Show Summary Details

Page of

Printed from Oxford Research Encyclopedias, Linguistics. Under the terms of the licence agreement, an individual user may print out a single article for personal use (for details see Privacy Policy and Legal Notice).

date: 29 June 2022

Sex-Denoting Patterns of Word Formation in the Romance Languagesfree

Sex-Denoting Patterns of Word Formation in the Romance Languagesfree

  • Franz RainerFranz RainerInstitute for Romance Languages, Vienna University of Economics and Business


Since sex distinctions are a basic fact of nature and society, any natural language must make available means to refer separately to males and females of humans as well as animals, to the extent that sex is salient or relevant with reference to animals. In each language, these means comprise a peculiar mix of the patterns used in enriching the lexicon, ranging from syntax to compounding, affixation, conversion, and sometimes devices that are even more exotic. In a minority of the languages of the world, such as Latin and its daughters, the distinction between the sexes has even been built into the grammar in the form of gender systems whose rules of gender assignment rely heavily on it for animate nouns. In these languages, the gender system itself can also be put to use in the creation of designations for males and females.

As is well known, the Indo-European gender system in origin reflected the animate/inanimate distinction, while the classification of animates along the feminine/masculine axis was a later development whose gradual expansion can still be observed in Latin and Romance. The demise, in spoken Latin, of one central pillar of feminization, namely, the suffix -trix, as well as other disruptive factors such as sound change and language contact, brought instability into the system. Each Latin and later Romance variety therefore had to adapt its system in order to cope with communicative needs concerning the expression of the male/female distinction. Different varieties did so in different ways, creating a large array of systems of sex-denoting patterns. In principle, it would be desirable to deal with each variety’s system on its own terms, describing as exactly as possible the domain of each pattern at the different stages of development as well as the mutual relationships among competing patterns and the mechanisms behind the changes. However, such an approach is unrealistic in the absence of detailed descriptions for many varieties, most notably the dialects.


  • Historical Linguistics
  • Language Families/Areas/Contact

1. Terminological Preliminaries

The area of word formation covered in this article has no agreed-upon name in the English literature. In (neo-)Latin grammar, the derivation of designations for females from those for males, or more rarely the other way round, was called motio substantivorum, a term that has been adopted, among others, by German, Romanian, Italian, and Spanish scholars (Ger. Movierung, Motion; Ro. moţiune; It. mozione, cf. Sgroi, 2007; Sp. moción), but not by their English and French counterparts.1 For the male > female direction, feminization is widely used, in English as well as in the other European languages (Ger. Feminisierung, Femininbildung; Fr. féminisation, etc.), less so masculinization for its opposite. What is missing in English is an established cover term for both processes. Doleschal (2015) opted for gender marking, but the awkward polysemy of the term gender might easily lead one astray. For this reason, preference has been given to the term sex-denoting patterns in this article. One additional reason for this terminological choice is that some sex-denoting suffixes do not operate on a base noun referring to the opposite sex, but on a non-personal base (cf. meretrix in section 2, for a case in point).

Most of the time the patterns treated here simply denote the sex of the person or animal2 designated by the base; by sense extension, however, they can also refer to the wife of the male person (or animal, in fables) designated by the base (e.g., Latin fratria ‘brother’s wife’, from frater ‘brother’). This secondary meaning featured quite prominently in former times but has receded into the background since the 20th century in the wake of women’s emancipation. In some Romance varieties, a family name with a feminizing suffix attached to it refers to a woman (wife, daughter) of that family: rural French ThibaudThibaude (Spitzer, 1923, p. 648), Occ. CravighacCravinhaga (Sibille, 2015, p. 59), Romanian IonescuIoneasca, DeleanuDeleanca, MironMironoaia, LonghinLonghineasa, PapadopolPapadopolina (stressed on -do-), and so forth (Ciobanu, 1959). As far as I know, Fassano (Central Ladin; Elwert, 1943, p. 184) is the only dialect to distinguish formally female workers from the wives of workers. The former receives the suffix -áa (< Lat. -aria; e.g. uráa ‘female day labourer’), the latter -éa (-a attached to -é < -ariu; e.g., molinéa ‘miller’s wife’).

2. The Starting Point: Sex-Denoting Patterns in Latin

Feminine marking with -a on nouns was less developed in old Latin than at later stages of the language.3 Lupus ‘wolf’, for example, was still of common gender: one could say hic ‘this-m’ and haec ‘this-flupus. In order to refer to a she-wolf, Ennius resorted to the appositional structure lupus femina ‘lit. wolf female’ instead of lupa, and Plautus has civis femina ‘female citizen’. Over time, however, the association between -a and feminine gender became closer: it replaced -us on designations for females (e.g., nura, for nurus ‘daughter-in-law’), and was even added to consonant stems (e.g., clienta ‘female client’ for cliens, hospita ‘female guest’ for hospes; cf. Löfstedt, 1963, pp. 61–62). Occasional backformations (Brender, 1920, pp. 59–60) based on designations for females (e.g., concubinus ‘catamite’, from concubina ‘concubine’) or epicene nouns in -a (e.g., columbus ‘he-pigeon’, from epicene columba) also testify to the status of this inflectional ending as a sex-denoting marker in appropriate contexts. Apart from this secondary use of the gender system for the purpose of word formation, Latin also had a series of dedicated derivational suffixes.

The most prominent one was -trix, which could either directly follow a verb (e.g., meretrix ‘courtesan’, from merere ‘to earn (money)’) or replace masculine -tor (e.g., genitor ‘father’ / genitrix ‘mother’, related to gignere ‘to create’).4 Although this suffix is frequent even in Late Latin texts, its scant representation in the oldest stages of Romance indicates that it must have receded in use in spoken Latin. According to Olcott (1898, p. 118), in inscriptions “2/3 of the forms are found only in Italy (with Rome)”.

The suffix -ia is of Indo-European descent, but was rare: auia ‘grandmother’ (from auus ‘grandfather’), fratria ‘brother’s wife’, nepotia / neptia ‘granddaughter’ (in inscriptions, Olcott, 1898, p. xiv), Late Latin ceruia ‘hind’ (REW, nº 1844; from ceruus ‘deer’). The Romance languages (Rainer, 2019, pp. 128–130) allow reconstruction of a few more of them, for example, *cania ‘bitch’ (REW, 1584a; from canis ‘dog’). In the case of tutoria ‘female guardian’, attested in a Roman inscription of the 2nd century (Olcott, 1898, p. 193), one might hesitate as to whether to analyze it as composed of tutor and the suffix -ia or as a nominalized adjective in -oria. The suffix -ria of pairs like citharista ‘male lyre player’ / citharistria ‘female lyre player’, which entered Latin in loans from Greek (André, 1971, pp. 91–92), is etymologically related to this suffix.

Even rarer were designations of females with the suffix -ina: gallina ‘hen’ (from gallus ‘cock’), regina ‘queen’ (from rex ‘king’), heroina ‘heroine’ (from heros ‘hero’), concubina ‘concubine’ (related to concumbere ‘to lie together’).

At the end of the 3rd century, Christian Latin borrowed the suffix -issa from Greek, but it remained relatively rare until the 6th century (André, 1971, p. 110): abbatissa ‘abbess’, diaconissa ‘deaconess’, prophetissa ‘prophetess’, and so forth.5 From the 6th century onwards, this suffix also spread to lexical domains outside the religious sphere, especially civil ranks (Stotz, 2000, §37): fratrissa ‘sister-in-law’ (Isidor of Seville), maiorissa ‘female administrator’ (Lex Salica), ducissa ‘duchess’ (8th century), imperatissa ‘empress’ (8th century), comitissa ‘countess’ (9th century), senatorissa ‘senator’s wife’ (9th century), and so forth.6

Occasionally, Latin also resorted to diminutive suffixes to express female sex: Drusus / Drusilla (names), adulescens ‘young man’ / adulescentula ‘young woman’, sacerdos ‘priest’ / sacerdotula ‘priestess’, and so forth.

3. Common Gender and Epicene Nouns

In gender languages like Latin and its daughters (Loporcaro, 2018), nouns of common gender are nouns that, without change in form, can refer to both males and females and control feminine and masculine agreement accordingly (e.g., Fr. le/la journaliste ‘them/f journalist’, European Pt. o/a polícia ‘them/f policeman/woman’, Pt. o/a chanceler ‘them/f chanceler’). This capacity to control both target genders distinguishes them from epicenes, which can also refer to males and females alike but only have one gender (Fr. la/**le victime ‘the victim’).7 We have already seen in section 2 that nouns of common gender were more frequent in old Latin than at later stages of the language, when several such nouns received an explicit mark signaling feminine gender (lupa instead of lupus, clienta instead of cliens, etc.). This trend continued in Romance, as will become apparent in section 5. However, there were also cases where nouns that once had separate forms for referring to females became of common gender.

One case in point are words ending in the suffixes -ista / -istria, which have become of common gender in most Romance languages: Lat. sophista (or sophistes) ‘sophist’ / sophistria versus Fr. le/la sophiste, Sp. el/la sofista, and so forth. In Sursilvan, though, where -ista appears as -ist, the suffix can be feminized by adding -a (e.g., emprendist/a ‘male/female apprentice’). In the western varieties of Catalan, masculine -iste contrasts with feminine -ista (Casanova, 1984). In Italian, males and females remain formally distinct in the plural (i sofisti vs. le sofiste), and in colloquial varieties of Spanish new masculine forms in -o have occasionally been created via backformation (e.g., modisto ‘male fashion designer’ vs. modista ‘dress maker’, Am. Sp. pianisto ‘male pianist’).8 Similar formations also abound in texts that parody “non-sexist” language use, both in Spanish and Italian.

Another important Latin suffix that became of common gender in at least some early Romance varieties is agentive -tor. After the demise of -trix in the spoken varieties of many parts of the Latin-speaking world, the use of the widowed suffix -tor for both sexes and genders was one way of filling the lacuna. Old Pt. pastor ‘shepherd’, according to Connors (1970–1971, pp. 584–585), was of common gender, while present-day Portuguese distinguishes pastor and pastora. Nouns of common gender of this kind persisted until the 16th century, adjectives even until the 17th century (Frias e Gouveia, 2005, pp. 529–530). In a Spanish document of 1253 we read “la cantor mayor” (Corpus Diacrónico del Español [CORDE]), with cantor ‘singer’ used instead of later cantora. Innovative nouns in -dora, at that time, were already gaining in frequency, but would not become definitively established in Spanish until the 15th century (Pattison, 1975, pp. 110–111; England, 1984, 1987).

A third novel kind of nouns of common gender is of more recent vintage. When, around 1900, the first women obtained university degrees and entered prestigious professions (medicine, pharmacy, law, architecture, etc.), many women preferred using the masculine title or name of the profession because they felt that the corresponding feminine variants sounded less distinguished.9 As neologisms, a jocular or even derogatory ring could easily accompany them, at least during the incipient stage of use. Emilia Pardo Bazán, for example, the first woman to receive a chair at a Spanish university in 1916, did not want to be called catedrática ‘female professor’ and used to write after her signature, Catedrático de Literatura contemporánea en la Universidad de Madrid. In that way, a new set of nouns of common gender like el/la catedrático entered the Romance languages. This innovation was frowned upon by Spanish grammarians (Aliaga Jiménez, 2011), who disliked it because it infringed the robust correlation feminine gender ≈ marker -a versus masculine gender ≈ marker -o of Spanish. Alternatively, such nouns also occur as epicenes in expressions like el catedrático de literatura, doña Emilia Pardo Bazán, used particularly in contexts that are more formal or where the speaker deems function more relevant than sex (Sgroi, 2018). In French and Italian, this latter usage became part of the norm (Fr. Madame le ministre, It. la signora ministro), a norm that has been challenged from the end of the 20th century onward. In France, Ségolène Royal was the first woman who, in 1992, insisted on being called Madame la ministre, a usage that has gained wide acceptance since (Cerquiglini, 2018, pp. 156–159; Burnett & Bonami, 2018). In Italy, language change is less advanced.

This issue of the feminization of names of titles and professions is still with us more than a century after it first emerged. Its history and present situation have been told several times (e.g., Stehli, 1949; Schafroth, 1998; Baudino, 2001; Haase, 2010; Cerquiglini, 2018; Robustelli, 2018) and need not be repeated in detail in an article primarily concerned with word formation. Some hints may suffice. Initially, the discussion was mainly concerned with grammatical correctness and uncertainties of usage, as we have already seen in the case of Spanish. The same is also true of the letter addressed by Louise Gagneur in 1891 to the Académie française, which is considered the official beginning of the debate. It was simply not obvious in some cases what the “correct” feminization of a particular word should be. Was Fr. auteur ‘author’ supposed to follow the model of acteur/actrice ‘actor/actress’, of prieur/prieure ‘prior/prioress’, or of péteur/péteuse ‘farter’? The matter was further complicated by the fact that a language question tied to a controversial issue like the emancipation of women could not remain unaffected by social, psychological, and political factors. These have always been present to a certain degree, but received a new boost when language issues became popular among feminists from the 1970s onward. Further complications arose from the fact that discussions and linguistic practices took a different, although not completely independent, course in each country, as a natural consequence of modern societies’ being embedded in nation-states. This problem is particularly pressing in the French- and Spanish-speaking world. It seems that, in the initial phase, the propensity to feminize masculine nouns of professions was lowest in France, which was probably due to the greater weight of normative grammar in this country, but partly also to the more complicated grammar of feminization in French. The transformation of epicene masculine nouns into nouns of common gender, as in la professeur ‘professor’, la ministre ‘minister’ or la bébé ‘baby’, was still rare in the 1970s and remained confined to colloquial usage (Boel, 1976, pp. 42–43). Shortly afterward, official guides on “non-sexist” language use were written in several French-speaking countries. In the beginning, these did not contain uniform recommendations with respect to the use of common gender: while the French and Belgian guides, for example, first recommended une ingénieur ‘engineer’, those of Canada and Switzerland preferred une ingénieure. From the early 21st century onward, however, French usage has adopted Quebec’s -eure, despite the opposition of the Académie française (Cerquiglini, 2018, pp. 94–98). Spanish grammarians, as we have seen, endorsed feminization right from the beginning, at least in those cases where clear models exist, as with nouns in -o, but usage went its own way: common gender nouns in -o continue to be used. Rull (2004, pp. 285–287) observes that, due to its status as a mainly spoken language unaffected by normative pressure (it was not taught at school during the Franco dictatorship), Catalan has always feminized quite freely. He rejects forms with common gender like una advocat ‘a female lawyer’ (instead of una advocada) as naked Castilianisms.10Lepschy (1989, p. 74) writes that, as far as he remembered, feminine forms were used more freely in Italian in the 1940s and 1950s also for titles and prestigious professions, while from the 1960s onward the use of masculine forms as epicenes tended to prevail, like in French. By contrast, the use of the masculine nouns in -o as common nouns, of the type la deputato ‘female MP’, seems to be rare in Italian (Thornton, 2004, p. 225n2).

4. Syntax and Compounding

The denomination of new concepts is often entrusted to noun phrases or “syntagmatic compounds” in the Romance languages. The concept ‘business man’, for example, yields Fr. homme d’affaires, Ro. om de afaceri, It. uomo d’affari, Sp. hombre de negocios, and so forth. And just as the corresponding female concept is created in English by replacing the lexeme for ‘man’ with that for ‘woman’, in the Romance languages we find Fr. femme d’affaires, Ro. femeie de afaceri, It. donna d’affari, Sp. mujer de negocios, and so forth. Besides constructions of the type N-Prep-N, combinations of adjectives and nouns are also well represented: It. gentiluomo ‘gentleman’ / gentildonna ‘lady’, and so forth. Complications can arise if the expression is highly lexicalized. For example, there is no parallel word for gentilhomme in French: **gentilfemme, **gentillefemme. The feminization of Fr. sage-femme ‘midwife’, literally ‘wise woman’, also created problems when the first men entered this profession. While the Swiss opted for the replacement strategy, creating sage-homme, the Académie française gave birth to maieuticien, which, however, sounded too stilted to gain currency. The French now use sage-femme as a generic term, adding homme when the sex of the person is relevant: homme sage-femme. The term can also be found with common gender: un sage-femme.

As we have seen in section 2, appositions were already used in old Latin (lupus femina), but this construction went out of use with the extension of the feminine marker -a. In medieval Romance varieties, appositions played no role for the denotation of sex (Bagola, 1988, p. 263). They reappear in French in the 17th century (Schafroth, 1998, pp. 86–94): “Et les femmes docteurs ne sont point de mon goût.” (Molière, Les femmes savantes) [and women doctors are not to my liking]. However, this type of construction only established itself firmly in French during the 20th century, mainly as a facile way to get around problematic cases, when speakers are unsure about which form to use. Formations of the type femme député ‘female MP’ (more rarely the other way round, député femme) are still in current use, although feminists and the official guidelines recommend not using them. Similar remarks apply to the other Romance languages, which were certainly influenced by French: Ro. femeie deputat / deputat femeie, It. donna deputato / deputato donna, Sp. mujer diputado / diputado mujer, and so forth.11 The status of such constructions as appositions or compounds is disputed.

5. Conversion: The Irresistible Spread of Feminine -a

Meyer-Lübke (1894, §§363, 365) had already pointed out that Romance is distinguished from Latin by a wider distribution of feminine -a (used here as shorthand for the different shapes that this ending received through sound change in the different branches). Even in Romance, however, -a did not become a default marker for feminization in word formation. Its exact, and quite intricate, distribution has to be stated separately for each variety. Among the most important innovations we must mention the addition of -a to some suffixes that were part of Latin third-declension nouns of common gender.

According to Fisch (1890, pp. 135–136), Latin personal nouns in -o, -onis were of common gender, although constructions like haec latro ‘this-f robber’ must have been exceedingly rare, if we are to believe the dictionaries. On the other hand, the Oxford Latin Dictionary (Glare, 2010) documents caupona ‘landlady’ as a feminine counterpart of caupo ‘innkeeper’. However that may be, the Romance languages, have the -a right from the beginning: It. mangiona ‘gluttonous woman’, Sursilvan patruna ‘female boss; boss’ wife’, Fr. gloutonne ‘gluttonous woman’, Sp. comilona ‘gluttonous woman’, and so forth.12

Another suffix of the third declension that has received -a in many Romance varieties is -tor: It. pastora ‘sheperdess’, Cor. cugidora ‘seamstress’, Sursilvan imperatura ‘emperess’, Prv. chantairo ‘female singer’ (from chantaire < Lat. cantator), Cat. professora ‘female professor’, Sp. tejedora ‘female weaver’, Pt. observadora ‘female observer’, and so forth.13 As we will see in section 6, some varieties (also) opted for other solutions: -toria, -torissa, -trissa, or the continuation of -trix.

The addition of -a to bases ending in -ntem already began in Latin (clienta), but was extended further in Romance, without ever reaching full productivity in any variety: Ro. studentă ‘female student’, Sursilvan penitenta ‘female penitent’, (written) Fr. amante ‘female lover’, Sp. giganta ‘giantess’, Pt. governanta ‘female housekeeper’, and so forth. Modern Italian shuns -a with this suffix or final sequence (cf. la presidente, presidentessa), the only exception being infanta ‘the ruling monarch’s daughter (in Spain)’, an obvious Hispanicism. In Spanish, this kind of feminization is more widespread in colloquial varieties and in Latin America than in the standard language of Spain, where some formations, like debutanta ‘female debutante’, are considered vulgarisms. A similar asymmetry between colloquial and standard language exists in Catalan (Riba i Viñas, 1992, p. 32). Also in Portuguese texts of the 16th and 17th century, feminizations like comedianta ‘female comedian’ and giganta ‘giantess’ are attested, but these are still not accepted as part of the standard language (Frias e Gouveia, 2005, pp. 528–529).

In some Romance languages, -a has also been added to nouns descending from Lat. -alem: (written) Fr. générale ‘head of a religious community; obs. general’s wife’, Sp. Pt. generala ‘id.’, while Italian prefers generalessa. The only case of -a added to a noun descended from Lat. -ensem seems to be Sp. marquesa ‘marchioness’ (from marqués ‘marquis’ < MLat. (comes) marchensis), which was the model for It. marchesa.14 Other at first sight similar cases do not belong here, because -a goes back to an older -o/-a contrast: Fr. charlatan ‘quack; fraud’ / charlatane (from It. ciarlatano/a), citadin ‘citizen’/citadine (from It. cittadino/a), loup ‘wolf’ / louve, and so forth.15 In the passage from OSp. nodriz ‘wetnurse’ to nodriza, final -a has a purely pleonastic function.

6. Dedicated Suffixes

The dedicated derivational suffixes will be treated here in alphabetical order.

Ro. -: The Romanian suffix - (Popescu-Marin, 2007, pp. 81–82; Gherman, 2015) was extracted from transparent Slavic loanwords such as tătarcă ‘Tatar woman’ (< Bulg. tatarka). They have been attested since old Romanian (e.g., ţigancă ‘gipsy’, 1533). It started declining in use in the 18th century but is still present on many words (e.g., ciobancă ‘shepherdess’, catârcă ‘she-mule’) and even occasionally occurs in neologisms (e.g., miliţiancă ‘militiawoman’). The suffix is also part of the compound suffix -oaică (-oaie + -).

Fr. -ette: The French diminutive suffix -ette (Hasselrot, 1972, pp. 71–76) has been used secondarily as a feminizing suffix. The older formations, which reach back to the Middle Ages, are neutral designations of female animals (e.g., chevreuil ‘roe deer’ / chevrette, lévrier ‘greyhound’ / levrette, with suffix exchange) or proper names (e.g., Pierre / Pierrette), while modern neologisms that can be found in colloquial French and the media tend to have a jocular overtone (e.g., député ‘MP’ / députette, gendarme ‘police officer’ / gendarmette). Hasselrot thinks that suffragette, taken over from English at the beginning of the 20th century, boosted this modern series. Note that -ette in these formations has no counterpart in masculine -et, but the origin of the pattern certainly has to be sought in masculine/feminine pairs like brunet / brunette ‘dark-haired male/female’. Apart from French, the use of this suffix for feminization seems to be rare; but compare Brazilian Pt. garçonete ‘waitress’, from garçon/garçom ‘waiter’. The secondary use of diminutive suffixes as feminizing suffixes, which also existed in Latin, is a natural consequence of the fact that females tend to be smaller than males among humans and animals alike.

Fr. -euse: This French suffix descends from Lat. -osam, the feminine form of the possessive adjectival suffix -osum. When Fr. -eur (< Lat. -atorem) and -eux (< Lat. -osum) became homophonous as /ø/, the French coopted -euse for expressing also female sex. Instead of fileresse ‘female spinner’, they started saying fileuse, and so on. Lindemann (1977, pp. 65, 70) found first examples of this use as early as the 12th century. However, it only became frequent in the 15th century, and replaced -eresse, the dominant suffix of medieval French, around 1600. Several hypotheses have been put forth to explain this striking suffix exchange. Lindemann (1977, pp. 171–173) claims that the decisive factor was the demise of the agentive suffix -ere (< Lat. -ator), the favourite base of -esse, in the standard language during the 15th century. In Modern French, -euse is mainly used with deverbal nouns in -eur, but some exceptions can be found (e.g., denominal footballeuse ‘female soccer player’). In recent Canadian recommendations on feminization (Arbour & de Nayves, 2014) the use of -euse is proscribed in names of higher professions because it is said to have a negative tinge. In Canada, chercheure ‘female researcher’ is therefore preferred to chercheuse. A feminizing suffix -eusa is also used in Piedmontese, which was strongly influenced by French (Regis, 2013, p. 278).

Lat. -ĭa: The suffix -ia was rare in Classical Latin but must have had a wider currency in spoken Latin (cf. section 2). As an unstressed suffix, it became opaque on the way to Romance: auia > OFr. aive, neptia > Fr. nièce, cervia > OFr. cierge, *cania > It. cagna, and so forth. A suffix -ia is also found in some Romance varieties, but this has nothing to do with the Latin suffix just mentioned. The Mareo dialect of Ladin (Videsott & Plangg, 1998) has four nouns denoting females ending in a suffix -ia: dotūria ‘woman doctor’, traditūria ‘traitress’, superbia ‘arrogant woman’, strömia ‘deaf-mute female’. The origin of this suffix is unknown as yet (Paul Videsott, personal communication). Lloyd (1960, p. 158) found one old Spanish example, ladria ‘female thief’, which he was also unable to explain. For the compound suffixes -toria and -onia, see following.

Fr. -ière: Meyer-Lübke (1921, p. 41) and Roché (2006, p. 89) mention the existence, in old and middle French, of some designations for females in -iere, like prisoniere ‘female prisoner’ (vs. prison ‘male prisoner’) and rechatiere ‘female buyer’ (vs. rechateor ‘male buyer’, with suffix exchange). The origin and fate of this suffix, which formally corresponds to the outcome of Lat. -ariam, remains to be studied. In Sursilvan, nouns in -ader (< Lat. -ator) are also occasionally feminized by replacing this suffix with -iera: darsentader ‘spendthrift’ → darsentiera, sfarlatader ‘spendthrift’ → sfarlatiera, and so forth.

Lat. -īna: This suffix was confined to a couple of bases in Latin (cf. section 2). Most of them were transmitted to or borrowed again by the Romance languages: It. regina, Fr. héroïne, Ro. găină, Sp. gallina, and so forth. The suffix has not become productive, but some recent formations can be found, mostly stemming from outside sources: It. zarina ‘czarina’ (from German Zarin), Sursilvan hermerina (from German Krämerin ‘female grocer’), tiladrina ‘female snuff-taker’ (corresponding to tilader ‘snuff-taker’), Cat. voivodina ‘voivode’s wife’ (Rull, 2004, p. 289), Fr. laborantine ‘female laboratory assistant’ (from German Laborantin), It. speakerina ‘female radio speaker’, and so forth.16 In It. crocerossina ‘Red Cross nurse’ (attested since 1918, while the masculine counterpart crocerossino is more recent), filandina ‘female worker of a spinning mill (filanda)’ and similar formations, we probably have to do with the feminine form of the relational suffix -ino, at least in diachronic terms (Thornton, 2004, pp. 224–225).

Lat. -ĭssa: This suffix of Greek origin came into Late Latin via the Christian literature (cf. Section 2). It was passed on to Romance (cf. Meyer-Lübke, 1935, nº 10, abbatissa, nº 2081 *comitissa), but was also borrowed again from medieval Latin later on. In some languages, these later borrowings are formally distinguished through the maintenance of the vowel /i/: Sursilvan diaconissa ‘deaconess’ (possibly influenced by German Diakonisse), Fr. pythonisse ‘prophetess’, Cat. clarissa ‘Poor Clare’, Sp. poetisa ‘poetess’, Pt. sacerdotisa ‘priestess’, etc. In Portuguese, the suffix even occurs in four different shapes (Nunes, 1951, p. 383). Besides the versus <e> contrast, there is also a contrast in sonority between <ss> and <s> (phonologically /s/ vs. /z/): diaconissa ‘deaconess’, profetisa ‘prophetess’, condessa ‘countess’, prioresa ‘prioress’. Romanian -easă is generally considered of Latin descent, but could well have come directly from Greek, since the ties with Western Latin were already quite loose when -issa was established there (André, 1971, p. 110).

The fate of this suffix was different from variety to variety. In Aromanian, Capidan (1908, p. 58) could only find one autochthonous formation, draksă ‘she-devil’, shortened from drakisă. In Romanian, the suffix -easă has been there since old Romanian and has remained moderately productive (Popescu-Marin, 2007, pp. 95–97): împărăteasă ‘empress’, băneasă ‘ban’s wife’, mireasă ‘bride’ (from mire ‘bridegroom’), and so forth. In Italian, -essa was widely used throughout the language history (e.g., sacerdotessa ‘priestess’, contessa ‘countess’, leonessa ‘lioness’), but has had only low productivity since the 20th century (e.g., professoressa ‘female professor’, studentessa ‘female student’, campionessa ‘female champion’). Possibly inspired by the discussion of the corresponding English suffix -ess in American feminist circles, the Italian feminist Alma Sabatini wanted to ban It. -essa altogether (Sabatini, 1985, p. 71), arguing that it had derogatory overtones (e.g., una vigilessa ‘a police woman’, instead of neutral una vigile). Meyer-Lübke (1890, p. 270) surmised that these overtones stemmed from the incongruence that arose from attaching an originally noble suffix to trivial bases. The alternative explanation in Lepschy, Lepschy, and Sanson (2002, p. 404), who attribute the jocular use to “una tradizione culturale maschilista” [a male chauvinist cultural tradition], seems unable to explain why the same overtones did not arise with -trice. In present-day Italian, we find neutral derivatives such as contessa or studentessa alongside ludic or derogatory ones like vigilessa (Cortelazzo, 1995; Lepschy, Lepschy, & Sanson, 2001, pp. 17–18, 2002). The suffix is also marginally present in Sardinian (Wagner, 1952, §137) and Raeto-Romance (De Leidi, 1984, pp. 99–100; Melcher, 1924, pp. 145–146).17 In old and middle French, it even became the dominant feminizing suffix (Connors, 1970–1971, pp. 575–582; Lindemann, 1977). Due to its frequent attachment to bases in -ere (< Lat. -ator), it eventually turned into a compound suffix -eresse (e.g., Middle Fr. mire ‘doctor’ / mireresse). Around 1600, the suffixes -euse and -trice pushed it into the background in the standard language, but it remained alive longer in the dialects, in popular French and slang (e.g., gonzesse ‘chick (girl)’, 1811; cf. Bonami & Boyé, 2019, pp. 176–177). This sociolinguistic background could probably explain the antipathy with which the suffix is viewed by the feminists. The suffix is less frequent in Occitan (e.g., OOcc. fauresa ‘smith’s wife’, Prv. pastresso ‘shepherdess’), where it has important competitors (-airitz, -airo, -arela). In Catalan, according to Rull (2004, pp. 283–284), the suffix has no negative connotation (e.g., metgessa ‘woman doctor’, jutgessa ‘judgess’). The same is true of Spanish and Portuguese -esa (e.g., Sp. alcaldesa ‘mayoress’, Pt. tigresa ‘tigress’).

Rom. -iţă: The Romanian feminizing suffix -iţă, which originated from Slavic (Bulgarian) -ica, came into the language with transparent loans and has been around since old Romanian times (Popescu-Marin, 2007, pp. 148–151): călugăriţă ‘nun’, împărătiţă ‘empress’, and so forth. The suffix is productive in Daco-Romanian (e.g., barmaniţă ‘barmaid’, chelneriţă ‘waitress’), less so in Aromanian (e.g., puľitsă ‘hen’; Capidan, 1908, pp. 62–63) and Istro-Romanian (e.g., lupiţa ‘she-wolf’; Kovačec, 1971, p. 163). The suffix -iţă is also used as a diminutive suffix. Both functions were drawn from Slavic.

Rom. -oaică: This compound suffix, which arose by pleonastic coalescence of the two feminizing suffixes -oaie and -, has been attested since the 18th century (Popescu-Marin, 2007, p. 82): vulpoaică ‘vixen’, drăcoaică ‘she-devil’, chinezoaică ‘Chinese woman’, and so forth.

Rom. -oaie: This feminizing suffix is believed to descend from Lat. -onia (Pascu, 1916, pp. 112–113; Graur, 1929, p. 98; Popescu-Marin, 2007, pp. 159–161), a compound suffix composed of -onem and -ia. As such, it has been attested since old Romanian times: porumboaie ‘she-dove’, păşoaie ‘pasha’s wife’, turcoaie ‘Turkish woman’, and so forth. It did not survive into modern Romanian as a productive suffix, however, except in deonymic expressions designating a man’s wife (e.g., Rădoaia ‘Radu’s wife’) and in the compound suffix -oaică. By contrast, the suffix survived in Aromanian, Istro-Romanian, and some Daco-Romanian dialects in a form where the /n/ is still preserved (Capidan, 1908, pp. 68–69; Kovačec, 1971, p. 165): Aro. lupoańe ‘she-wolf’, pikurăroańe ‘sheperdess’, Petroańe ‘Petru’s wife’, Banat unguroańe ‘Hungarian woman’, IRo. lupóńe ‘she-wolf’, and so forth.

Occ. -rela/-rello: In some varieties of Occitan the suffix -airitz/-eiritz has been replaced by -arela/-erela (also written -arello/-erello; Ronjat, 1937, p. 376): manjarela ‘gluttonous woman’ (from manjar ‘to eat’, cf. masc. manjaire), dormerela ‘woman asleep’ (from dormir ‘to sleep’, cf. masc. dormeire), and so forth. A detailed diachronic-synchronic treatment of this suffix seems to be missing.18

Lat. -tŏria:In Rainer (2019), it has been shown that many Romance suffixes denotYou could also place it directly after this word, if that makes things females go back to Lat. -toria: Ro. dansatoare ‘female dancer’ (Graur, 1929, pp. 110–111), Sardinian partordza ‘puerpera’, Friulan sopressadòrie ‘ironing woman’, Gardenese kaluniadoia ‘slandererous woman’, Piedmontese filoira ‘female spinner’ (Meyer-Lübke, 1894, p. 416; Regis, 2013), Ligurian pescavuira19 ‘fisherwoman’, Occ. pecadouiro ‘female sinner’, Picard pondouére ‘layer (hen)’, filoire ‘female spinner’, Aragonese llavadera ‘washerwoman’, OSp. cantadera ‘female singer’, Asturian mazadoria ‘woman agitating the cream in the butter churn’, Mirandese bailadeira ‘female dancer’, Pt. tecedeira ‘female weaver’, and so forth. This pattern was one of the means used to overcome the lacuna left by the demise of -trix. What is more difficult to decide is whether the sequence -toria should be analyzed as a nominalized adjective in -toria or as a compound suffix consisting of -torem plus -ia. The adjectival hypothesis certainly had an important role to play, as the many adjectival uses show (e.g., ARo. feată mărtătoare ‘marriageable girl’ (Capidan, 1908, p. 73), Prv. foulo pedacoiro ‘sinning crowd’, OFr. geline couveoire ‘laying hen’), but one cannot exclude that the feminizing suffix -ia also contributed its share. In synchrony, these designations for females show the “same” suffix as instrument nouns, which also descend from Lat. -toria; however, there is no immediate relationship between the two, neither in diachrony nor in synchrony.

Lat. *-trĭssa: In a celebrated article, Ascoli (1886–1888) showed that the feminizing suffixes -(d)ressa of old Venetian (e.g., serviressa ‘female servant’) and -dresse of Friulan (e.g., filadresse ‘female spinner’) were the result of a contamination of the Latin suffixes -trīcem and -ĭssam, which yielded *-trĭssam. Ascoli shows that these cases have to be distinguished from cases that go back to combinations of -tor (nominative) and -ĭssa, like OFr. troveresse ‘female troubadour’, derived from masc. trovere. Such formations cannot go back to *-trĭssa, since -tr- yielded -rr- in protonic position in French, witness Lat. latrone > Fr. larron. This explanation, however, is not applicable to Friulan, where -tr- yielded -dr- in pretonic position, but -r- in posttonic position (cf. Lat. amátor > *amari (the regular outcome according to Ascoli) vs. *amatrĭssam > madresse). Old Venetian behaved in the same way. Also distinct, of course, are those cases where -ĭssa was added to oblique -torem (e.g., Lat. *traditorissam ‘traitress’ > OOcc. trachoressa, OVnt. sartoressa ‘seamstress’). In present-day Friulan, according to De Leidi (1984, p. 138), -dresse is no longer productive. Siller-Runggaldier (1989, pp. 59–60) documents the suffix in Gardenese (e.g., sejladrëssa ‘female reaper’), but Mareo follows a different pattern, as mentioned in note 17.

Lat. -trīx: A number of Latin and Romance innovations, as we have seen, were a response to the loss of favour of -trix in many varieties of spoken Latin. Nevertheless, -trix did not disappear completely. It certainly survived in Occitan, in the forms -airitz and -eiritz. Adams (1913, pp. 51–54) contains old Occitan formations like cantairitz ‘female singer’ and menteiritz ‘female liar’. As shown by enfantairitz, derived from the verb enfantar ‘to bear children’, such designations of females, like in Latin, could also be derived directly from a verb stem and not only via suffix substitution from the corresponding masculine nouns in -aire and -eire. The suffix is still productive in some modern dialects, but the -airis of modern literary Provençal is considered by Ronjat (1937, p. 376) a “restauration artificielle” [artificial restoration]. Old French, by contrast, had only preserved few words (essentially emperer(r)iz ‘empress’ and pecherriz ‘female sinner’), but the suffix was later restored with Latinate stems in the form -trice in the process of re-Latinization, from around 1600 onward, as a counterpart of -teur (e.g., spectateur ‘spectator’ / spectatrice; Lindemann, 1977, pp. 95–101). Among the modern Romance standard languages, the suffix features most prominently in Italian (e.g., cantatrice ‘female singer’), where the suffix was restored early on, probably on inconspicuous remains of the Latin pattern (Migliorini, 1956). The dialects prefer other solutions, and where today corresponding formations are in use there, they are due to the influence of the standard language (e.g., Piedmontese -tris, -triss; Regis, 2013, pp. 277–278). REW nº 9194 (Meyer-Lübke, 1935), however, derives venderígola (Veneto, Istria, Romagna) and vendríkola (Marche) from *venditricula, a diminutive of venditrix ‘saleswoman’. Sardinian -driži is also considered an Italianism by Wagner (1952, §103), while Meyer-Lübke, in REW nº 6530, derived the Logudorese fish name piskadriži directly from Lat. piscatricem. These areal data, together with the indirect survival of -trix in the blend -trissa and Olcott’s remark on the predominance of -trix in inscriptions of the Italian peninsula, might lead one to assume that the suffix still enjoyed a limited popularity in Late Latin in Italy and southern France. Catalan, according to Rull (2004, pp. 145–148), has a few learned reflexes of -trix, such as actriu ‘actress’, emperadriu ‘empress’, and institutriu ‘female primary school teacher’. Similar remarks apply to the other Romance languages (on Spanish, cf. Huertas Martínez, 2014). Occasional neologisms, in these languages, have a jocular tinge (e.g., Sp. amatriz ‘female lover’, C. J. Cela, from Suárez Solís [1969, p. 475]).

7. Masculinization

Designations for females in gender languages tend to have bases referring to males. This, however, is not necessarily so. They can also be on a par with their masculine counterpart in the same lexeme, as Bonami and Boyé (2019) argue for French. Furthermore, they may be derived directly, for example from verbs (Lat. metretrix, OOcc. enfantairitz, etc.), or receive the feminine marker -a without there being a masculine counterpart (e.g., Lat. concubina, It. casalinga ‘housewife’). Relatively rare are cases where a designation for males has a base referring to a female. Two scenarios must be distinguished.

First, new designations for males may come into being via backformation. In some instances, the two patterns involved have the same degree of complexity. This notably obtains for -o/-a pairs, that is, when the gender markers are used in conversion. In the few instances where, for biological or sociocultural reasons, a word denoting a female exists without a masculine counterpart, the latter can easily be supplied by replacing -a with -o. Here are some examples: It. mammo ‘father taking care of small children and the family’ (from mamma ‘mother’), Venetian filaresso ‘male spinner’ (from filaressa ‘female spinner’; Ascoli, 1886–1888, p. 259), Sp. azafato ‘flight attendant’ (from azafata ‘air hostess’), Pt. lavadeiro ‘laundryman’ (from lavadeira ‘washerwoman’), and so forth.20 Such formations often have a jocular tinge in origin, when speakers are still aware of the derivational history. As long as this awareness subsists, furthermore, these masculine nouns cannot function as generic terms. In other instances, backformation involves subtraction. In one type, a derivational suffix is subtracted: Ro. lenjer ‘garment worker’ (from lenjereasă ‘seamstress’; Grossmann, 2016, p. 2738). More often, the subtracted element is the feminine marker -a: Ro. zân ‘heathen deity’ (from zână ‘fairy’ < Lat. Diana), Fr. prostitué ‘male prostitute’ (from prostituée ‘prostitute’), Cat. mestrès de casa ‘house husband’ (from mestressa de casa ‘housewife’; Rull, 2004, pp. 291–292).21

The second group comprises derivational suffixes for deriving designations of males from bases referring to females. They often occur where feminine nouns for animals function as epicenes: Ro. gâscan ‘gander’ (from gâscă ‘goose’; Popa, 1967), Fr. dindon ‘turkey cock’ (from dinde ‘turkey’), Cat. abellot ‘drone’ (from abella ‘bee’), cabrot ‘billy goat’ (from cabra ‘goat’; Rull, 2004, pp. 290–291), and so forth. In Fr. canard ‘he-duck’ (from cane ‘she-duck’), the designation of the male serves as the generic term. Human bases are rarer: It. stregone ‘sorcerer’ (from strega ‘witch’), Fr. compagnon ‘male partner’ (from compagne ‘female partner’, in synchrony), Cat. parterot ‘male midwife’ (from partera ‘midwife’), etc.22 We have seen that feminization is sometimes carried out by means of diminutive suffixes.23 For that same reason, augmentative suffixes sometimes serve the purpose of masculinization. Cat. -ot is a case in point, and the same seems to obtain for ORo. blendău ‘male bird of prey’, from bleandă ‘(female) bird of prey’ (Popescu-Marin, 2007, pp. 80–81).24 It is not the case, however, that masculinization necessarily involves augmentatives. French -on, for example, is not an augmentative, but a diminutive suffix (among other uses). The case of Romanian -oi (e.g., cioroi ‘he-crow’, from cioară ‘crow’, vulpoi ‘he-fox’, from vulpe ‘fox, vixen’) is particularly intricate. The suffix -oi is actually also used for forming augmentatives, but Graur (1929, p. 98) claims that, in diachronic terms, the augmentative function is derived from the masculinizing function and not the other way round.

8. A Synthesis by Language Varieties

As has become apparent in sections 3 to 7, the Romance languages and dialects reacted in different ways to the challenge posed by the demise of -trix and additional complicating factors. In Romanian, -trix disappeared completely.25 The main substitutes at the Latin stage were -toria and -onia (a solution unique to Romanian), as well as -a, which was extended all over the Romània. Romanian also resorted to borrowing, taking -easă from Greek (under André’s hypothesis) and - and -iţă from Slavic. In Italian, -trix has been re-established early on in the literary language as -trice, which would then spread to the dialects to a certain degree. Late Latin -issa was also used abundantly in standard Italian, receding somewhat in the 20th century, while the use of -toria remained confined to a few dialects. In Venetian, Friulan, and Gardenese, a new suffix -trissa arose from a blend of -trix and -issa. In Occitan, -trix survived, in some dialects until the present time, but in others it had to compete with, or was ousted by -issa, -toria, -rela, and of course -a. In old French, by contrast, only two words in -trix (emperer(r)iz, pecherriz) seem to have survived. This opened the playing field for -issa, which became the dominant suffix in old and middle French. Around 1600, however, this suffix in turn succumbed to -euse, a solution unique to French that was taken over from adjectives in -eux, and Latinate -trice. On the Iberian Peninsula, -trix is only present in Latinisms. The vacuum was first filled there by the use of -tor as a suffix of common gender and by -toria. The first of these strategies disappeared completely already during the Middle Ages, while -toria survived to a limited extent in Portuguese, Leonese, Asturian, and Aragonese (for examples, see section 6, under -toria). In Catalan, -toria became homophonous with -dora, which is why they cannot be distinguished formally. In all Ibero-Romance varieties, -issa is also present in a couple of formations. The marker -a has witnessed a much greater extension in Ibero-Romance than, for example, in Italian. Apart from the derivational means just mentioned and the marker -a, all Romance standard languages also use appositions as a kind of backup strategy for otherwise difficult cases with prestigious titles and professions.26 This is a relatively recent development, however, due to the necessity to refer to women entering these professions in the wake of women’s emancipation. This same necessity has also brought along a new wave of epicene masculine nouns or nouns of common gender of the type Sp. la médico, which gave rise to heated, but often not very well-informed discussions among feminists, grammarians, and the public at large. In part, governments and other institutions tried to settle the question by issuing official recommendations, but language use is still in a state of flux.

9. Critical Analysis of Scholarship

Sex-denoting patterns of word formation in Romance have never been the object of a systematic general overview, beyond the succinct remarks in Meyer-Lübke (1894). Even for single varieties, a solid diachronic monograph only exists for French (Lindemann, 1977). Similar studies would be highly welcome for the other varieties. The issue that, due to its societal relevance, has absorbed most of the attention over the last decades is the feminization of titles and names of (prestigious) professions. The gist of the discussions and relevant literature on this topic can be found in sections 3, 4, and 6 (under -ĭssa).


  • Adams, E. L. (1913). Word-formation in Provençal. New York, NY: Macmillan.
  • Aliaga García, F., & Lázaro Mora, F. (2007). Reflexiones en torno a ‘modisto / modista’. In I. Delgado Cobos & A. Puigvert Ocal (Eds.), Ex admiratione et amicitia: Homenaje a Ramón Santiago (Vol. 1, pp. 63–79). Madrid, Spain: Ediciones del Orto.
  • Aliaga Jiménez, J. L. (2011). El género de los nombres de profesión en un manuscrito inédito de Juan Moneva y Pujol (1929). In M. E. Vázquez Laslop, K. Zimmermann, & F. Segovia (Eds.), De la lengua por sólo la extrañeza: Estudios de lexicología, norma lingüística, historia y literatura en homenaje a Luis Fernando Lara (Vol. 1, pp. 75–97). México, D.F.: El Colegio de México.
  • André, J. (1971). Emprunts et suffixes nominaux en latin. Geneva, Switzerland: Droz.
  • Arbour, M.­È., & de Nayves, H. (2014). Féminisation linguistique: Étude comparative de l’implantation de variantes féminines marquées au Canada et en Europe. Langage & Société, 148, 31–51.
  • Ascoli, G. I. (1886–1888). Di -tr-ĭssa che prenda il posto di -tr-īce. Archivio Glottologico Italiano, 10, 256–260.
  • Bagola, H. (1988). Zur Bildung romanischer Berufsbezeichnungen im Mittelalter. Hamburg, Germany: Buske.
  • Baudino, C. (2001). Politique de la langue et différence sexuelle: La politisation du genre des noms de métier. Paris, France: L’Harmattan.
  • Boel, E. (1976). Le genre des noms désignant les professions et les situations féminines en français moderne. Revue Romane, 11, 16–73.
  • Bonami, O., & Boyé, G. (2019). Paradigm uniformity and the French gender system. In M. Baerman, O. Bond, & A. Hippisley (Eds.), Perspectives on morphology: Papers in honour of Greville G. Corbett (pp. 171–192). Edinburgh, UK: Edinburgh University Press.
  • Brender, F. (1920). Die rückläufige Ableitung im Lateinischen. Lausanne, Switzerland: La Concorde.
  • Burnett, H., & Bonami, O. (2018). Linguistic prescription, ideological structure, and the actuation of linguistic changes: Grammatical gender in French parliamentary debates. Language in Society, 49(1), 65–93.
  • Capidan, T. (1908). Die nominalen Suffixe im Aromunischen. Leipzig, Germany: Barth.
  • Casanova, E. (1984). Sobre la moció del suffix -ISTE/-ISTA. Boletín de la Sociedad Castellonense de Cultura, 60, 259–267.
  • Cerquiglini, B. (2018). Le ministre est enceinte, ou la grande querelle de la féminisation des noms. Paris, France: Seuil.
  • Ciobanu, E. (1959). Observaţii asupra formării numelor de familie feminine de la masculine. In Studii şi materiale privitoare la formarea cuvintelor in limba romană (Vol. 1, pp. 137–143). Bucureşti: Academia Republicii Populare Romîne, Institutul de Lingvistică din Bucureşti.
  • Cirillo, C. (2000). Gender and feminine agentives in Italian dictionaries: 1612–1917. In G. Lepschy & P. Shaw (Eds.), A linguistic round-table on dictionaries and the history of the language (pp. 11–23). London, UK: University College London.
  • Connors, K. (1970–1971). Studies in feminine agentives in selected European languages. Romance Philology, 24, 573–598.
  • Cortelazzo, M. (1995). Perché non si vuole la presidentessa? In G. Marcato (Ed.), Donna e linguaggio: Convegno Internazionale di Studi: SAPPADA/PLODN (Belluno) 1995 (pp. 4–7). Padova, Italy: CLEUP.
  • De Leidi, G. (1984). I suffissi nel friulano. Udine, Italy: Società Filologica Friulana.
  • Doleschal, U. (2015). Gender marking. In P. O. Müller, I. Ohnheiser, S. Olsen, & F. Rainer (Eds.), Word-formation: An international handbook of the languages of Europe (Vol. 2, pp. 1159–1171). Berlin, Germany: De Gruyter.
  • Elwert, W. T. (1943). Die Mundart des Fassa-Tals. Heidelberg, Germany: Winter.
  • England, J. (1984). Observaciones sobre las nuevas formas femeninas en el castellano del siglo XIII. In J. M. Ruiz Veintemilla (Ed.), Estudios dedicados a James Leslie Brooks presentados por sus colegas, amigos y discípulos (pp. 31–44). Barcelona, Spain: Puvill.
  • England, J. (1987). New feminine forms in Old Spanish: The fourteenth and fifteenth centuries. Bulletin of Hispanic Studies, 64, 205–214.
  • Ernout, A. (1908). Remarques sur l’expression du genre féminin en latin. In Mélanges de linguistique offerts à M. Ferdinand de Saussure (pp. 211–222). Paris, France: Champion.
  • Fisch, R. (1890). Die lateinischen nomina personalia auf “o, onis”. Berlin, Germany: Gaertner.
  • Fradin, B. (2019). Competition in derivation: What can we learn from doublets in -age and -ment. In F. Rainer, F. Gardani, W. U. Dressler, & H. C. Luschützky (Eds.), Competition in inflection and word formation (pp. 67–93). Cham, Switzerland: Springer.
  • Frias e Gouveia, M. C. de. (2005). A categoria gramatical de género do português antigo ao português actual. In G. Rio-Torto, O. M. Figueiredo, & F. Silva (Eds.), Estudos em homenagem ao Profesor Doutor Mário Vilela (pp. 527–544). Porto, Portugal: Facultade de Letras da Universidade do Porto.
  • Gherman, C. (2015). ᶦ-că (ᶦ-ca). In M. Rădulescu Sala (Ed.), Formarea cuvintelor în limba română: Vol. 4/1. Sufixele: Derivarea nominală şi adverbială. Bucureşti: Editura Academiei Române.
  • Glare, P. G. W. (Ed.). (2010). Oxford Latin dictionary. Oxford, UK: Clarendon Press.
  • Graur, A. (1929). Nom d’agent et adjectif en roumain. Paris, France: Champion.
  • Grossmann, M. (2016). Romanian. In P. O. Müller, I. Ohnheiser, S. Olsen, & F. Rainer (Eds.), Word-formation: An international handbook of the languages of Europe (Vol. 4, pp. 2731–2751). Berlin, Germany: De Gruyter.
  • Haase, P. (2010). Feminisierung im spanischen Sprachraum: Berufs-, Amts- und Funktionsbezeichnungen – el juez, la juez, la jueza. Hamburg, Germany: Dr. Kovać.
  • Hasselrot, B. (1972). Étude sur la vitalité de la formation diminutive française au XXe siècle. Uppsala, Sweden: Alqvist & Wiksells.
  • Huertas Martínez, S. (2014). El sufijo -triz/-driz en la lexicografía preacadémica. In C. Grande López, L. Martín Aizpuru, & S. Salicio Bravo (Eds.), Con una letra joven: Avances en el estudio de la historiografía e historia de la lengua española (pp. 151–158). Salamanca, Spain: Ediciones Universidad de Salamanca.
  • Kovačec, A. (1971). Descrierea istroromânei actuale. Bucharest: Editura Academiei Republicii Socialiste România.
  • Lepschy, A. L., Lepschy, G., & Sanson, H. (2001). Lingua italiana e femminile. Quaderns d’Italià, 6, 9–18.
  • Lepschy, A. L., Lepschy, G., & Sanson, H. (2002). A proposito di -essa. In L’Accademia della Crusca per Giovanni Nencioni (pp. 397–409). Florence, Italy: Le Lettere.
  • Lepschy, G. (1989). Lingua e sessismo. In G. Lepschy, Nuovi saggi di linguistica italiana (pp. 61–84). Bologna, Italy: Il Mulino.
  • Leumann, M. (1977). Lateinische Laut- und Formenlehre (5th ed.). Munich, Germany: Beck.
  • Lindemann, M. (1977). Zum Suffixwechsel von -eresse zu -euse und -trice im Französischen. Tübingen, Germany: Narr.
  • Lloyd, P. M. (1960). A linguistic analysis of Old Spanish occupational terms (Dissertation). University of California.
  • Löfstedt, B. (1963). Bemerkungen zum Problem Genus: Sexus im Lateinischen. Symbolae Osloenses, 38, 47–68.
  • Loporcaro, M. (2018). Gender from Latin to Romance: History, geography, typology. Oxford, UK: Oxford University Press.
  • Melcher, F. (1924). Furmaziun nominala nel idiom d’Engadin’ota. Annalas da la Societad Retorumantscha, 38, 119–179.
  • Meyer-Lübke, W. (1890). Italienische Grammatik. Leipzig, Germany: Reisland.
  • Meyer-Lübke, W. (1894). Grammatik der romanischen Sprachen: Vol. 2. Formenlehre. Leipzig, Germany: Reisland.
  • Meyer-Lübke, W. (1921). Historische Grammatik der französischen Sprache: Vol. 2. Wortbildungslehre. Heidelberg, Germany: Winter.
  • Meyer-Lübke, W. (1935). Romanisches Etymologisches Wörterbuch (3rd ed.). Heidelberg, Germany: Winter.
  • Migliorini, B. (1956). A proposito dei nomi in -trice. In B. Migliorini, Saggi linguistici, (pp. 129–134). Firenze, Italy: Le Monnier.
  • Nunes, J. J. (1951). Compêndio de gramática histórica portuguesa (4th ed.). Lisboa, Portugal: Teixeira.
  • Nyrop, K. (1903). Grammaire historique de la langue française: Vol. 2. Copenhagen, Denmark: Nordisk Forlag.
  • Olcott, G. N. (1898). Studies in the word formation of the Latin inscriptions: Substantives and adjectives; with special reference to the Latin Sermo Vulgaris. Rome, Italy: Sallustian Typography.
  • Pascu, G. (1916). Sufixele româneşti. Bucureşti, Romania: Socec.
  • Pattison, D. G. (1975). Early Spanish suffixes: A functional study of the principal nominal suffixes of Spanish up to 1300. Oxford, UK: Blackwell.
  • Popa, E. (1967). Sufixe moţionale pentru formarea masculinului. In Studii şi materiale privitoare la formarea cuvintelor in limba romană (Vol. 4, pp. 163–176). Bucureşti: Academia Republicii Populare Romîne, Institutul de Lingvistică din Bucureşti.
  • Popescu-Marin, M. (Ed.). (2007). Formarea cuvintelor în limba română din secolele al XVI-lea – al XVIII-lea. Bucureşti: Editura Academiei Române.
  • Rainer, F. (2019). The benefit of the pan-Romance perspective: A new attempt to solve the tecedor/tecedeira puzzle. Word Structure, 12(1), 127–151.
  • Regis, R. (2013). I suffissi agentivi in piemontese: Fattori esterni e fattori interni. Lingua e Stile, 48, 249–284.
  • Riba i Viñas, C. (1992). El gènere en càrrecs, professions i oficis. In J. Solà (Ed.), Sobre lexicografía catalana actual (pp. 11–39). Barcelona, Spain: Empúries.
  • Robustelli, C. (2018). Lingua italiana e questioni di genere: Riflessi linguistici di un mutamento socioculturale. Canterano, Italy: Aracne.
  • Roca, I. (2005). La gramática y la biología en el género español. Revista Española de Lingüística, 35(1), 17–44.
  • Roché, M. (2006). La dérivation en -ier(e) en ancien français. Lexique, 17, 55–96.
  • Ronjat, J. (1937). Grammaire istorique des parlers provençaux modernes: Vol. 3. Morphologie et formation des mots: Notes de syntaxe. Montpellier, France: Société des Langues Romanes.
  • Rull, X. (2004). La formació de mots: Qüestions de normativa. Vic, Spain: Eumo.
  • Sabatini, A. (1985). Occupational titles in Italian: Changing the sexist usage. In M. Hellinger (Ed.), Sprachwandel und feministische Sprachpolitik: Internationale Perspektiven (pp. 64–75). Opladen, Germany: Westdeutscher Verlag.
  • Schafroth, E. (1998). Die Feminisierung von Berufsbezeichnungen im französischen Sprachraum: Mit einem vergleichenden Blick auf das Deutsche und andere Sprachen. (Habilitationsschrift). Universität Augsburg.
  • Sgroi, S. C. (2007). La mozione: Un revival? Lingua Nostra, 68(1–2), 56–59.
  • Sgroi, S. C. (2018). Il genere grammaticale e la teoria sessista della lingua. In V. L. Castrignanò, F. De Blasi, & M. Maggiore (Eds.), “In principio fuit textus”: Studi di linguistica e filologia offerti a Rosario Coluccia in occasione della nomina a professore emerito (pp. 651–665). Florence, Italy: Cesati.
  • Sibille, J. (2015). Description de l’occitan parlé à Sénaillac-Lauzès (Lot) et dans les communes voisines. Lemòtges, France: Lambert-Lucas.
  • Siller-Runggaldier, H. (1989). Grödnerische Wortbildung. Innsbruck, Austria: Institut für Romanistik.
  • Spitzer, L. (1923). Das Feminin als Ableitungstypus im Frz. Zeitschrift für romanische Philologie, 43, 644–654.
  • Suárez Solís, S. (1969). El léxico de Camilo José Cela. Madrid, Spain: Alfaguara.
  • Stehli, W. (1949). Die Femininbildung von Personenbezeichnungen im neuesten Französisch. Bern, Switzerland: Francke.
  • Stotz, P. (2000). Handbuch zur lateinischen Sprache des Mittelalters: Vol. 2. Bedeutungswandel und Wortbildung. Munich, Germany: Beck.
  • Thornton, A. (2004). Mozione. In M. Grossmann & F. Rainer (Eds.), La formazione delle parole in italiano (pp. 219–227). Tübingen, Germany: Niemeyer.
  • Thornton, A. (2012). Quando parlare delle donne è un problema. In A. Thornton & M. Voghera (Eds.), Per Tullio de Mauro: Studi offerti dalle allieve in occasione del suo 80° compleanno (pp. 308–316). Rome, Italy: Aracne.
  • Thornton, A. (2015). Paradigmatically determined allomorphy: The “participial stem” from Latin to Italian. In P. O. Müller, I. Ohnheiser, S. Olsen, & F. Rainer (Eds.), Word-formation: An international handbook of the languages of Europe (Vol. 1, pp. 780–802). Berlin, Germany: De Gruyter.
  • Tommaseo, N., & Bellini, B. (1861–1879). Dizionario della lingua italiana. Turin, Italy: Unione Tipografico-Edictrice.
  • Videsott, P., & Plangg, G. (1998). Ennebergisches Wörterbuch. Innsbruck, Austrai: Wagner.
  • Wagner, M. L. (1952). Historische Wortbildungslehre des Sardischen. Bern, Switzerland: Francke.


  • 1. Thornton (2004, p. 219) reports that in Italian 69.2% of human bases have special forms for females, but only 14.5% of animal bases. For helpful comments on an earlier version, I would like to thank Maria Grossmann, Graça Rio-Torto, Anna Thornton, as well as two anonymous reviewers.

  • 2. Some scholars, like Meyer-Lübke (1894, p. 409, 1921, p. 39), use the term in the more restricted sense of feminization/masculinization by means of inflectional gender markers, as opposed to derivational suffixes.

  • 3. On sex-denoting patterns in Latin, compare Ernout (1908), Löfstedt (1963), Leumann (1977, §269).

  • 4. For masculines in -sor, some feminines in -strix are documented (e.g., tonsor ‘barber’ / tonstrix ‘female barber’, related to tondere ‘to cut (hair)’). The Romance languages did not adopt this type of formations in the process of relatinization. Italian and French therefore still struggle with the feminization of agentive nouns in -sore/-seur; compare Thornton (2012, 2015, pp. 784–786, 797–798), Fradin (2019, pp. 69–70).

  • 5. Scythissa ‘Scythian woman’ is already in Cornelius Nepos, a contemporary of Caesar.

  • 6. As far as the form of the base is concerned, compare Ro. împărat ‘emperor’.

  • 7. In the feminist literature, epicene is sometimes used in a non-technical sense. So, for example, the use of “non-sexist” language is called in French rédaction épicène ‘epicene way of writing’.

  • 8. As Aliaga García and Lázaro Mora (2007) observe, in cases such as modisto / modista, where the masculine and the feminine form do not denote exactly the same professional role, the masculine form cannot act as a generic designation. For the same reason, monje ‘monk’ cannot act as a generic englobing the meaning of monja ‘nun’ (Roca, 2005, p. 27). This shows that encyclopedic knowledge interferes with the mechanism that is responsible for generic reference.

  • 9. It certainly was no easy matter for a woman to establish herself in the liberal professions at that time, judging from the following entry under medichessa ‘woman doctor’ in Tommaseo-Bellini (1861–1879): “S.f. di Medico. Ora le donne si addottorano medichesse. Ma una suora di Carità risica di saper far meglio” [Feminine form of medico. Now women receive doctor’s degrees as physicians. But a Sister of Mercy will probably be more efficient] (quoted in Cirillo, 2000, p. 19).

  • 10. A detailed account of how contentious cases have been dealt with by Catalan lexicographers and grammarians can be found in Riba i Viñas (1992).

  • 11. Note that Ro. femeie goes back to Lat. familia, not femina. Maria Grossmann (personal communication) points out to me that femelă / femeiuşcă are used with female animals and mascul / bărbătuş for male animals. (Analogous remarks apply for the other Romance languages.) When the male sex of a masculine noun is highlighted, bărbat ‘man’ can be used: medic-bărbat ‘male doctor’.

  • 12. On Romanian -oaie, compare section 6.

  • 13. Fr. prieure ‘prioress’ is not etymologically an agent noun, but a comparative. Nyrop (1903, p. 278), however, quotes intercesseure, inventeure, and promoteure from a medieval text, formations that foreshadow the modern feminines in -eure recommended for prestigious professions by some guides.

  • 14. The case of OFr. duchoise ‘duchess’, which was later completely ousted by duchesse, is somewhat different. The suffix -oise also goes back to a combination of -ensem and the feminine marker -a, but there was no corresponding masculine noun duchois ‘duke’, as far as we know. Possibly the formation took place already at the Latin stage.

  • 15. Romance varieties that underwent major sound changes present all kinds of allomorphic complexities beyond the change that affected -a itself. These cannot be addressed here.

  • 16. Masculine laborantin is a later backformation.

  • 17. In Mareo, agentives in - (< Lat. -torem) are flanked by feminines in -dëssa (e.g., forladú ‘thresher’ / forladëssa). This pattern recalls Med. Lat. imperatissa, Ro. împărăteasă, and It. procuratessa ‘procurator’s wife’. Gardenese, by contrast, has the Venetian-Friulan pattern -trissa (Siller-Runggaldier, 1989, pp. 59–60).

  • 18. Compare also Sursilvan pastur ‘shepherd’/ pastorella ‘shepherdess’. This seems to be the only formation of its kind in Sursilvan.

  • 19. Tuscan designations of females in -tora could also theoretically descend from Lat. -toria in those regions where -torium yielded -toro. Alternatively, it can be seen as the result of adding -a to -tore.

  • 20. In jocular formations, the semantic relationship can be more indirect, as in Sp. cocacolo ‘frivolous teenager’ (from Coca Cola) or aspirino ‘medical orderly’ (from aspirina ‘aspirin’).

  • 21. Ro. privighetor ‘male nightingale’ (from privighetoare ‘nightingale’) and turturel ‘he-turtle dove’ (from turturea ‘turtle dove’ are cases of affix substitution (Grossmann, 2016, p. 2038).

  • 22. In old French, the Latin nominative companio yielded compain, while oblique companionem became compagnon. Compagne originally was a feminization of compain. Compain has become copain in modern French. It was also written as copin, which explains why the corresponding feminine form is copine.

  • 23. Fr. mulet ‘he-mule’ is surprising from this perspective, since the diminutive suffix -et serves the purpose of masculinization. It is synchronically related to epicene-feminine mule. In old French, mule (< Lat. mulam) was the feminine form of mul ‘mule’ (< Lat. mulum), of which mulet was a regular diminutive.

  • 24. In modern Romanian, the meaning is ‘scarecrow’ (also figuratively).

  • 25. However, as one reviewer points out, there are loans from French like actriță, possibly understood as the Slavonic suffix -iţă.

  • 26. The same strategy is also used with epicene animal names (e.g., It. volpe maschio ‘he-fox’).