Google translates "the eye" into "l'oeil", which should be according to my other sources "l'œil". How common is it to find "oe" instead of "œ"? Very rare or quite common?

  • 2
    Seulement un problème de clavier et de logiciel. Gogole utilise des claviers anglais et affiche oe, un traducteur francophone utilise un logiciel qui aurait corrigé en œ qui est la seule graphie admise officiellement, l'autre peut être tolérée dans les SMS et brouillon, mais pas dans un écrit officiel/publicitaire/documentaire …. – Personne Dec 27 '20 at 17:59
  • Worth reading: wiesmann.codiferes.net/wordpress/?p=757 – jlliagre Jan 3 at 1:55

You do find "oe" regularly in the TLFi (œil,bœuf), albeit the TLFi should be one among the foremost sources from which we are in right to expect a scrupulous upholding of the traditional norms. In spite of this stunning choice of spelling in the TLFi, the ligature is fairly well preserved by all sources that have some authority in the domain of language; in fact, most spell checkers provide automatically the ligature. It is mostly the text of users of the language left to their own devices for the production of a ligature by means of a keyboard that shows neglect of the ligature.
According to some, however, the ligature hangs by a thread: « chronique d’une mort annoncée ».

  • Mort annoncée pour les ligatures “entre deux « f », entre un « f » et un « l », un « s » et un « t » … ” probable, ligatures réservées aux puristes et aux bons logiciels francophones, mais que dit « la nouvelle orthographe » de œ et æ ? – Personne Dec 28 '20 at 8:24
  • 2
    @Personne Non, il ne s'agit pas du tout des ligatures de nature esthétique, celles-là ne sont apparemment pas menacées. Il y a eu des éliminations de ligatures orthographiques (œ et æ) d lors de la Réforme 90 et elles me semblent raisonnables à première vue : j'ai trouvé un compte-rendu à ce propos mais je n'ai pas eu le temps de m'y intéresser de près ; en particulier je ne me suis pas attardé sur la question de savoir si la réforme était allée trop loin ou pas assez en ce qui concerne ces ligatures. – LPH Dec 28 '20 at 8:40
  • … alors on ne va plus comprendre pourquoi Gainsbourg chante "Elaeudanla Teïtéïa" (youtube.com/watch?v=5Kkc--aYns8 :-) – Personne Dec 28 '20 at 10:33
  • In the TLFi, the presence of "oe" vs "œ" is not a choice of spelling but a OCR related typo. It has been listed in this document: Un autre cas remarquable est celui de Œ et œ. Dans une tranche alphabétique qui s’étend de la fin de la lettre l à la fin de la lettre p, les digrammes ligaturés Œ et œ ont été dissociés en OE et oe. – jlliagre Dec 31 '20 at 15:22

The normative usage in French is that “Œ” and “OE” are distinct spellings. The ligature is treated like a diacritic (i.e. like an accent or a cedilla): it is sometimes omitted when there are technical difficulties (French keyboards don't have an Œ key), but this is considered an approximation of the real spelling, not a correct spelling. Spelling *oeil instead of œil is no more correct than spelling *ete instead of été or *ca instead of ça. Serious typography (such as professionally typeset books and newspapers) always uses the ligature.

In practice, not making the ligature is moderately jarring, because it only has a minor effect on the pronunciation. In most common words spelled with œ, it's part of the digraph œu, and the trigraph oeu doesn't exist in French but is close enough that it could be pronounced in the same way ([œ] or [ø]). Apart from words with œu, I think the only common word with œ is œil (and family), where likewise *oeil* would not be too wild. A œ without a following vowel is often pronounced [e] or [ɛ], which it would be jarring to spell oe, but I think all the words where œ isn't followed by a vowel are relatively rare (and several have an alternative é spelling). Not making the ligature is not as jarring as leaving out a cedilla or an accent that changes the prononciation, but more so than an accent that doesn't change the pronunciation or doesn't change it much. But any educated person would consider it wrong, just perhaps not enough to matter.

A sign that Œ is a distinct spelling from OE and not just a typesetting variant is that when a word is typeset with an initial capital, the whole Œ is uppercased, and not just the O. For example see the title pages¹ of L'Œil tourné and La poule et l'œuf, as well as that of Langue de vipère et œil de biche (note how the whole Œ letter is in blue):

L'Œil tourné La Poule et l'Œuf

Langue de Vipère et Œil de Biche

Looking online through the covers of books titled “(L')Œil …” or “(L')Œuf …” (I went through the top 50 of https://www.google.com/search?tbm=bks&q=oeuf and https://www.google.com/search?tbm=bks&q=oeil), I did find a few where the œ word was not spelled with an Œ or œ ligature:

  • OEIL, L'OEUF, Oeuf are self-published books with otherwise evidently low production values (and the last one isn't even in French).
  • “Allez vous faire cuire un œuf” is typeset as oeuf on the cover page but œuf inside, so this seems to be a case of not putting the effort (maybe the font used in the title didn't have an œ character) or perhaps a deliberate stylistic effect to avoid an unusual letter.
  • “L'Œil de Shiva” and L'Œil de Caine, both from mainstream publishers, make a non-standard ligature with an uppercase O and a lowercase e. This seems to be a stylistic effect; at least “L'Œil de Caine” is spelled with a regular Œil inside and another edition has ŒIL on the cover. Note that even though the letters are not spelled in the standard way (where the E in the capital Œ ligature is an uppercase E), the typesetter or graphic artist went through the effort of attaching the e to the O, so it's still a ligature and not separate letters.

My conclusion from this non-scientific sample is that not spelling the ligature properly is rare, and is either due to technical limitations, not bothering, or a stylistic effect. There are far too few occurrences to consider separate letters a standard spelling variant.

Technically French also has an Æ/æ ligature, but it's only present in a few very rare words and in some Latin imports, and as far as I know an alternative spelling using ae, e or é instead of æ is always acceptable. In contrast, œ is present in several common words (œil, œuf, bœuf, cœur, …) for which there is no standard spelling variant.

¹ Traditionally, title case in French means typesetting the first word in uppercase, and if that's an article, then typesetting the following noun as well as any adjective preceding it. However modern title typesetting often only capitalises the very first letter. Cover page typesetting often has additional stylistic effects such as all caps (very common), all lowercase (not very common), or occasionally cosmetic effects to highlight some words or just to stand out.

  • Parmi les quelques recherches effectuées après la réponse de LPH à mon commentaire, je n'ai pas trouvé de règle officielle pour la graphie de æ et œ (vive les claviers BÉPO [bepo.fr/wiki/Accueil] :-) proposée par la réforme de l'orthographe. – Personne Dec 28 '20 at 17:40

According to Google Ngram, there was a time, 1709 to be precise, when the two spellings œil and oeil were almost equally used, at least in print (who knows why?). The results are, of course, subject to the availability of the linguistic corpus. There was another, less dramatic convergence in 1994. Now, of course, œil is almost always used. The sudden divergence after 1999 is another mystery (at least to me)...

Ngram attached

  • Impressive chart! The encoding ISO/IEC 8859-15 (ASCII-based standard character encodings), which includes the ligatures, was published in 1999, so it became easy to use them in the Internet and computer-generated publications. That's probably the reason for the divergence. – UweD Jan 2 at 6:33
  • 2
    No, this diagram is worthless. What you're seeing is what optical character recognition reported. If you check the actual images for thel hits that are supposedly for the oe spelling, they're pretty much always œ. My guess for the divergence in recent years is that most books became available as text which had œ from the start. – Gilles 'SO nous est hostile' Jan 2 at 9:18
  • @Gilles Checking the hits shows a lot of true positives. In any case, why would OCR accuracy vary depending on the publication date? – jlliagre Jan 2 at 9:36
  • 2
    @jlliagre I didn't check the early 1700s, but I did check recent hits when I wrote my answer, and almost all hits reported for oe were in fact OCR mistakes. I don't know if OCR accuracy varies by publication date, although it is possible due to printing quality and font choice. My hypothesis is that many recent books are not OCRed at all. – Gilles 'SO nous est hostile' Jan 2 at 10:43
  • @Gilles Granted A closer look shows many false positives too, so the rates are indeed worthless. There seems to be more separated "o e" in the past, but a different aproach should be used to get the actual figures. – jlliagre Jan 2 at 14:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.