OK, that may be too apocalyptic a question, but I’m astonished by the results of a study conducted by Dalila Ayoun of the University of Arizona and reported on by Heidi Harley at Language Log: “Fifty-six native French speakers, asked to assign the gender of 93 masculine words, uniformly agreed on only 17 of them. Asked to assign the gender of 50 feminine words, they uniformly agreed [on] only 1 of them. Some of the words had been anecdotally identified as tricky cases, but others were plain old common nouns.”

There’s an even more interesting twist in Ayoun’s native-speaker results. Her native speakers fell into two groups: 14 adult speakers and 42 teenage speakers. On most grammatical tasks, for all intents and purposes, teenagers’ native-language abilities are identical to adults’ abilities. But when she broke down the gender-assignment task results by age, she found that teenagers showed considerably more variation than the adults. On the 50 feminine nouns, for example, the 14 adults all agreed on 21 of them, while the 42 teenagers agreed on only one: cible, ‘target’. Of the 93 masculine nouns, the adults agreed on 51 of them, while all adults and teenagers agreed on only 17 (of 93!!)

Heidi reproduces one of Ayoun’s tables “illustrating significant differences in the rates at which adults and teenagers agreed on the gender of 10 feminine nouns”; it’s well worth the look. I wouldn’t have known that primeur ‘early fruits and vegetables’ (often used metaphorically: avoir la primeur d’une nouvelle ‘to be the first to hear a piece of news’) was feminine without looking it up, but that only one of the French teenagers did is amazing. Mme Ruegg (my high school French teacher) would be wielding her ruler vigorously and/or emptying the bottle of booze she kept in the bottom drawer of her file cabinet.


  1. I assume this is shocking because one assumes that the French Academy and French tradition is generally more successful at keeping archaic elements of the language alive. Wouldn’t one expect that since the gender markers were mostly lost through phonetic change centuries ago, French (and German) would eventually follow the trend of English and Dutch towards much simplified gender structures? Is there any evidence in French dialects or colloquial speech of gender beginning to fade away?

  2. A while back my brother was with two travelling French Canadians who were trying to figure out the gender of the name of a certain tree they’d never seen before. They used this as an excuse to make a lot of smutty comments about the poor tree. Gendered languages can be fun. If you’re not a tree, anyway.

  3. I assume this is shocking because one assumes that the French Academy and French tradition is generally more successful at keeping archaic elements of the language alive.
    Well, actually it’s shocking because it’s the first indication I’ve seen that gender is an archaic element. As far as I knew it was as alive as it is in German or Russian.

  4. In Yiddish nowadays everything has the same article: “duh.” I don’t know whether this is male or female and it doesn’t seem to matter anymore. There goes Gloria Mundi!

  5. fimus scarabaeus says:

    IT shows minimum adequacy survives and is well.

  6. I’ve always felt that gender is somewhat archaic in German – it doesn’t feel to me like an intrinsic part of the language the way it is in Russian, Italian or Spanish. Partly it’s because gender of a given noun in Hochdeutsch can be completely different than the gender of the same noun in the local dialect – which makes the whole structure seem a peculiarly arbitrary formality. Partly because in all the other Germanic languages, and a lot of German dialects, gender is much simpler than Hochdeutsch, which to me implies that the current gender structure is an artificial imposition. And mostly because despite speaking pretty decent German, I personally have an awful time remembering the gender of nouns.

  7. Bruno van Wayenburg says:

    @Vanya, I know that we haven’t got much to boast about compared to Russian or Swahili, but how is Dutch gender simpler than French gender? They got to of them (masculine ‘le’, feminine ‘la’), we’ve got two (masculine/feminine ‘de’, neuter ‘het’), and some southerners/Flemings even keep track of separate masculine and feminine genders.
    Also, we pretty much agree on the genders of most words, as a n=2 study with my wife shows.
    Nice post, anyway.
    -Leiden, the Netherlands

  8. Sorry for the confusion – I was thinking of Dutch and Swedish in relationship to German (or Latin, Greek or the Slavic languages) i.e. you’ve already merged masculine and feminine. But of course you’re right, it is no simpler than French.

  9. Interesting observation about German, vanya. I never had the sense that gender is archaic (I’m a native speaker myself). As far as I know – I also happen to be a native speaker of the Swabian dialect – gender mutations are a relatively rare phenomenon. The only thing that comes to mind is “die Butter” (Hochdeutsch) and “der Butter” (Schwaebisch/Swabian). I would love to hear some other examples.

  10. Bavarian has dozens of differences – da Zwife, da Aschn vs. die Zwiebel, die Asche, or s’Egg vs. die Ecke, as Tunnoi vs. der Tunnel, da Schbiez vs. die Spitze, I could go on. I also thought Plattdeutsch often followed Dutch in obscuring masculine vs. feminine, although I’m not an expert.
    In any case, if native German (or French) speakers don’t find gender artificial and archaic, then it clearly isn’t. I’m sure it’s an Anglocentric bias I have – if we don’t need it in English any more, etc. etc. It just seems to me that once the word final gender markers erode, as they did in English, and have in German and French, (but haven’t in Russian, Polish or Spanish), grammatical gender should become more and more confused to the point where it eventually disappears. But apparently gender does have some utility to German, Dutch and French speakers. I’d be interested to hear from a native speaker what that might be.

  11. Vanya-
    1-Leaving aside Creoles and advanced cases of language death, there is no variety of French, or indeed any Romance language/dialect, no matter how non-standard/colloquial it is, which has lost grammatical gender.
    2-A majority of the feminine nouns listed on LANGUAGE LOG are highly literary in flavour, and if the teenage francophone informants were American-born or raised, I suspect the extreme variability is indicative of reduced exposure to standard registers of French: more in any case than of variability in French gender in general.
    3-As a native speaker, I can tell you that grammatical gender exhibits little variability as far as basic vocabulary is concerned, and indeed is acquired quite early by native French-speaking children.
    4-The rules of gender assignment, which native speakers apply consistently (albeit unconsciously), are rather complex, however, and rare/unfamilar words whose ending does not clearly indicate gender can cause native speakers problems (I remember a native Polish-speaking acquaintance of mine who was amazed native French speakers could hesitate about this).
    4-The chief impact of learned influence upon Standard French grammatical gender is that it has “embalmed” in the Standard a number of rare, low-frequency nouns with irregular (in terms of the link between form and gender) or variable gender which, had there been no such learned influence, would have been regularized long ago.
    Hope this helps.

  12. This doesn’t prove anything. The result was found in the native-speaker control group in a study of non-native speakers. You wouldn’t get 56 native speakers of any language to agree on the genders of a bunch of words, or on any other grammatical quiz. As the the second graf of the original Heidi post says, this just proves that native speakers don’t unanimously agree on gender, not that they don’t know it. The only way to say whether gender is disappearing from French would be to take a random sample of spoken French in all kinds of venues, and analyze it to measure the rate of correct gender assignment. I suspect you’d find it pretty high.

  13. Siganus Sutor says:

    Yes, Étienne, even “learned” French native speakers can get confused I think. Un ou une échappatoire ? And oasis? albâtre? acné? obélisque? immondices? amiante? astérisque? après-midi? interview?
    Those are words that could be used daily in ordinary conversation.
    grammatical gender exhibits little variability
    Maybe from one place to the other, but with time quite a few words underwent some kind of… er… surgical operation.
    Pour finir, un petit joke* :
    Deux hommes se promènent en montagne. L’un deux regarde au loin et dit : “Tiens, une hélicoptère.” L’autre : “C’est pas une hélicoptère, c’est un hélicoptère.” Le premier : “Tu vois ça d’ici, toi ? T’as vraiment de bons yeux !”
    * masculine in Mauritius, feminine in Quebec I believe, just like plug, job and some other words used while speaking French…

  14. You wouldn’t get 56 native speakers of any language to agree on the genders of a bunch of words, or on any other grammatical quiz.
    I doubt that’s true even literally, and I’m quite sure you wouldn’t get the level of disagreement displayed in the survey. If that many native speakers disagree that substantially about that many words, there’s something going on.
    Sutor: Great joke!

  15. You wouldn’t get 56 native speakers of any language to agree on the genders of a bunch of words
    Not true. Consider most Slavic languages where gender is – to a certain extent – still marked morphologically. To use Siganus’ example, I cannot imagine a native speaker of Polish who would identify “helikopter” as female (note the word-final consonant), just as there is no speaker of Slovak who would identify “helikoptéra” as male (note the word final -a). An occasional borrowing may be a source of confusion (I for one can never figure out where Polish “satelita” falls – male?), but in the vast majority of cases, the results would be unanimous.

  16. I assume that Siganus’s joke is the kind of joke my brother’s French-Canadian buddies were making about that poor innocent tree, but as I recall the story they went on at some length.
    Levi-Strauss and Roman Jakobsen’s analysis of one of Baudelaire’s poems involved understanding masculine and feminine rhymes, masculine and feminine grammatical gender as indicators of actual maleness and femaleness. It made a big splash when it came out, but if I’m not mistaken the consensus now is they they were lost in the weeds.
    Is there a copy-editor in the house? Should the previous sentence have begun “Levi-Strauss’s and Roman Jakobsen’s”? Because that sounds just awful to me, but it seems that it might be required.

  17. I have a weak but usable reading knowledge of German, and I ended up concluding that readers don’t need to learn the noun declensions. Very often the article or the context tells you what’s singular and what’s plural, and often enough it doesn’t make any real difference. So you can afford to just look words up whenever you think you need to know.
    Which leads on to my main point, which is that I think that the German noun declensions are the most horrible mess I’ve ever seen in any language. The Chinese writing system is a mess, but it’s a beautiful mess that gave us Chinese calligraphy and also provides us with an endless supply of jokes, nuances and cultural insights (which can be real, frivolous, or wrong). But the German noun declensions gave us only Hitler.
    The declension “system” is also a good touchstone for prescriptivist or pedagogical insanity. There exist in this world German teachers who explain to you that their system of five rules, two hundred exceptions, and twenty or more rules for applying the rules to new words is perfectly logical once you understand it.

  18. Not a copy editor, but you got it right the first time: for joint possession an apostrophe goes with the last element in a series of names (according to Bryan Garner at least).

  19. michael farris says:

    “I for one can never figure out where Polish “satelita” falls – male?)”
    I hadn’t thought of that but quick googling brought up someone complaining about recent occurences of it being treated as feminine ‘ta satelita’ where as it’s traditionally been ‘ten satelita’. For this non-native speaker neither ta satelita nor ten satelita sound right.
    There are also two separate towns in Poland called Ostrów (from an older word for island) To distinguish them, regional adjectives are added, but the adjectives differ by gender so there’s Ostów Wielkopolski (masc) but Ostów Mazowiecka (feminine).
    My favorite weird gender case in Polish is por (leek) which is officially masculine but at least where I live also occurs in the feminine even though there’s nothing phonologically feminine about it (a consonant that ends a feminine noun should be soft or at least softish but that r is hard as a rock).
    The wikipedia page:
    seems to use to treat it as masc and fem in different places (it’s just possible that the feminine references are supposed to be for the word roślina ‘plant’ but it doesn’t really look that way to me)

  20. Siganus Sutor says:

    Thank you Steve. A great (borrowed) little joke, maybe, but a joke in which I have shamefully forgotten an apostrophe…
    Apostrophe? Un or une apostrophe, by the way? (I’d say “une” but…)
    Bulbul, words (and names) ending with an -a tend to be seen as feminine, but there is some hearsay that it’s not a universal rule. Not even a Latin one. Some mad Latinist around to confirm or deny?

  21. Sutor:
    Nice joke. Actually, in Quebec French there has long been a trend whereby vowel-initial nouns become feminine, so that “une helicoptere” would sound natural to many speakers. However, the impact of the standard yields some rather interesting complications in gender assignment: for example, in Montreal (where I grew up), even speakers for whom “autobus” is masculine would consistently refer to a bus number with the feminine article, i.e. “la dix-sept” for “bus number seventeen”: but when referring to a train or plane flight number one would always say “le dix-sept” in both instances (because “train” and “vol”, in spoken Canadian as in standard French, have both always been masculine). So variability in gender assignment needn’t mean that the system is dying: on the contrary, in this case it makes the system more complex (“autobus” is masculine as a noun but elliptically feminine, whereas “vol” and “train” are masculine nouns that are elliptically masculine as well).

  22. LH: “If that many native speakers disagree that substantially about that many words, there’s something going on.”
    Well, maybe. If teenagers in France are getting genders wrong at the rates in this study, though, they’d all be flunking out of school But my point, again, wasthat these results don’t prove anything, or give any strong answer to your question “Is French losing gender?” That’s because (a) the native speakers sample is only 56, (b) the sample of nouns is only 143, (c) the result was found in the control group, not in the group being studied. We have no indication how representative the 56 native speakers were of the French population (only 14 were adults, so obviously they were not; and it appears they may not have been living in France), or how representative a sample the 143 nouns are of the language. Were the nouns chosen at random? The best you could conclude from the information given is that maybe a larger, more random, study of French speakers in France is warranted.

  23. When I first heard the “helicopter” joke, the cast of characters included an unclear-on-the-concept English businessman calling attention to a fly in his soup, and a smug Parisian waiter correcting his usage. The punch line went something like, “By Jove, you’ve got bloody good eyesight.”

  24. Bulbul, words (and names) ending with an -a tend to be seen as feminine, but there is some hearsay that it’s not a universal rule.
    Sure. For example, in Slovak, Polish and Czech, there is an entire class of masculine nouns ending in -a – SK “hrdina” = “hero”, CZ “sluha” = “servant” and PL “kolega” = “colleague”. And let’s not forget suffixes, like Slovak/Polish/Serbo-Bosnian-Croatian -ca (SK sudca, radca; PL chodowca, wyborca; SBC izbeglica, ubojica) which is used to create masculine nouns (“judge, counsellor, breeder, voter, refugee, killer”, respectively). But all of these are animate nouns with female counterparts formed using suffixes – “hrdinka, služka, koleżanka” (note the assimilation in the latter two).
    The -a = feminine rule applies only to inanimate nouns, like the “helikoptéra” above.

  25. michael farris says:

    Does Slovak (or Czech) have the pattern of noun that takes masculine agreement (like adjectives and verbs) but what look like feminine case endings?
    for example dobry kolega (good colleague, friend)
    nom. dobry kolega
    acc. dobrego kolegę
    gen. dobrego kolegi
    dat. dobremu koledze
    ins. dobrym kolegą
    That never feels right to me somehow.

  26. michael,
    this is what it looks like in SK:
    N dobrý kolega / chlap
    G dobrého kolegu / chlapa
    D dobrému kolegovi / chlapovi
    A dobrého kolegu / chlapa
    L o dobrom kolegovi / chlapovi
    I s dobrým kolegom / chlapom
    As you can see from the declination of “chlap” = “man”, it’s all good sound male suffixes, especially -ovi. Basically you have here a bunch of originally masculine a-stems behaving like masculine o-stems.

  27. michael farris says:

    Okay, I like that pattern better than the Polish one (which always makes me think I’m making a mistake somehow) but is -u a normal (animate) masculine acc. gen. suffix? It looks so … non-animate somehow.

  28. Not even a Latin one.
    Correct. agricola ‘farmer’ is masculine.
    Here‘s a list.

  29. but is -u a normal (animate) masculine acc. gen.

    It only shows up in this class. The gen. suffix used to be -y, same as in Czech, but some time after the 15th century these two forms merged. The accusative suffix is left from the old a-stem declination, see žena – ženu. So it’s not non-animate, but rather non-masculine :)

  30. This seems normal to me. As words like professeur begin to take both le and la, it seems natural that many of the other words, sentenelle, and even victime could get `gender reassignment.’ As for words like oasis, I thought it was well known that an oasis could be masculine or feminine in French. And no mention of the words which are masculine when singular but feminine when plural? There is also some dialectical variation, as in Meridional French la serpent, la lievre, etc. I don’t think gender is going away in French, its just that, as always, French speakers don’t agree on which is which.

  31. Bruno van Wayenburg says:

    @Vanya, I’m not sure there’s much utility in our somewhat castrated gender system. You just have to say ‘de’ and ‘het’ according to the rules, doesn’t seem to mean much (except that most living beings and persons are probably ‘de’, however: ‘het meisje’: the girl, and ‘het wijf’: the woman (warning: not a nice word)).
    As a native speaker, it doesn’t feel useful or even logical, anyway, just something that sounds wrong when misapplied.
    People of Moroccan descent, even those born here, often get this wrong, giving mixing up genders a certain urban street credibility, along with the accent. So who knows how long it will hold out. As a copy editor, I don’t get to see many wrong articles, though.

  32. Did someone ask for a Latinist?
    1. Do French speakers really have to worry about the genders of trees (John Emerson, 2nd comment)? In Latin, all trees (and islands*) are feminine, though at least half of them have typically masculine endings, e.g. ‘quercus, fagus, pinus’ (“oak, beech, pine”) and all the islands whose Greek names end in -os and Latin names in -us (‘Samus, Lesbus, Chius, Ceus, Tenedus’, dozens more). Have some trees changed gender in French?
    2. MMcM’s linked list of Latin masculines ending in -a is useful, but mixed way too many very rare words with the common ones. I always tell my Latin I classes that the way to remember the commoner first-declension masculines ‘poeta, agricola, incola, nauta’ (“poet, farmer, inhabitant, sailor”) is that their initials spell PAIN.
    The list is also incomplete. The section on Greek words ending in -ES is missing ‘satelles’ (pl. ‘satellites’), “henchman, bodyguard, attendant, member of an entourage”, masculine because it referred to human beings before it was applied metaphorically to heavenly bodies. I assume Polish ‘satellita’ is masculine with a feminine ending because it comes straight from a Latin masculine word with a feminine ending. The same goes for Polish ‘kollega’, which must come straight from Latin ‘collega’, M, “colleague”.
    3. Most Latin nouns ending in -a belong to the first declension, have plurals ending in -ae (the only one I can think of that keeps its Latin plural in English is “alumna, alumnae”), and are feminine by default, masculine if they refer to male human beings, and never ever neuter. However, there is also a class of neuter 3rd-declension nouns ending in -a (plural -ata) taken from Greek, and some of these have turned masculine in (some? all?) modern Romance languages, following the rule that neuters turn masculine in preference to the rule that words ending in -a are feminine. Example: Spanish ‘el problema, el drama, el magma’. These do not keep their Greco-Roman plurals in English, but we do see the stem -T- in the adjectives ‘problematic, dramatic, magmatic’. Other such words in English are ‘schema’ (which I’ve only seen in DataBase Management jargon, though ‘schematic’ is standard English), ‘pragma’ (again usually only in the adjective ‘pragmatic’), ‘sigma’, and ‘dogma’ (but not ‘karma’). I don’t know the linguistic jargon, but these are nouns formed from verb stems and expressing the result of the action of the verb: ‘mazo’ (“knead”) –> ‘magma’ (“dough”), ‘drao’ (“do, act”) –> ‘drama’ (“deed, action”), ‘sizo’ (“hiss”) –> ‘sigma’ (“S”), and so on.
    - – - – - – - – - – - -
    *Well, all but one of the islands. Someone once wrote a whole article on the exception: J. R. Trevaskis, “A masculine island in Lucan II, 610-627″, PCPhS, 181, 1950-1951, pp. 15-16.

  33. vanya_6724 says:

    Just as the general term l’arbre became masculine, most common trees are also masculine in French – le pin (pine tree), le chêne (oak), l’érable (maple), l’orme (elm), le marronier (chestnut tree)- actually I can’t think of any French trees that are feminine, which kind of wrecks the joke if you think too hard.
    Same is true in Italian and Spanish, although Italian does still have quercia (oak). Did the -us ending get reinterpreted in vulgar latin as a masculine ending? Anyone know when latin arbor got its sex change?

  34. It sounds plausible that masculine-looking nouns would have tended to turn masculine, and vice versa, but, as I already noted, feminine-looking neuters such as ‘problema’, ‘magma’, and ‘drama’ have managed to resist going in the opposite direction over the last 2000 years of classical and vulgar Latin and now Castilian, and they are far less numerous in dictionaries and in the wild.
    Just to complicated things, many of the Latin tree-names that look masculine but are feminine have corresponding fruit-names that look and are neuter. So ‘pirus’ (pl. ‘piri’) is a (feminine) pear-tree, while ‘pirum’ (pl. ‘pira’) is a (neuter) pear. Neuters tended to turn masculine when the three Latin genders were reduced to the two modern Romance genders, so it’s conceivable that the genders of the fruit-names influenced the genders of the tree-names. Then again, most trees don’t bear fruit.
    The word ‘arbor’ is interesting, in that it also looks masculine and is the only feminine noun I can think of ending in -or. In fact, words ending in -or provide, as I recall, the only exception to the rule that abstract ideas are feminine. Unlike ‘avaritia’ (“greed”), ‘libertas’ (“freedom”), ‘virtus’ (“manliness, courage, virtue”) and dozens of other feminine abstract ideas in various declensions, ‘error’, ‘labor’, ‘stupor’, and similar words (most with obvious meanings) are masculine. (Also ‘amor’, but he’s a god as well as an abstract idea.) I’ve always assumed that they are masculine because the main class of nouns ending in -or is the very numerous (3rd-declension) class of male agent nouns, many of which have come into English unchanged in form (but not necessarily in meaning), e.g. ‘actor’, ‘assessor’, ‘auditor’, ‘cantor’, ‘confessor’, ‘doctor’, ‘monitor’, ‘visor’, ‘divisor’. Since these are masculine (the feminine equivalent is of course -trix, as in ‘aviatrix’ and ‘dominatrix’) they might well have pulled the other words ending in -or into the masculine gender — except for ‘arbor’, of course.

  35. French trees that are feminine
    arbres de genre féminin.
    when latin arbor got its sex change
    Pliny: inter duas arbores
    Inscription in Lyon: inter duos arbores
    That’s the earliest (per here).

  36. feminine noun I can think of ending in -or
    -ōris is okay, right?
    soror, uxor.
    -ŏris is harder.

  37. I love the erudite turn this thread has taken. Gratias ago.

  38. David Marjanović says:

    I live in Paris, and I think I’d have noticed… (But, no, I certainly don’t get out enough.)
    However, I correctly guessed that primeur was feminine before I finished reading the sentence. That’s because all abstract nouns in -eur (douleur, couleur…) seem to be feminine (in stark contrast to their Latin ancestors in -or, which were without exception masculine).

    A while back my brother was with two travelling French Canadians who were trying to figure out the gender of the name of a certain tree they’d never seen before.

    That kind of thing can happen in German, though it seems to be mostly limited to river names one doesn’t know, off the top of my head.

    In Yiddish nowadays everything has the same article: “duh.”

    In American Yiddish, that is, right?
    (The same has happened to Pennsylvania “Dutch”.)

    s’Egg vs. die Ecke, as Tunnoi vs. der Tunnel, da Schbiez vs. die Spitze

    All these are cases of gender reassignment surgery: apocope has removed the final -e, which was the feminine ending. This turned die Ecke into das Eck (ends in a definite [k] even in the dialects), die Spitze into der SpitzDas Tunnel was once standard, but keep in mind that this form is stressed on the last syllable. Der Tunnel is stressed on the first, and the last has a syllabic l instead of a vowel.
    I haven’t seen any signs of gender disappearing in German. Part of the reason probably is that gender, number, and case are all three indicated on the article and inseparable — marking one without the other would lead to confusion because there are fewer article endings available than gender-number-case combinations.

    masculine in Mauritius, feminine in Quebec I believe, just like plug, job and some other words used while speaking French…

    We get that a lot with English words in Germany vs Austria. Some even get more complex distributions.

    Which leads on to my main point, which is that I think that the German noun declensions are the most horrible mess I’ve ever seen in any language.

    Bingo. That’s why the trend for centuries now has been to move them to the article.

    Actually, in Quebec French there has long been a trend whereby vowel-initial nouns become feminine

    Initial? Wow. We’re really leaving Standard Average European here.

    even speakers for whom “autobus” is masculine would consistently refer to a bus number with the feminine article, i.e. “la dix-sept” for “bus number seventeen”:

    No, “line number seventeen”. For flights you don’t refer to lines but to flights…
    BTW, tramway lines are feminine in Linz (agreement with “line” or with “tramway”) but masculine 200 km to the east, in Vienna (the number per se is nominalized, and nominalized numbers are by default masculine). Bus lines are masculine in both places.
    (WTF? “Bingo” is “questionable content”?)

  39. All–
    1-The “sex change” of Latin ARBOR is not pan-Romance: Portuguese to this day still has feminine ARVORE, so that the Lyon inscription INTER DUOS ARBORES is also interesting as a sign of the fragmentation of Vulgar Latin.
    2-I haven’t the relevant reference tools with me, but I am almost one hundred per cent certain that Spanish PROBLEMA, MAGMA and DRAMA are learned loanwords (Late Middle Ages/Early Renaissance, I suspect) from Latin, and hence did not survive (gender and all) from Latin to Spanish for two thousand years (unlike Portuguese ARVORE).
    3-I am also almost one hundred percent certain that Latin POETA, AGRICOLA, INCOLA and NAUTA all failed to survive in any Romance language (Spanish POETA and similar forms in other Romance languages are definitely learned borrowings). Hence, before the Greek and Latin loans made their way into Spanish (and other Romance languages) the form/gender correlation must have been perfect in this instance: all nouns ending in -A were feminine, without exception. It fits in nicely with what I wrote above regarding the impact of learned influence, i.e. that it often introduced/maintains irregularities in gender assignment.
    4-While in the singular Latin neuters indeed typically merged with masculines, neuter plurals were very often re-analyzed as feminine singulars (because of the final -A): thus French FEUILLE, Spanish HOJA, Italian FOGLIA, and Portuguese FOLHA “leaf” all go back to Latin FOLIA, the plural of FOLIUM, treated as a feminine singular.
    5-However, in Italian and (especially) Romanian, the process was not completed: Romanian has nouns such as TIMP/TIMPURI, from TEMPUS/TEMPORA “time/times”: again because of the similarity between neuter plural and feminine singular endings, singular TIMP triggers masculine and plural TIMPURI feminine agreement (one of many features that makes Romanian, in this non-native’s opinion, the toughest of the Romance languages): such nouns are sometimes called “ambigeneric” in Romanian grammars.
    6-David: colloquial French, with its subject- and object-marking verbal prefixes (they’re not really clitics anymore), agreement-marking floating prefixes and prefixed or infixed pluralizer, strong preference for Verb-initial syntax, and the like, is already *quite* UN-Standard Average European-like.

  40. 4&5-neuter plurals … re-analyzed as feminine singulars
    Because ligna ‘firewood’ is semantically more than just the plural of lignum ‘wood’, you get pairs for it, one singular of each gender: Sp. leña / leño, Pt. lenha / lenho, It. legna / legno. I think (but could be mistaken) the feminine ‘firewood’ is much more common in Spanish and Portuguese and the masculine ‘wood’ is somewhat more common in Italian. I think Romanian has closer to the original state, like you say: lemn ‘wood’ with plural lemne ‘firewood’.

  41. (WTF? “Bingo” is “questionable content”?)
    Sorry about that; I’ve removed it from the blacklist now. What happens is I get a bunch of spam comments linking to “” or “” or whatever and I get fed up and say “screw it, who needs to talk about bingo in a LH comment?” and just add “bingo” tout court to the list. Then of course someone wants to use the word and gets the nanny note. Mea maxima culpa. Acedia is one of the seven deadly sins.
    * * *
    Hahahaha! When I hit “Post” I got:
    Your comment submission failed for the following reasons:
    Your comment could not be submitted due to questionable content:
    Please correct the error in the form below, then press Post to post your comment.
    I’ve replaced i with 1 so it will allow me to post that. I’m learning to think like a sp@mmer!

  42. David Marjanović says:

    [...]is already *quite* UN-Standard Average European-like.

    Good point — but I wonder if SAE is actually moving in the same direction. Below is “does [personal pronoun] have it/she?” and does [personal pronoun] have him?” conjugated in Standard German and in my dialect:
    habe ich es/sie /ˈhɒvis/
    hast du es/sie /ˈhɒstɐs/
    hat er es/sie /ˈhɒdɐs/
    hat sie/es es/sie /ˈhɒdsis/
    haben wir es/sie /ˈhɒmˑɐs/
    habt ihr es/sie /ˈhɒbtsis/
    haben sie es/sie /ˈhɒmsis/
    habe ich ihn /ˈhɒvin/
    hast du ihn /ˈhɒstn̩/
    hat er ihn /ˈhɒdɐn/
    hat sie/es ihn /ˈhɒdsn̩/
    haben wir ihn /ˈhɒmˑɐn/
    habt ihr ihn /ˈhɒbtsn̩/
    haben sie ihn /ˈhɒmsn̩/
    Isn’t that like a Samoyedic-style subjective-objective conjugation, except that it’s not agglutinating but more fusional like in Navajo…?
    For example, /ɐ/ occurs in total five times above: first as an epenthetic vowel, then as a pronoun, then as half a pronoun, then as a pronoun, and then as half a pronoun again.

    I’ve replaced i with 1 so it will allow me to post that.

    I just put <i></i> into forbidden words. Works every time and is completely invisible.

  43. Fuck.

  44. Socialism.

  45. Pottymouth!

  46. Etienne says:

    Interesting examples! One major difference between colloquial French and German is that subject pronoun-verb inversion is basically extinct in most (all?) varieties of colloquial French, so that “She has it” would be ELLE L’A, where the sequence ELLE L’ (like all clitic sequences!) is practically ALWAYS pre-verbal, and where nothing may separate the clitics from the verb; whereas (all varieties of?) spoken German allow inversion, so that “She has it” would be SIE HAT ES, versus interrogative HAT SIE ES? (Right?); I am inclined to think, therefore, that German pronouns are probably still best analyzed as clitics, and that French “clitics” are probably (because of their fixed pre-verbal position) best analyzed as prefixes.
    And I believe the first scholar who compared such phenomena in Western European languages to the morphology of an “exotic” language was a fellow countryman of yours, a linguist whose name I cannot recall: he compared European Portuguese clitic-doubling of definite objects and the objective conjugation of Hungarian (I *might* be able to dig up the name + reference if you or anybody on this thread asks me to do so).

  47. Dr. Weevil:
    There are many other English plurals in -ae, though most of them have competing English plurals in -as. However, algae, antennae (in biology; antennas in radio and television), larvae, and minutiae always appear in those forms in edited English prose, as far as I know. As for schema, it has three competing plurals: schemas, schemata, and (unfortunately) simply schema.

  48. David Marjanović says:

    whereas (all varieties of?) spoken German allow inversion, so that “She has it” would be SIE HAT ES, versus interrogative HAT SIE ES? (Right?);


    I am inclined to think, therefore, that German pronouns are probably still best analyzed as clitics, and that French “clitics” are probably (because of their fixed pre-verbal position) best analyzed as prefixes.

    Point taken.
    Thanks for the offer, but I don’t think I could do anything with a reference, at least not anytime soon, unless it happens to be on Google Books. I do not have JSTOR access.

  49. Siganus Sutor says:

    Étienne: in Montreal (where I grew up), even speakers for whom “autobus” is masculine would consistently refer to a bus number with the feminine article
    Nothing so wrong there if one considers that some serious dictionaries say that the similar automobile could be either feminine or masculine:

    Remarque. Nouveau Larousse illustré, Académie 1932, ROB., Lar. encyclop., DUB. font ce mot féminin, toutefois, Nouveau Larousse illustré, Larousse encyclopédique et Larousse de la Langue française signalent que ce mot (abrégé ou non) s’est également employé au masculin (cf. ex. 4, 7). QUILLET 1965 donne les deux genres mais avec une préférence pour le féminin.

  50. marie-lucie says:

    When I was a child (in Western France), we sometimes ate licorice, something my parents did not buy but that was shared among the neighborhood children. All the children around us called it du réglisse (masculine), but my mother insisted it should be de la réglisse (feminine). Decades later, I am still self-conscious when I have to use the word – the feminine does not sound right to me.
    About vowel-initial words becoming feminine:
    At one time in the evolution of French (about the 1500′s), some phonetic changes occurred which (among other things) resulted in the current differenciation between nasal vowels (as in bon ‘good (m)’) and a sequence of non-nasal vowel + nasal consonant (as in bonne ‘good (f)’). However, the phonetic change in itself had nothing to do with a need to preserve the distinction between masculine and feminine forms of the same word but everything to do with whether the nasal consonant was followed by a vowel or not, regardless of context. As a result, a word like bon is pronounced with a nasal vowel when followed by a consonant, as in un bon garçon ‘a good boy’ but exactly like the feminine bonne if followed by a vowel, as in un bon ami ‘a good (male) friend’, un bonhomme (not ‘a good man’ but more something like ‘a little old man’ or as ‘man’ in ‘a snowman’). Similarly the word un ‘one (m)’ was pronounced just like une ‘one (f)’ in front of a word beginning with a vowel, as in un homme ‘a/one man’. This pronunciation is still used in some varieties of French, at least in rural Western France and Québec. In standard French the resulting gender ambiguity was solved by using the nasal vowel together with the nasal consonant at least in the case of un and mon ‘my’ (which is used for both masculine and feminine nouns beginning with a vowel).
    With the old pronunciation it is not surprising that vowel-initial nouns which are not extremely common should cause some hesitation as to their gender, and sometimes a total switch.
    How do francophone children learn the gender of a noun? by hearing it in context, where the clues are in part in the noun itself but mostly in the other words (articles, adjectives, possessives, etc) which are used with the noun. With words beginning with a vowel, the prefixed clues can be misleading, as the definite article is not le or la but simply l, and depending on the spoken variety the indefinite article un may be identical in the masculine and feminine forms. The plural does not help as articles and possessives are the same for both genders. Many adjectives also sound identical when used before a vowel (eg un petit arbre ‘a little tree’, where petit sounds the same as petite in une petite fille ‘a little girl’). So the only clue for the gender of a vowel-initial noun is a singular adjective occurring after a noun, but the most common adjectives occur before the noun, as in the examples above, and many of the after-noun adjectives have the same form regardless of gender, as in admirable. Very common nouns are likely to be used in many contexts, so that at some of these contexts will give clues to their gender, but less common ones might not. No wonder some people wonder if a word like hélicoptère is masculine or feminine.

Speak Your Mind