Preface II.7), probably a reference to his fellow French Academicians. As for reason, he notes that the majority of grammar in fact operates ‘par raison’ (‘with reason’), notably features such as the agreement of an adjective with its noun or a subject with its verb (1647: Preface V.3). However, since his aim is to treat, above all, peculiarities of usage, the questions on which, he considers, even the best people make mistakes, his observations frequently discuss examples which are ‘contre raison’ (‘against reason’). In the Remarques whilst analogy is associated with the processes of reason, it is nevertheless defined as a type of usage:

Cette Analogie n’est autre chose en matiere de langues, qu’un Usage general et estably que l’on veut appliquer en cas pareil à certains mots, ou à certaines phrases, ou à certaines constructions, qui n’ont point encore leur usage declaré, et par ce moyen on juge quel doit estre ou quel est l’usage particulier, par la raison et par l’exemple de l’Usage general. (1647: Preface IV.4)

This analogy in terms of language is nothing but a general and established usage which is applied in a similar case to certain words or expressions or constructions, whose usage is not established, and in this way one judges what must be or what is the particular usage, through reason and through the example of general usage.

In practice, Vaugelas’s conception of analogy relies on morphological regularity and patterning, rather than logical rules. The term raison (‘reason’) has a number of meanings in his work. Whilst Vaugelas states that language mirrors not just thought but the thing being represented (1647: 160),^[22] his belief in the link between language and thought does not lead to the same conclusion as the Port-Royal grammarians, authors of the Grammaire générale et raisonnée (Arnauld and Lancelot 1676 [1660]), seen by Chomsky (2009 [1966]) as key figures in his history of Cartesian linguistics. In their work, first published thirteen years after the Remarques, to use Kukenheim’s terms (1966: 35), ‘le langage étant fondé sur la pensée, les modalités du langage sont également celles de la pensée’ (‘language being founded on speech, the modalities of language are also those of thought’). The desire to explain grammatical processes in terms of operations of the mind dictates, for instance, their elaboration of two main categories for the parts of speech (Arnauld and Lancelot 1676 [1660]: 30), the ‘objects of our thoughts’ (nouns, articles, pronouns, participles, prepositions, adverbs), and the ‘form or manner of our thoughts’ (verbs, conjunctions and interjections). Conversely, for Vaugelas language operating par raison (‘with reason’) refers to the rule-governed nature of language and a common-sense attitude to the way language functions. His most frequent use of the term raison, especially in the expression avecque raison (‘with reason’) is simply ‘having an explanation’, based on a range of considerations (Ayres-Bennett 1987: 26–34), both language internal, such as clarity (notably the avoidance of ambiguity), semantics or indeed etymology, and language external, such as the rationalization that the masculine gender is ‘le plus noble’ (‘the most noble’, Vaugelas 1647: 83), based on contemporary views on gender in society, or the social imperative not to mention unpleasant things, which militates against the use of an expression like vomir des injures (literally ‘vomit insults’, Vaugelas 1647: 127–128), likely to offend in polite society.

Two other ideologies are associated with standardization: prescriptivism and purism.^[23] Whilst descriptive texts describe what is ‘normal’, ‘regular’ or ‘frequent’ in language usage without making a value judgment about it, prescriptive texts prescribe what should be said, or more usually written, based on value judgments, a notion of what is right and wrong, correct and incorrect (Ayres-Bennett 2020: 187–191).^[24] In Joseph’s words (1987: 18), ‘the prescriptive-descriptive dichotomy – or better, continuum – reduces essentially to the matter of conscious value judgment’. Descriptions of purism may be either narrow or broad (Ayres-Bennett 2020: 191–193). In narrow definitions, purism is essentially restricted towards negative attitudes towards foreign elements or language contact (Brincat 2003), and may be associated with nationalism and an attempt to maintain national identity (Hansen et al. 2018: 366). In broader definitions, purism aims to protect the (standard) language from all kinds of ‘elements held to be undesirable (including those originating in dialects, sociolects and styles of the same language’ (Thomas 1991: 12)). Prescriptivism and purism obviously have elements in common, especially the notion that some elements of language are desirable whilst others are not, but Walsh (2016: 9) usefully distinguishes the two: purism goes further than prescriptivism in introducing two additional ideas: the language (or language variety, usually the standard language) is currently pure and, therefore, to change it equals contamination, corruption or decline of some sort; and it must be protected from contamination and preserved in its current state (or, alternatively, if the language has already begun to be corrupted, the corrupted part must be removed). Purification, the result of purist ideologies, frequently leads to a reduction in linguistic resources ‘as they are de-authenticated by speakers and experts as contaminated’ (Woolard 2021: 977).

Prescriptivism and purism are enshrined in the founding statutes of the French Academy:

La principale fonction de l’Académie sera de travailler avec tout le soin et toute la diligence possibles à donner des règles certaines à notre langue et à la rendre pure, éloquente et capable de traiter les arts et les sciences. (Article 24)^[25]

The main task of the Academy will be to work with as much care and diligence as possible to give fixed rules to our language and to make it pure, eloquent and capable of treating arts and sciences.

Nicolas Faret in his preface to the Academy’s statutes (Faret 1983) equally uses highly prescriptive and purist language in his description of the Academy’s role, introducing ideas of cleansing, errors and corruption. The Academy’s principal function is to:

Nettoyer [la langue] des ordures qu’elle a contractées, ou dans la bouche du Peuple, ou dans la foule du Palais, & les impuretez de la Chicane, ou parmy les mauvais usages des Courtisans ignorans, ou par l’abus de ceux qui la corrompent en l’escrivant, & de ceux encore qui disent bien dans les Chaires ce qu’il faut, mais qui le disent autrement qu’il ne faut. (Faret 1983: 53)

Cleanse [the language] of the rubbish it has picked up, either in the mouth of the people, or in the throng of the law court, and the imperfections of legal jargon, or in the bad usages of ignorant courtiers or in the errors of those who corrupt it when writing, and also of those who indeed say in the pulpit what one should, but express it otherwise than it should be.

In statutes 25 and 26, the Academy makes it clear how it intends to fulfil its mission, by examining and commenting on the usage of the best authors, and by producing a grammar, a dictionary, a work on poetics, and a rhetoric. The idea of creating the latter two seems to have been abandoned early on, and the first edition of the dictionary only appeared in 1694 and its grammar in 1932. Whilst the Academy did not publish its own official grammar until the twentieth century, it was closely associated with the activities of individual Academicians. For the promised grammar, this meant Vaugelas’s Remarques (Vaugelas 2018 [1647]). Since Vaugelas has at times been considered the very incarnation of purism and prescriptivism, my research has focussed on examining the extent to which he and his fellow remarqueurs were in fact prescriptive or purist, that is, whether they tried to shape and change the evolution of French or simply recorded usage (as Vaugelas himself claims). I will return to this question in Section 6 below. Suffice it to say here that many of his observations are apparently couched in the language of prescription such as il faut dire (‘you must say’, of which there are more than 200 occurrences in the Remarques), faute or more rarely abus (two words for ‘fault’ or ‘error’), and the adjectives vicieux (‘wrong, defective’), mauvais, pas bon (‘bad, not good’), etc.

3 Hierarchization of variants and the qualities of good French

The creation of a standard requires the hierarchization of variants. The expression ‘pas françois’ (‘not French’) – implying a model of usage as one of exclusion rather than inclusion – is employed by Vaugelas to criticize a whole range of words, expressions and constructions which he considers unacceptable. This may refer to Latin or regional usages, but it may also include material from different registers or constructions deemed by Vaugelas non-standard and therefore not part of good usage. Allied to this is the key concept of barbarisme (‘barbarism’) which Vaugelas evokes, not to refer to the other outside of French, the foreign other, which is its traditional usage, but to what is deemed unacceptable within French. As the century progressed, attitudes in general only hardened. In 1671 Dominique Bouhours expresses similar ideas, interestingly associating the idea of removing material not just with purism but also with ‘embellishing’ the language: ‘Ainsi pour polir, pour épurer, pour embellir nôtre langue, il a fallu necessairement en retrancher tout ce qu’elle avoit de rude & de barbare’ (‘So to polish, to purify, to embellish our language it was absolutely necessary to remove from it everything which was rude or barbarous’, Bouhours 1671: 80). In short, more and more of French is excluded, to be avoided by the ‘honnête homme’ (see below) in polite society. As a result, for Bouhours (1671: 78), ‘Le beau langage ressemble à une eau pure & nette qui n’a point de gouft, qui coule de source, qui va où sa pente naturelle la porte’ (‘Fine language resembles a pure and clear water that has no taste, which runs from its source and goes where its natural slope leads’).

An ideological analysis, in Silverstein’s terms (1985: 223), considers how doctrines of correctness/incorrectness are rationalized and how they relate to ‘doctrines of inherent representational power, beauty, expressiveness, et cetera of language as a valued mode of action’. One of the dominant myths about the French language is that it is an inherently clear or logical language (Lodge 1998), but a range of other aesthetic qualities such as brevity or euphony are elaborated as part of what comes to be known as the ‘genius of the language’ (Ayres-Bennett and Seijido 2011: 80–82).

The terrain for the seventeenth-century discussion is mapped out at the end of Vaugelas’s Remarques. Having devoted two remarks to purity and clarity, the two principal qualities guiding his judgments, he concludes:

A la pureté, et à la netteté du stile, il y encore d’autres parties à ajouster, la proprieté des mots et des phrases, l’elegance, la douceur, la majesté, la force, et ce qui resulte de tout cela, l’air, et la grace, qu’on appelle le je ne sçay quoy, où le nombre,^[26] la briefveté, et la naïfveté^[27] de l’expression, ont encore beaucoup de part. (Vaugelas 1647: 593).

To purity and clarity of style, there are other considerations to add, appropriateness of words and expressions, elegance, sweetness [euphony], majesty, force, and what results from all this, the air and grace which is called the je ne sais quoi, in which harmony of word order, brevity, and natural simplicity of expression play a large role.

This means that a whole series of corresponding adjectives are used to indicate preferred forms – pur, net, propre, elegant, doux, etc. (‘pure, clear, appropriate, elegant, sweet/euphonous’, etc.). It is striking that the treatment of purity and clarity focusses above all on ways in which these qualities can be infringed: through the two principal faults against purity – barbarisms (1647: 568–572) and solecisms (572–577)^[28] – and those against clarity including infractions of word order or transpositions, incorrect structures and constructions (577–585), and ambiguity, the worst violation of clarity, which includes the incorrect usage of relative, demonstrative and possessive pronouns as well as overlong and what Vaugelas terms ‘boss-eyed’ constructions (‘constructions lousches’, 585–593). Purity is said to apply to words, expressions, particles and syntax, whereas clarity applies to constructions and structures, word order and everything which contributes to precision of expression (1647: 567). The choice of the appropriate word or expression (‘la proprieté des mots et des phrases’) – a notion related to La Bruyère’s concept of the ‘mot juste’ (I, 17)^[29] – concerns the ‘right’ choice of word or expression according to the genre, context, register, and style of the work. The other terms enumerated by Vaugelas are not discussed in detail in these final observations, but they are invoked elsewhere in the volume to justify his preference for one form over another. The criterion of elegance is mentioned in around fifty observations,^[30] typically to differentiate two expressions or constructions which are both deemed acceptable but of which one is adjudged more elegant, as in the case of à l’impourveu and à l’improviste (‘unexpectedly’), the latter borrowed from Italian but considered so naturalized in the French language that it is deemed more elegant. As Vaugelas himself notes, euphony or sweetness is a subjective notion, founded above all on what is familiar (1647: 73, 212, 424). Douceur is used to discriminate between different pronunciations, for instance, on a number of occasions Vaugelas judges that the vowel written as e is ‘plus doux’ than a (1647: 141, 146, etc.), but it is also invoked to differentiate between lexical items (rival and desbarquer are judged more euphonous than corrival and desembarquer (358, 467)) and even alternative word orders (pour aviser avec Quintus is judged ‘plus doux’ (‘sweeter’) than pour avec Quintus aviser (63)). Majesty and force are rarely cited in the Remarques, but grace is, as is natural simplicity, or naivety, a quality often associated with women’s usage since their usage of and judgments on French are deemed more natural and authentic because they are uncorrupted by knowledge of Greek or Latin. In an interesting methodological remark Vaugelas thus concludes that it is better to consult women on matters of doubtful usage (1647: 503–505).

This, then, gives us a set of values and metalanguage for evaluating usage and establishing a hierarchy of acceptability such as bon, tres-bon, meilleur, excellent (‘good, very good, better, excellent’). These are usage labels^[31] which can indicate the genres (burlesque, historique, épistolaire…, ‘burlesque, historical, epistolary’), registers (familier, mediocre, moyen, soutenu…, ‘familiar, mediocre, middle, elevated’), domains (style de pratique, de palais, de l’Église …, ‘style of legal practice, of the law court, of the church’) or style (magnifique, majestueux, naturel, néligé … , ‘magnificent, majestic, natural, neglected’) in which a particular term may be used, the characteristics of the speaker (le peuple, la lie du peuple, le peuple de Paris… ‘the people, the dregs of the people, the people of Paris’), or changing usage (vieux, moderne, vieillir, ‘old, modern, to age’).

These qualities and the associated terminology dominate much of the metalinguistic writing in the seventeenth century (Ayres-Bennett and Seijido 2011: 77–97), and indeed beyond (Ayres-Bennett 2015), suggesting the general assimilation of the associated ideologies in the powerful and influential layers of society. Another manifestation of the strength of the ideologies is linguistic insecurity felt by those who considered themselves unable to live up to the ideals of good usage, which I will discuss in the next section. The ideologies therefore create, as we might expect, inequalities of power and access to the organs and instruments of power.

4 Linguistic insecurity

An ideological allegiance to the standard language may generate feelings of linguistic insecurity in speakers who feel under pressure to maintain the linguistic norms but may in practice struggle to do so.^[32] Turpin (2002: 77) characterizes linguistic insecurity as a more or less generalized feeling of malaise when using a language or a variety. This lack of confidence in using one’s own language is underpinned by an asymmetric – or even conflictual – relationship between languages or, in the case of seventeenth-century France, between different language varieties.

A constellation of factors came together in the second quarter of the seventeenth century to make French society a particularly fertile ground for breeding linguistic insecurity. On the one hand, as we have seen, there was the symbolic strengthening of the norm through the founding of the Academy and the production of metalinguistic texts. On the other hand, France was in a period of instability at home, and war abroad: there was a series of revolts by the nobility in the 1620s and 1630s, including that of the King’s brother, Gaston d’Orléans in 1631–1632, and rebellions by the Huguenots, with the siege of La Rochelle (1627–1628) marking the last conflict with the Huguenots of Louis XIII’s reign. From 1635 France, unhappy with the terms of the Peace of Prague, entered the Thirty Years’ War. Consequently, Richelieu urgently required finances to support the war effort, raising loans, selling government offices, and placing intendants in the provinces to supervise the venal royal bureaucracy. These efforts to increase the monarchy’s authority and to raise the money it needed to pay its armies provoked the Fronde (1648–1653), the last major noble rebellion before Louis XIV laid the foundations for a system of government after 1660 (Asch 2003: 116).

The historian Asch (2003: 37) records that when royal commissioners examined claims to noble status in Paris in 1696, they discovered that 76 % of all noblemen came from families newly ennobled after 1560, and that of the 460 new families 51 % owed their status to the acquisition of a royal office. The principal period of social mobility was the sixteenth and early seventeenth century, the number of royal offices increasing from around 4,000 in 1515 to more than 45,000 in 1665 (Asch 2003: 37). Members of this new noblesse de robe were, because of their bourgeois background, initially disdained by the noblesse d’épée, nobles who had derived their status from military service, and the noblesse de race, whose rank derived from long-standing possession. Above all, they resented the social pretension which offices gave to ‘rich upstarts’ (Doyle 1996: 8).

The desire to curtail the venality of offices underpinned the debates of the Estates General of 1614 (ibid.). Richelieu had initially been in favour of the abolition of venality, but finance was required on a scale previously unknown in the 1630s, and the King also came to see an advantage to the richest and most powerful nobles being his clients, reluctant to defy his authority (Doyle 1996: 9–10). By the time of Richelieu and Mazarin, every office in the judicial hierarchy had become venal. The fast-tracked nobility which came with many of these roles^[33] meant that there was a new class of French citizens who found themselves needing to acquire rapidly the habits and social graces of the old nobility, including linguistic integration. Their linguistic insecurity was acute – as Molière’s parody in his play Le Bourgeois gentilhomme so wittily demonstrates – and this created a need for etiquette guides and metalinguistic texts to help them learn to behave, dress and speak in accordance with their newly acquired social status. My research has shown that it was above all this class who purchased Vaugelas’s Remarques (Ayres-Bennett 1987: 192–194).

5 Three personae: the honnête homme, the honnête femme and the précieuse

In this section I discuss one of the ways in which the linguistic ideologies elaborated – and reflected – in the volumes of observations and in other metalinguistic texts of the period become embedded in polite society through the creation of personae, who appear across a range of text types, thereby reinforcing the underpinning ideologies and the qualities of good language usage to be promoted as standard language style (Figure 1). In becoming recognized aspirational models to be emulated, they illustrated how adhering to the norms of good linguistic usage had social and cultural benefits.

Figure 1:

Ideologies and personae.

The notion of linguistic insecurity is crucial to our understanding of the success of the remarqueurs and of the relationship between good usage and the honnête homme. In its day Vaugelas’s volume of observations was something of a bestseller, with those aspiring to integrate into the court and polite society among the principal purchasers of the work (Ayres-Bennett 1987: 191–194), thereby spreading the ideology of the standard to new readers and new social classes. A good command of language and the skills of conversation were crucial to success as an honnête homme at the court and in the salons. In the previous century Castiglione, writing in Il libro del cortegiano (Castiglione 1981), had already set out some of the qualities required of a courtier, including sprezzatura, a nonchalance or easy facility in what one says or does, hiding the effort that has gone into it.

Norms governing the behaviour of French noblemen changed considerably between the late sixteenth century and the 1660s, especially at court (Asch 2003: 84–85). The persona of the honnête homme was elaborated in a number of key texts including L’Honneste-homme ou, L’Art de plaire à la court by Faret (1630),^[34] like Vaugelas, a founding member of the French Academy and, later, by Antoine Gombaud, chevalier de Méré.^[35] In Faret’s text, the honnête homme seeks social advancement and adopts behaviour that is polite, civil and unpretentious.^[36] Alongside the discussion in courtesy books and etiquette manuals and the satirical representation of the bourgeois gentilhomme in Molière’s theatre,^[37] the honnête homme is given linguistic characteristics and guidance in the metalinguistic texts of the period. For instance, in Des Mots a la mode, et des nouvelles façons de parler (1693 [1692]), François de Callières elaborates the qualities that constitute an honnête homme:

cette douceur, cette grace naturelle qui est répanduë sur ses moindres actions, comme sur les plus grandes, ces manieres honnêtes dont il les accompagne et qui l’ont toûjours fait regarder comme le plus civil de son Royaume, cette grande justesse qui regne dans toutes ses pensées, et cette politesse avec laquelle il les exprime. (Callières (1693 [1692]: 236–237))

this sweetness, this natural grace which permeates his most trivial actions, as it does the greatest, these ‘honest’ manners with which he accompanies them and which have always led to him being considered the most civil in his Kingdom, this great justice which reigns in all his thoughts, and this politeness with which he expresses them.

Vaugelas (1647: 123) points to the need for honnêtes gens to be constantly vigilant in their use of language in his observation on the expression des mieux:

Il n’y a rien de si commun, que cette façon de parler, il danse des mieux, il chante des mieux, pour dire il danse fort bien, il chante parfaitement bien; mais elle est tres-basse, et nullement du langage de la Cour, où l’on ne la peut souffrir; Car il ne faut pas oublier cette maxime, que jamais les honnestes gens ne doivent en parlant user d’un mot bas, ou d’une phrase basse, si ce n’est par raillerie; Et encore il faut prendre garde qu’on ne croye pas, comme il arrive souvent, que ce mauvais mot a esté dit tout de bon, et par ignorance plustost que par raillerie. Il ne faut laisser aucun doute, que l’on ne l’ayt dit en raillant.

There is nothing so common as this expression il danse des mieux, il chante des mieux, to say il danse fort bien, il chante parfaitement bien [‘he dances very well, he sings perfectly’], but it is very low register and not at all part of the language of the Court, where it cannot be tolerated: For this maxim should never be forgotten, that honnêtes gens must never use a low-register word or expression when speaking, unless it is in jest; And even then it is essential to take care that the interlocuter does not think, as often happens, that this bad word has been said in earnest, and through ignorance rather than in jest. There must be no doubt that it has been said in jest.

Good use of language was deemed essential not only to please the listener, but to avoid censure and ridicule – a parlous fate to befall someone at court – since poor language is more obvious than bad reasoning:

Il ne faut qu’un mauvais mot pour faire mespriser une personne dans une Compagnie, pour descrier un Predicateur, un Advocat, un Escrivain. Enfin, un mauvais mot, parce qu’il est aisé à remarquer, est capable de faire plus de tort qu’un mauvais raisonnement, dont peu de gens s’apperçoivent, quoy qu’il n’y ait nulle comparaison de l’un à l’autre. (Vaugelas 1647: Preface IX.2)

It only needs a bad word to make a person be despised in good company, to disparage a preacher, a lawyer, a writer. In short, a bad word, because it is easily noticed, is capable of doing more damage than a bad reasoning, which few people notice, although there is no comparison between them.

In this, women are particularly hard judges and extremely difficult to please. Discussing the use of the term expedition in the sense of a military expedition to a foreign country, Vaugelas observes the ill effect that use of just one unacceptable word can have on women:

J’ay bien pris garde, que des Dames d’excellent esprit lisant un livre, où ce mot estoit employé au sens dont nous parlons, s’estoient arrestées tout court au milieu d’un des plus beaux endroits du livre, perdant ou du moins interrompant par l’obscurité d’un seul mot le plaisir qu’elles prenoient en cette lecture. (Vaugelas 1647: 370).

I have noticed that women with excellent minds reading a book where this word was used in the sense we are discussing had stopped dead in the middle of one of the loveliest parts of the book, losing or at least interrupting the pleasure of reading it because of the obscurity of a single word.

In his Pensées, Pascal underlines the need for the honnête homme not to stand out in any way, to practice self-effacement, and not to show off their linguistic skills:

Il faut qu’on n’en puisse [dire] ni: il est mathématicien, ni prédicateur, ni éloquent, mais il est honnête homme. Cette qualité universelle me plaît seule. Quand en voyant un homme on se souvient de son livre, c’est mauvais signe. Je voudrais qu’on ne s’aperçût d’aucune qualité que par la rencontre et l’occasion d’en user, Ne quid nimis, de peur qu’une qualité ne l’emporte, et ne fasse baptiser. Qu’on ne songe point qu’il parle bien, sinon quand il s’agit de bien parler. Mais qu’on y songe alors. (Pensée 532; Pascal 1976: 282)

It shouldn’t be that one can say, he’s a mathematician, he’s a preacher, he’s eloquent, but that he is an honnête homme. It is this universal quality alone which pleases me. If, when seeing a man one remembers his book, it’s a bad sign. I would prefer that no quality is noticeable until you encounter it and have occasion to use it, Nothing in excess, for fear that one quality dominates, and gets you labelled. It shouldn’t be the case that you think he speaks well, except when it is the moment to speak well. But then you should pay attention to it.

This is mirrored in Vaugelas’s Remarques, with its insistence on linguistic conformity, the need to choose the word that the interlocutor would have chosen in order to please (Ayres-Bennett 1987: 197). As D’Onofrio (2020: 4) asserts, ‘the ability of a listener to understand the intended social meaning relies on the speaker and listener having a shared association between this set of linguistic features, the persona indexed, and the personality traits […] linked with that persona’. In the seventeenth century, the emphasis on conformity, on not standing out, is instrumental in promoting the linguistic norm as codified in metalinguistic texts. The social consequences of not choosing the right word are stressed and the reader is continually warned not to use any personal peculiarities of language, but to conform:

Je sçay bien qu’elle ne se trouvera pas tousjours conforme au sentiment de quelques particuliers [sc. la decision rapportée dans ces Remarques], mais il est juste qu’ils subissent la loy generale, s’ils ne veulent subir la censure generale, et pecher contre le premier principe des langues, qui est de suivre l’Usage, et non pas son propre sens, qui doit tousjours estre suspect à chaque particulier en toutes choses, quand il est contraire au sentiment universel. (Vaugelas 1647: Preface III.2)

I know that it [the decision registered in these Remarques] will not always conform to the opinion of a few individuals, but it is right that they are subjected to the general law, if they do not want to be generally censured and sin against the first principle of languages which is to follow Usage, and not one’s own judgment, which must always be suspicious to each individual in all matters, when it goes against the universally held opinion.

The overriding concern to please explains Vaugelas’s dislike of any syntactic ambiguity, since this distracts the reader and causes displeasure. The Remarques and the work of many of the other authors of metalinguistic texts of the period then offer a means for the newly ennobled to adopt the correct manners for their society and to establish themselves in polite society through following the precepts of good usage: this then explains Vaugelas’s subtitle, ‘utiles à ceux qui veulent bien parler & bien escrire’ (‘useful to those who wish to learn how to speak and write well’). It plays into the feelings of insecurity of those wanting to integrate themselves linguistically and socially into the dominant court milieu.

The important role attributed to women both as models of good usage and as arbiters of good linguistic usage^[38] meant that soon after the publication of Faret’s (1630) exposition of the male persona of the honnête homme, volumes appeared presenting the female persona of the honnête femme (or the honnête fille) (Du Bosc 1632–1636; Grenaille 1639–1640). François de Grenaille presents women as naturally good speakers, arguing that ‘le bien parler n’est pas tant un ornement convenable aux filles, comme une proprieté attachée à leur nature’ (‘speaking well is not so much as an adornment suitable for girls, as an inherent quality of their nature’, Grenaille 1639–1640: III, 207). The same praise of women’s language found in the metalinguistic texts is also echoed in works discussing the role of women and men in contemporary French society and indeed more generally by men of letters. Interestingly, very often women’s language is praised, not for specific linguistic features, but for the general qualities of their language.^[39] The qualities of women’s language praised in the other text types is precisely that found in the metalinguistic texts. For instance, Gilbert (1650: 15–16), writing in his work in praise of women, praises their ‘politesse’ (lit. ‘politeness’) and the clarity of their speech, and Poullain de la Barre (1679: 39, 42), in his work entitled ‘On the equality of the two sexes’, points to their dislike of linguistic ambiguities which ‘blessent leurs oreilles’ (‘hurt their ears’) and their avoidance of over-specialized vocabulary which may hinder comprehension. Other qualities associated with the good usage of women include grâce (‘grace’, Grenaille 1639–1640: III, 75) and the ‘sweetness’ (douceur) of their pronunciation and their language usage (Grenaille 1639–1640: III, 209). As Inoue (2004: 50) observes, women’s language is ‘rationalized as the natural outcome of women’s intrinsic nature. Because women’s nature is such, they prefer, or, are socialized to use feminine speech’. In this case – just as in her example of Japanese “schoolgirl” speech – it is the society’s ‘gender ideology’ that ‘naturalized the symbolic association between femaleness and softness’ (Inoue 2004: 50).^[40]

The emphasis on the natural quality of the honnête femme means that affectation of knowledge and pedantry are deemed even less acceptable for women. There thus emerges in different text types another persona, the femme savante (‘the learned woman’), who is a figure of criticism or ridicule.^[41] Guez de Balzac, writing to Jean Chapelain, comments that he would rather see a woman with a beard than ‘une femme qui fait la sçavante’ (‘a woman who plays the savante’, Balzac 1661: 138–39; cf. Andry de Boisregard 1692 [1689]: 376–377). Similarly, Andry de Boisregard argues that it is worse if women adopt a ‘stile Pédantesque’ (‘pedantic style’, 1692 [1689]: 375). A balance must be maintained between women having the appropriate social skills, such as how to express herself clearly and politely or how to dance (Callières 1693: 124), and not appearing overtly ‘savante’.

One of the most important arenas for social integration, where those aspiring to rise socially could learn to ‘live nobly’ (Lougee 1976: 212) and debate questions of language were the salons (cf. Ayres-Bennett 2004: 118–120). They were also a place for the development of feminine culture, at least for women of the upper echelons of society. From around 1610 to 1650, salon culture was dominated by the aristocratic salon of Mme de Rambouillet (Timmermans 1993: 95), who is said to have withdrawn to her hôtel because of her dislike of the coarseness of the court of Henri IV. In the salons, discussion of literature and the French language, social behaviour and taste took place, and conversation was all important.^[42] Naturalness and lack of pedantry were highly valued (Timmermans 1993: 82–83). Among the habitués of her salon were not only Vaugelas (2018 [1647]: 11), other authors of metalinguistic texts (e.g. François de Malherbe, Gilles Ménage) and important academicians (e.g. Valentin Conrart and Jean Chapelain who occupied the first two seats of the Academy), but also the leading writers of the age (e.g. Pierre Corneille, Mme de Sévigné). It was thus a place where, typically for the period, the lines between ‘professional’ and ‘lay’ linguistic opinions were blurred. In a second period (c. 1650–c. 1665), the number of salons increased dramatically. While most salons remained Parisian and aristocratic, there was an increasing role for bourgeois women in salon culture, thereby spreading the values and ideologies of the précieuses to new echelons of French society. Foremost among these salons was Madeleine de Scudéry’s ‘samedis’, Saturday meetings at her salon in the early 1650s. Whilst the demography of her salon might have changed, she nevertheless continued to stress the civilizing role of the salons, maintaining the ‘la politesse ne s’acquiert que dans le monde’ (‘politeness is only acquired in polite society’, Pekacz 1999: 27). Lougee (1976: 53) argues that the ‘salons played a central role in the process of social assimilation because within the salons ladies taught the social graces which covered the new rich with the “parfum de l’aristocratie”’. Having analysed the background of the women on Somaize’s list of précieuses, she notes (1976: 122) that about 40 % of the titles held by the families of these women were first-generation and nearly one-fifth of the total number of titleholders were new nobles.

In the salons, another key persona emerges, that of the précieuse (Ayres-Bennett 2004: 133–143). From the earliest mention of their existence in a letter written by the Chevalier de Sévigné on 3 April 1654 to Christine de France, an association is made between them and their use of a particular jargon: ‘Il y a une nature de filles et de femmes à Paris que l’on nomme ‘Précieuses’, qui ont un jargon et des mines’ (‘There is a type of women and girls that are called ‘Précieuses’, who have their own jargon and expressions’, Maître 1999: 59). There has been considerable debate about the exact nature of préciosité, since unfortunately many of the key sources for préciosité are satirical in tone.^[43] Maclean (1987), nevertheless, argues that some basic facts about it can be established: there was a distinct group of girls and women between 1654 and 1660 called précieuses who had their own gestures and jargon; they held new ideas on love and marriage; and they were found in the ruelles ^[44] where they discussed not only love and marriage, but also language and literature. Many of the features which typify seventeenth-century polite society seem to form a natural basis for préciosité: the rise of the bourgeoisie, the ideal of the honnête homme and the honnête femme, the refinement of society and concern for taste and good manners, the role of the salons and of ‘mondain’ culture, the rejection of humanist learning and pedantry, and above all the civilizing role attributed to women (cf. Lathuillère 1966: 14, 675). In the literature of the period, the term is found with both positive and negative senses, with the latter usages particularly associated with satirical texts (Ayres-Bennett 2004: 133–136).^[45] Viewed positively, they share many of the preoccupations of the authors of contemporary metalinguistic texts, favouring use of the mot juste (‘the correct word’), the avoidance of words considered too old, too low in register or too ‘realistic’. In short, they sought to ‘bannir de la société l’impurité des mots aussi bien que des choses’ (‘ban from society the impurity of words as well as of things’, Pure 1938: 380). Viewed more negatively – and here contrary to the doctrine of the remarqueurs – they favoured the creation of new words and expressions, paraphrases, ‘fashionable’ and striking expressions and lively metaphors, to the point that they were said to have their own jargon. This seems to result from a desire to ‘se distinguer du commun’ (‘to distinguish themselves from the ordinary’, Bouhours 1671: 144) which, according to contemporary opinion, at its worst led to excess, affectation and an ‘artificial’ language, in contrast with the ideal of women’s language as being natural and pleasant.^[46] Whilst the positive persona of the précieuse reinforces the idea of the model female speaker, the negative persona of the précieuse ridicule underlines the dangers of doing anything to excess and not conforming.

These, then, appear to be the dominant linguistic ideologies and related personae of seventeenth-century France. As with most language ideologies, they ‘shed light on the tight but often invisible connections between speaking and wielding – or being excluded from – power’ (Cavanaugh 2020: 56). I now want to turn briefly to discussion of the linguistic practices of the time. As we might expect, whilst the lens of linguistic ideologies offers new perspectives on seventeenth-century French linguistic thought, ideologies and linguistic practice do not always coincide.

6 Linguistic practices

Starting with the link between language and nation, if the French Academy came to embody this association par excellence in seventeenth-century France, in reality from the outset the relationship between the Academy and Richelieu was tense, and the first task it was given, to critique Pierre Corneille’s Le Cid (Académie française 1638; Civardi 2004), was carried out under duress, since the Academy had determined not to criticize living authors (Caron 2002: 30).^[47] Moreover, the manuscript of its commentary contains corrections in Richelieu’s hand, suggesting his direct – and likely unwelcome – involvement in the process (Caron 2002: 31). Richelieu pressed the Academy for tangible signs of its work, but it was not until 1694 and the appearance of the first edition of the French Academy’s dictionary that its next publication appeared, followed four years later by Tallemant’s collection of the ‘remarks and decisions’ of the Academy (Tallemant 1698), a work which sits in the tradition of the remarqueurs (Ayres-Bennett and Seijido 2011: 33–34). Whilst the Academy stood as a powerful symbol of linguistic control, it remained just that – largely symbolic – for much of the century.

When we come to the other ideologies, detailed analysis, on the one hand, of seventeenth-century metalinguistic texts and, on the other hand, of usage as attested in a large corpus of written language, demonstrates that, while the ideologies of purism and prescriptivism are clearly circulating in seventeenth-century France, metalinguistic texts and their authors are, perhaps unsurprisingly, more complex in reality. Studies of linguistic ideologies have often pointed out that even apparently descriptive works have a prescriptive dimension. For instance, Milroy (2001: 531) argues that ‘what is now clear is that the idea of what is believed to constitute a “language” can hardly escape the influence of a standard ideology’ (Milroy 2001: 539), while Amstrong and MacKenzie (2013: 106) suggest that the notion of grammaticality is ultimately a projection of standardization. What we find in the seventeenth century case is frequently the reverse: observations couched in prescriptive language, when compared with actual usage as attested in the corpus Frantext,^[48] which for the seventeenth century alone comprises over 21 million words, very often provide a good description of contemporary usage and variation and change in it. This means that, in practice, frequently the justifications added to support one usage over another, such as an appeal to clarity, elegance or euphony, are post hoc, as we might expect given they are ideologically motivated. A good example of this is the observation entitled ‘Print, prindrent, prinrent’ (Vaugelas 1647: 204):

Tous trois ne valent rien, ils ont esté bons autrefois, & M. de Malherbe en use tousjours, Et d’elle prindrent le flambeau, dont ils desolerent leur terre, &c. Mais aujourd’huy l’on dit seulement, prit, & prirent, qui sont bien plus doux.

All three are worthless, they used to be correct, and M. de Malherbe [one of his authorities] always used them [example with prindrent]. But today only prit and prirent are used, which sound much better.

The remark records language change, which comparison with usage in Frantext shows is entirely accurate (Ayres-Bennett 2016: 113), in a mixture of descriptive statement (‘today only prit and prirent are used’), the language of prescription (‘worthless’, ‘correct’) and evaluative language which apparently provides a justification for the change (‘which sound much better’). A second example is provided by the observation which prescribes the use of the preposition à and not de after the verb commencer (‘to begin’), an observation which is expressed in highly prescriptive and purist terms (‘in the purity of our language’, ‘to speak French well’):

Ce verbe dans la pureté de nostre langue demande tousjours la preposition à, apres soy, et pour bien parler François il faut dire par exemple il commence à se mieux porter, et non pas il commence de se mieux porter. (Vaugelas 1647: 424)

This verb in the purity of our language always requires the preposition à after it, and to speak French well you must say [example with à and not de].

Vaugelas adds that this requirement is so absolute that the preposition à is required even in the expression il commença à avouer, with three successive [a], which might seem to contravene euphony if this were not trumped by usage and ‘la naïfveté de l’expression’, that is, what is natural, since what is natural can only have ‘good grace’ (1647: 425). Here again, he seems to have captured the dominant usage of the decade, even if both constructions remained and remain in usage (Ayres-Bennett 2016: 113–114). And if we review the example des mieux cited above, we find that of the eleven attested examples of the construction for the period 1600–1650, eight of them appear in burlesque, comic or satirical texts. Once again, despite the long excursion on the need for honnêtes gens to avoid an expression which is ‘tres-basse’ (‘very-low register’) and ‘nullement de la Cour’ (‘in no way of the Court’), the observation seems to record contemporary usage pretty well.^[49]

A second point which emerges from the above examples is that there is variation as to which rationalization for the observation trumps others, both within the work of a single remarqueur and, indeed, across the remarqueurs corpus.^[50] Whereas in our first example, euphony was employed to justify one choice over another, in the second it had to give way to naturalness of expression.

This may affect even the choice between favouring usage or reason, to use the terminology of the period. As we have seen, the relationship between usage and reason is complex in Vaugelas’s (1647) Remarques, as is illustrated in a passage in section V.2 of its Preface. He argues that usage is like faith, in that it has to be believed simply and blindly, without reason shedding any light on it; nevertheless, we do continue to reason about faith and find reasons for things which are above and beyond reason. Consequently, although usage must be submitted to entirely, ‘il n’en exclut pas la raison ny le raisonnement, quoy qu’ils n’ayent nulle authorité’ (‘it does not exclude reason or reasoning, even if they have no authority’).^[51] Conversely, the Port-Royal grammarians, authors of a ‘general and rational grammar’, admit a role for usage, claiming that all those who work on a living language must respect the maxim (Arnauld and Lancelot 1676 [1660]: 87):

Que les façons de parler qui sont autorisées par un usage general et non contesté, doivent passer pour bonnes, encore qu’elles soient contraires aux regles et à l’analogie de la Langue

That expressions which are authorized by a general and uncontested usage must be considered correct, even if they are contrary to the rules and the analogy of the language.

However, they go on to say that these expressions must not be used to question rules or analogy, or to authorize other expressions which usage has not authorized. The reason they give for not doing this and for following the maxim is that the language otherwise will never be fixed and will lack principles, this being one of the key aims of prescriptive grammarians.

7 Conclusions

The lens of language ideologies and personae helps bring together and explain some of the dominant linguistic, political and societal currents in seventeenth-century France. Linguistic ideologies – linking language and nation; advancing standardization, prescriptivism and purism; and requiring the hierarchization of variants and the promotion of certain linguistic qualities – shape both behaviour and metalinguistic practice which are, in turn, reflected in them. They provide explanations for language behaviour and underpin the kinds of justifications offered for preferring one linguistic form over another. As the work of D’Onofrio and others has shown, they create or maintain power relations in the evolving dynamic of the historical and cultural context. Standardization, prescriptivism and purism, and the hierarchization of variants can all engender linguistic insecurity, especially in the newly ennobled, and allow those able to conform to linguistic standards to despise or mock those unable to attain these norms. Three personae – the honnête homme, the honnête femme and the précieuse – whose linguistic features index them, are also symptomatic of the ideological and sociohistorical context of seventeenth-century France. Whilst we have to substitute close textual analysis for the type of ethnographic work conducted by Woolard and others, in a historical context we can still conduct research on language ideology ‘to comprehend change and stasis in linguistic life within one integrated model’ (Woolard 2021: 978). Or to put it differently, ‘the concept of linguistic ideology … has opened up a way to look at language as the lived experience of and local response to power in historically specific contexts’ (Inoue 2004: 43).

Language ideologies do not, of course, provide the full picture of seventeenth-century linguistic practices, and the simple dichotomies which are often generated by an ideological analysis may misrepresent a much more complex reality. Just as apparently descriptive statements can veil prescriptive judgments, so prescriptive metalanguage can conceal sound linguistic description, as becomes evident when the prescriptive judgments are measured against contemporary usage. Nevertheless, what Woolard (1998: 4) terms the ‘explicitly social-theoretical frame of ideology analysis’, when used sensitively, is likely to continue to be a fruitful lens through which to read standardization and codification in seventeenth-century France and language variation and change in the past more generally.