Consistency of word gender between Romance languages

[yong321]

We know most nouns or adjectives in Spanish are easy in terms of identifying their gender, but not so in French. If a person knows Spanish, what is the approximate success rate in guessing the gender of this word said in French? For example, he knows Spanish clave is feminine so he guesses French clé is also feminine. If he has significantly higher than 50% of chance to guess it right, this rule will be useful. Is there such a study? If not, it won't take us long to check the first 1000 or so most frequently used words.

When someone posted a message about this on Facebook, one comment follows: "The exceptions to this rule are mostly words that have neuter gender in Latin, such as "le lait" - "la leche"." I think that's quite sensible. I also want to add a trivial comment that the two words in the two languages (e.g. Spanish and French) must be cognates for this rule to work.

Any comment is welcome.

[reineke]

Chung mentioned a study. I (almost) emailed the author for a copy.

While there's a tendency for Romance cognates of a given Latin noun to have the same grammatical gender, it does get screwed up by what happened to the neuter nouns of Classical Latin, in addition to analogy and reanalysis / reinterpretation of a noun's grammatical gender on the strength of its ending in a given case. The reassignment of grammatical gender (especially the neuter nouns) was already there in Vulgar Latin with the plebes distinguishing just masculine and feminine.

I still stand by what I posted on this topic on HTLAL and hold that it's not that great of an idea to generalize grammatical gender assignment on the basis of using cognates from just one or two Romance languages. I guess that I'm a bit of a spoil-sport on the subject.*

The reinterpretation of Latin gender in the modern Romance languages doesn't seem to follow a recognizable pattern. Compare the terms for "milk" and "sea" to see how neuter Latin words are represented in French, Italian, and Spanish.

There is a recognizable pattern: words that were feminine in Latin are generally feminine in the modern Romance languages, words that were neuter or masculin are masculine in today's Romance languages.

The exceptions you cited are exactly that: exceptions. For the vast majority of nouns the above rule is valid.


PS: This is especially true for large classes of nouns sharing certain cognate endings: -tion, -cao, -cion, -zione almost always indicate feminine nouns.

Somehow I'm not comfortable with writing off these differences as just exceptions. It seems that many of the modern languages had differing handling for cognates that were originally neuter in Classical Latin (it reminds me somewhat of how declension of numerals in most Slavonic languages numerals got "screwed up" when the dual faded away).

Here are a couple of sources which show that a simple generalization of cognates' gender to all Romance languages using the example of just one or two Romance languages is far less useful than one would think.

http://dspace.nitle.org/handle/10090/8323 (scroll down the screen to click on a .pdf link that is called "s10inde2008washington.pdf" in the box "Files in this item")

The study's author finds that gender incongruency for cognates in the modern Romance languages is affected by idiosyncratic developments (e.g. analogies, phonological changes to old endings) in each of the descendants of Vulgar Latin. In addition she finds that modern cognates originating from the third declension in Latin (which also was the largest class of nouns) had their genders reinterpreted the most among the modern Romance languages. It has partially something to do with how the endings were starting to be mixed up even in Vulgar Latin which had already reduced the three genders of Classical Latin into two. It is illuminating how the study explains why Romance cognates for "flower", "colour", "milk", "sea" and many others appear with genders that vary from one language to the next (the analysis of the inconsistencies in cognates from the third declension begins on Chapter 4 or p. 38 of the study). Pages 52-55 have tables showing the inconsistency of gender assignment in commonly-used Romance cognates. The largest list of irregularities comes from the cognates of the 3rd declension in Latin.

http://accurapid.com/journal/42gender.htm

This is a translator's "cheat sheet" for managing to keep track of gender for cognates in Romance. It shows example cognates from Portuguese, Spanish, French, Italian and Latin (sorry, no Romanian).

The "femininity" of nouns ending in -tion, -cao, -cion, -zione is well-known, and this trait has also passed on into German with loanwords ending with the suffix also being feminine. However, many nouns in Romance do not come with such a sure-fire marker of grammatical gender like -tion/-cao/-cion/-zione, and it's not as if a noticeable part of every-day lexicon is dominated by words ending in this way.

I have found that a better approach for learning gender in Romance languages is to try first to detect patterns applicable to the Romance language in question, rather than try to "cheat" by applying gender assignment in the related languages. When I was taking Latin in high school, I tried to cheat a bit by assigning French genders to Latin words if I couldn't remember the gender in the Latin ancestor. Now, if I could have got a dime for every half-point lost on those Latin tests because I was assigning the wrong gender to Latin words by relying on my background in French... Learning Romanian was a similar story for me. I pretty much had to learn genders one-by-one for most new words because of all of the development and external influence since it evolved from Vulgar Latin, with genders of cognates not usually matching what I was used to in French or from what I could remember in Latin.

See here for the whole thread on HTLAL.

Since I posted the preceding in 2010, I've dabbled in Spanish and haven't had reason to change my original conclusion. However, I did feel (rather subjective, I guess) that predicting grammatical gender assignment of Spanish nouns on a given noun's form alone was slightly easier than in French. However it was nowhere near as consistent as what I've experienced with Slavonic languages where grammatical gender there is strongly predictable based on knowing the noun in its citation form (i.e. nominative singular).

The first link which holds the study on incongruent grammatical gender assignment in Romance languages is now here at the archive of Middlebury College.

The second link which holds the translator's "cheat sheet" is now here. This is probably a bit of what luke was looking for originally when it comes to a chart comparing nouns in the familiar Romance languages. Romanian and Catalan aren't in the picture.

There's also this explanation by a linguist which shows how incongruencies in gender assignment across Romance began with the plebes' reanalysis and reinterpretation of Classical Latin forms.

* Out of fairness, I do note that Teschner's study "The Genders of French and Spanish Noun Cognates: Some Statistics and a List." published in 1987 compares cognate nouns in Spanish and French from a corpus of 15,000 words. According to Hannah B. Washington, the author that study in Middlebury College's archive, Teschner found 14,966 noun cognates
between French and Spanish, of which 13,825 have the same grammatical gender. In other words, the grammatical gender of 92.4% of the cognates is the same in French and Spanish. It's a shame that I can't find Teschner's study anywhere since I'm interested in the 7.6% of cognates that don't share grammatical gender. I'm especially interested in the etymology as well as the frequency of each divergent pair in question.

https://forum.language-learners.org/vie ... =14&t=5273

[reineke]

Re: "easy"

Granularity and the acquisition of grammatical gender: How order-of-acquisition affects what gets learned

"Why do adult language learners typically never achieve the same level of language mastery as native speakers? We examine how prior knowledge and experience might influence the size of the linguistic units employed in learning, and as a consequence, what gets learned. We hypothesize that adult learners tend to learn from more segmented representations, and that these can hinder learning about the relations between segments (e.g. between articles and nouns). In particular, we focus on the acquisition of grammatical gender, an aspect of language later learners show difficulty with compared to native speakers.

Grammatical gender provides a good test case for studying differences between L1 and L2 learning..."

"Native and non-native speakers show different patterns of learning grammatical gender. Children master grammatical gender relatively early (see Slobin, 1985 for cross-linguistic reports), and make few mistakes in spontaneous speech. In contrast, L2 learners have
persistent difficulty with grammatical gender even after extensive exposure (Scherag et al., 2004). Native and non-native speakers also differ in their ability to use the gender information conveyed by the article in real time processing. Native speakers (adults and children) can use this information to guide lexical access; they anticipate a feminine noun following a feminine article (Lew-Williams & Fernald, 2007) and slow down if there is a gender mismatch between the article and the noun (Grosjean et al., 1994; Dahan et al., 2000)."

" Non-native speakers do not show these effects (Guillelemon & Grosjean, 2001; Scherag et al., 2004). These findings suggest that native speakers treat the article and the noun as a more cohesive unit than do non-native speakers; this allows them to select the correct article in production, and use it to facilitate recognition in comprehension..."

"Several suggestions have been advanced to explain these different patterns of grammatical gender learning, including: that grammar cannot be fully mastered in adulthood (Clahsen & Muyksen, 1986), or at least not aspects of it that are not found in the learner’s native
tongue (Hawkins & Chan, 1997); and that L2 learners form more shallow grammatical representations that are hard to access in real time (Clahsen & Felser, 2006). These accounts describe the difficulty L2 learners have with grammatical gender, but they do not fully explain why this difficulty arises..."

"While none of children’s early language input is written, adults are likely to learn from written input in which the distinction between the article and the noun is explicit and visually salient. Finally, there is evidence that while adults can use cognitive control to selectively attend to particular aspects of the input, children may largely lack this facility (Ramscar & Gitcho, 2007). In other words, adults not only know that articles and nouns are separate, but they can also ‘choose’ to focus their attention on one or the other."

https://www.google.com/url?sa=t&source= ... 1UWCNHkJ6L

[yong321]

Thank you for posting the January's discussion. I apologize for not having read that before. The discussion is quite informative. The only thing missing is the statistics, which is very likely in Dr. Teschner's article "The Genders of French and Spanish Noun Cognates". I wish at least the statistics part was made public.

I suspect that as the corpus gets smaller, the consistency / congruency between Spanish and French word gender decreases, because the 15000 words Teschener studied must have included many compound or derived words, which drives the consistency rate up to 92.4%. But even the most frequently 5000 or 1000 words, I suppose, should have much greater than 50% of the consistency. And that's helpful to a student. In spite of that, the student of course has no reason to not study and try to be lucky. That's not my intention of posting this message.

[tastyonions]

Makes me wonder whether speakers of unrelated (but nevertheless gendered) languages would master Romance gender better than native speakers of non-gendered languages, e.g. thinking not of Italian -> Spanish or French -> Portuguese but maybe Arabic -> Romance language.

Interestingly, I have witnessed native speakers of one Romance language getting genders wrong in another Romance language they are learning in cases where the word is "shared" between the two languages and has the same gender in both.

[Querneus]

Since Spanish and French were explicitly mentioned, this looks like a good time to post the notes on differences in terms of noun gender between French and Spanish that I've been expanding, intended for French speakers learning Spanish (although Spanish speakers learning French could use it too).

Mots de genre différent en espagnol

Première édition publiée ici (LLorg) : 2017-10-23
Édition actuelle : 2023-11-26.

Quoique les grammaires et les livres pour apprendre l'espagnol remarquent souvent qu'il existe des mots de genre différent, je n'ai pas eu de la bonne chance en cherchant une liste détaillée de tels éléments lexicaux. J'ai fait l'effort de ramasser autant de ces paires de mots que j'ai pu, bien qu'une liste exhaustive est impossible de faire : le français familier a beaucoup de termes de genre incertain, spécialement des mots qui commencent par une voyelle. Par exemple, « avion », lequel est masculin en français standard, peut être féminin dans la langue familière (« une avion »). Cet article contient seulement des mots français selon son genre standard. Les mots qui finissent par « ‑eur » en français et ‑or en espagnol se trouvent au-dessous du reste dans son propre groupe.


Sommaire
1 Cas individuels
2 Mention d'honneur
3 Mots qui finissent par « ‑eur »
4 Catégories de mots


1 Cas individuels

agenda (m.) - agenda (f.)
analyse (f.) - análisis (m.)
annonce (f.) - anuncio (m.)
asperge (f.) - espárrago (m.)
attaque (f.) - ataque (m.)
barbecue (m.) - barbacoa (f.)
calme (m.) - calma (f.)
coma (m.) - coma (m. ou f.)
(Note: coma est masculin s'il veut dire « un coma », féminin s'il veut dire « virgule ». Les deux significations ont de différentes étymologies.)
comète (f.) - cometa (m. ou f.)
(Note: cometa est masculin s'il veut dire « une comète », féminin s'il veut dire « cerf-volant ».)
cour (f.) - corte (m. ou f.)
(Note: corte est masculin s'il veut dire « une coupe », féminin s'il veut dire « la cour ». Les deux significations ont de différentes étymologies.)
dent (f.) - diente (m.)
disquaire (m.) - disquera (f.) (compagnie qui vend des disques)
doute (m.) - duda (f.)
enclume (f.) - yunque (m.)
énigme (f.) - enigma (m.)
étude (f.) - estudio (m.)
équipe (f.) - equipo (m.)
fin (f.) - fin (m.)
front (m.) - frente (m. ou f.)
(Note: frente est féminin s'il veut dire « visage », masculin s'il veut dire « front de guerre ».)
fruit (m.) - fruta (f.)
guide (m.) - guía (m. ou f.)
(Note: guía est féminin s'il s'agit d'un livre, d'indications ou d'un outil pour guider, mais s'il s'agit d'une personne le mot est masculin ou féminin selon le genre.)
horloge (f.) - reloj (m.)
lait (m.) - leche (f.)
légume (m.) - legumbre (f.)
lièvre (m.) - liebre (f.)
limite (f.) - límite (m.)
marge (f.) - margen (m. ou f.)
(Note: margen est masculin s'il veut dire « marge », mais féminin s'il veut dire « rive », c'est-à-dire le côté d'une rivière ou d'un fleuve.)
massacre (m.) - masacre (f.)
méthode (f.) - método (m.)
mer (f.) - mar (m.)
(Note: en espagnol poétique mar est souvent féminin.)
minute (f.) - minuto (m.)
miel (m.) - miel (f.)
moratoire (m.) - moratoria
oasis (f.) - oasis (m.)
ordre (m.) - orden (m. ou f.)
(Note: orden est masculin s'il veut dire le contraire de désordre, féminin s'il veut dire « commandement » ou « ordre réligieux ».)
origine (f.) - origen (m.)
orque (m.) - orca (f.)
paire (f.) - par (m.)
paradoxe (m.) - paradoja (f.)
pédale (f.) - pedal (m.)
pelvis (m.) - pelvis (f.)
peluche (f.) - peluche (m.)
période (f.) - período/periodo (m.)
phoque (m.) - foca (f.)
planète (f.) - planeta (m.)
police (f.) - policía (m. ou f.)
(Note: policía est féminin s'il veut dire l'institution de la police, mais s'il s'agit d'une personne (un policier ou une policière/femme-policier) le mot est masculin ou féminin selon le genre: el policía, la policía.)
rat (m.) - rata (f.)
rencontre (f.) - encuentro (m.)
sang (m.) - sangre (f.)
sel (m.) - sal (f.)
seconde (f.) - segundo (m.)
serpent (m.) - serpiente (f.)
sourire (m.) - sonrisa (f.)
tarif (m.) - tarifa (f.)
tomate (f.) - tomate (m.)
touche (f.) - toque (m.)
vidéo (f.) - video/vídeo (m.)
visa (m.) - visa (f.)
vodka (f.) - vodka (m. ou f.)
(Note: vodka est presque toujours masculin, rarement féminin.)
yucca (m.) - yuca (f.)

2 Mention d'honneur

En particulier, l'espagnol cura est masculin s'il veut dire le curé, mais féminin s'il veut dire la guérison. Notez aussi le mot capital, lequel peut être masculin s'il veut dire « le capital » (de l'argent) ou féminin s'il veut dire « la capitale » (une ville principale). On peut mentionner également choléra/cólera (m.) et colère/cólera (f.).

augmentation (f.) - aumento (m.)
bague (f.) - anillo (m.)
bijou (m.) - joya (f.)
carte (f.) - mapa (m.)
cigarette (f.) - cigarro (m.)
coffre-fort (m.) - caja fuerte (f.)
couple (m.) - pareja (f.)
dos (m.) - espalda (f.)
épaule (f.) - hombro (m.)
exploit (m.) - hazaña (f.), proeza (f.)
lit (m.) - cama (f.)
matin (m.) - mañana (f.)
nuage (m.) - nube (f.)
nez (m.) - nariz (f.)
soupçon (m.) - sospecha (f.)
souris (f.) - ratón (m.)
voiture (f.) - coche (m.)

3 Mots qui finissent par « ‑eur »

Excepté fleur/flor (les deux féminins) et les mots ci-dessous, les mots qui finissent par ‑eur sont masculins en français et en espagnol. Exemples : congélateur/congelador, extérieur/exterior, facteur/factor, honneur/honor, intérieur/interior, ordinateur/ordenador, projecteur/proyector, réflecteur/reflector, rupteur/interruptor et secteur/sector, tous masculins.

« Aspirateur » est un cas spécial, puisqu'il se peut traduire par aspirador (m.) ou aspiradora (f.), dépendant de la variété de l'espagnol utilisée. On peut mentionner aussi dans cet espace la paire amour/amor, les deux masculins. « Largeur » n'a pas de cognat en espagnol mais se traduit par ancho (m.) ou anchura (f.), de forme similaire à « longeur » qui se traduit par largo (m.) ou longitud (f.).

Féminin en français - Masculin en espagnol

chaleur (f.) - calor (m.)
couleur (f.) - color (m.)
douleur (f.) - dolor (m.)
erreur (f.) - error (m.)
faveur (f.) - favor (m.)
fraîcheur (f.) - frescor (m.)
fureur (f.) - furor (m.)
horreur (f.) - horror (m.)
liqueur (f.) - licor (m.)
odeur (f.) - olor (m.)
rigueur (f.) - rigor (m.)
saveur (f.) - sabor (m.)
sueur (f.) - sudor (m.)
vapeur (f.) - vapor (m.)

Masculin en français - Féminin en espagnol

chou-fleur (m.) - coliflor (f.)
labeur (m.) - labor (f.)

4 Catégories de mots

Il y a aussi des catégories entières de mots de genre différent entre le français et l'espagnol. Ce sont notamment les lettres de l'alphabet, les rivières et les pays qui finissent par ‑a. Les noms des professions présentent plusieurs difficultés.

As you can see, I still have to write a detailed discussion in part 4... Basically, alphabet letters are masculine in French but feminine in Spanish (le a - la a, le bé - la be, le ache - la hache, le i grec - la ye...). In Spanish, most if not all rivers are masculine (el Sena, el Mosela, el Amazonas, el Mekong, el Nilo...), but in French, there is a distinction between rivières (which are feminine and tend to be narrower, e.g. la Seine, la Moselle) and fleuves (which are masculine and tend to be wider, e.g. le Mékong, le Nile, le majestueux Amazone). In both French and Spanish, the names of countries are gendered, but in French the ones that end in -a tend to be masculine (un meilleur Uganda, le Venezuela). Gender marking in the names of professions is complex and controversial, due to reasons of tradition and modern political issues.

EDIT: 2019-08-26: fixed a mistake about marge in French. Then fixed the dates at the top.
EDIT: 2021-10-16: fixed a mistake about faim and *savor...
EDIT: 2022-09-05: add pelvis
EDIT: 2022-11-29: add disquaire
EDIT: 2023-05-11: add oasis and pédale
EDIT: 2023-11-26: add énigme, rencontre, visa, vodka, yucca

[Via Diva]

Can't speak for a Romance language, but when it comes to Russian/German, there are just no parallels except for the already mentioned -tion and the like nouns, that are feminine. There are some common rules, like, generally the nouns that end with -e are feminine in German, much like generally the nouns ending with -a are feminine in Russian, etc, but one just can't always rely on this: мама is feminine, but папа isn't. More than that, in German diminutives become neutral, hence the famous die Magd - das Mädchen, in Russian
diminutives get to keep their gender intact, the very same дева - девочкa, for example.

I guess if the source languages are very different you just can't rely on anything from one language while trying to learn another.

But since I'm studying Czech again now I wonder whether the same studies on grammatical gender in Slavic languages exist xD So far all the words I can think of match if they sound simular. And if they don't... час is masculine, hodina is feminine, and that's enough already.

Edit: Corrected. That's what working with no days (well ok 2 out of 24...) off does to you, lol.

[Josquin]

More than that, in German diminutives become neutral, hence the famous die Made - das Mädchen

Thanks for giving me quite a good laugh!

"Made" means "maggot" in English. "Mädchen" is derived from "Magd" ("maid"). The g fell silent and wasn't even written any more. We don't call little girls "maggots" in German!