Translating English WOMAN IS AN ANIMAL metaphors: Spanish native speakers’ associations with novel metaphors

Animal metaphors are prevalent across languages and convey a variety of, oftentimes negative, meanings – more so for women than men. In English, for example, both lion and lioness refer to a sexually active, dominant man or woman respectively, but while the former is endowed with positive connotations (courage, strength), the latter evokes negative associations (danger, voracity). There are some animal terms, however, that do not feature in animal metaphors in a certain language, posing the question as to which associations are evoked by those animal terms that are not part of conventional animal metaphors. This paper explores Spanish speakers’ interpretations of mappings of the WOMAN IS AN ANIMAL metaphor that are documented to exist in English but not in Spanish. This was tested with two online questionnaires, one employing open questions and the other one Likert scales presenting possible traits (e. g. quarrelsome, kind, promiscuous), in which Spanish speakers had to judge the animal metaphors which were translated from English. The results show that the novel animal metaphors are mainly associated by Spanish native speakers with negative features, first and foremost with ugliness. Additionally, most of the animal terms convey different meanings in English and Spanish. For example, musaraña, the Spanish equivalent of shrew, is not associated with bad temper and quarrelling, but instead with ugliness and muddleheadedness. Furthermore, the findings reveal significant insecurities in the interpretation of the translated metaphors by the Spanish speakers. These results might be an indication for both the arbitrariness and the stableness of associations with different animal species, depending on the speakers’ culture. It also seems that novel animal metaphors mainly provide mental access to unattractiveness as it is a concrete physical feature and might therefore be more accessible than abstract personality traits such as kindness or quarrelsomeness.


Introduction
It has long been argued by cognitive linguists that metaphor is not just a figure of speech for the purpose of embellishing language but instead is integral to human thinking (cf. Lakoff/Johnson 1980Kövecses 2000;Deignan 2010). Thus, through metaphor, humans conceptualize one concept (target) in terms of another concept (source). Examples of such conceptualizations include the metaphors ARGUMENT IS WAR, LOVE IS A JOURNEY (cf. Lakoff/Johnson 1980), and PEOPLE ARE ANIMALS (cf. Kövecses 2010: 153). English examples of the latter conceptual metaphor are chick and cow to refer to a young and coarse woman respectively, as well as wolf and lion denoting a sexually aggressive and notable man respectively. These examples highlight the role that the PEOPLE ARE ANIMALS metaphor can play in reinforcing gender stereotypes by likening women and men to different types of animals.
While animal metaphors can be considered universal as they are likely prevalent in all languages (cf. Kövecses 2005: 25), the associations conveyed by the animal terms can sometimes vary greatly between languages. For example, both rabbit and the Spanish equivalent coneja refer to a woman who has given birth to many children, but in another English/Spanish pair, namely bird and pájara, the associations differ greatly: Bird is used affectionately in English to denote a girl or young woman, whereas the Spanish equivalent pájara refers to a cunning woman or a prostitute (cf. López-Rodríguez 2009: 88). Such variations are due to cultural differences between the speakers of different languages, with culture serving as a filter: While the body is a potentially universal source for emerging metaphors, culture functions as a filter that selects aspects of sensorimotor experience and connects them with subjective experiences and judgments for metaphorical mappings. That is, metaphors are grounded in bodily experience but shaped by cultural understanding. (Yu 2008: 247) Animal metaphors are "pre-established by a long tradition" (Bisschops 2019: 1) so the associations that they convey are rather stable. However, in each language only a certain, although large set of animal species features in animal metaphors. This poses the question as to which associations are evoked by animal terms that are not part of conventional animal metaphors in a given language. Contrastive studies have generally focussed on the comparison of animal metaphors that are documented to exist in the respective studied languages (cf. Fernández Fontecha/Jiménez Catalán 2003;Reza Talebinejad/Vahid Dastjerdi 2005;Hsieh 2006;López-Rodríguez 2009). However, it is the goal of the present study to explore the extent to which the associations conveyed by conventional animal metaphors in one language (English) correspond to those conveyed by novel animal metaphors in another language (Spanish) when they are not fixed by tradition. Thus, the present study tests Spanish speakers' judgements about mappings of the WOMAN IS AN ANIMAL metaphor that are documented in English but not in Spanish. For example, the English animal term shrew is conventionally used to denote an ill-tempered, quarrelsome woman, whereas the Spanish equivalent musaraña is not documented to be used in mappings of the WOMAN IS AN ANIMAL metaphor.
So far, research on animal metaphors has had various foci. There have been contrastive studies (cf. Nesi 1995), cognitive linguistic analyses (cf. Sanz Martin 2015), combinations of those two approaches (cf. Fernández Fontecha/Jiménez Catalán 2003), studies on the translation of animal metaphors (cf. Dobrotă 2017), diachronic analyses (cf. López-Rodríguez 2014), and studies on the behavioural impact that animal metaphors can have (cf. Bock/Burkley 2018). However, to the author's knowledge, no study has yet tested speakers' associations with animal metaphors that are not documented in the speakers' native language but in another language. It is the goal of the present study to contribute to filling this research gap. For that, using English dictionaries (The Oxford English Dictionary (OED), The Oxford Dictionary of Modern Slang (Ayto/Simpson 2013), etc.), English animal terms referring to different types of women (e. g. promiscuous woman, old woman, good-looking woman) were collected which are not documented to exist in Spanish. Then, in two separate online questionnaires, native speakers of Spanish judged the translated animal metaphors in terms of their meaning. In the first one, participants rated the animal metaphors on Likert scales, whereas the second questionnaire employed open questions, allowing for a multi-methodological approach to detect Spanish speakers' associations with undocumented animal metaphors. Accordingly, participants appear to be more unsure about the meanings of animal metaphors when confronted with open questions compared to when being able to choose from several options on Likert scales.
The present study, thus, aims to contribute to the contrastive investigation of animal metaphors by researching animal metaphors that are novel in one language, instead of comparing animal metaphors that are established in both languages as has been the focus of many previous studies. This approach should help shed light on native speakers' associations with certain animal species when they are not conventional source concepts in the WOMAN IS AN ANIMAL metaphor.
In the following, the theoretical framework of this study is outlined, which comprises Cognitive Metaphor Theory followed by an overview of the research that has been done on animal metaphor. Next, the methodology employed for the creation and analysis of the questionnaires is introduced, and finally, the results obtained from the two questionnaires are presented and discussed.
2 Theoretical framework

Metaphor in Cognitive Linguistics
According to cognitive linguists, metaphor is more than simply a figure of speech. Instead, it is fundamental to human thought and cognition (cf. Lakoff/Johnson 1980Kövecses 2000;Deignan 2010). Metaphor is "understanding and experiencing one kind of thing in terms of another" (Lakoff/Johnson 1980: 5), meaning that, through metaphor, a target domain is conceptualized in terms of a source domain. This correspondence between the two domains is understood of as conceptual mapping. The following sentences highlight a well-known example of metaphor, namely the ARGUMENT IS WAR metaphor in which a mapping occurs from the source domain WAR to the target domain ARGUMENT (cf. Lakoff/Johnson 1980: 4): (1) I've never won an argument with him.
(2) He attacked every weak point in my argument.
As Lakoff/Johnson (1980) state, "we don't just talk about arguments in terms of war […] [but] the ARGUMENT IS WAR metaphor is one that we live by in this culture: it structures the actions we perform in arguing" (ibd.: 4).
When a target domain is conceptualized in terms of a source domain, certain aspects of the target are highlighted while others remain hidden. The ARGUMENT IS WAR metaphor, for example, "highlights the adversarial nature of argument but hides the fact that argument often involves an ordered and organized development of a particular topic" (Evans/Green 2006: 304 (1) I need to get rid of this emotional baggage. (cf. Cameron/Deignan 2006: 678-680) (b) Novel metaphors that follow the mapping principle but are novel usages. They require the activation of an underlying connection.
(2) I need to get rid of this emotional luggage. (c) Novel metaphors that do not follow the mapping principle. They require the listener to make a new connection between the source and the target.
(3) I need to get rid of this emotional handbag. (d) Anomalous metaphors, i. e. novel metaphors that use a source-target domain pairing that never occurs in the language, e. g. ECONOMY IS FOOD.
(4) The two sides are digesting natural resources. (cf. Ahrens et al. 2007: 164) While previous research has only distinguished between conventional and novel metaphors, Ahrens (2010) has shown in on-line and off-line psycholinguistic experiments that there are indeed differences in terms of metaphoric understanding between all four types of metaphor (a-d). Accordingly, declining conventionality and increasing novelty equal a decline in metaphor acceptability and interpretability. For that reason, it is important to distinguish between the different types of metaphors one is dealing with.
It has long been argued by cognitive scientists that metaphors which originate from human physiology and embodied experiences are universal (cf. Neumann 2001: 124;Kövecses 2005: 3;Evans/Green 2006: 308). For example, the metaphor AFFECTION IS WARMTH (cf. Kövecses 1986: 101) -which is based on "the correlation in our childhood experiences between the loving embrace of our parents and the comforting bodily warmth that accompanies it" (Kövecses 2005: 2f.) -exists in many languages and can be considered a universal metaphor. However, in metaphor research, emphasis has also been put on the cultural and contextual nature of metaphor. This approach, which distinguishes between cross-cultural and within-culture variation, is described by Kövecses (2008)

The PEOPLE ARE ANIMALS metaphor
When dealing with the PEOPLE ARE ANIMALS metaphor, the question arises how it happens that animals are used to talk about humans and human behaviour. As Kövecses (2010: 152) states, animals do not inherently possess human qualities. Instead, humans ascribe human characteristics to animals, thus personifying them, before reapplying them to humans, hence animalizing them. For example, a female dog is not spiteful per se, but humans attribute spitefulness to female dogs and apply it to a woman when referring to her as bitch.
In order to make sense of the conceptualization of humans as animals, it is essential to consider the so-called Great Chain of Being (cf. Lakoff/Turner 1989). This cultural folk model places humans, animals, plants and inanimate objects on a vertical hierarchy, representing the supposed world order of "higher" versus "lower" entities. Accordingly, the lowest entity are inanimate objects, which are nothing but substance, while plants additionally have life. Animals have all these properties, but they also show animal behaviour such as instincts. Lastly, on top of all these properties, humans also have human-specific attributes such as reason, morality, and consciousness. However, each level contains further sublevels, so while they are both inanimate objects, a chair is more complex than a stone, and while they are both animals, a cat is more complex than a bug, for example (cf. Lakoff/Turner 1989: 167f.). Furthermore, sublevels also exist within the category of humans. Accordingly, "men rank above women because traditionally the former are believed to be ruled by reason whereas the latter by their heart, which seems to bring the female sex closer to the animal kingdom" (López-Rodríguez 2016: 94; cf. also Fernández Fontecha/Jiménez Catalán 2003: 794). When referring to a human as an animal -as is the case in the PEOPLE ARE ANIMALS metaphor -the target HUMAN is downgraded to a member of the lower category ANIMAL, hence depriving the person of their human-specific characteristics (cf. López-Rodríguez 2016: 77; Tipler/Ruscher 2019: 110).
However, as Haslam/Loughnan/Sun (2011) have shown, animal metaphors are not always inherently offensive. Instead, there are factors that contribute to the offensiveness of animal metaphors while others mitigate it: Offensiveness derives both from the transfer of reviled characteristics from taboo animals to metaphor targets and from the positioning of the target as literally less than human, even when the animal in question is not taboo.
[…] It may be possible to distinguish two kinds of offensive animal metaphors: those that are disgusting and those that are degrading. (ibd.: 318) Furthermore, the context of the animal metaphor plays a crucial role in its offensiveness. It seems that offensiveness is greater when the metaphor is used in a hostile manner addressed to women and in intergroup contexts, especially when referring to someone's appearance (cf. ibd.: 322). In any case, "animal metaphors may be insulting in part because they are -or are intended to be -dehumanizing" (ibd.: 312). As Kiełtyka/Kleparski (2007) point out -referring to research conducted by Thornton (1988) -, in English animal metaphors, humans are conceptualized as mammals far more often than as any other animal category. Table 1 shows the ranking of animal categories (source domain) used in the conceptualization of humans (target domain).
Rank Animal category used in the conceptualization of humans 1 mammals 2 insects, reptiles, birds 3 fish 4 arachnids, amphibians 5 crustaceans The preference for mammals as source concepts is "due to their widely-understood similarity, familiarity and closeness to mankind" (Kiełtyka/Kleparski 2007: 89). Additionally, as suggested by Hines (1999a: 15), common animals, such as monkey, rat and sheep, appear more often in animal metaphor than rarer animals, such as polar bear, porcupine and zebra (cf. also Sanz Martin 2015: 380).
In line with the finding that men tend to rank above women on the Great Chain of Being (see above), it has been shown that women are conceptualized as animals more often than men and in more variety regarding the types of animals (cf. López-Rodríguez 2009). Furthermore, animal metaphors referring to women are overall more derogatory than those referring to men (cf. Fernández Martín 2011). This is especially the case for animal metaphors denoting sexual behaviour, such as promiscuity (cf. Fernández Fontecha/Jiménez Catalán 2003). For example, while both lion and lioness refer to a sexually active, dominant man or woman respectively, the former is endowed with positive associations of male courage, ferocity, and strength, whereas the latter equates the woman with a dangerous and voracious animal, implying a threat to man's hegemonic masculinity (cf. Crespo-Fernández 2015: 147f.). López-Rodríguez (2009) explains the gender imbalance when it comes to (English and Spanish) animal metaphors as follows: Drawing a clear boundary between the rational human and the instinctual beast, animal metaphors are often used in English and Spanish to degrade particular social groups that are regarded as inferior or marginal. Obviously, taking into account that within the English-and Spanish-speaking world, the male white heterosexual is assumed to be the norm, that is, "the self", any other social group deviating from this, such as women … will fall into the category of "the other". Belittlement of such "other" […] is often carried out via animal metaphors, as though implying the inappropriateness of their behavior. (ibd.: 94f.) López-Rodríguez (2007b) shows that women tend to be conceptualized as animals belonging to three main categories, namely pets (bitch, kitten, cat), farm animals (mare, hen) and wild animals (vixen, crow). While the first two categories tend to carry positive connotations, the latter category is tied to unfavourable associations: Obviously, from the human perspective, pet and farmyard animals are domesticated and tamed, depend on man for their survival and do not pose any threat. Wild animals, by contrast, are not subject to man's control and are menacing. Hence, by portraying women in the guise of pets and farm animals, the idea of domesticity is being highlighted, evoking the patriarchal view that a woman's place should be confined to the domestic arena. Leaving their designated domestic sphere, however, is linguistically castigated by endowing the names of wild animals with negative associations. (López-Rodríguez 2009: 95) In her study on animal metaphors used by the written media to speak about women and their relationship with food, López-Rodríguez (2016) points out that an animal's size also plays a crucial role regarding the associations portrayed by the conceptualization. Hence, as suggested by Hines (1999a: 9, 17), the metaphor DESIRED WOMAN IS A SMALL ANIMAL likens young and attractive women to small animals, such as bird, mink and bunny. Furthermore, Hines (1999a: 12) argues that in most cases small animals are only mapped if they remain small in size when mature (e. g. cat, chicken), otherwise the immature animal is used (e. g. filly instead of mare). Sometimes though, both the immature and mature animal coexist, which can be seen in the pairs kitten/cat and chick/chicken. When it comes to larger animals, they tend to be perceived as menacing and negative (e. g. mare, nag, seal, walrus or coyote). Thus, when applied to women, these animal metaphors are offensive and pejorative (cf. López-Rodríguez 2016: 95). This goes hand in hand with perceived weakness and strength of certain animals so that referring to women as weak animals (e. g. chicken, Sp. pollita 'chick') appears to have positive connotations, whereas when it comes to strong animals (e. g. vixen, Sp. zorra 'vixen') the associations are negative (cf. López-Rodríguez 2009: 95).
In her study of metaphorical lexicalization, Hines (1999a) makes the interesting discovery that the animals in her list of active central terms denoting women considered sexually -bird, bunny, canary, cat, chick, filly, fox, goose, grouse, kitten, partridge, pigeon, plover, pussy (cat) and quail -seem to not only be chosen based on their semantic properties but also their phonetic features. Thus, "[e]very central term begins with a labial or velar obstruent" (ibd.: 11). 2 Hines (1999a: 11) argues that this sound symbolism is explicable as labials and velars have been shown to be associated with derogation in English. However, while animal terms like coyote and flamingo match the phonetic, but not the semantic prototype, hamster and stork fit semantically, but not phonetically (cf. ibd.: 14). This interplay of phonetics and semantics is not unique to this specific metaphor but instead also exists in the metaphors WOMAN IS A DESSERT, manifested, for example, in cookie and cupcake denoting an attractive woman (cf. Hines 1999b), and WOMAN IS A FEMME FATALE, manifested, for example, in harpy to refer to a grasping, unpleasant woman (cf. Hines 1996).
In her analysis of English and Spanish animal metaphors, López-Rodríguez (2009) shows that the two languages share many patterns and similarities when it comes to the conceptualization of women as animals. In both languages, men tend to predominantly be conceptualized as big, strong, and wild animals (e. g. wolf, Sp. toro 'bull'), while women are mainly referred to as a variety of pets and farmyard animals or as wild animals. For example, both dog and Sp. perro are used to denote an ugly woman. Parakeet and the Spanish equivalent periquita both refer to a young woman, and rabbit and its Spanish counterpart coneja denote a woman with many children. However, there are also clear differences between the two languages: Bird, for example, is used affectionately in English to denote a girl or young woman, whereas Sp. pájara refers to a cunning woman or prostitute. Furthermore, vixen is applied to an ill-tempered quarrelsome woman, while Sp. zorra denotes a cunning or promiscuous woman or a prostitute (cf. López-Rodríguez 2009: 83, 86, 88, 91f.). Striking differences as well as interesting similarities were also found by Fernández Fontecha/Jiménez Catalán (2003) in their contrastive-cognitive analysis of two male/female examples in English and Spanish. Investigating the animal pairs fox/vixen and bull/cow as well as their Spanish counterparts zorro/zorra and toro/vaca, the authors found semantic derogation in both languages and more so for women compared to men, which seems to be evidence for the hypothesis of semantic imbalance in language (cf. ibd.: 792). Regarding the derogation conveyed by the female animal terms, there are clear differences between the two languages: [W]ith regard to the animal pair fox/vixen-zorro/zorra, Spanish is more derogatory to women than English in view of the fact that in Spanish, women are related to promiscuous sexual behavior, whereas this is not so in English. On the other hand, regarding the animal pair bull/cow-toro/vaca, English proves to be more derogatory since, among other things, in the main metaphor from cow, women are associated not only with negative physical aspects, i. e. unattractiveness or large size, but also with negative behavioral aspects such as coarseness.
(ibd.: 793) Thus, Fernández Fontecha/Jiménez Catalán (2003: 789) argue that, when it comes to attempting to measure derogation, a negative behavioural aspect weighs heavier than a negative physical aspect and among negative behavioural aspects female promiscuity is considered the worst.
The many above-mentioned findings from previous research on animal metaphors are of utmost importance for the analysis of the results of the present study whose methodology is introduced in the next section, followed by the presentation and discussion of the results.

Methodology
In order to detect native speakers' associations with documented and undocumented Spanish animal metaphors, two online questionnaires were created using the web application SoSci Survey. Participants only took part in one of the two questionnaires by means of randomization. In both questionnaires, the participants were shown 15 Spanish sentences of the type Ana es una musaraña 'Ana is a shrew', with one sentence per page and only the animal term changing on each page. In order to test Spanish speakers' associations with undocumented animal metaphors and to compare them to their English meanings, ten animal metaphors referring to certain types of women which are documented in English but not in Spanish were translated into Spanish (see Table 2, white cells). Additionally, three well-documented Spanish animal metaphors referring to certain types of women were included in the questionnaire (see Table 2, light grey cells) as well as two animal metaphors that are undocumented in both languages (see Table 2, dark grey cells). This design was chosen in order to be able to draw conclusions about animal metaphor interpretation depending on the degree of metaphor conventionality.  Some of the animal terms of the study have masculine gender (ratón, castor, semental, murciélago, cuervo), some feminine gender (codorniz, musaraña, cucaracha, trucha, perdiz, zorra, vaca, lagarta, cigüeña, nutria). To determine which animal terms would sound unknown or unacceptable to the participants, three native speakers judged all animal terms for which both masculine and feminine forms exist (e. g. castor vs. castora, ratón vs. ratona) in advance and unanimously decided on the above arrangement.
The ten animal metaphors that are documented to exist in English but not in Spanish were chosen based on an English dictionary search of the following dictionaries: The English animal terms were then directly translated into Spanish and -in order to determine if the Spanish animal terms were documented to exist as metaphors for certain types of women -they were subsequently cross-checked with a Spanish dictionary search of the following dictionaries: -Diccionario de la lengua española de la Real Academia Española (DRAE), online version -Diccionario de uso del español, CD-ROM version (Moliner 2008) -Diccionario del sexo y el erotismo (Rodríguez González 2011) Additionally, the process was supported by a Google search in both languages. By including not only animal terms that are documented in English and undocumented in Spanish (a), but also well-documented animal terms in Spanish (b) as well as animal terms that are undocumented in both languages (c), it is possible to put the results obtained from (a) in relation to likely highly accepted animal terms (b) as well as likely highly unaccepted animal terms (c). However, as indicated in Table 2, it is important to keep in mind that the metaphorical animal terms in (b) could be classified as conventional metaphors, while those in (a) and (c) could be classified as novel metaphors that follow mapping principles as animal metaphors of the respective classes of animals do exist in Spanish.
In the first questionnaire, the participants were asked to rate each animal term on a separate page on ten individual 7-point Likert scales from 0 to 6. The value 0 corresponded to no significa lo mismo en absoluto, no sería posible usar esta expresión en este sentido 'does not mean the same at all, it would not be possible to use this expression in this sense', the value 6 corresponded to totalmente significa lo mismo, sería posible usar esta expresión en este sentido 'totally means the same, it would be possible to use this expression in this sense'. The scales remained the same on every page, i. e. for every animal term. For every single scale the participants had the option to indicate no sé 'I don't know'. The following scales were included (in Spanish only): In the second questionnaire, for every animal term the participants were asked the same type of open question, for example, ¿Cómo describirías el tipo de mujer que es una musaraña? 'How would you describe a woman who is a shrew?', which they answered in an empty text box. For each animal term, the participants could instead indicate no sé 'I don't know' (see Figure 2).
All participants were Spanish native speakers from Spain. They were aged between 22 years and 72 years, with a mean age of 36.15 years. Overall, 112 participants took part in the study, with 59 in the first questionnaire (Likert scales) and 53 in the second (open questions). Among the participants of the first questionnaire were 34 women, 24 men, and one intersexual participant. Among the participants of the second questionnaire were 37 women, 15 men, and one participant who indicated otro 'other' when asked about their gender. Additionally, the participants were asked to indicate their level of English on a 5-point Likert scale, with 1 corresponding to very bad English skills and 5 corresponding to very good English skills. The mean value of the participants' English skills was 3.92 in the first questionnaire and 4.15 in the second questionnaire. The participation in the study was voluntary and the participants did not receive any compensation. The study was conducted in September and October of 2020.

Results
In this section, the results of the study are presented. Firstly, the results obtained from the first questionnaire (Likert scales) are shown and analysed, followed by the results obtained from the second questionnaire (open questions). This allows for both theoretical and methodological conclusions. It is important to note that no significant differences were found between the overall results of the participants of either questionnaire regarding the social variables (gender, age, English skills). However, in some cases, the participants' English skills do provide additional meaningful information. Table 3 shows the 20 animal terms with the highest mean values for a single character trait. Put differently, vaca 'cow' is the animal term which was rated the highest out of all the animal terms as referring to a fat woman more than any other one of the remaining nine character traits. As can be seen, with a mean rating of 5.71 it is closest to the high end of 6, which corresponds to 'totally means the same, it would be possible to use this expression in this sense'.  Unsurprisingly, the three well-documented Spanish animal terms (zorra 'vixen', vaca 'cow', lagarta 'lizard'; see 3) appear in the five highest ranks. Additionally, the two animal metaphors that neither exist in English nor Spanish (cigüeña 'stork', nutria 'otter') do not feature at all in the 20 highest entries. When looking at the entries in Table 3, it is striking that some character traits seem to clearly feature more than others. Accordingly, while vaca 'cow' to refer to a fat woman is the highest-ranking entry, the character trait fat does not appear apart from that. Instead, the single most frequently character trait appears to be ugly (8/20), followed by quarrelsome (4/20), promiscuous, and sexy (both 3/20). Another character trait with only one instance among the 20 highest entries is lesbian. Thus, the character traits good-looking, kind, old, and young do not feature at all in the 20 highest entries. This tendency also reveals itself when considering the mean values of all ten character traits for all 15 animal terms per character trait, as shown in Table 4.  As mentioned above, good-looking, young, old, and kind score the lowest mean values, indicating that the participants perceive the animal terms tested in this study to overall not refer to women who inherit these character traits. Instead, they perceive them to mostly refer to ugly, promiscuous, and quarrelsome women, followed by fat, sexy, and lesbian women. However, the overall mean values are relatively small, i. e. rather than approaching the high end of 6, they approach the low end of 0 which corresponds to "does not mean the same at all, it would not be possible to use this expression in this sense".

Rank
Apart from the most common character traits, Table 3 also gives an indication of the animal terms that are most highly rated as referring to a woman of a specific character trait. As mentioned above, the three well-documented Spanish animal terms (zorra 'vixen', vaca 'cow', lagarta 'lizard') appear in the highest ranks. This tendency remains the same when looking at the mean value of all the animal term data obtained from this questionnaire.  In other words, the participants perceive zorra 'vixen', lagarta 'lizard', and vaca 'cow' to be the three animal terms that most likely refer to certain types of women overall, while cigüeña 'stork', castor 'beaver', and perdiz 'partridge' are considered the least likely animal terms to refer to certain types of women. Other relatively highly rated animal terms include semental 'stud', cucaracha 'roach', cuervo 'crow', and trucha 'trout' -all animal terms that are documented to refer to certain types of women in English but not in Spanish (see 3).
It is worth looking at each animal term and its respective character traits in more detail. Figure  3 shows the participants' ratings of every animal term based on all ten character traits. It aligns with the data in Table 3. As can be seen, while some animal terms, e. g. zorra 'vixen', vaca 'cow', lagarta 'lizard', cucaracha 'roach', and semental 'stud', have high mean values for certain character traits, not a single character trait of the bird terms codorniz 'quail', perdiz 'partridge', and cigüeña 'stork' scores a mean value of 1 or higher. While none of the three birds seem to make for potential source concepts, there are interesting differences between them, after all: Based on the ten character traits, the participants rate codorniz 'quail' to most likely refer to an ugly (0.67), but also sexy (0.57) and good-looking woman (0.56), and least likely to a fat (0.30) and lesbian woman (0.30); perdiz 'partridge' to most likely refer to a young (0.60) and kind woman (0.58), and least likely to a lesbian (0.22), quarrelsome (0.22), and old woman (0.22); and cigüeña 'stork' to most likely refer to an ugly (0.59) and interestingly also a goodlooking woman (0.54), and least likely to a quarrelsome (0.12) and lesbian woman (0.13). In English, only quail and partridge are documented out of those three bird metaphors, with the former referring to a young woman and the latter to an attractive woman. Thus, the associations evoked by the English conventional metaphors and the Spanish translations differ in these cases.
Apart from the three well-documented animal metaphors zorra 'vixen', vaca 'cow', and lagarta 'lizard', the animal terms which score a mean value of 2 or more for at least one character trait are semental 'stud', cucaracha 'roach', trucha 'trout', and cuervo 'crow' (see Table 3 and Figure 3) -all documented to exist in English but not in Spanish. In comparison with the associations in English, the Spanish associations reveal interesting results: Both cucaracha 'roach' and cuervo 'crow' seem to align to a great extent with their English counterparts when it comes to the perceived character traits. The former refers to an unpleasant, despicable, ugly, and licentious woman in English, and is rated as referring to a quarrelsome and ugly woman in Spanish. The latter refers to an old and ugly woman in English, and is considered to refer to an ugly, quarrelsome, and old woman by the Spanish participants.
When it comes to semental 'stud' and trucha 'trout', however, the Spanish participants' associations seem to differ from those documented in English. While stud refers to a homosexual woman of stereotypically masculine identity or appearance in English, the Spanish participants perceive semental 'stud' to predominantly refer to a promiscuous and sexy woman. However, the character trait with the third highest mean value is in fact lesbian, although it is far behind promiscuous and sexy (see Figure 3). A possible reason for the Spanish participants' focus on sexual activity and attractiveness -instead of homosexuality -might be the association of a stud with frequent mating and high sex hormone levels. This topic is discussed in more detail in 5. Regarding trucha 'trout', the animal term is used in English to refer to an ugly, old, and ill-tempered woman. However, while the character trait ugly does have the second highest mean value, it is lesbian that scores the highest mean value by far (see Figure 3). This high rating likely occurs due to the fact that trucha 'trout' is a documented expression for a homosexual man -but not for a homosexual woman. It seems, then, that the participants extended this association to female homosexuality. This case is discussed in more detail in 5.
ISSN 1615-3014 Animal terms and character traits fat good-looking promiscuous sexy kind old quarrelsome lesbian young ugly Some additional interesting findings include the following: Apart from the character trait promiscuous, both zorra 'vixen' and lagarta 'lizard' -two well-documented animal metaphors in Spanish -also have high ratings for the character trait quarrelsome (see Figure 3). In fact, in the case of lagarta 'lizard', the scores of the character traits promiscuous (3.24) and quarrelsome (3.22) are nearly identical. Interestingly, the English vixen is documented to refer to an ill-tempered, quarrelsome woman. However, for both zorra 'vixen' and lagarta 'lizard', this part of the meaning seems to be novel and, so far, undocumented, as the following dictionary entries highlight: zorra 'vixen' (cf. Rodríguez González 2011: 1089) Referido a la mujer fácil, promiscua, muy laxa en lo tocante a la moral sexual; prostituta. 'Referring to an easy, promiscuous woman, very lax regarding sexual morality; prostitute.' -lagarta 'lizard' (cf. DRAE, s. v. lagarta) Persona pícara, taimada; prostituta. 'Naughty, cunning person; prostitute.' A Google search for zorra 'vixen' and lagarta 'lizard' in combination with cizañera 'quarrelsome' did not yield fruitful results either, suggesting that this character trait is in fact undocumented as part of the meaning of the two animal metaphors.
Furthermore, Figure 3 also nicely visualizes that the character trait ugly has the overall highest mean value (see Table 4). Out of the 15 animal terms, it scores the highest mean value in eight of them (codorniz 'quail', musaraña 'shrew', cucaracha 'roach', ratón 'mouse', castor 'beaver', murciélago 'bat', cuervo 'crow', cigüeña 'stork'). For three additional animal terms, ugly has the second highest score (vaca 'cow', trucha 'trout', nutria 'otter'). The question arises whether this tendency is due to the specific animals featuring in this study or whether this might be a more general tendency in Spanish animal metaphors. This is further discussed in 5.
Apart from vaca 'cow', the only other highly rated animal term regarding the character trait fat is nutria 'otter' -an undocumented animal metaphor in both English and Spanish. For both animal terms, fat has the highest and ugly the second highest rating, implying that being corpulent is regarded an unattractive and therefore negative feature. This has been previously shown to be the case for vaca 'cow ' (cf. López-Rodríguez 2009: 88).
While it is, of course, important to consider in detail the ratings that each animal term received regarding the character traits, the instances of participants indicating that they did not know which character traits to assign to an animal, i. e. where they chose no sé 'I don't know', offer some interesting insights too. Thus, while the three well-documented Spanish animal metaphors zorra 'vixen', vaca 'cow', and lagarta 'lizard' each unsurprisingly received only one indication of no sé 'I don't know' (each one for the character trait quarrelsome), the results for the remaining animal terms suggests significantly more insecurity among the participants (and for all character traits), as Figure 4 shows. x 10 (character traits), there are 654 instances of no sé 'I don't know', which equals around 7.39% of all data points. As can be seen in Figure 4, the participants indicated the most insecurity in the case of codorniz 'quail', with 139 instances of no sé 'I don't know', which means that nearly one quarter (23.56%) of all data points for this animal term are no sé 'I don't know'. For the Spanish translation musaraña of the highly conventionalized English shrew as a metaphor for a quarrelsome woman, participants indicated the second highest amount of no sé 'I don't know'. In fact, English skills seem to not have been beneficial in this case -quite the contrary: Participants who stated to have good or very good English skills indicated no sé 'I don't know' 1.73 times per musaraña 'shrew', compared to 1.05 times for those with bad English skills. What is striking in Figure 4 is the fact that the two animal metaphors that neither exist in English nor Spanish, nutria 'otter' and cigüeña 'stork', do not have the highest amounts of no sé 'I don't know', with nutria 'otter' even featuring in mid-range. This suggests that otter and stork might be equally suited for source domains in animal metaphor as other animals that are indeed documented to exist in English -or even more so than some of them, e. g. quail and shrew. In Japanese folklore, for example, otters are believed to be tricksters and shapeshifters: "They are said to emerge from the water and take human form, often that of a beautiful woman, in order to bewitch unwary travelers" (Toriyama/Yoda/Alt 2016: 17). It seems, then, that animal metaphors are to some degree arbitrary: Amount of no sé in Questionnaire I Certainly the real world provides a starting point for metaphor, but the choice of salient feature, and the significance attached to that feature, varies to such an extent as to appear arbitrary. […] [C]onventional metaphor is not natural but cultural. Once a perceived similarity between two entities is codified, that similarity may even cease to exist, yet the metaphor will remain meaningful. (Nesi 1995: 276) Apart from the three well-documented Spanish animal terms zorra 'vixen', vaca 'cow', and lagarta 'lizard' (no sé: 0.17%), the participants were most sure about cucaracha 'roach' (no sé: 2.37%), followed by semental 'stud' (no sé: 3.56%), trucha 'trout' (no sé: 3.73%), and ratón 'mouse' (no sé: 4.24%). Interestingly, when it comes to the insecurity about the animal terms featured in this study, it appears not to matter which animal class the animal belongs to, whether the animal is a farm animal/edible animal or a wild animal/inedible animal, and whether or not the animal metaphor is documented in English.
After having highlighted the most noteworthy results obtained from the Likert scales of the first questionnaire, it is time to explore the findings from the second questionnaire employing open questions instead.

Questionnaire II: Open questions
Starting, in this section, with the participants' indication of no sé 'I don't know' per animal term, it becomes clear that the results overlap with those of the first questionnaire to a large extent, though not in all cases, as can be seen in Figure 5, in which the ascending order of Figure  4 has been kept. It is important to note that the overall numbers are smaller in this case as the participants in the second questionnaire could only indicate no sé 'I don't know' once per animal term (instead of answering the question in the empty text box).
Again, unsurprisingly, the participants are least unsure about the well-documented Spanish animal metaphors zorra 'vixen' (no sé: 0), vaca 'cow' (no sé: 0), and lagarta 'lizard' (no sé: 2). This is followed by cucaracha 'roach' (no sé: 8) and semental 'stud' (no sé: 17), just like in the first questionnaire. However, the next animal term is not trucha 'trout' (no sé: 27), but cuervo 'crow' (no sé: 18) and then ratón 'mouse' (no sé: 19). A possible explanation could be that in the first questionnaire the participants were presented with the ten character traits, including lesbian which they indicated to be most prominent. Without any leads as to what type of woman could be referred to as trucha 'trout', it appears more participants were unsure about the animal term. This is discussed in more detail in 5.
When comparing the results of this questionnaire with the meanings of the ten animal metaphors that are documented in English but not in Spanish, it is apparent that most animal metaphors convey different meanings in the two languages, as Table 8 indicates.  Thus, the Spanish participants' associations with the animals quail, shrew, mouse, beaver, trout, crow, and partridge do not match those documented in English. In fact, in some cases the disparity seems striking: For example, while in English a woman referred to as shrew is considered to be ill-tempered, malignant, aggressive, and quarrelsome, the Spanish associations evoke thoughts of a rather harmless woman (quiet, small, sleepy). Similarly, while the English beaver refers to a sexy woman -a metonymic extension of beaver referring to the female genitals or pubic area due to the supposed similarity in hairiness -, the Spanish equivalent seems to remind the participants of an ugly woman with big teeth. There are, however, three close or partial matches between the two languages. Firstly, a woman referred to as roach is seen as an unpleasant, despicable, ugly, and licentious woman in English, while the Spanish participants associate similar character traits with this animal (despicable, disgusting, bad, dirty). Secondly, a homosexual woman with a stereotypically masculine identity or appearance is referred to as stud in English. Although the Spanish participants do not indicate associations with homosexuality, they do list stereotypically masculine traits, e. g. sexually potent, competent, 12 coarse, and masculine. Finally, among other things, the English animal term bat refers to an ugly woman or a prostitute who walks the streets at night, which corresponds to the Spanish participants' associations of murciélago with an ugly and nocturnal woman. An interesting finding concerns the participants' English skills. Each of the following character traits -representing a match between the English and the Spanish animal term -was in each case mentioned by one participant with very good English skills: ugly and licentious (cucaracha 'roach'), homosexual (semental 'stud'), ugly and unkind (trucha 'trout'), ugly and old (cuervo 'crow'), and elegant and dolled up (perdiz 'partridge'). This could be a sign of possible L2 lexical inference.
Overall, there does not appear to be a clear correlation between the type of animal and whether or not the Spanish associations correspond to the English meanings of the respective animal metaphors. Thus, the three animals with matches between English and Spanish are one insect (roach) and two mammals (stud, bat), even though a bat is certainly a highly non-prototypical mammal as it is the only mammal capable of flying, which is heavily associated with birds. There are no matches between English and Spanish for the remaining three mammals (shrew, mouse, beaver), any of the birds (quail, crow, partridge) or the fish (trout). Furthermore, when it comes to matches between English and Spanish, it seems not to matter whether the animal is a farm animal/edible animal (stud) or a wild animal/inedible animal (roach, bat). The same seems to be true regarding the participants' degree of insecurity about the possible animal metaphors (see Figure 5).

Discussion
In this section, the most interesting results and patterns detected in section 4 are discussed in more detail in order to draw both theoretical and methodological conclusions. Firstly, when it comes to the three well-documented Spanish animal metaphors zorra 'vixen', vaca 'cow', and lagarta 'lizard' to refer to certain types of women, it is little surprising that the participants in both questionnaires showed no insecurity about the respective meanings and indicated them in accordance with the respective documented meanings. However, the results of both questionnaires revealed that even well-documented animal metaphors might have more multidimensional meanings than those recorded. Accordingly, while both zorra 'vixen' and lagarta 'lizard' are documented as metaphors for a promiscuous and cunning woman and were indeed rated as such by the participants, they were both indicated to refer to additional types of women. In the Likert scale questionnaire, both animal terms were indicated to refer to a quarrelsome and sexy woman, and in the questionnaire with the open questions, zorra 'vixen' was reported to refer to a bad, malicious, selfish, intelligent, perverse, hurtful, lewd, unscrupulous, and disrespectful woman, and lagarta 'lizard' to a bad, selfish, exploitative, mysterious, unreliable, careerist, false, malicious, deceitful woman who talks behind people's back and is interested in unavailable men. On the one hand, this shows that "[a]nimal metaphors […] are multi-faceted" (López-Rodríguez 2016: 83) and that personality and character traits consist of more than only one or a few dimensions (cf. Corr/Poropat 2016: 27). On the other hand, it might suggest that dictionaries tend to display only the most prominent feature(s) of multidimensional personality types, disregarding features that form part of a certain personality type to a lesser degree. Methodologically, the two questionnaire types used in this present study -but especially the open question one -can help uncover the many facets of personality and character traits expressed through animal metaphor as well as provide more detail in this regard compared to dictionaries.
When it comes to the animal terms referring to different types of women that are documented to exist in English but not in Spanish, semental 'stud' is an interesting case. As shown in 4, participants of the Likert scale questionnaire rated this animal term as referring to a promiscuous and sexy woman (see 4.1), while participants of the second questionnaire indicated that this animal term is either not used to refer to women or that they associate it with promiscuity, reproduction, attractiveness, competency, masculinity, transsexuality, coarseness, and sexual appetite, potency, and activity (see 4.2). In English, stud can refer to both a sexually successful, particularly masculine young man and a lesbian who adopts a stereotypically masculine identity or appearance. Overall, the participants seem to associate masculinity as opposed to homosexuality with this animal -even though semental 'stud' is the animal term that has the second highest score for homosexual in the Likert scale questionnaire, after trucha 'trout' in first place. An explanation for the participants' focus on masculinity over female homosexuality could lie in stereotypicality: The stereotypical person of masculine appearance and behaviour, which the animal stud alludes to, tends to be a heterosexual man rather than a homosexual woman. In Spanish, this tendency is intensified further by the fact that semental 'stud' is derived from semen 'semen'.
The animal term that was clearly indicated to be referring to a homosexual woman in both questionnaires is trucha 'trout'. In English, trout is not documented to refer to homosexuality. As mentioned in 4, the general association of that animal with homosexuality in Spanish is likely due to the common usage of trucha 'trout' to refer to a homosexual and effeminate man. This suggestion is supported by the fact that two participants in the second questionnaire indicated that this animal metaphor is only used to refer to men. Fish and seafood are known to serve as source concepts for the metaphorical conceptualization of women and their genitalia, arguably due to the supposed similarity in slipperiness and smell (cf. Allan/Burridge 2006: 195), and that by metonymic extension male homosexuality is associated with femininity (cf. Crespo-Fernández 2015: 174). When it comes to the specific case of trucha 'trout', it has been suggested that its usage as an expression for a homosexual man has been popularized by the Spanish television series Los Serrano in 2003 and that it might be motivated by the association of the rainbow trout with the rainbow flag as an LGBTQ+ symbol (cf. Moscas de Colores 2020). In any case, it seems that in the present study, by a second metonymic extension, male homosexuality provides mental access to female homosexuality and causes the participants to disregard most of the other options -regardless of which questionnaire the participants were presented with.
However, while the participants of both questionnaires indicated that trucha 'trout' refers to a homosexual woman, this trend was much higher in the Likert scale questionnaire compared to the open questions -and this is not only the case for trucha 'trout'. Overall, the results reveal a clear discrepancy when it comes to the participants' insecurities about the animal metaphors.
When presented with possible character traits, it seems that the participants are more inclined to accept one or multiple options. On the contrary, when required to fill in an empty text box, it appears the participants are generally more unsure about the potential meanings of the animal metaphors. When comparing the amount of times that the participants indicated no sé 'I don't know' out of all data points, it is 7.39% in the Likert scale questionnaire and 39.62% in the questionnaire with the open questions. Of course, the approach using the Likert scales provides more quantifiable and comparable data and less blank values, but the open questions approach offers a wider picture of the multidimensionality of personality types and appears to sway the participants less. Something that both methodologies have in common, though, is that they highlight how multifaceted the associations conveyed in animal metaphors are and how difficult it is to narrow them down for concise dictionary entries.
All in all, this present study found that the animal metaphors that are documented to exist in English but not in Spanish as referring to certain types of women are for the most part not meaningful to the Spanish participants (see Figures 3 and 5) and do, to a large extent, not match with the English meaning when translated into Spanish. There are some partial or close matches (roach, stud, bat) that seem to largely convey the same meaning in both languages, but the remaining animal metaphors do not appear to correspond. However, from a broader perspective, there is some congruence between the English and Spanish animal metaphors in that -specifically in the questionnaire employing open questions -animals that are associated in English with what is stereotypically considered positive features in women, such as small size, young age, inferiority (cf. López-Rodríguez 2009: 95), tend to also be considered positively in Spanish; and vice versa in the case of negative associations, such as large size, old age, strength, promiscuity. Accordingly, mouse, quail, and partridge seem to convey positive and crow negative connotations in both English and Spanish.
Another important finding concerns the fact that the big majority of the tested animal metaphors refers to women pejoratively. This is the case for the documented and undocumented animal terms as well as for both types of questionnaires (Likert scales and open questions). Of course, for the well-documented animal metaphors zorra 'vixen', vaca 'cow', and lagarta 'lizard' this is to be expected as their associations are known to be negative (promiscuity, obesity, ugliness). When it comes to the undocumented animal metaphors, though, there is a clear pattern in the Likert scale questionnaire to predominantly associate various animal terms with ugliness: Out of the twelve undocumented animal terms (ten documented in English, two undocumented in English and Spanish), the participants rate ugliness as the number one trait in eight animals and as the number two trait in two animals. This applies for both 'prototypically' unsightly animals like roach, beaver, and bat, as well as animals that are not generally tied to unsightliness such as quail, stork, and mouse. It seems then that, first and foremost, novel animal metaphors provide mental access to ugliness, mostly regardless of the animal species. This might be due to the fact that ugliness is a concrete physical and visual factor which tends to be easier to grasp than abstract character traits denoting a woman's personality. In other words, ugliness might be the default association of novel animal metaphors applied to women, while other associations (e. g. kindness, quarrelsomeness, homosexuality) might be less accessible and establish with increasing metaphor conventionality. Additionally, as suggested by the Great Chain of Being (see 2.2), animal metaphors are in any case more likely to link women to negative features.
Accordingly, in the second questionnaire (open questions), seven out of the twelve undocumented animal metaphors are endowed by the participants with rather negative traits (e. g. despicable, promiscuous, malicious, scatter-brained), while four are generally rated as positive (ratón 'mouse', cigüeña 'stork', codorniz 'quail', perdiz 'partridge') and one as neutral (trucha 'trout'). Similarly, in the Likert scale questionnaire the character traits that were most often indicated to be associated with the animal terms are generally negative (ugly, promiscuous, quarrelsome, fat), with the positive ones featuring less (kind, young, good-looking, sexy) (see Table 4).
While some results of the present study conform to findings of previous studies, others do not. For example, as argued by López-Rodríguez (2009: 95), it seems to be the case that names of wild animals -as opposed to pets and farmyard animals -are consistently endowed with negative connotations. This claim can be substantiated when looking at the results from the Likert scales questionnaire, but not in its entirety when looking at the answers to the open questions.
In the former case, the highest-rated character trait of each animal is one that conveys negative connotations, with ugly as the most common one. However, it must be mentioned at this point that, in the Likert scale questionnaire, the highest-rated character trait of all 15 animal terms (including farmyard and edible animals), except for perdiz 'partridge', is a negative one. This, then, seems to partially contradict López-Rodríguez' (2009) overall findings. Additionally, when it comes to the open questions, the results do not fully support her claims either as all animal terms are indicated as having mostly negative character traits, with the exception of the two edible animals codorniz 'quail' (small) and perdiz 'partridge' (happy) as well as the two inedible wild animals ratón 'mouse' (small) and cigüeña 'stork' (tall). Again, it seems that the methodology chosen in this present study involving questionnaires employing Likert scales and open questions tends to yield results that differ significantly from dictionary-based approaches.
An interesting observation from this present study concerns one participant's associations with the animal term murciélago 'bat' in the second questionnaire (open questions). When presented with this animal metaphor and asked about the type of woman it could refer to, the answer was the following: The participant is, of course, alluding to the global COVID-19 pandemic that began in late 2019 and is believed to have spread from bats to humans. The participant's association of a woman referred to as bat with the pandemic could be an indication as to how influential current affairs and developments in our environments can be on our understanding of animal metaphors. However, considering the study was conducted at the height of the pandemic but only one participant addressed it, it shows how stable people's associations with animal species tend to be: "In our metaphorical language [animals] stand for a limited number of slots, that is to say features which can be projected upon the target. They are pre-established by a long tradition" (Bisschops 2019: 1).

Conclusion
It has been shown in this paper that animal metaphors for different types of women that are documented to exist in English but not in Spanish convey, to a large extent, different meanings in the two languages when translated into Spanish and judged by native speakers. This appears to be the case regardless of the type of animal (mammal vs. bird etc.; farm/edible vs. wild/inedible animal). Additionally, it seems that novel animal metaphors denoting women are by default interpreted as referring to ugliness, regardless of whether the animal, i. e. the source concept, is generally considered an unsightly animal. This appears to be a new finding and it would be interesting to test if it holds true when investigating a larger amount of novel animal metaphors.
Furthermore, while the well-documented Spanish animal metaphors revealed no insecurities among the participants regarding their meaning, there were significant insecurities in several cases of the undocumented metaphors. However, since the documented English animal metaphors were chosen for this study based on a dictionary search, it would be interesting to investigate how well-established they really are. It is possible that English native speakers would in fact associate some of the animal terms with different character traits than those documented in the dictionaries consulted for the present study, revealing thus more similarities between Spanish and English speakers' associations. Additionally, it seems likely that the ten Likert scales in the first questionnaire were not numerous enough for the participants to precisely indicate their associations with the different animal terms as personalities are extremely multidimensional -something that the answers provided in the second questionnaire employing open questions highlighted. Such methodological refinements, then, could yield interesting divergent results. In any case, the combination of the two types of questionnaires (Likert scales and open questions) seems to be a fruitful design to approach novel animal metaphors multi-methodologically. Furthermore, instead of focussing on animal terms, follow-up studies could employ animal pictures to be presented to Spanish and English native speakers and have them assign the animals to a variety of character traits. This would shed light on potential differences between an animal concept and the respective animal term. Moreover, as opposed to investigating one or two languages, future research could focus on large-scale cross-cultural studies on novel animal metaphors, incorporating a variety of different languages and cultures.