Normative data for 148 Spanish emotional words in terms of attributions of humanity

Título: Datos normativos de 148 términos afectivos en dimensiones relacionadas con la atribución de humanidad Resumen: Las investigaciones sobre la infrahumanización del exogrupo se apoyan en la distinción sutil y no deliberada que existe entre los sentimientos, una emoción exclusivamente humana, y las emociones, que son compartidas por animales y humanos. De acuerdo con esos estudios, las personas atribuyen más sentimientos al endogrupo que al exogrupo al que niegan o restringen la capacidad para experimentarlos. Este estudio presenta los datos normativos relativos a 148 términos afectivos en siete dimensiones relacionadas con las evaluaciones de humanidad. El Análisis de Componentes Principales calculado sobre las respuestas dio lugar a dos factores. El primero agrupó las dimensiones que diferencian los términos afectivos en función de las demandas cognitivas requeridas (cognición, naturaleza moral y duración), mientras que el segundo las diferenció en función de su perfil expresivo (visibilidad, edad a la que se adquiere, universalidad y foco causal). Estas dimensiones se analizaron en relación a la valencia, la familiaridad y la humanidad explicita. Palabras clave: infrahumanización; emociones; endogrupo; estudio normativo; exogrupo; sentimientos; términos afectivos. Abstract: Research on outgroup infrahumanization is based on the subtle and not deliberate distinction of secondary emotions, an exclusively human emotion, and primary emotions, which are shared by animals and human beings. According to prior studies, people attribute more secondary emotions to the ingroup than to the outgroup which they deny or restrict the ability to experience them. This study presents normative measures for 148 emotional words viewed by Spanish people in seven dimensions related to humanity assessments. Two factors were revealed by the principal components analysis (PCA). The first component was loaded on dimensions that differentiate the emotions depending on the cognitive demands (cognition, moral quality and duration) whereas the second one was loaded on their expressive profile (visibility, age at which they are acquired, universality and causal locus). These dimensions were analyzed in relation to desirability, familiarity and explicit humanity.

Participants in studies on infrahumanization are shown a list of emotional terms and have to associate them with the ingroup and the outgroup, and their answers are recorded.The analysis of these responses consistently converges in the same pattern of results: participants associate emotional terms that are uniquely human (secondary emotions) with the ingroup significantly more than with the outgroup, while they associate terms that apply to both humans and animals (primary emotions) to not only the ingroup, but also the outgroup.Researchers can then use the distinction between secondary emotions and primary emotions as a criterion of social discrimination.
In this sense, the aim of this study is to provide scores about a large group of emotional terms in relevant dimensions for research on infrahumanization for the use of the Spanish speaking scientific community.This database could be added to the set of instruments collected by Pérez-Sánchez, Campoy-Menéndez and Navalón-Vila (2001) and, more recently, by Davis and Perea (2005), Redondo, Fraga, Padrón and Comesaña (2007), and Sebastián-Gallés, Martí, Carreiras and Cuetos (2000) that contribute to the replication of results and enable comparative and cultural studies in the Spanish language.
The study of the differential attribution of primary and secondary emotions is interesting because it deals with a form of social discrimination that is available to all groups regardless of their social status.This is relevant because ethnocentrism has long been considered a reaction of the cognitive and moral superiority of the high-status groups.For instance, high status groups attribute to themselves intelligence and talent, uniquely human characteristics, and they do not tend to attribute such characteristics to lower status groups, (see Gould, 1981for a critical history of the ideas about human differences as biological determinism which speculates on the evolutionary proximity of non-Caucasian ethnicities to primates).However, accepting that groups can only consider themselves to be superior to others in intellectual skills means that those who consider themselves less competent cannot have prejudices towards the most competent outgroups.
Then, how can low status groups and stigmatized minorities have prejudices towards high status groups, with greater scientific and technological development and more efficient social organization?Take the case of the relationship between the Canarians and the Catalonians.Several studies on stereotypes show that the Canarians see themselves as less competent than the Catalonians and, in turn, that the latter share this assessment.Now if the Catalonians, anales de psicología, 2014, vol. 30, nº 3 (octubre) according to the Canarians themselves, are more competent, what are features which make the islanders feel superior?According to Leyens et al. (2000), this role is played by what is known as secondary emotions, a typically human competence that does not concur with the status and provides the group with a strong moral value and social superiority.Thus, the minority groups with less power are able to show an ethnocentrism of the same magnitude as the more powerful groups.
The antecedent of this research is the study by Demoulin et al. (2004), where they obtained normative data from a sample of Spanish, Belgian, American and Dutch people who provided numerical values for 13 dimensions.They carried out a principal components analysis and extracted two factors accounting for over 44% of the variance in the four national and language samples used.The first factor grouped seven dimensions concerning humanity.Specifically, secondary emotions are distinguished from primary emotions by their visibility (primary emotions are more visible than secondary emotions), the moral information (secondary emotions give more information about the morality of people than primary emotions), the cognitive dimension (secondary emotions require more cognitive resources than primary emotions), the causal focus (primary emotions have an external cause and secondary emotions have an internal cause), duration (primary emotions dissipate before secondary emotions), age (secondary emotions appear much later) and cross-cultural applicability (secondary emotions are more culturally specific than primary emotions ).The second factor consisted of two evaluative dimensions: desirability and acceptance.
However, Demoulin et al. (2004) only presents the normative values of humanity for the English version of the questionnaire.Therefore, the purpose of the present study is to provide the corresponding scores for seven relevant criteria which distinguish between secondary and primary emotions when such terms are given in Spanish.Specifically, data on duration, visibility, need for cognitive resources, moral information, causal focus, cross-cultural applicability, age and humanity.
Additionally, the same assessment scale used by Demoulin et al. (2004) was used to study the valence of each term and another dimension about familiarity was added.The valence dimension shows the desirability of the emotional terms and the familiarity dimension shows the subjective frequency of use.The four dimensions of intensity, sensitivity, acceptability and association with gender, that were found to be the least relevant in the study by Demoulin et al. (2004) were also eliminated.
Besides providing the scores for the seven criteria mentioned above, the present study also checks whether the normative data on these criteria converge into a single factor and whether this correlates with the explicit dimension of "humanity".By confirming this expectation, a single empirical dimension would be able to reflect the difference that infrahumanization studies find between secondary emotions and primary emotions.

Method Participants
Five hundred and forty university students from Madrid (N = 76), Valencia (N = 104) and Santa Cruz de Tenerife (N = 360), took part in the study.The age range of participants was 17-51 years of age (M age = 21.03,SD = 4.01) and there were 125 men and 415 women.The participants were divided into nine samples of 60 people so that the proportion concerning gender and place of residence were roughly the same.

Material and procedure
A total of 148 emotional terms were tested.Many of these terms were taken from the study by Demoulin et al. (2004), while others came from ad hoc pilot studies on people with similar characteristics to the definitive sample who were asked to indicate terms reflecting emotional reactions.The 148 terms were distributed in nine questionnaires, each one having 16 to 17 terms.We tried to keep an equal balance between positive and negative terms in each questionnaire.Such terms were evaluated regarding 10 dimensions, seven of which were relevant in the work by Demoulin et al. (2004) to differentiate between secondary emotions and primary emotions.Specifically, the seven dimensions were presented to the participants with the following wording: 1. Duration."When a person has an emotional experience it can last a long time or be fleeting.All the emotional experiences we know are to be found between these two extremes.The way somebody behaves when he or she experiences an emotion provides us with signs about whether we are dealing with a good or bad person.There are other emotional experiences that do not provide any information in this respect" (1= Does not give moral information; 7= Gives a lot of moral information).5. External cause."The emotional experience is usually a reaction to something.In some cases, the cause is outside the person, this can be an event or another individual.In others, however, the cause might be inside the person, something that has been generated without any apparent and immediate cause to trigger it" (1= Internal causes; 7= External causes).6. Cross-cultural applicability."There are emotional experiences that occur in the same way everywhere in the world regardless of cultural singularities.Others, however, heavily rely on cultural factors such as traditions, customs, beliefs and values" (1= Different according to the culture; 7= Same in all cultures).7. Age."All human beings are capable of experiencing emotions.However, not all occur in the same developmental stage of the person.Some emotional experiences are possible from the moment when we are born and a few months afterwards.Others, however, require more advanced stages of development and maturity" (1= Appears very early; 7= Takes time to appear).
In addition, three complementary dimensions have been included.Firstly, a dimension about the humanity of the emotional terms as a criterion to distinguish between secondary emotions and primary emotions.Secondly, a dimension concerning familiarity with the aim of verifying whether the terms to do with secondary emotions are more familiar, and subjectively more frequent than those relating to primary emotions.Finally, a valence dimension has been added in order to clarify the like or dislike of each emotional term.The following questions have been included in the normative study: 1. Humanity."Emotional terms differ in their ability to express physical and mental states.Thus, some emotions can be experienced by animals and human beings, and others are more specific to humans.In your opinion, to what extent are the emotions listed below unique to humans or can they also be experienced by animals?" (1= Animals and humans; 7= Only humans).2. Familiarity."Words differ in their familiarity.Some words are very familiar and others, however, are almost un-known.The intention of this question is to assess whether these terms are commonly known or used by you "(1= Very unfamiliar; 7= Very familiar).3. Valence."Not all the emotions expressed by the terms listed below evoke a pleasant state.In your opinion, to what extent is it agreeable to personally experience each emotion?"(1=Very unpleasant, 7= Very pleasant).
The participants were presented with each dimension followed by emotional terms which they had to rate on a Likert scale from 1 to 7. The questionnaires were answered collectively and according to the following written instructions."On the sheets that you will be given, you will find a question followed by several scales regarding different emotions.Although the questions are different, the emotions to be evaluated are always the same.It is important for our research that you do not omit any response and read each statement carefully to prevent the answers to one question from influencing the answers to the following question.

Results
The results were analyzed according to the objectives of the study.Firstly, a list of average scores was made.Secondly, the underlying structure of the seven dimensions relevant to the distinction primary emotion vs. secondary emotion was analyzed.Finally, the relationship between the scores for "Humanity", "Valence" and "Familiarity" and the aforementioned seven dimensions were considered.
The mean and standard deviation and the correlation between them were calculated to determine the normative values of each emotional term for the ten questions (see Appendix).As can be seen in Table 1, there are close correlations between the seven indirect dimensions of humanity.However, the most relevant results of this analysis concern the correlation between the explicit criteria of humanity and the implicit dimensions.Specifically, the results suggest that people perceive the emotional terms which are less visible as being more human, likewise with those requiring more cognitive resources and providing more information about the moral profile of the person.Furthermore, people perceive those emotional terms which are manifested in different ways in different cultures, which are not caused by external causes and those which appear later in the development of a person as being more human.
Inter-correlation analysis also provides information about the relationship between the humanity and the valence dimensions.Taken together, it appears that the longer lasting the emotional term is, the more visible it is and the more moral information it provides, the more likely it is to be positively valued.Nevertheless, it is more likely that the terms are valued more negatively when they require cognitive resources and appear in more advanced stages of development.
Finally, the familiarity dimension is positively associated with those dimensions which refer to the duration, visibility, cognitive resources, moral information and cross-cultural applicability, whereas the age of acquisition dimension tends to have a negative association.A principal components analysis (PCA) was then carried out with the means of the seven relevant dimensions regarding the primary vs. secondary emotion distinction.The factor solution of the mean scores of the 148 terms in the seven dimensions gave a KMO = .64,and a Chi-square in the Bartlett sphericity test of χ 2 (21 ) = 411.1;p < .001,which shows the suitability of the data to perform the PCA.
Two components with an eigenvalue greater than 1 were revealed by the PCA and these two components jointly account for 65.9 % of the variance.The first, with an eigenvalue of 2.98, accounts for 42.63% of the variance while the second, with an eigenvalue of 1.63, accounts for 23.23 % of the variance.The loadings of the dimensions in the two components are shown in Table 2.As can be seen in Table 2, the dimensions are almost evenly distributed between the two components.Thus, while the first is sensitive to the variances showing the cognitive demands of the emotional terms (processes such as cognitive resources, moral information and duration), the second presents its expressive components (visibility, age, cross-cultural applicability and external causality).Specifically, the emotional terms can be differentiated depending on, on the one hand, whether or not they require higher mental processes, that is, whether they require more memory, cognitive resources and moral components and, on the other hand, according to their expressive nature, they can be differentiated depending on when they appear in the development, their visibility, their external cause and their cross-cultural applicability.
Before determining the reliability of the dimensions corresponding to the two factors, and in order to clarify the mean scores, the responses were recalculated to guide them in the same direction, so that the higher the score in the factor the more exclusively human the emotional term is.Specifically, the scores from the "visibility ", "cross-cultural applicability" and "external cause" were inverted.The Cronbach's alpha computed on dimensions of the first factor was α = .84and α = .74for the second factor.
Finally, a correlation and regression analyses were conducted between the two factors and the complementary dimensions investigated keeping the inverted scale for the "visibility", "cross-cultural applicability" and "external cause" dimensions.The correlation showed that both the factor relating to higher processes and that relating to the expressiveness correlated significantly with the score obtained for the "Humanity" dimension included in this study as a criterion factor.Furthermore, it is interesting to note how the second one (r (148 ) = .68)has a stronger correlation than the first (r (148) = .19).In addition, we carried out a multiple linear regression analysis to determine to what extent the humanity of an emotional term could be predicted by higher cognitive requirements and low expression.The result of this analysis was R 2 = .46;(F (2,147) = 61.3;p < .001),which means there was a significant relationship between the predictor variables and the criterion variable "Humanity " and that the two predictors included in the equation account for 46% of the relative variance for "Humanity".The standardized partial regression coefficients representing the relative importance of each predictor in the equation show that both factors make a statistically significant contribution to the prediction (t = .54;p < .012for cognitive requirements factor and t = 10.55;p < .001for the expressiveness factor) without colinearity occurring, since the tolerance values are greater than .90. anales de psicología, 2014, vol.30, nº 3 (octubre) Since the humanity of an emotional reaction is also determined by its frequent use to communicate the mood of human beings, a correlation between the two factors and the "Familiarity" dimension was performed.As shown in Table 3, two significant correlations were found although in opposite directions.Thus, on the one hand, the more higher mental resources the emotional terms require, the more familiar they are and the more commonly used they are (r (148) = .50).On the other hand, the less visible and the more internally caused the terms are, the less familiar they are (r (148) =-.31).As with the previous variable, the multiple linear regression gave an R 2 = .48;(F (2,147) = 66.9; p < .001), indicating the existence of a significant relationship between the predictors and the criterion variable, and that both the predictors included in the equation account for 48% of variance for familiarity.Furthermore, the standardized partial regression coefficients show that the two factors make a statistically significant contribution to the prediction of familiarity (t = 6.78; p < .001for cognitive requirements factor and t = -8.19;p < .001for the expressiveness factor), with tolerance values above .90and, therefore, no co-linearity.
Finally, the correlation analysis between the two factors and the valence of the emotional terms gave two results with a low profile (r (148 ) = 0.10 for the first factor and r (148) = -.20 for the second factor, only significant in the latter case).The aforementioned differential association suggests that those emotional terms reflecting the expressive function have a moderately negative valence.

Discussion
The aim of this study was to establish the norm of 148 emotional words in terms of their degree of humanity (see Appendix).Although the shortest way to reach this goal is by directly questioning people about the degree of humanity of the terms presented, the data obtained in pilot studies as well as the research by Demoulin et al. (2004) recommend the use of various dimensions that can serve as indirect indicators.Accordingly, the study included the seven dimensions that Demoulin et al. (2004) found most relevant in their normative research.A PCA was conducted in our research on the seven relevant dimensions to qualify the terms according to their uniquely human quality.The results show the existence of two underlying evaluations in the responses.On the one hand, there is an evaluation emphasizing the role of higher cortical processes in emotional experiences.Seen from this perspective, an emotional response is human if it is long lasting and requires memory, as occurs with serenity (serenidad) and sorrow (pesadumbre) that are more human than fleeting instant responses such as terror (terror) or irritation (irritación).Besides, emotional terms which provide information about the morality of the people who experience them also demand cognitive resources.This is the case of emotional responses such as guilt (culpa), regret (arrepentimiento) or gratitude (gratitud) which are perceived to be more human than pleasure (placer), surprise (sorpresa) or confusion (desconcierto).Finally, the cognitive demand of the emotional experience also serves as a criterion for differentiation.This places remorse (remordimiento) and guilt (culpa) at the human end of the distribution while pleasure (placer) and excitement (excitación) are placed at the end that characterizes humans and animals.
The other underlying evaluation from the PCA regards the external profile of the emotional term, how it is expressed and the ease with which it is perceived.The visibility component has the greatest load in this dimension.Thus, participants considered that the less externally visible an emotional term is (e.g., remorse or self-pity; remordimiento or autocompasión) the more unique to humans it is, whereas the more visible ones (e.g., joy or surprise; alegría or sorpresa) are common to both animals and humans.The age at which an expression appears turns out to be a discriminatory dimension in this direction due to the similarities between the early stages of human development and the animal kingdom.Thus, the participants considered that while children and animals share joy (alegría) and surprise (sorpresa) reactions, only mature human beings can experience demoralization (desmoralización) or regret (arrepentimiento).The third relevant dimension to this factor is the cross-cultural applicability.Emotional terms which are expressed in the same way in all cultures are more likely to be shared with animals as is the case of sadness (tristeza) and joy (alegría).In contrast, emotional terms which incorporate idiosyncratic elements of each culture such as modesty (pudor) and lamentation (lamentación) are uniquely human.Finally, the reactive nature to internal or external cues constitutes the fourth relevant dimension.Thus, the emotional terms that are activated in responses to external stimuli such as horror (horror) or surprise (sorpresa) are less exclusive to humans than serenity (serenidad) or self-pity (autocompasión) that are understood as responses to internal demands.
In addition, our study included three complementary dimensions.The first, the degree of humanity, was used as an explicit index and valuation criterion for the emotional terms.The response of the participants showed that this issue is a fairly reliable and more economical way to obtain differential values between emotional terms.In fact, the correlation analysis of these responses with the average obtained for the 148 words in the two components obtained was significant and especially high with respect to the second component.
Familiarity data were more ambiguous since the correlation showed that the emotional terms which scored high on the first factor and low on the second factor and vice versa were more familiar.We have no explanation for this outcome beyond that which might arise from the frequency of use of some terms compared to others.This is the case, with primary emotions such as joy (alegría), friendship (amistad), affection (cariño) or love (amor) in contrast to amazement (estupor), jubilation (alborozo), dismay (desaliento) or apprehension (aprensión) which are not so commonly used in everyday language.
Finally, as reported in the study of Demoulin et al. (2004), the valence dimension does not correlate with the human quality of the emotional terms.In other words, regardless of whether they are uniquely human or shared with animals, primary emotions can be positive or negative.
In summary, the results of this normative study suggest that speakers of Castilian Spanish are able to differentiate between emotional terms that are uniquely human and those we share with animals.However, the discrepancies observed in the hierarchy of each dimension and in the humanity dimension is evidence that we are not referring to an explicit and systematic distinction here.Therefore, the fact that this distinction is not explicit or systematic means that this distinction can be used as a subtle way to study intergroup prejudice and the infrahumanization of others.This implicit distinction has been used in studies by Leyens et al. who coined the term infrahumanization to refer to the association between secondary emotions and ingroup and the systematic dissociation between secondary emotions and outgroups (for a review, see Leyens et al., 2000 andLeyens et al., 2007).This differential association only occurs with secondary emotions but not with primary emotions.In other words, people reserve for themselves those uniquely human emotional responses (secondary emotions) whereas they are able to recognize that others may feel primary emotions, which both animals and humans experience.In addition, according to the data in the above mentioned studies, this tendency to associate secondary emotions to the ingroup also extends to secondary emotions of negative valence, i.e., all secondary emotions, both positive and negative, since both are equally human (Delgado et al., 2012;Leyens et al., 2001).

Table 2 .
Factor loadings for the dimensions in the two factors.

Table 3 .
Correlations for the two factors of PCA with complementary dimensions.