Where is grammatical gender encoded in german

The notion that language affects cognition has a long history. This idea was popularized due to the development of cultural linguistics at the turn of the 19th century and the work of such researchers as F. Boas, E. Sapir, and B. Whorf, who perceived grammatical structures as the source of differences in the way of thinking, learning and experiencing reality Whorf, — the principle of linguistic relativity.

Nowadays, numerous studies confirm that language has an impact on cognition Lucy, , and one of the factors influencing it is grammatical gender see Boroditsky et al. However, apart from a strong version, which assumes that language strictly determines thinking and that language categories limit and determine cognitive categories, there is a weaker version of the linguistic relativity hypothesis, which states that language influences cognitive processes only under certain circumstances, especially in tasks where grammatical coding is necessary Montefinese et al.

These two versions of the linguistic relativity hypothesis also apply to the issue of the influence of grammatical gender on cognitive processes Vigliocco et al.

Therefore the literature review presented in this paper needs to take into account studies in which grammatical gender effects were not found, as these effects may depend on circumstances such as certain types of cognitive tasks. Prewitt-Freilino et al. In case of the last language group, nouns always have a particular grammatical gender assigned to them, which influences — due to grammatical agreement — the declension and conjugation of other parts of speech.

Grammatical gender languages are of great interest in the field of psycholinguistics, addressing the impact of grammatical gender on various cognitive processes, including memory, categorization, personalization, assigning properties linked to the biological sex of objects, etc.

For example, grammatical gender can lead to a transfer of male and female attributes onto inanimate objects which do not have a biological sex e. In general, research on grammatical gender effects focuses mainly on whether it influences various cognitive processes, especially the semantic processing of nouns. It is emphasized that interlingual differences in grammatical gender systems influence its cognitive effects. For example, numerous studies which used categorization tasks provided evidence that grammatical gender effects occur only in languages with two grammatical genders, such as Arabic Clarke et al.

Inanimate objects may have any of the grammatical genders e. Additionally, nouns used to describe females sometimes have a masculine grammatical gender [e. On the other hand, Polish has a lot of grammatical gender markers: personal pronouns, demonstrative pronouns, nouns, adjectives, verbs, numerals — all containing morphological information about grammatical gender, consistent with the grammatical gender of the noun.

Please see Appendix 1 for a summary of word types carrying information on grammatical gender in Polish. The primary goal of our research was to further investigate the effects of grammatical gender in Polish. Our research concerns the influence of grammatical gender on the categorization of words and attribution of gender-related traits to objects.

We are also interested in whether this influence goes beyond verbal material — whether the grammatical gender effects will occur when objects are presented as images, rather than words.

The aim of our research was also to test two groups of mechanisms by which the grammatical gender can influence selected cognitive processes, referred to as the similarity and gender hypothesis and the sex and gender hypothesis Vigliocco et al.

In addition, the aim was to check whether an implicit influence of grammatical gender on the categorization process will occur in the Polish grammatical gender system. Our research implemented three different research procedures. The first experiment used triadic similarity judgments based on the research by Vigliocco et al. The second experiment used a modified procedure based on the Implicit Association Test IAT , where participants performed a task of semantic categorization of words into selected categories.

The third experiment concerned the assignment of male and female voices to animals and inanimate objects, presented either as words or pictures, in order to test the link between masculine and feminine properties of referents and grammatical gender of nouns.

On the one hand, when learning a language, one could assume that the differences in grammatical gender of words provide some information about the world and are not accidental. In linguistics it is often postulated see Fodor, that grammatical gender of inanimate objects does not carry semantic connotations.

While the grammatical gender of animate personal and some of the animate non-personal nouns has semantic foundations, assigning other nouns to particular grammatical genders may be completely arbitrary Boutonnet et al. Indeed, in various grammatical gender languages particular words have the same or different grammatical genders. Contemporary empirical research in psycholinguistics provides evidence of the influence of grammatical gender on various cognitive processes, related to lexical and semantic levels of words, as well as mental representations of their referents.

There are many ways in which grammatical gender influences speakers of gendered languages. A very important effect of grammatical gender is its impact on semantic categorization. Vigliocco et al. These words were nouns referring to animals and artifacts. It turned out that grammatical gender affected triadic similarity judgments in Italian speakers more than English speakers, but this effect was present only for words referring to animals and not artifacts.

This study was the basis for Experiment 1 in the present paper and will be discussed in detail further on. Similarly, Cubelli et al. Responses to gender-consistent pairs were significantly faster for Italian and Spanish speakers, but not English speakers.

A study with the use of EEG showed that there is spontaneous and unconscious access to grammatical gender when people are asked to perform a semantic evaluation of individual relationships Boutonnet et al. As part of our research Experiment 2 we tested implicit access to grammatical gender using a modified Implicit Association Test IAT and measuring response times as indicators of the influence of grammatical gender.

Grammatical gender also affects how we imagine and personify inanimate objects and ideas. Artists and research participants personify such ideas as war and death Segel and Boroditsky, or days of the week Deutscher, as women or men, in accordance with their grammatical gender in a given language.

For example, death der Tod - masculine tends to be represented by German artists as a male, while Spanish artists depict death la muerte — feminine as a female. A similar effect in a general population was reported by Sera et al. This procedure was used in Experiment 3 of the present study and will be discussed in detail there.

Research also shows that grammatical gender affects the way we describe objects. Boroditsky et al. Participants described grammatically masculine objects using more stereotypically masculine qualities, and grammatically feminine objects with stereotypically feminine ones. Gender effects also occurred when participants were asked to assign certain features stereotypically related to one sex to nouns e.

It turned out that voice gender was consistent with grammatical gender with an odds ratio of It should be noted that there was a significant main effect for language i. In the second experiment, the same participants were asked to assign gender-stereotypical adjectives to artifacts and natural objects varying in grammatical gender.

Moreover, directly comparing two languages i. One group of participants was presented with objects as pictures or nouns and asked to use three adjectives to describe each one. Afterward, another group of participants was asked to rate whether these adjectives are suitable for masculine and feminine properties of objects.

In accordance with Experiment 2 by Boroditsky et al. Other studies on grammatical gender effects show that it also affects memory. It turned out that pairs of words which shared grammatical gender were remembered more correctly than pairs of words with different genders. In another paper, it turned out that fragrances were remembered significantly better when the grammatical gender of the nouns describing them agreed with the type of fragrance masculine or feminine; Speed and Majid, There are some unresolved theoretical issues regarding grammatical gender effects.

Primarily, it is not clear how they vary in languages with different gender systems. In various studies, grammatical gender effects seem constrained to certain languages — Vigliocco et al. The same paper showed no grammatical gender effects for inanimate objects and when objects were presented as pictures, rather than words.

Moreover, there is not enough clear evidence on whether grammatical gender effects vary for different classes of items most of the presented studies concern inanimate objects, sporadically including animals, and none of them investigates, e. Finally, while a large body of research shows that grammatical gender effects can be elicited using visual stimuli pictures , few studies directly compare visual and verbal stimuli in this aspect.

The present study aims to add to this body of research, while further investigating grammatical gender effects in Polish and applying a single-language framework in contrast with existing research which mostly uses comparisons across different languages. There is no clear consensus whether grammatical gender effects occur in languages with three genders — research in German shows no such effects, while the two existing studies in Polish seem to confirm the influence of grammatical gender on cognitive processes.

It is assumed that the influence of grammatical gender on cognitive processes must be a by-product of the process of language acquisition. By learning the grammatical gender of a given noun, a person concentrates on select properties of the object consistent with its gender [e. The need to refer to masculine or feminine properties may make them more important in the mental representation of the particular object, and thus more easily accessible Boroditsky et al.

In the similarity and gender hypothesis it is assumed that words with similar syntactic and morphological properties usually have a similar meaning. Nouns with the same grammatical gender are used in the same linguistic context because in a sentence they require gender agreement with prepositions, adverbs, pronouns, etc.

Grammatical gender effects are therefore a by-product of inferring semantic similarity from the linguistic context alone. This hypothesis predicts that the effect of gender will occur both in languages with two genders like Italian and in languages with more than two genders German or Polish.

In morphologically rich languages i. If grammatical gender effects are based on the similarity of linguistic context, one can expect grammatical gender effects to occur also for inanimate objects. The sex and gender hypothesis explains the effects of grammatical gender based on the formation of relationships between the grammatical gender of nouns and the biological sex of their human or animal referents.

When learning a grammatical gender language one can notice the relationship between grammatical gender and biological sex, i. It is also assumed that there is greater semantic similarity between nouns belonging to the same gender category Vigliocco et al. According to this hypothesis, grammatical gender effects may appear mainly for animate nouns in languages with two grammatical genders. The effects will be weaker or even absent in languages with more than two grammatical genders. This prediction was confirmed in research conducted in German which has three gender categories Sera et al.

It is commonly assumed see Vigliocco et al. Grammatical gender effects, especially in a categorization task, were mainly confirmed for languages with two grammatical genders, in which it is easier to perceive a link between sex and grammatical gender. The aim of our research was to verify whether grammatical gender influences categorization processes in Polish — a language which has three genders in singular form and additional two in plural, with a large number of grammatical gender markers.

We also tested grammatical gender effects on abstract nouns, in addition to inanimate and animate nouns. To our knowledge there is no research regarding grammatical gender effects on abstract nouns, at least not in Polish. We present a series of three experiments in which we examined grammatical gender effects in Polish.

In Experiment 1 based on the research paradigm by Vigliocco et al. The goal was to check whether under such conditions grammatical gender effects would be present for abstract words, names of animals and inanimate objects. In Experiment 2, based on the Implicit Association Test IAT paradigm, participants performed a semantic categorization task under severe time pressure.

The aim was to check whether grammatical gender was implicitly accessible when performing a task that runs at the lexical and semantic level. In Study 3, we examined the attribution of male and female voices to animals and inanimate objects, presented either as words or pictures, in order to verify whether the influence of grammatical gender goes beyond the lexical and semantic levels and into visual representations, and directly comparing its effects for verbal and visual stimuli.

We assumed that the obtained results will allow us to test the assumptions of the sex and gender hypothesis and the similarity and gender hypothesis in the context of the Polish language. Based on the sex and gender hypothesis, grammatical gender effects are expected to be stronger in languages with a high consistency between the grammatical gender of nouns and the biological sex of their referents. In Polish, this consistency is restricted by the fact that there are three genders in the singular form masculine, feminine, neutral and two genders in the plural form masculine-personal and non-masculine-personal , which have quite a complicated relation to their singular counterparts.

Grammatical gender of nouns varies between their singular and plural forms — the non-masculine-personal plural form encompasses all singular form nouns except for masculine nouns referring to people.

The strict version of the sex and gender hypothesis would therefore predict a lack of grammatical gender effects in Polish, not unlike many published studies in German. A less constrained version of the sex and gender hypothesis would predict there to be grammatical gender effects only for nouns referring to animals and humans — with a biological sex.

In the three experiments presented in this paper, we included nouns referring to animals Experiment 1 and 3 , inanimate objects all experiments and abstract ideas Experiment 1. Grammatical gender needs to be considered when using verbs, numerals, adjectives, pronouns, etc. Therefore the similarity and gender hypothesis would predict that grammatical gender effects be present in all the presented experiments, using both animate and inanimate nouns.

Observing grammatical gender effects for nouns referring to objects which do not have a biological sex would provide strong support to the similarity and gender hypothesis.

The presented research also aims to test two more specific research questions. Experiment 2 uses a modified implicit association test to investigate whether grammatical gender effects in Polish are present on the lexicosemantic level, and not just on the conceptual level.

Whenever referring to lexicosemantic and conceptual levels of processing, we use the differentiation as in Vigliocco et al. Conversely, the conceptual level is based on the mental representations of the referents and their qualities, which is non-linguistic in nature.

Using language without referring to mental representations of objects is strictly lexicosemantic, while processing mental representations without using language is strictly conceptual. The above hypotheses will be discussed in more detail in the description of each experiment.

Please note that these hypotheses are, to a degree, independent — it is possible that grammatical gender reflects biological sex of animals and humans, while the similarity and gender hypothesis is true for inanimate objects.

Moreover, grammatical gender effects may be present both on the lexicosemantic and conceptual levels. Participants performed a categorization task in Italian and German languages with formal gender systems using the triadic similarity procedure. The task was to assess which of the three presented words were the most semantically similar. The triads always contained two words with the same grammatical gender, and a third one with a different gender.

Words referred either to animals or artifacts. Pairs of words chosen by the participants were classified as same-gender or different-gender. The authors tested whether there was an effect of grammatical gender on similarity classification by comparing those results with the results of an English group, which was presented with the English version of the Italian or German nouns used in the study.

In Experiment 1 by Vigliocco et al. The Italian language has two gender classes masculine or feminine. It turned out that the effect of grammatical gender on the categorization process occurred only for animate nouns i. In other words, pairs of animals with the same grammatical gender in Italian were chosen significantly more often than the same pairs of animals in English; however, this effect did not occur for inanimate objects. The experiment was then replicated in German, which is a three-gender system masculine, feminine, and neutral and has a less transparent correspondence between the sex of referents and the grammatical gender of nouns that refer to them.

It turned out that in this language grammatical gender effects did not occur for any type of nouns. In their next experiment conducted on Italian language users image triads were used, instead of word triads, to illustrate the same nouns as in Experiment 1. There was no effect of grammatical gender in this case.

It was concluded that grammatical gender effects do not go beyond the domain of language into the conceptual level. Results of these studies suggest that the influence of grammatical gender on the categorization process is limited to languages with two grammatical genders, it occurs in tasks requiring verbalization and is limited to animate nouns Vigliocco et al. These results suggest that gender effects are based on a generalization of the established relationship between the gender of nouns and the sex of human referents, expanding to other gender-defined entities.

Such effects occur in languages? In turn, the lack of effect for pictorial stimuli means that these effects appear on the lexicosemantic level, not on a conceptual level.

In Experiment 1, we used the above similarity judgment task to investigate grammatical gender effects in Polish. The study described above raises doubts about factors which may have influenced the results. The authors did not control the semantic criteria used by participants when assessing the similarity of words.

For example, in the triad of Italian words taken from the paper by Vigliocco et al. That is, participants would often choose fork with knife because they match semantically, although they differ in terms of their grammatical genders.

It can also happen that the same grammatical gender goes hand in hand with another criterion. For example, in the triad zebra la zebra - F , giraffe la giraffa - F and deer or wolf il cervo; il lupo - M , the place of origin i.

Similarly in the triad: tiger la tigre - F , lion il leone - M and goat la capra - F the predominant taxonomic criterion is being a predator, and again not grammatical gender. Research by Vigliocco et al. Another doubt is the direct comparison between Italian and German vs. One could argue that language differences other than grammatical gender word length, usage frequency, pronunciation, cultural context, etc. In our research presented below, we took extreme care to select items which do not have obvious properties linking two words in any given triad.

In addition, we modified the instruction so that it did not suggest an explicit search for a semantic criterion; instead, there was a request to choose two nouns which are associated the most with each other, based on the first impression. Moreover, we included nouns referring to abstract ideas, which should facilitate grammatical gender effects on a lexicosemantic level, as abstract ideas are difficult to imagine in a conceptual form.

On the other hand, words referring to animals would be the most likely to generate grammatical gender effects on a conceptual level, as gender is a relevant trait of animals.

Finally, rather than compare different languages, we conducted one-sample analyses based on a theoretical expectation of indifference and randomness, using only the Polish language. Participants were native Polish speakers 96 women, 50 men recruited from various faculties of the Jagiellonian University in Krakow. All participants volunteered to take part in the study upon informed consent, and they did not receive compensation financial or otherwise for their participation.

All experiments presented in the paper were approved by the Ethics Committee of the Jagiellonian University Institute of Applied Psychology. We used a similarity judgment task such as in Vigliocco et al. The nouns forming the triads belonged to one of three categories: abstract nouns e. Each triad consisted of nouns from a single category, i. The complete list of stimuli is presented in Appendix 2.

Each triad consisted of two nouns of the same grammatical gender two feminine or two masculine and one noun of another grammatical gender.

This means that each triad had either the MMF structure two masculine words and one feminine word or FFM structure two feminine words and one masculine word — in randomized order. The nouns forming the triad belonged to one of three categories: 16 abstract nouns, for example: time M in Polish , function F in Polish ; 16 inanimate objects, for example: key M , needle F ; 12 words referring to animals, for example: bison M , zebra F.

As can be noted, we used fewer animals This resulted from the fact that we adopted strict selection criteria. Firstly, these were animal names with a certain grammatical gender, but ones which do not indicate the biological sex of the animal, meaning that the same word refers to both the male and female of the particular animal.

Secondly, we chose animals which do not have distinctive features which could be a clear criterion for choosing two objects in a triad e. Similarly, if there were two mammals and one reptile in the triad, then this could become the criterion of choice. Therefore we chose animals which all belonged to the category of exotic even-toed ungulates. The selection criterion was the similarity in appearance, behavior and the habitat of the animals. For other types of nouns, we applied a similar rule.

As such, we avoided situations in which two of the items in a triad would have an obvious common feature, which could be used as a categorization criterion. We used counterbalancing in order to ensure that different word combinations would be used. Triads were constructed using all the possible three-noun combinations with repetitions within each category, yielding triads for abstract nouns and artifacts each, and 91 triads for animals.

Out of these available triads, we created 7 sets for the abstract words and artifacts 64 triads each and 4 sets for animals 45 triads each. In case of animals there were fewer sets of triads, because fewer words passed the strict selection process described above 6 masculine gender words and 6 feminine gender words.

Each participant evaluated one set of triads within one category of nouns abstract, artifacts, or animals; the category of nouns was a between-group variable. We ask you to make decisions quickly and rely on your first impression.

The aim was to check whether participants would mention grammatical gender as the basis of their decisions, hence not meeting the assumption that grammatical gender effect is implicit. We predicted that the influence of grammatical gender would occur primarily for nouns referring to animals.

In Polish there are cases where the grammatical gender does not indicate the biological sex of the animal — but in general, the assignment of grammatical gender to animals in Polish has semantic foundations, and gender information is often consistent with the biological sex of the referent. This can lead to a generalization of the relationship between grammatical gender and the biological sex of the referents and facilitate the grammatical gender effects.

Predicting the occurrence of these effects in animal names is consistent with the sex and gender hypothesis and was confirmed in other studies, where the influence of grammatical gender on the categorization process was observed only for animate nouns Boroditsky et al. In case of abstract nouns, it is more difficult to access mental representations, which could facilitate lexicosemantic processing. Therefore the categories of nouns used in this experiment vary in two main aspects: 1 whether the referent has a gender and 2 whether a mental representation is easily accessible.

In the post-experimental questionnaire, none of the participants indicated that they used grammatical gender as the conscious criterion for categorizing words.

Words chosen by each participant were classified into same-gender and different-gender pairs. Grammatical gender consistency of individual triadic similarity judgments was converted into an average consistency measure for all judgments made by each participant, which could range from 0 none of the judgments was consistent with grammatical gender to 1 all of the judgments were consistent with grammatical gender. The dependent variable was essentially the percentage of same-gender pairs selected from the triads which contained two words of one gender and one word of the other gender.

In the studies by Vigliocco et al. If grammatical gender does not influence categorization, and other qualities are randomized, participants should select each of the three possible pairs within a triad equally often.

Polish speakers chose pairs of animal words with the same grammatical gender significantly more often than what would result from random chance. Results are presented in Figure 1. Figure 1. Grammatical gender consistency across triad types and noun categories.

Grammatical gender of nouns influenced the choice of words as similar in case of animals, but not inanimate objects, and it was reversed for abstract nouns albeit the selection frequency differed from random chance only by about 1. The result by Vigliocco et al. Such a result is consistent with the sex and gender hypothesis, according to which grammatical gender affects the semantic representations of gendered entities.

This hypothesis also stated that in languages in which personal nouns are assigned to more than two gender categories, the influence of gender on cognitive processes is unlikely to occur and this notion was supported in a study by Vigliocco et al.

However, in Polish, where personal nouns belong to one of three grammatical genders, the grammatical gender effect was present. We assume that this is linked to the inflection richness of this language, where nouns require gender agreement with prepositions, adverbs, pronouns and other grammatical elements of the language.

We also tested the role of grammatical gender in the case of abstract nouns, referring mainly to physical categories, such as movement M , temperature F , time M. Their grammatical gender is arbitrary perhaps even more than for inanimate objects , and additionally there is no easy access to the mental representations of the corresponding referents and their traits.

There was no grammatical gender effect which can be predicted based on the sex and gender hypothesis. However, in this group of nouns a reverse effect occurred; the mean proportion of selected same-gender pairs was significantly lower than random.

Numerous words with masculine gender e. For example, the word temperature temperatura — F can be included in the broader category condition stan — M. Therefore the observed effect could be a result of factors other than grammatical gender. The fact that grammatical gender effects were present only for animals, and not for artifacts and abstract nouns, could be considered support for the sex and gender hypothesis.

Even though we used nouns which do not directly state the sex of the animals, most nouns referring to animals in Polish do state whether it is a male or a female, making gender a relevant category. Therefore it is safe to assume that the category of gender is highly cognitively activated when thinking about animals. In the similarity judgment task, researchers have no control over the selection criteria used by participants.

This applies to both Vigliocco et al. Therefore we decided to check whether in Polish the effect of grammatical gender will occur for specific inanimate nouns, in a task to categorize words to unambiguously named semantic categories under strict time constraints, which facilitate implicit cognitive processing.

To achieve this, we designed a task based on the Implicit Association Test IAT paradigm, where the categorization process takes place at the lexicosemantic level. In recent years, an interesting study was conducted on the hidden impact of grammatical gender on semantic processing using the category decision task. Cubelli et al. In their study, participants were presented with pairs of pictures containing objects the names of which had the same or different grammatical genders.

They were asked to decide whether these objects belong to the same semantic category. It turned out that participants Italian and Spanish speakers responded faster to pairs of stimuli with the same grammatical gender than ones with inconsistent gender.

Compatibility of the grammatical gender of object names accelerated reactions to both semantically related pairs positive responses and semantically unrelated pairs negative responses. Results indicated that the grammatical gender of object names affects semantic processing in tasks requiring categorical judgments based on visual stimuli, facilitating processing of the meaning of gender-consistent noun pairs.

According to the authors, the lack of gender effects under articulatory suppression Experiment 3 means that the categorization of objects requires the processing of lexical representations and depends on the level of activation of object names. Additionally, results confirmed that grammatical gender is activated implicitly when making lexical and semantic decisions.

Boutonnet et al. Spanish—English bilingual participants were presented with three pictures of objects in Spanish on the screen. Their task was to assess whether the image of the third object belongs to the same or to a different semantic category as the first two, while measuring event-related brain potentials ERPs.

In half of the triads the name of the third picture had the same grammatical gender as the first two, and in the other half — the opposite one. When reading the table, it is important to bear in mind that in Section 1, only the fields that are relevant, depending on the category to which a language belongs, have been filled. The next three sets of features — described in Sections 2 to 4 of the Table — are common to all languages and therefore always filled in see Supplementary Table S1.

They pertain to various aspects of linguistic structures, the lexicon and language use, 2 and have not been described in detail in previous taxonomies, specifically:. Morphology esp. Which derivational processes are relevant and where may one find negative connotations attached to certain forms?

Some of the derived feminine forms may carry a negative connotation such as the suffix - ette in gendarmette n. In French, generically used forms are found both in nominal forms e. Asymmetries: What types of asymmetric forms or semantic features can be observed in the lexicon? For example, address terms may not be symmetrical between women and men e.

Note that in some cases more fine-grained distinctions based on usage have not been exhaustively documented in the table see Supplementary Table S1 for practical reasons.

For example, a given form may exist, but its usage may be infrequent or fading. We still qualify it as present but urge that researchers interested in these particular features should always carefully control for its usage. For each feature, the following classification has been used:. Note that this tag was only used in reference to usage in Sections 2—4.

When cells in the table concern other groups of languages in Section 1, these are filled in with the indication not applicable. Psycholinguistic investigations of the way people perceive gender have shown different biases associated with the particulars of grammatical gender.

Not surprisingly, since many languages possess grammatical gender, these investigations have been conducted across a wide range of languages. However, between language comparisons — as rich as they may be — face intrinsic questions of legitimate comparability. In the present language index, we present different grammatical gender dimensions that might be of special interest for those interested in cross-languages comparisons in the way grammatical gender constrains our mental representations of women and men.

However, constructing a language index raises some important issues that also need to be taken into consideration in order to document how grammatical gender is encoded across languages.

While the classification of languages into one of the five main categories that we established genderless, natural gender, etc. One such issue is the necessity to determine whether some features are truly productive in a language. This question can hardly be answered based on the intuitions of native-speaking informants alone, as it requires the use of quantitative analyses. This implies that for every feature in every language, a correct estimation of its prevalence would require extensive studies of language use in corpus data.

Conducting such empirical analyses is beyond the scope of our index. While conscious of the limitations of our approach, we had to content ourselves with an estimation of usage provided by native speaker informants that we divided into three intuitive categories i. These categories provide an estimation that should therefore be treated with caution, and are best used as a starting point for researchers who are interested in one particular aspect of gender differences.

Another limitation of our index is that the usage of feminine forms has evolved over the past decades in many languages. As a result, many forms that are attested may now be falling out of use. For example, the use of the word le minister to designate a female government minister in French is now declining, following an official decision from the French government in to feminize occupation names for women Cerquiglini, Thus, even though some naming practices might be recognized by informants as existing in their language, it does not mean that they still correspond to current practices, or would not be recognized as sexist by its speakers.

Here again, extensive studies of languages use that go well beyond the scope of our index would be needed to determine the nature of current practices. Yet again, we believe that our index represents a useful starting point for researchers who want to investigate these questions.

Another issue is that naming practices often vary from country to country, even when those countries share the same language. French is a case in point. While in France, the feminization of occupation names for women is a recent phenomenon, the use of feminine names was already current practice decades earlier in other French-speaking countries such as Canada, Belgium, and Switzerland. Finally, our index contains a sample of 15 languages, representing mostly the Indo-European family.

However, grammatical gender distinctions are widespread across the languages of the world. Adding languages from other families that fall into this category would therefore bring valuable enrichments to our index, allowing us to move beyond Western cultural representations of sex and gender, as cultural differences have an impact of the representation of gender.

For example, Corbett b reports that in Lak, a language spoken in the central Dagestan highlands, girls were not classified within the category of rational females, which for example applied to grown-up women, but in the category of other non-male and non-female animate beings.

This classification led to an evolution of usage concerning the terms of address for young women. Using the gender marking for animate but not females when addressing young women became a sign of politeness. Aside from such anecdotal examples, documentation of gender-related usage for these languages is to a large extent lacking.

We hope, however, to be able to enrich the present database in the future with more publications on languages for which gender-related usage can be collected.

PG prepared the first draft of the manuscript and coordinated the work among all authors. DE and SZ completed the index and collected the final data for it. AG worked on the final draft of the manuscript. FB, JO, and AG created the first version of the index and organised the collection of the initial data for the languages presented in the Supplementary Material.

All authors worked on their specific language to complete the index. The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Revisiting masculine and feminine grammatical gender in spanish: linguistic, psycholinguistic, and neurolinguistic evidence.

Bodine, A. Cerquiglini, B. Le Ministre est Enceinte. Paris: Seuil. Google Scholar. Corbett, G. Cambridge: Cambridge University Press. Dryer and M. Fleischer, E. Rieken, and P. Widmer Berlin: De Gruyter Mouton , doi: Dixon, R.

Where Have All the Adjectives Gone? Berlin: Walter de Gruyter. Doleschal, U. Das Generische Maskulinum im Deutschen. Esaulova, Y. Cross-linguistic evidence for gender as a prominence feature. Fedden, S. Non-Canonical Gender Systems. Oxford: Oxford University Press. Formanowicz, M. Verbs as linguistic markers of agency: the social side of grammar. Foundalis, H. Gabriel, U. Giles and A.

Maas Bern: Peter Lang Publishers. Neutralising linguistic sexism: promising but cumbersome? Group Process. Intergroup Relat. Gygax, P. The masculine form and its competing interpretations in French: when linking grammatically masculine role names to female referents is difficult. Generically intended, but specifically interpreted: When beauticians, musicians, and mechanics are all men. Hannover, B. Spinath and E. Heise Hamburg: Kovac , 37— PubMed Abstract Google Scholar.

Hegarty, P. Maass Pieterlen: Peter Lang Publishing , — Hellinger, M. Gender Across Languages. Amsterdam: John Benjamins Publishing Company. Horvath, L. Does gender-fair language pay off? Kesebir, S. Word order denotes relevance differences: the case of conjoined phrases with lexical gender. Sato, S. Altering male-dominant representations: a study on nominalized adjectives and participles in first and second language German.

Gender inferences: grammatical features and their impact on the representation of gender in bilinguals. Bilingualism 16, —