Gustave Flaubert, in a letter to his publisher of October 1862, and after two other grumbles about the typesetting on the page-proofs of his new book: ‘3º The circumflex accent on Salammbô has no profile. Nothing could be less Punic. I demand a more open one.’ To demand with Flaubert was to get: within a few days he had an accent that straddled its underlying vowel in the comprehensive way that he wanted, and gave the name of the heroine of his Carthaginian novel a suitably Punic appearance on the title-page. Or it would be truer to say that it gave her a suitably alien appearance, because to a French reader Salammbô’s terminal chapeau comes as a shock, occurring as it does in a position where no circumflex has any business being.
A circumflex looks strange on a vowel coming right at the end of a word, since, characteristically, this accent marks the spot from which a consonant once present in the spelling before another consonant has been dropped. We were taught in school that where now there is a circumflex in a French word there was generally in the old days a pre-consonantal s, and that when this s was suppressed it was replaced by an accent set to the left over the preceding vowel. Whence such familiar modern French words as bête, île or croûte, which were at one time written as beste, isle or crouste. And whence, also, the title of Bernard Cerquiglini’s thoroughly enjoyable little book, in which he traces the history of the circumflex and the mysteriously strong feelings it has aroused, for and against, over the past 450 years.
It is tempting to say, after reading Cerquiglini, that there is more to the circumflex than meets the eye, except that meeting the eye is so very much part of this assertive signifier’s function. As Flaubert well knew: the circumflex that he gave to Salammbô is conspicuous, a puzzle, because it clearly can’t be there for the usual reason – set up, that is, in memory of the loss of an s. It could hardly be there either out of any wish on Flaubert’s part to imitate the eccentric use of the circumflex over the passionately vocative O in French – as in the tragic hero’s ‘ô ciel’ or ‘ô dieux’ – since that would make his character’s name look as if it was ending too soon, its invocation suspended.
The circumflex is there in fact for quite a simple reason: to make a strange name look stranger still. Whoever or whatever Salammbô may turn out to be, whether place or person, he, she or it cannot be French. There is a first evidence of that in the unnatural combination of a double m followed by a b; and the final circumflexed ô clinches it. Salammbô is a Punic name, and some measure at least of its Punicity is conveyed in that barbaric accent. It seems that the Punic language itself didn’t know the circumflex, which was a Greek, not a Phoenician, invention, but the Romans had it and Cerquiglini cites its use among their grammarians in the accenting of such alien, not to say Punic, names as Hannibal and Hadsrubal. Salammbô’s circumflex would have been at home in Latin, which is where the obsessively well-prepared Flaubert no doubt took it from when he was reading up for his novel.
The Romans, like the Greeks before them, used the circumflex for phonetic purposes, as a graphic index to how the speaker’s voice was to behave in pronouncing particular vowels – the accented vowel in Hannibâlem was to be pronounced long. The circumflex may look like an arbitrary sign, as if it could have taken any shape but just happens to have taken the shape we know. Not so, however: it is in its way iconic. The Greeks had had both an acute and a grave accent, and used them to indicate whether the voice should be going up (acute) or down (grave) at a given point in a word. On occasion, the Greek voice needed to go both up and down within the space of a single syllable, so that acute and grave accents were required in rapid succession, a succession so rapid that eventually the two written signs were amalgamated, giving birth to the circumflex, the upside-down chevron or ‘hat’, which was flattened out in classical Greek into a sinuous, tilde-like line, perhaps because that seemed a more suitably melodic representation of the vocal ups and downs involved. This phonetic inheritance has remained with the circumflex throughout its history, complicating it no end and leading in recent centuries in France to a chronic, at times angry, disagreement about what the accent is really for. That disagreement is at the heart of Cerquiglini’s book, and he gives an admirably neat account of it, accessible even when it touches on the technicalities of linguistic change.
The issue is a large one: of what, after years of letting the arguments of Jacques Derrida sink fully in, we now know as phonocentrism. Phonocentric is what we almost all of us are but none of us should be, according to Derrida: we look on the written language simply as the representation of the spoken, rather than as a medium with rights and freedoms of its own. The original applications of the circumflex were phonocentric, since they meant using a graphic sign purely in order to represent certain inflections of the voice required in speaking. And the same phonocentrism was behind the gradual – the remarkably gradual – installation of the circumflex in the written language in France between the middle of the 16th century, when a pro-circumflex campaign began, and the middle of the 18th, when the campaign at last succeeded. The incentive to introduce the accent was the recognition that French as currently spoken and as currently written had diverged to a degree that some found intolerable: a nonsense, and a handicap when it came to teaching the language. The key example around which the generations of reformers and conservatives subsequently contended was that of the s which came before a t (or, less often, another consonant) in words such as maistre, masle or beste. In French as spoken this s was silent, a dead letter; according to Cerquiglini, it had been so in many cases since as early as the tenth or 11th century, which was when the question of how French should be written was first posed. Language reformers five hundred years later, during the Renaissance, thought it was high time written French recognised that the s was dead; they wanted it removed. Simple abolition, however, wouldn’t do, because over the long period of its redundancy the s had had its effects on the pronunciation of the vowel that preceded it, ensuring that this was a long or open sound, not a short or closed one. Take the s away and who would know the proper value to give to the bereaved vowel? The answer was found in the circumflex accent, which could henceforth serve as a phonetic indicator to the pronunciation of the vowel, in accordance with the ancient tradition.
This was the progressive point of view, maintained by, among others, the poet Ronsard. It did not prevail. The alternative point of view was that the s should be kept, on the grounds that it was a valuable reminder of an earlier form of the French word in question, a reminder of the essential Latinity of the French language, which was in danger of being lost to view as the result of so many centuries of phonetic erosion. That isle should be written incongruously as isle, rather than brought into line with speech as île (or ile: the circumflex itself is redundant over an i because it doesn’t change the pronunciation), was a small lesson in cultural history, since in the s form of the word it was possible to retrace the Latin noun insula from which isle had come.
The Renaissance was not a good time for the reformers to have set about modernising French orthography, because it was a time also when the vocabulary was being enlarged by imports, from classical languages as well as from other Romance languages, with the consequence that learned terms came to co-exist with popular ones: the popular île, for example, with the learned insulaire, where the un-eroded Latin original was there for all classically educated persons to enjoy. To replace the s of isle with a circumflex, therefore, would disconnect the word doubly, synchronically from the scholarly form insulaire, and diachronically from its Latin etymon, insula. Latin, indeed, was making a comeback, to the point where, says Cerquiglini, esses before tees that hadn’t been sounded for centuries began to be revived in the spoken language and to undermine the reformist case.
The argument nevertheless had been joined once and for all, between the phonocentrists who wanted words to look the way they sounded, and the graphocentrists, who saw virtue in preserving a degree of autonomy for the written language. The graphocentrists argued well, pointing out that practised readers read with the eyes, recognising words by how they look and without pronouncing them silently to themselves as they go, so that there was no logic in the demand that the written language should be the closest possible represention of the spoken. There was the good argument, too, that the written forms that words take are conditioned by grammar as well as by phonetics and can be seriously impoverished as bearers of information by the removal of what look like superfluous letters.
The graphocentric party held on for two whole centuries, élitist to the last, before losing out. The circumflex was finally imposed by the authority of the third edition of the Dictionary of the Académie Française, which appeared in 1740. The two previous editions of this work, of 1694 (an earlier edition was so full of mistakes it was pulped) and 1718, had stuck to the etymological line, the 1694 Preface making no bones about where its loyalties lay: ‘The Academy has followed the old orthography accepted by all men of letters, because it helps in the understanding of the origins of words.’ But the decisions about spellings that the immortals had taken did not necessarily go through on the nod, and Cerquiglini quotes one or two jolly snippets from the minutes of the deliberations:
Boyer: ‘I don’t want any s in flute.’
Tallemant le Jeune: ‘It needs one.’
Segrais: ‘I am for fluste and for composte.’
Doujat: ‘So am I.’
Bossuet: ‘And I.’
By 1740 a more democratic view of the written language was suddenly being taken. What was the point, as one earlier reformer had asked, of conserving a built-in reminder of the Latin origin of words when so many people were incapable of recognising it? Those who knew Latin would know without having to be reminded that île derived from insula: to those who didn’t know Latin the reminder was useless. And there were large classes of people who did not know Latin because their sex, age or nationality was against them: women, children and many foreigners. Why make it harder for them to learn the correct forms of written French than it needed to be?
Women especially: they had for too long been shut out by the exclusiveness of those at whom Cerquiglini has a well-justified dig as the ‘masculine Latinists’, whose ‘virility was measured by the number of superfluous consonants’. The first women who tried to lobby against this were the 17th-century Précieuses, who were far from being as ridicule as Molière chose to portray them on the stage. Precious or not, they were in favour of a major spelling reform, as comes out in a brief scene quoted by Cerquiglini from Somaize’s Grand dictionnaire des Précieuses: ‘Roxalie said that they must so go about things as to make it possible to write the same [de mesme in the French, inevitably!] as one spoke and, in order to execute this plan, Didamie took a book, Claristène took a pen, and Roxalie and Silenie prepared to decide what needed to be adjusted or reduced in words so as to make them easier of use and more convenient to spell.’ (An echo, here, of an earlier grammarian’s advice to his fellow men: ‘Make it simple, Messieurs. Write like your charming spouse.’)
When the new orthography officially came, there was a wholesale slaughter of superfluous letters, leading to much cutting of new fonts in the type foundries as printers adjusted to the shotgun marriages between vowels and accents. Five thousand out of the eighteen thousand or so words contained in the third edition of the Académie’s Dictionary were affected, though by no means all of these were cases of a circumflex being brought in to replace an s or other unnecessary letter. Inconsistencies remained, but this new word-list marked a large, irreversible advance in simplicity and regularity of spelling. The circumflexes it authorised are, all but a few, with us still.
Not that everyone is happy about that: the spelling reformers never sleep. At the end of 1990 their chance came when a quangoid body called the Conseil Supérieur de la Langue Française proposed a modernisation of French orthography. This had mainly to do with neologisms, and not least with how to treat the dreaded invaders from English, whose new imperium is a source of perennial anxiety to linguistic nationalists in France. As a sideshow, the circumflex came under attack, too, on the old, phonocentric grounds that it occurs on vowels that don’t require it, since it has no effect on their pronunciation (as with île). In those cases it should go, advised the Conseil. It didn’t, and it won’t, go, however: once again, the circumflex found intelligent defenders, among them the then President, François Mitterrand, who quashed the proposal and was able afterwards to claim: ‘I have saved a few accents.’
Cerquiglini is on the conservative side, almost lyrically so at moments, as when he describes his favourite accent as a ‘graphic sign that belatedly took flight to be added as an adornment to the line of type, a pure icon of the attention that we give to the character’. The circumflex deserves to find friends abroad as well, not least among a deprived and envious people like ourselves whose written language contains no accents at all. Accents are an enrichment, as supplements to a script; they look good and they have things to tell us. To want to do away with them in the cause of simplicity is insulting. And in case that, remembering the Précieuses, sounds like the old ‘virility’ all over again, let a French woman writer, Colette, say it: ‘I adore words that are complicated to spell.’ And I, as Bishop Bossuet might have added.