Variation in embodied metaphors: a contrastive analysis of taste metaphors in Spanish and English

Whereas in the late 90s the universal character of many embodied conceptual metaphors was overemphasised, in the last years some authors have claimed that culture plays a crucial role in the motivation of all kinds of conceptual metaphors, including those grounded on universal bodily experiences. In order to shed some light on this issue, we carry out a contrastive analysis of conceptual metaphors with basic tastes as a source domain in Spanish and in English. To this end, we employ a mixed approach, combining data from dictionaries and linguistic corpora. Our analysis reveals that variation is higher at the level of linguistic expression and lower, but still significant, at the conceptual level. Although most taste metaphors are shared by Spanish and English, a few language-specific conceptual metaphors are also found, proving that food culture has an influence on the motivation of conceptual metaphors.

tended to overemphasize the universal character of some embodied conceptual metaphors. For example, the conceptual metaphor understanding is seeing, which was classified as a primary metaphor (Grady 1997), was also considered universal, since sight is the main source of objective data for every human being (Sweetser 1990: 38-39). However, later works demonstrated that a significant number of languages from America and Oceania do not present this conceptual metaphor, but understanding is hearing (Evans & Wilkins 2000;Ibarretxe-Antuñano 2008).
In the 2000s, several cross-linguistic and contrastive studies on conceptual metaphors showed that culture plays an essential role in conceptual metaphors, even on those motivated by sensorimotor experience. Some well-studied conceptual metaphors across languages are those with body-part vocabulary as a source domain (Gutiérrez Pérez 2008;Yu 2009;Maalej & Yu 2011), perception verbs as a source domain (Ibarretxe-Antuñano 1999;Viberg 2008) and emotions as a target domain (Pavlenko 2002;Kövecses 2003;Soriano 2005). In the last years, the interest on variation of conceptual metaphors has increased, both across cultures (Ogarkova, Soriano & Gladkova 2016;Belkir 2019) and within cultures (Callies & Onysko 2017;Littlemore 2019). The new evidence has led up to claim that culture contributes to shape every conceptual metaphor. In this sense, Caballero and Ibarretxe-Antuñano (2009) coined the term cultural sieve, since they defend that culture models and validates the sensorimotor experience grounding conceptual metaphors. This model allows explaining the fact that universal bodily experiences can originate different conceptual metaphors across cultures, as in the case of understanding is seeing and understanding is hearing (Ibarretxe-Antuñano 2013).
Adopting a strong understanding of embodiment (Sweetser 1990;Grady 1997), conceptual metaphors with taste terms as a source domain might be a clear case of primary metaphor, since they are grounded on a given physical experience, which correlates directly with an emotional experience (Grady 1997: 89). From this perspective, taste metaphors are good candidates to be universal, because all human beings share the same physiological characteristics that allow us to taste (Bagli 2016). However, from a culture-defending point of view, as the one in Caballero and Ibarretxe-Antuñano (2009), the gastronomic experience in one culture might give rise to culture-specific conceptual metaphors. In order to shed some light to this debate, in this work we analyse the variation of conceptual metaphors and metaphorical expressions with taste descriptors as a source domain in Spanish and English. Although there is already a significant number of cross-linguistic studies on conceptual metaphors with perception-related terms, such as colours (Soto Nieto 2011), temperature terms (Valiulienė 2015) and perception verbs (Ibarretxe-Atuñano, 1999), the cross-linguistic variation of metaphors with taste descriptors as a source domain remains an under-explored area. It might be due to the fact that taste, as well as smell and touch, has been traditionally considered a "lower" sense (Speed & Majid 2019). However, there is currently an incipient interest on taste metaphors. Marco Bagli (2016; carries out an exhaustive analysis on taste metaphors in English, showing the intricate polysemy of taste descriptors and the difficulties to explain it systematically. To our knowledge, the only cross-linguistic study on taste metaphors is authored by Zawisławska and Falkowska (2018), who compare the metaphorical meanings of the basic taste adjectives in English and Polish, showing a certain degree of variation in the semantic frames in which metaphorical expressions are used. Inspired by Zawisławska and Falkowska, with this work we attempt to find out to what extent culture influences the conceptual metaphors with taste descriptors as a source domain in Spanish and English.

Levels of variation in conceptual metaphor
Since conceptual metaphors are complex cognitive associations showing hierarchic structures, contrastive analyses must consider the fact that variation can take place at different levels. However, variation does not occur similarly in every kind of conceptual metaphor, but asymmetrically, depending on the degree of specificity of the metaphorical mapping analysed. According to Onysko (2017), variation is expected to be lower in primary metaphors mapping basic and experienced concepts, while the degree of variation increases in complex metaphors and elaborations of conceptual metaphors. The maximum degree of variation is assumed to occur in the surface of language, in metaphorical expressions. Figure 1 shows the continuum of variation in conceptual metaphor, according to Onysko (2017). A set of parameters of variation of metaphorical language was first proposed by Barcelona (2001) and then developed by  and Kövecses (2005: 117-162), who offers an extensive description of them. In our study, we cover variation both at the conceptual and at the linguistic level, in order to check if the variation of taste metaphors in Spanish and English fits in the model proposed by Onysko (2017). At the conceptual level, we consider 1) the existence or not of the same conceptual metaphor in both languages (Barcelona, 2001: 137) and 2) the saliency of conceptual metaphors, that is, the productivity of a given mapping in each language . At the linguistic level, we consider 3) the grammatical status and features of the metaphorical expressions of a given conceptual metaphor (Barcelona 2001: 138) and 4) the semantic frames in which metaphorical expressions are employed, which can be defined as chunks of knowledge structure (Fillmore 2006). In studies on conceptual metaphor, frames are considered more specific conceptual structures than cognitive domains (Dancygier & Sweetser 2014: 18-19). Thus, the use of frame semantics in our analysis pursues a deeper understanding of the use of metaphorical expressions and makes our results comparable to those in Zawisławska and Falkowska (2018).

Methodology
From the physiological point of view, taste is a so-called chemical sense that allows us to distinguish only a few stimuli, mainly sweet, salty, sour and bitter, although several authors have recently argued for the existence of other tastes, such as umami and oleogust (see Bagli 2018). However, when eating or drinking, human beings feel a multimodal experience, grounded on taste, smell, temperature, texture and spiciness. This is why, in this study, when we use the term taste descriptors, we refer to adjectives usually employed to speak about the general experience of savouring, without considering the different senses involved in this activity. Marco Bagli (2018) carries out an experimental study with a group of informants on the categorization of taste descriptors in English. In light of the results, he argues that the five basic taste terms in English are sweet, salty, bitter, sour and spicy (2016: 62). In our study, we decided to analyse these five lexical units, as well as their Spanish equivalents: dulce, salado, amargo, agrio, ácido and picante. While in English sour is the main adjective encoding one basic taste, in Spanish we find two equivalent adjectives: ácido and agrio, which have a similar meaning. In the absence of an experimental study similar to Bagli's on the Spanish language, we decided to cover both lexical units in our analysis.
Methodologically, this study uses a mixed approach combining corpus data and information from dictionaries, since it allows us to benefit from the advantages from both approaches (Kövecses et al. 2019). On one hand, we analysed 700 occurrences from the Corpus of Contemporary American English (COCA) and 800 occurrences from the Corpus del Español del Siglo XXI (CORPES XXI), concretely 100 occurrences of each selected taste descriptor. The two corpora contain synchronic texts of a variety of genres, all of which have been considered. On the other hand, 8 prestigious dictionaries have been examined. For the English language, we employed the Oxford English Dictionary, the Cambridge English Dictionary, the Collins Online English Dictionary and the Macmillan English Dictionary. For the Spanish language, we used the Diccionario de la Lengua Española, the Diccionario de Uso del Español, the Diccionario del Español Actual and the Clave: Diccionario de Uso del Español Actual. Both the corpora and the dictionaries are referred at the end of this paper.
In the dictionary-based analysis, we attempted to identify the conceptual metaphors existing in Spanish and English, and we also took into account the information related to the varieties of use of the metaphorical meanings, such as register or dialect. In our corpus-based analysis, we classified and quantified the metaphorical expressions according to the conceptual metaphor to which they responded, and according to the semantic frame in which they were used. We also compared qualitatively some grammatical features of the metaphorical expressions in our corpora. In order to tag the occurrences as metaphorical or not, we applied the MIPVU procedure (Steen et al. 2010). By this procedure, we took into consideration the contextual meaning of the lexical unit and then checked if there was a more basic meaning (related to the senses, historically older, etc.).

Sweet taste metaphors
Our data show that sweet taste is employed very frequently in metaphorical expressions both in Spanish and in English. Firstly, we present in table 1 the quantitative results of our corpus-based analysis of sweet taste metaphors. In other words, we show the number of occurrences corresponding to a conceptual metaphor, and we also quantify the occurrences of that given conceptual metaphor which are employed in a certain semantic frame. All the metaphorical expressions identified in Spanish and in English respond to the conceptual metaphor Pleasure is sweet taste, which links the sweet taste to likeable emotions and feelings. Although this conceptual metaphor is quite productive in both languages, our corpus analysis suggests that it might be slightly more salient in English than in Spanish, since we have identified 74 occurrences in our corpus in English and 65 occurrences in our corpus in Spanish.
At the linguistic level, the conceptual metaphor Pleasure is sweet taste is most often employed in the semantic frame PeoPle both in Spanish and in English, expressing the quality of having a kind character, e. g. sweet boy / chico dulce. However, according to our data, this is noticeably more frequent in American English (34 occurrences) than in Spanish (16 occurrences). We also noticed that, when referring to people, sweet is usually employed in English next to a noun in vocative statements, as shown in example 1, as well as in fixed expressions, e. g. sweet Jesus!, while in Spanish this seems to be rare, considering that we do not find any similar case in our corpus.
(1) She held me close for moment. Then she said, "My sweet boy -I will miss you more than you know." 2010 Caldwell, Bo: City of Tranquil Light.
Sweet taste is also employed metaphorically in other frames both in Spanish and in English, such as mental aCtivities, e. g. sweet memories / recuerdos dulces and CommuniCation, e. g. sweet words / palabras dulces, among others. However, even in those cases, the adjectives sweet and dulce do not always behave similarly, from the perspective of grammar. For example, both in English and Spanish, sweet taste is metaphorically used to describe events, e. g. the victory was sweet / la victoria fue dulce. In English the adjective sweet is often employed independently to describe an event which is happening in the situational context where the speech takes place, as shown in example 2, while in Spanish this does not seem to be possible.
Sweet taste metaphorical expressions are usually employed in a given frame more frequently in one language than in the other. An example is the case of Body Parts and FaCial exPressions, e. g. sweet look / mirada dulce, which seems to be more salient in Spanish (10 occurrences) than in English (3 occurrences). Moreover, our analysis reveals that sweet taste is metaphorically employed by Spanish speakers in some frames that English speakers do not, and vice versa. For example, while in Spanish it is frequent to use dulce to speak about the weather, e. g. clima dulce, our data show that in contemporary English this is rare. Another case is the frame of mediCal interventions. Our dictionary-based analysis reveals that, in Spanish, dulce is employed to describe medical techniques, indicating that they are less harmful than conventional ones, as we can see in example 3, whereas in English sweet is not used similarly.
(3) La acupuntura fue la primera técnica dulce extendida en occidente. Retrieved from: Diccionario del español actual. Lit. 'Acupuncture was the first sweet technique widespread in the Western world'.

Salty taste metaphors
According to our data, salty taste functions as a source domain in metaphorical expressions in Spanish and in English. However, compared to the rest of the tastes analysed, salty taste is much less productive in terms of metaphorical language. In order to illustrate this, in the table below, we show the quantitative data of our corpus-based analysis of salty taste metaphors. As we can see in table 2, we have found only 6 occurrences in English and 3 in Spanish in which the salty taste is employed in metaphorical expressions as a source domain, which is a very small amount in comparison to other tastes. All the metaphorical expressions found in our corpus in English respond to the conceptual metaphor anger is salty taste, which allows us to speak about the emotion of anger, usually considered to be negative, by using gustatory terms. On the other hand, in all the metaphorical expressions identified in Spanish, salty taste refers metaphorically to hilarity, understood as a complex emotional phenomenon caused by humorous stimuli, thanks to the conceptual metaphor hilarity is salty taste. In this case, variation takes place at the conceptual level, giving rise to two culture-specific conceptual metaphors, even if none of them is particularly salient.
Our corpus analysis shows that, in English, anger is salty taste is expressed linguistically in two semantic frames. The most frequent is the frame CommuniCation, e. g. salty words. In this case, salty indicates the quality of being rude or potentially offensive, and it usually refers to insults, as we can see in example 4. On the other hand, salty also appears in the frame PeoPle, e. g. salty character, referring to someone's irritated mood. The Cambridge English Dictionary (online) specifies that it is specially employed when the angry mood is unreasonable. In this sense, salty is usually preceded by get or jump, as we see in example 5, indicating the beginning of the angry mood.
(4) And then he says Cabron, which is kind of a salty word in Spanish. 2010, NPR TalkNat (radio program).
He got salty with me because I wouldn't go out with him. Retrieved from: Cambridge English Dictionary (online).
Regarding the Spanish language, all the occurrences of the conceptual metaphor hilarity is salty taste in our corpus are located in Spain, so it is likely to be exclusive of the European Spanish, although the dictionaries do not indicate any geographical limitation. Most metaphorical expressions of hilarity is salty taste take place in the frame PeoPle, e. g. persona salada, expressing the quality of being funny and amusing, as we can see in example 6. Another possibility is the frame CommuniCation, e. g. comentario salado. In this case, salty indicates that someone's intervention is funny, as in example 7.
(6) Pedrita, que es una pulga muy salada a la que Gloria Fuertes acude a menudo, un buen día conoce al perro Picatoste y es feliz con él. 2010 Sáiz Ripoll, Anabel: Espéculo. Lit. 'Pedrita, who is a very salty flea to which Gloria Fuertes usually turns, one fine day meets the dog Picatostes and is very happy with him'.

Bitter taste metaphors
Both in Spanish and in English, metaphorical expressions with bitter taste as a source domain are very frequent, even more than sweet taste metaphorical expressions. In table 3, we show the quantitative results of our corpus-based analysis, that is, the number of occurrences responding to a conceptual metaphor and the quantity of occurrences that are employed in each frame. Our analysis reveals that English and Spanish share the conceptual metaphor aFFliCtion is Bitter taste, by means of which we conceptualise emotional suffering and, to a lesser extent, physical pain, in terms of a bitter gustatory perception. The high number of occurrences of aFFliCtion is Bitter taste in our corpora shows that this conceptual metaphor is quite salient in both languages, although it is slightly more salient in English (86 occurrences) than in Spanish (68 occurrences).
However, there seems to be subtle differences in the target domain activated by the conceptual metaphor aFFliCtion is Bitter taste in Spanish and in English. While in Spanish the target domain that we have called aFFliCtion is restricted to sadness, in English the target domain is wider, since bitter can be employed to speak about sadness, but also about resentment and even anger, whenever it is a deep emotion causing intense suffering. A representative metaphorical expression of aFFliCtion is Bitter taste in Spanish is shown in example 7, in which the adjective amarga is employed to describe a sad life. Example 8 illustrates a similar use of bitter in English, since the adjective refers to a sad and disappointing moment in history. On the other hand, in example 9, the adjective bitter refers to a feeling of resentment and rage that is typical in revenge and that is shown through the character's aggressive intervention, whereas in Spanish we do not find similar examples.
Habría sido una vida como la de muchas otras personas en el mundo: algo amarga, tristona, pero una vida normal al fin y al cabo. We also observed similarities and differences on the combination of aFFliCtion is Bitter taste with the conceptual metaphor aCCePting is swallowing. Both in Spanish and in English these conceptual metaphors are often combined, giving rise to the complex metaphor a diFFiCult exPerienCe to aCCePt is a Bitter swallow. However, while Spanish speakers employ the general noun trago ('swallow') (3 occurrences), in English this metaphor is always concretised by the more specific term pill (5 occurrences). Examples 10 and 11 illustrate two metaphorical expressions of a diFFiCult exPerienCe to aCCePt is a Bitter swallow in Spanish and in English, respectively.
Lit. 'Maybe we are unconsciously trying to sweeten the bitter swallows that life gives us'. (11) Not being able to tag him was a bitter pill to swallow. 2010: OutdoorLife (magazine).
Regarding the level of linguistic expression, our analysis shows that most frames in which bitter taste metaphorical expressions are employed are the same in both languages. For instance, the adjectives amargo and bitter are employed to describe events, e. g. bitter experience / experiencia amarga, which is the most frequent frame both in English (29 occurrences) and in Spanish (19 occurrences). In the case of the frame PeoPle, e. g. bitter man / hombre amargo, we appreciate that it is significantly more frequent in English (19 occurrences) than in Spanish (5 occurrences), although it might be due to the existence in Spanish of the derivate form amargado, which we have not considered in this work. Our analysis also shows that bitter taste metaphorical expressions are employed in some frames only in English and not in Spanish, and vice versa. One example is the frame weather, e. g. bitter cold, in which the adjective bitter indicates an extreme cold, whether in Spanish amargo is not used in such way. Another case is the frame hostile enCounters, e. g. bitter fight, which appears only in English, probably because in this frame bitter taste is associated to angry feelings, which is not conventional in Spanish.
Lastly, we found some metaphorical expressions which are typical of only one language. This is the case of until the bitter end (4 occurrences), which refers to the fact of standing an emotionally painful situation, but does not present any equivalents in Spanish. Another example is the metaphorical expression dejar algo un sabor amargo en la boca ('to leave something a bitter taste in the mouth'), which is relatively frequent in Spanish (4 occurrences), but is rare in English, according to our data.

Sour taste metaphors
For the analysis of source taste as a source domain, we have considered one adjective in English, sour¸ and two in Spanish, ácido y agrio, since both of them are frequently employed to describe the basic taste that in English is encoded by sour. Our analysis shows that these three adjectives are employed in metaphorical expressions, but not to the same extent. Table 4 contains the quantitative results of our corpus-based analysis. According to our data, all the metaphorical expressions with the sour taste as a source domain respond to the conceptual metaphor disPleasure is sour taste, since they express unpleasant emotions and feelings. In Spanish, ácido and agrio appear both in metaphorical expressions, but ácido is employed metaphorically more often (73 occurrences) than agrio (36 occurrences). In general terms, our corpus analysis suggests that disPleasure is sour taste is more salient in Spanish than in English (34 occurrences).
At the linguistic level, we can affirm that sour taste metaphors are employed in the same semantic frames in some cases, such as the frame PeoPle, e. g. sour man / hombre ácido / hombre agrio, in which sour taste descriptors indicate a disagreeable mood or personality. However, there are also other frames that are exclusive of one language. It is the case of the frame eConomy, e. g. sour economy, in which sour indicates a time of economic depression, as we can see in example 12, while in economic-financial Spanish sour taste descriptors are not employed in a similar way. The case of the frame CommuniCation is also very relevant, e. g. comentario ácido / comentario agrio (lit.: sour comment), since in Spanish it appears very frequently in our corpus (12 occurrences of ácido and 29 occurrences of agrio), but we have not found any similar case in our corpus in English. In this frame, ácido usually refers to criticisms (5 occurrences) and jokes (3 occurrences) that can offend sensibilities, as we show in example 13, while agrio usually refers to arguments (9 cases) or debates (4 cases), as illustrated in example 14. Lit. 'It was a sour criticism, with a lot of black humor, of machismo'. (14) Eso mismo he preguntado yo, y nos hemos enzarzado en una discusión muy agria. No quiero entrar en detalles, pero te aseguro que la reunión ha sido muy desagradable. 2010 Bordas, Jordi: Sin cobertura. Lit. 'This is what I asked, and we got involved in a very sour argument. I do not want to go into any detail, but I assure you that the meeting was very disagreeable'.
There are also some differences on the combination of the sour taste descriptors with other lexical units in metaphorical expressions. In English, we can find very often the collocations go sour (5 occurrences) and, to a lesser extent, turn sour (1 occurrence), expressing the deterioration of a situation, as we illustrate in example 15. In Spanish, however, we only found one similar case, concretely volverse ácido ('turn sour'), that we show in example 16, in which the collocation volverse ácido refers to the process of becoming an unkind person. The greater tendency of collocate with verbs of change in English might be due to the existence of the derivate verb agriar ('become sour') in Spanish, although this form is excluded of our work.

Spicy taste metaphors
Spicy is the only flavour that cannot be considered a taste from the physiological point of view, since it is not perceived by the tongue. More concretely, it is an irritation of the mucosae (Bagli 2016: 27), so it can be classified as a kind of tactile perception. However, according to Bagli, spicy it is the fifth most salient taste descriptor in English (2016: 62), and it is linguistically encoded as a taste, as well as sweet, salty, sour and bitter. Therefore, for the purpose of this study, we refer to the spicy flavour as a taste. Our analysis reveals that spicy taste descriptors are employed as a source domain in metaphorical expressions both in Spanish and in English. Table 5 shows the quantitative results of our corpus analysis. Our analysis indicates that Spanish and English share the conceptual metaphor exCitement is sPiCy taste, by means of which spicy taste descriptors can express the quality of being somehow offensive or obscene and, for this reason, at the same time exciting or attractive. Although this conceptual metaphor was already identified by Bagli in the English language (2016: 89), our analysis reveals that it is significantly more salient in Spanish (37 occurrences) than in English (8 occurrences). On the other hand, our data suggests that only the Spanish language holds the conceptual metaphor vulgarity is sPiCy, which is analysed at the end of this subsection.
Regarding the conceptual metaphor exCitement is sPiCy taste, there are similarities and differences in the way it is linguistically expressed in Spanish and in English. In both languages, the spicy taste descriptors are employed in synaesthetic metaphorical expressions to describe smells, e. g. spicy perfume / perfume picante, indicating the quality of being provocative or exciting, as illustrated in examples 17 and 18 in English and in Spanish, respectively.
Spicy taste descriptors appear also in both languages in the frame PeoPle, e. g. spicy man / hombre picante, and CommuniCation, e. g. spicy words / palabras picantes, among others. In those cases, spicy taste descriptors both in Spanish and in English usually indicate the quality of being exciting in a sexual sense, as in example 19, but sometimes also in the sense of being controversial and challenging, as in example 20. On the other hand, our analysis suggests that the Spanish taste descriptor picante is employed in some frames where the English spicy does not, such as artworks (11 occurrences), e. g. película picante (lit.: spicy film), and oBjeCts (1 occurrence), e. g. coche picante (lit.: spicy car). Regarding the conceptual metaphor vulgarity is sPiCy, it is observable in two occurrences of our corpus, both from Chile, that we show in examples 21 and 22. Since the general dictionaries of Spanish do not contain any meaning of picante related to vulgarity, we consulted the Diccionario del habla chilena, where we found the following entry (Academia Chilena 1978: 169): "Hombre grosero, vulgar, ordinario // Bohemio de ínfima clase // adj. fam. despec. Andrajoso, desaseado. U.t.c.s." (Rude, vulgar, ordinary man // Bohemian of the lowest class // Ragged, dirty. Also employed as a noun). In example 21, picante is employed in the frame PeoPle, expressing contemptuously that someone is from a low social class or holds an ordinary kind of life. In example 22, picante is used in the frame CommuniCation, indicating vulgar or rude kind of language. Therefore, we can affirm that vulgarity is sPiCy is a conceptual metaphor typical of Chilean Spanish, whether in English there is not any similar conceptual metaphor. (21) […] hago planes sobre qué vamos a hacer con la plata. Siempre, lo primero es el departamento. Y con calefacción central, ¡cueste lo que cueste! Después me imagino tomando aviones, nunca me he subido arriba de un avión, por la puta madre, cómo es posible, si hasta los más picantes compran paquetes pa Cancún. 2011 Serrano, Marcela: Diez mujeres. Lit. ' […] I make plans about what we are going to do with the money. Always, the department comes first. And with central heating, whatever the cost! Then I imagine taking airplanes, I've never gotten on a plane, for the hell of it, how is it possible, if even the spiciest buy packages for Cancun'.

Discussion
In our work, we have analysed the variation in several conceptual metaphors with specific tastes as source domain in English and Spanish. Most of them, concretely Pleasure is sweet taste, anger is salty taste, hilarity is salty taste, aFFliCtion is Bitter taste, disPleasure is sour taste and exCitement is sPiCy taste respond to the generic-level conceptual metaphor emotions are tastes. emotions are tastes is a conceptual metaphor at the generic level, since it does not link particular cognitive domains, but abstract concepts that can show up on a variety of cognitive domains (Kövecses 2010: 44-45). Thus, we can affirm that taste metaphors in Spanish and English do not present variation at the generic level, given that emotions are tastes is shared by both languages. Moreover, most conceptual metaphors with specific tastes as a source domain show low variation at the conceptual level in Spanish and in English, since they are shared by both languages. It is the case of Pleasure is sweet taste, disPleasure is sour taste and exCitement is sPiCy taste. This outcome is expected, since these metaphors are grounded directly on universal sensory experiences, which are determined by the physiological properties of the human body (at least in the case of sweet, salty, bitter and sour tastes), so they can be classified as primary metaphors, in Grady's terminology.
However, our analysis reveals that there are also some conceptual metaphors with specific tastes as a source domain that show a certain degree of variation at the conceptual level. In the case of the bitter taste, Spanish and English share the conceptual metaphor aFFliCtion is Bitter taste, but we have showed that there is a slight difference on the scope of the target domain, which is wider in English. As regards the salty taste, we have proved that it gives rise to different conceptual metaphors, concretely anger is salty taste in English and hilarity is salty taste in Spanish. Lastly, there are some differences on the saliency of the conceptual metaphors shared by English and Spanish, particularly in the case of exCitement is sPiCy taste, which is more salient in Spanish than in English. All these cases of variation at the conceptual level can be explained by the role of food culture in influencing conceptualization. The new evidence of variation in taste conceptual metaphors, which are motivated by universal sensory experiences, supports the idea that culture influences and shapes all kinds of conceptual metaphors, even on basic and embodied ones. This is consistent with the model proposed by Caballero and Ibarretxe-Antuñano (2009), who explain the role of culture as a sieve, which models and validates the sensorimotor experience giving rise to conceptual metaphors.
At the level of linguistic expression, the degree of variation in Spanish and in English increases. All the taste conceptual metaphors shared by Spanish and English present a certain degree of variation, which takes the form in several ways. One of them is the semantic frames in which a given conceptual metaphor is employed. We have seen that a given frame can be significantly more salient in one language than in the other, or even can be exclusive of one of them. It is the case of the conceptual metaphor disPleasure is sour taste, which is often employed in English in the frame eConomy, e. g. sour recession, but not in Spanish. Another way of variation at the linguistic level concerns the grammatical features of the metaphorical expressions. One example is the independent use of sweet qualifying an event that takes place at the moment of the speech, which is ungrammatical in Spanish. Another example is the regular combination of the adjective sour with the verbs of change go and turn in metaphorical expressions in English, while in Spanish it is much less frequent.
Considering the degree of variation of taste metaphors at the different levels, we can affirm that it fits in the model proposed by Onysko (2017), since variation is higher at the linguistic level and lower at the conceptual level. However, in our opinion, it can be improved by adding a new level of analysis, the generic level, in which there is no variation, at least in the case of the taste metaphors considered in this work in Spanish and in English. Thus, in figure 2, we propose a new version of Onysko's model, illustrating the variation of aFFliCtion is Bitter taste in Spanish and in English, as an example.

Conclusions
The contrastive study of conceptual metaphors with tastes as a source domain in Spanish and in English offers new evidence about the fact that variation can take place both at the conceptual and the linguistic level asymmetrically. The degree of variation is higher at the linguistic level, in which a variety of parameters can be considered. Cross-linguistic variation is lower, but still significant, at the conceptual level. This is particularly relevant in the study of taste metaphors, since they are grounded on universal embodied experiences determined by the sensory abilities of human beings. Therefore, our study shows that culture plays an important role in the motivation of embodied conceptual metaphors. Lastly, we have argued that most taste metaphors respond to the generic-level conceptual metaphor emotions are tastes, which presents no variation in Spanish and in English. Thus, we have improved Onysko's model of variation in conceptual metaphor (2017: 13) by adding the generic-level metaphors stage, which shows the lowest degree of variation.
In our opinion, this study invites further cross-linguistic research on taste metaphors, because through them the way tension between sensory perception and food culture is resolved in the motivation of conceptual metaphors can be productively explored. Besides cross-linguistic research, we think that intralinguistic variation in taste metaphors deserves more attention, since we have found some interesting cases of variation at the conceptual level, such as the conceptual metaphor hilarity is salty taste, which seems to be typical of European Spanish, and the conceptual metaphor vulgarity is sPiCy taste, which is only found in Chilean Spanish. These cases of variation might also be explained by cultural differences among Spanish-speaking territories. Lastly, in future studies on taste metaphors it would be interesting to include derivate forms, such as the Spanish words amargado ('embittered') or agriar ('make sour'), in order to have a broader perspective of the linguistic expression of taste metaphors.