Champaigne & Champagne, fantaisie & fantasie, and similar French examples


According to wiktionary, Champaigne and Champagne are from campanea, campania (stressed long a), whereas fantaisie and fantasie are from phantasia (short a). Fantaisie and fantasie both seem borrowed, whereas Champaigne and Champagne seem inherited.

When inherited, is ai more likely from Latin long a (stressed and unstressed, e.g., maison < mansio), whereas a more likely from Latin short a (e.g., France < Francia)?
  • Swatters

    Senior Member
    French - Belgium, some Wallo-Picard
    Nothing to do with length in Latin, long and short /a/ have identical outcomes throughout Romance.

    The factors that influenced the outcome of /a/ in French are stressed, syllable structure in Vulgar Latin (open or closed), and various surrounding environnement factors: was it followed by a nasal in another syllable, by /s/, or a nasal in the same syllable, next to a palatal segment (whether in onset, coda, or the next syllable), or followed by velar vowel, glide, fricative or velarised /l/

    Any of those factor could affect the outcome or even combine. For example, open syllable /a/ usually became /e/ or /ɛ/ (spelled e, not ai: mare > mer), but before a lenited velar, it combines with it to form a /ɔw/ diphthong, which later evolved to /u/ (fagus > */'faɣo/ > /fɔw/ + suffix -et > fouet /fuɛ/), unless the velar was palatalised, in which case it forms a /aj/ diphthong, which evolved to /ɛ/ later (plagam > */'plagʲa/ > /'plæjə/ plaie), unless the coda before the /a/ is itself palatalised, in which case the resulting /jɛj/ diphthong simplifies to /i/ (cacat > */'kʲakʲat/ > */'tʃjɛjəθ/ > chie). And so on.

    In the case of the words you highlight, Champagne and Champaigne might be graphical variants (/ɲ/ used to spelled <ngn> or <ign> before the modern standard <gn> was generalised, see oignon /ɔɲõ/ or Jodoigne /ʒɔdɔɲ/ for a few words that kept the old spelling) or reflect a real difference in spelling. The expected outcome of stressed open syllable /a/ followed by a nasal should be /ɛ/ (or /ɛ̃/ if the nasal ended up syllable final), but in place names, it's often /a/ instead.

    In the case of mansionem > maison, the /n/ was lost before /s/ in Classical Latin (and often restored later, but not in this word), so the real etymon is masionem. This yields VL /ma'zjɔne/, then */maj'zone/ by yod-metathesis. Words with this /aj/ diphthong in Old French are usually still spelled with <ai>.

    As For Francia, the /a/ was in a closed syllable, so remained /a/ until it was nasalised by the following nasal, hence modern French /fʀɒ̃s/.

    The wiktionary entry for "fantasie" seems to be about a rare spelling of the loanword from English "fantasy" (usually left as is and pronounced /fãntezi/). The entry for "fantaisie" (and its source source the TLFI) mention a spelling "fantasie" in OF, which is simply a straight adaptation of medieval Latin fantasia. The modern form "fantaisie" seems influenced by words where a similar /asia/ sequence in the etymon wasn't stressed on the /i/ and where the resulting /a'zja/ was metathised to /aj'za/ (basi'are > baiser, phasi'anum > faisan). Most other borrowing in /a'sia/ ended up as /a'ziː/ instead, but they're more modern (aphasia > aphasie, colocasia > colocasie)

    Very simplified (haha) table of outcomes of /a/:

    Open syllable

    Generally: /e/ or /ɛ/, spelled e, é or è: allare > aller, scalam > échelle
    Followed by a consonant+yod, or a lenited palatalised stop: /e/ or /ɛ/, spelled ai: area > aire, baca > baie
    Followed by a nasal consonant: /ɛ̃/ or /ɛ(ː)/, spelled ain: sanum > sain, sanam > saine
    Preceded by a palatal and followed by a nasal consonant: /jɛ̃/ or /jɛ(ː)/, spelled ïen: paganus > païen, paganam > païenne
    Closed Syllable
    General outcome: /a/, spelled a: stabulam > */i'stabla/ > étable, quattuor > quatre
    Closed by a Latin cluster, evolved by early OF to /s/: /ɑ/, usually spelled a, sometimes â: capsam > */'kassa/ > */'kʲasa/ > OF /'tʃaːsə/ > châsse
    Closed by /s/ > /ɑ/, spelled a or â: passum > pas, pastor > pâtre
    Closed by a lenited velar: /ɛ/, usually spelled ai: factum > */fajto/ > fait
    Closed by a nasal: /ɑ̃/, usually spelled an: sanguinem > VL sanguem > sang
    Closed by /gn/ > /ɲ/: /ɛ̃/ or /ɛ(ː)/, spelled ain: stagnum > VL */i'staɲo/ > étain
    Closed by /l/, velarised in OF: /aw/ > /o/, spelled au: albam > aube
    Always reduced to /ə/ then lost: causam > chose (OF /tʃɔːzə/ MF /ʃoːz/)
    Unstressed, in initial syllables
    General outcome: /a/, spelled a: amorem > amour, parentem > parent
    Preceded by /ʃ/ or /ʒ/, in an open syllable: /ə/ spelled e: camminus > chemin
    Followed by a vowel: reduced to schwa then always lost, sometimes kept in the orthography as e: habutum > eu (OF /ə'y/ MF /y/)
    Closed by a velarised /l/, a nasal or /s/: same outcome as in stressed syllables: salimuria > *salmuira > saumure, antecessorem > ancêtre, castellum > château
    Unstressed, in middle syllables
    General outcome in open syllable: /ə/, spelled e: disgravare > dégrever
    In closed syllables, /a/ spelled a, with the same special cases as the initial syllables


    Besides metathesis, what are the other variables that result in unstressed a becoming ai?

    I saw a few irregular ones:

    captivus > caitif chaitif (quasi metathesis?)

    cathedra > chaire chaiere (lenition?)

    lactuca > laitue (influenced by lait?)


    Senior Member
    French - Belgium, some Wallo-Picard
    captivus > caitif chaitif (quasi metathesis?)
    The FEW (go forward a few pages to p. 332) makes it descend from a variant *cactivus > *caχtivus, influenced by the Gaulish cognate of capio, caχto (cf Old Irish cacht, bondage, captivity). Afaik, velar codas lenited to /x/ then /j/ in unstressed syllables too, so captivum > */kʲax'tiβo/ > /tʃaj'tif/

    This also explains lactuca > laitue
    cathedra > chaire chaiere (lenition?)
    Here the FEW (p 508b) says that the Gallo-Romance and Gallo-Italic forms it listed ("wie ach sp. cadera", instead of *cadiedra, I guess) "show perturbations, seemingly from the influence of the suffix -aria". This fittingly explains chaire (cathedram > *cat'dra > */'kʲarːa/ > */'kʲajra/ (by analogy with word - in -ariam) > /'tʃajrə/), but I'm at a loss about chaiere, which seems to preserve the medial posttonic e (pretty baffling!).

    EDIT: the TLFi cites chayere in 1404. It's plausibly the result of a stress displacement to the second syllable (*ka'tɛdra > *kʲadɛrːa > *kʲaðiɛ̯ra > tʃa'iɛ̯rə > tʃa'jɛrə), but then I don't know why the initial /a/ didn't shift to /ə/ or didn't get lost altogether, but at the very least, it's likely chaiere has to syllabified cha+ie+re and not chai+e+re
    Last edited: