Beyond saying Thanks : Compliment responses in American English and Peninsular Spanish

This study explores how American English and Peninsular Spanish speakers respond to a compliment. Participants completed an online discourse completion test with nine different complimenting scenarios. A total of 14 different strategies for responding to a compliment were found in the data. Based on verbal reports on language use, it was found that Peninsular Spanish speakers do not compliment as often as American English speakers do. The data analysis also revealed that both language groups clearly prefer to accept a compliment, but whereas American English speakers regard a simple ‘thank you’ as an appropriate compliment response, Peninsular Spanish speakers prefer to agree with the complimented assertion by making a semantically fitted comment. Other differences include the importance of returning a compliment in American English and the need to scale down the illocutionary force of the compliment among Peninsular Spanish speakers.


Introduction
Compliment responses (henceforth referred to as CRs) have been researched extensively across languages (see Chen 2010 for a comprehensive review). There is general agreement about what constitutes an acceptable response to a compliment across different languages and cultures. In fact, very early on children are taught to respond to compliments by saying 'thank you' as a sign of politeness, regardless of whether or not the child understood the nature of the compliment. This behavior becomes part of what polite verbal responses are supposed to be and thus, it is understood that one should always acknowledge the compliment. Although we tend to simplify this speech act for children by indicating that a simple 'thank you' is enough, speakers use a wide array of strategies to respond to a compliment. A simple 'thank you' may not always be enough. The aim of this paper is to explore the strategies used in Peninsular Spanish and American English to respond to a compliment and examine crosslinguistic differences in response preferences.
The interest in CRs research resides in the conflicting dilemma of acknowledging a compliment and simultaneously avoiding self-praise. In other words, there is a clash between Leech's (1983) maxims of modesty and agreement. On the one hand, one should agree with one´s interlocutor in responding to a compliment; on the other hand, in doing so, one may sound boastful and immodest. Rejecting a compliment may also appear as facethreatening for both complimenter and complimentee. For this reason, complimentees adopt a wide array of strategies to minimize or mitigate the potential face threat that their response to the compliment may represent. The strategies adopted in CRs have been classified under several taxonomies (Chen 1993;Cheng and Yang 2010;Felix-Brasdefer and Hasler-Barker 2012;Golato 2002;Herbert 1989;Holmes 1988;Huth 2006;Maíz-Arévalo 2010;Maíz-Arévalo 2012;Mack and Sykes 2009;Pomerantz 1978), but a common tri-partite system seems to have emerged across all investigations. This system divides CRs into Acceptance, Deflection/Evasion, and Rejection (Chen and Yang 2010). The three categories represent the conflicting nature of CRs, that is, the acceptance of the compliment, the need to avoid self-praise by deflecting or evading the compliment, and the choice to refuse the compliment or disagree with the complimenter.

CRs in American English and Spanish
Research in compliments and CRs in American English started with the pioneer works of Pomerantz (1978) and Manes and Wolfson (1981) and their followers. These early studies used an ethnographic approach that focused on the social relationship between interlocutors and the impact of social characteristics in speech act behavior (Chen and Yang 2010). This approach to complimenting research in American English became the benchmark against which to compare other languages and design data collection instruments such as the discourse completion test (henceforth referred to as DCT). Since the 90s, CRs in American English have mainly been studied from a comparative and/or second language learning perspective (Behnam and Amizadeh 2011;Brezolin 1995;Cedar 2006;Cheng 2011;Felix-Brasdefer andHasler-Barker 2012, Huth 2006;Rose 2001;Sharifian 2008;among others).
In comparison to English, research on Spanish CRs is scarce. Some research has looked at compliments and compliment responses in Latin American Spanish (Valdes and Pino 1981;Yáñez 1990) but very few studies have selected Peninsular Spanish in isolation or in comparison with other languages. Siebold's (2006) comparison of Peninsular Spanish and German CRs in role plays showed that Peninsular Spanish speakers accepted compliments more often (82%) than German speakers (33%), but German speakers opted for deflecting the compliment more frequently (50%) than Peninsular Spanish speakers (14%). Choi (2008) compared Peninsular Spanish and Korean CRs using a written DCT and found that although Spanish and Korean participants favored to express thanks in responding to compliments, Korean speakers also preferred questioning the sincerity of the compliment and using humor. Spanish speakers, on the other hand, frequently scaled down the object of the compliment and used more than one strategy in their responses. Mack and Sykes (2009) compared compliment-response sequences in Peninsular Spanish and Mexican Spanish on the basis of the presence or absence of positive irony in the compliment using an electronic oral response DCT. The results revealed that both Mexican and Peninsular Spanish participants showed acceptance as the most preferred strategy. Cross-linguistic differences appeared when looking at the second most preferred strategy: whereas Peninsular Spanish participants opted for ironic CRs, Mexican Spanish speakers favored self-praise avoidance strategies. With regard to irony, both groups accepted the use of positive irony as a compliment in about 45% of the responses. In cases where irony was recognized, Peninsular Spanish participants favored an ironic comment as a response, whereas Mexican Spanish speakers preferred to respond to ironic compliments with self-praise avoidance. Maíz-Arévalo (2010 used naturally occurring conversational exchanges to explore compliment sequences in Peninsular Spanish and English. The English corpus came from Holmes' (1986Holmes' ( , 1995 and Herbert's (1989Herbert's ( , 1990 studies and included New Zealand English, American English and South African English. Accepting a compliment is the most frequent response in both languages, although Peninsular Spanish speakers do not accept a compliment as frequently as English speakers do. Perhaps one of the most surprising differences is the use of rejections among Peninsular Spanish speakers (24.5% Peninsular Spanish vs. 8% English speakers) (Maíz-Arévalo 2012). However, as the author explains, this rejection is quite formulaic (i.e. ¡Qué va! 'No way') and not viewed as offensive by the complimenter (2012: 162). Finally, the Spanish corpus also revealed that overall Spanish CRs were more elaborated and involved more than one function resulting in longer conversational exchanges. Recently, Maíz-Arévalo and García-Gómez (2013) analyzed CRs in Facebook interactions in Peninsular Spanish and found that the asynchronous nature of computer-mediated interactions favors the option of not responding to a compliment without the danger of offending the complimenter. In addition, Facebook users in this study utilized other tools besides typing a response, such as clicking on 'like', which accounted for 38% of the CRs.
In her comparative study of naturally occurring CRs in Peninsular Spanish and Lebanese, Ramajo-Cuesta (2012) found that Peninsular Spanish participants preferred to offer one single strategy responses to the compliment (86% Spanish, 55% Lebanese) whereas Lebanese speakers favored responses with a combination of strategies (45% Lebanese, 15% Spanish). In both languages, the most preferred response was to accept the compliment but in Peninsular Spanish the option to mitigate the compliment was also slightly preferred (31% Spanish, 22% Lebanese). In addition, Peninsular Spanish speakers liked to offer long explanations in their responses whereas Lebanese speakers used more formulaic response types.
A contrastive analysis of CRs in Peninsular Spanish and British English was conducted by Lorenzo-Dus (2001) using an electronic written DCT. Besides asking for CRs, the participants were also asked to offer meta-comments on their responses. Results uncovered several cross-cultural and cross-gender patterns, although Lorenzo-Dus does not offer any descriptive statistics to support her findings. British participants questioned the sincerity of the compliment more often than Spanish speakers, as indicated in the meta-comments. Irony and humor were equally used by both language groups, although some gender differences appeared. Spanish male participants preferred to use irony to upgrade the illocutionary force of the compliment response. Something unique about the Peninsular Spanish participants was the use of a request for repetition or expansion, something Lorenzo-Dus called 'fishing for further compliments ' (2001: 118). For example, responses such as ¿Tú crees? ('Do you think so?') or ¿De verdad ('Really?'), were quite common among Peninsular Spanish participants.
The current study hopes to add to the small number of investigations on CRs in Peninsular Spanish in several ways. First, this is the first systematic study of American English and Peninsular Spanish CRs that combines a quantitative and qualitative analytical approach. Secondly, besides CRs, this study also examines speakers' own understanding and usage of compliments in everyday communication. Finally, a new taxonomy of CRs has emerged from the responses obtained, offering future investigations a more precise way to classify CRs.

Methodology
The data for the present study was collected by means of an electronic written DCT. Despite its well-documented limitations (Golato 2003), the DCT is reliable to measure off-line pragmatic knowledge (what the learners know) under constraint conditions (Felix-Brasdefer 2010). The focus is on production of forms, not on their use in social interaction. In that sense, the DCT elicits a user's explicit knowledge of speech acts.
A Spanish and English version of the nine-item DCT used by Lorenzo-Dus (2001) was used for this study. The items included nine different scenarios depicting compliments on appearance, possessions, skills and personality between interlocutors that differed in their power position: Scenario 1: A friend compliments another on his/her new haircut and says "That hair cut makes you look great. It makes you look younger!" Scenario 2: An employee compliments his/her boss on a new car and says "It's smashing! I love the model. And you've got good taste in choosing the color!" Scenario 3: A student compliments his/her teacher on his/her cooking skills and says "I didn't know you were such a talented cook. The food was wonderful!" Scenario 4: A coach compliments a trainee on his/her tennis skills and says "All the effort has been worthwhile. You have played brilliantly today!" Scenario 5: A friend compliments another on his/her written essay and says "It's an excellent essay. You've structured it in a very clear and concise way. If only I could write something half as interesting as that".
Scenario 6: A boss compliments an employee on his/her smart clothes and says "You look so elegant and that outfit really suits you".
Scenario 7: A friend compliments another on his/her organizational skills at a conference and says "You're the right person for this type of job. You're ever so nice to the others and know how to avoid disagreements with everyone".
Scenario 8: A teacher compliments a student on his/her computing skills and says "You're very intelligent and have a flair for computers. Besides, you show a lot of interest in what we do in lessons".
Scenario 9: An employee compliments a boss on his/her eyes and says "You've got beautiful eyes".
Two groups of university students from mid-size public universities in Spain and in the United States participated in the study. The American English native group (henceforth ENG) included 37 female and 8 male students (total: 45). The Spanish native group (henceforth SPA) consisted of 86 female and 14 male students (total 100). The participants were asked to voluntarily and anonymously participate in the study. The disparity in numbers between the Spanish and English groups is due to the volunteering nature of the study and, thus, beyond researchers' control. Due to the unbalanced number of participants between the English and the Spanish group and computer software limitations, gender was not included as a variable. For each contextual cue given, participants were asked to offer up to four possible socially appropriate CRs. Our corpus consists of 704 English responses and 2443 Spanish responses. In addition, participants responded to two open-ended questions at the end of the survey: 1) What is a compliment for you and under which circumstances do you use compliments? 2) How often do you use compliments? The purpose of these two questions is to gain a better understanding of how the two language groups view the complimenting act. For comparative purposes, all answers to these two questions from the English group (i.e. 45 respondents) and a random selection of another 45 respondents from the Spanish group were analyzed.
The taxonomy used to analyze our data derived from much of the current research in CRs discussed above, but ultimately the response strategies included under each category resulted from the responses obtained: 1. Acceptance: The purpose of these responses is to accept the compliment without any ambiguity so that the complimenter does not feel obliged to add anything else and the conversation can shift topics or end there. Accepting a compliment can be done in different ways: 1.1. Appreciation token: A brief conventional expression of gratefulness.
1.2. Acceptance with emphasis: An expression which emphasizes the degree of gratefulness of the speaker.
1.3. Agreement: A response that is semantically fitted to the compliment. For our purposes, we include responses with the adverb ´too´ or ´yes´ indicating explicit agreement with the complimenter, any positive comments directly related to the compliment, an upgrade, or an offer to the complimenter.
(3) A: Your car is smashing! B: Yes, it is a great car. B: I am proud too.
2. Self-Praise Avoidance: These responses are intended to minimize and/or mitigate the complimentary force of the compliment by implicitly accepting the compliment and, at the same time, avoiding self-praise. We distinguish seven different sub-categories within this group: 2.1. Scale down: An expression that presents the object of the compliment as ordinary or as unworthy of the compliment.
(4) A: That haircut makes you look great. B: Oh, it's just the same old thing.
2.2. Reassignment: Responses that redirect the praise to a third party.
(5) A: Your car is smashing! B: Thanks to Dad! 2.3. Return: An expression that redirects the praise or focus to the complimenter by returning the compliment (6) or offering an encouraging word (7).
(6) A: You look great. B: Thanks. You look wonderful too.
(7) A: It's an excellent essay. B: You can write a paper as good as this.
(8) A: You're ever so nice to the others. B: Well, arguments don´t get us anywhere.
(9) A: Your car is smashing! B: I never thought I'd be able to afford this but then I won the lottery! 2.5. Qualification/Uncertainty. An utterance that qualifies the praise (10) or expresses doubts or uncertainty about the assertion of the compliment (11).
(10) A: You played brilliantly. B: Really? But I think I could have done it better.
(11) A: That hair cut makes you look great. B: I am not sure about it yet.
"Yes, I am very depressed about having earned a terrible A+".
(13) A: You have beautiful eyes. B: Yeah I used to be an eye model! 2.7. Question: Responses intended to request expansion or repetition of the compliment and in some cases, question the sincerity or motives of the complimenter (Herbert 1989). For our analyses, these responses have to appear in isolation. When they appear followed by another utterance, they are coded based on the semantic content and illocutionary force of the utterance.
(14) A: That hair cut makes you look great. It makes you look younger B: Younger? Really? Did I go too short?
3. Non-Acceptance. These responses include explicit expressions of disagreement and refusal, but also, ambiguous responses where it is not clear whether the speaker heard and/or understood the compliment. We distinguish four main sub-strategies: 3.1. Disagreement: A negative comment that disagrees with the assertion of the compliment totally or partially.
(15) A: The food was wonderful! B: I thought it had too much salt. (16)  In cases where more than one strategy was used, it was decided that in order to facilitate the analysis, we would categorize the CR according to the strategy that carried the most transparent semantic value. An exception was made for CRs that included thank you (i.e. 'Appreciation token') in combination with other response strategy. We indicated the presence of thank you by adding 'a' to the code (see Table 1). For example, showing agreement as a CR, with (21) or without an 'Appreciation token' (22) were coded differently.

Results and discussion
This section is organized as follows. In 4.1, we summarize the different definitions of a compliment given by both language groups. In 4.2, we present and interpret the results of the overall frequencies obtained across all scenarios in the DCT. Finally, in 4.3, the use of specific strategies is further analyzed and explained according to the contextual variables displayed in the DCT scenarios.

Respondents' definition of compliments and frequency of use
In defining a compliment, participants focused on three main aspects: (a) the verbal nature of the compliment, (b) the object of the compliment and (c) the purpose for which the compliment is made. Our Spanish informants referred more frequently to the verbal nature of the compliment by introducing in their definition either a noun or pronoun such as halago (a synonym for a compliment), frase ('phrase'), comentario ('comment'), algo ('something') accompanied by a verb of saying. In most of the cases the definition includes an adjective indicating positiveness: agradable ('nice'), gratificante ('gratifying'), positivo ('positive'), as in example (23). (23) Un cumplido es un comentario positivo de carácter subjetivo acerca de algo/alguien. "A compliment is a subjective positive comment about something/someone".
The English definitions, on the other hand, mainly referred to the object of the compliment: an action, appearance or character (24-25).
(24) A compliment for me is when people notice a job well done.
(25) I use compliments when I like something I see (i.e. a friend's outfit, makeup, shoes).
When looking at the purpose of the compliment, we see a clearer emphasis on the complimenter in Spanish than in English. Thus, whereas our Spanish participants referred to "getting something from another", "something I like", or simply "to be nice", in the English responses we find more expressions focusing on the complimentee like "words spoken to you in appreciation for something", "acknowledgement of something that I do well" or "someone says something that makes me feel better".
Both language groups mentioned that a compliment is used to make someone feel better but a few Spanish participants questioned the sincerity of the compliment and expressed that sometimes a compliment is used to get something in return, to mock someone or to overcome a difficult situation as a polite formula.
(26) Lo uso en situaciones de confianza o para ganarla o acercarme a alguien, causar buena impresión, ya sea para ligar y para causar buena impresión a jefes, profesores. "I use it in familiar contexts or to win someone over or to get closer to someone, to make a good impression, as in flirting and to make a good impression on bosses, professors".
When asked how frequently compliments are used, the answers reveal that 62% of English informants 'often' use compliments as compared to only 25% in Spanish. However, 67% of Spanish respondents indicated that they 'sometimes' use compliments as compared to 27% in English. Therefore, although both American English and Peninsular Spanish speakers are comfortable using compliments in their daily interactions, American English speakers claim to use them much more frequently.
In Lorenzo-Dus' study (2001), her British participants questioned the sincerity of the compliment more often than her Spanish participants. Her results, as ours, do not come from CRs but from comments made by participants. Nonetheless, these comments lead us to tentatively conclude that American English speakers do not question the sincerity of the compliment as much as Peninsular and British speakers do. It is possible that the less frequent use of this speech act explains why Spanish speakers question the sincerity of the compliment. However, perhaps, speakers prefer not to use compliments, because of their hidden and insincere intentions. Further research on this matter is required to address this dilemma.

Strategy type: Frequency and preference
In this section, we first look at the three main strategies defined as possible CRs (i.e., Acceptance, Self-Praise Avoidance, and Non-Acceptance) and then, we present the most preferred sub-strategies used by both language groups.

INSERT TABLE 1 HERE
Our English and Spanish informants did not differ in the three main response strategies: Acceptance (ENG: 48.4%, SPA: 48.3%); Self-Praise Avoidance (ENG: 46.8%, SPA: 45.4%); and Non-Acceptance (ENG: 4.8%, SPA: 5.9%). These results show that American English and Peninsular Spanish speakers equally desire to accept the compliment. However, a closer look at the frequency and preference of sub-strategies reveals that while for English speakers responding with a simple thank you was the most preferred CR (1.1: 23.5%), in Spanish, expressing agreement with the complimenter was favored (1.3 + 1.3a= 26.8%). The evident preference for including thank you in a CR in English is also confirmed when we add up all CRs that include thank you (i.e. all the codes with an 'a'), the difference being 63% for English and 51% for Spanish. An appreciation token is a sign of acceptance; therefore, acceptance of a compliment appears to be a more expected act in American English than in Peninsular Spanish and, in any case, the presence of the 'Appreciation token' is more predictable for English than for Spanish as an element of the second-pair part of the complimenting adjacency pair.
In avoiding self-praise, the most preferred strategies by our English participants were to return the compliment (2.3 + 2.3a = 18.2%) or respond with humor (2.6 + 2.6a= 8.1%). Our Spanish informants found it similarly acceptable to offer an informative comment (2.4 + 2.4a = 11.3%) or to return a compliment (2.3 + 2.3a = 11.2%), but they also relied on scaling down the compliment's illocutionary force (2.1 + 2.1a = 8.2%). Some of these results are confirmed by previous research. For example, the use of explanatory comments and longer exchanges in Spanish was found in Ramajo-Cuesta's (2012) and Maíz-Arévalo´s (2010) studies. Although these studies used naturally occurring data, our DCT responses partly support these findings. In addition, Choi's results (2008) also revealed the option of scaling down among Spanish speakers in comparison with Koreans.
Overall, Spanish CRs showed a greater degree of creativity, verbosity and unpredictability. Whereas 74% of the English responses fell within the seven most selected sub-strategies, the Spanish participants used nine different substrategies to reach 75% of the responses. Furthermore, the Spanish group offered many more responses in the DCT (an average of 24 responses per participant) than the English respondents (16 responses per participant).
In their study of American English compliments, Manes and Wolfson (1981) point out the formulaic nature of many of the English compliments. Although the syntactic nature of CRs is not as formulaic as in the case of compliments, our results suggest that in comparison with Peninsular Spanish, American English CRs are more systematic and predictable. Peninsular Spanish speakers rely on more different ways to answer to a compliment, as evidenced by the higher number of responses from the Spanish participants, a result also found in Lorenzo-Dus' study (2001), and the wider number of strategies they choose from when responding to a compliment. The lack of frequency of compliments in Peninsular Spanish, as indicated by our participants, may trigger a diminished confidence level in how to respond, which explains the need to offer different response types and go beyond a simple thank you. On the other hand, the higher frequency of compliments in American English may explain why CRs are more predictable and thus, speakers heavily rely on thank you and in a more limited number of strategy types.
Other studies in CRs in Spanish have pointed out the positive politeness orientation of Peninsular Spanish culture (Lorenzo-Dus 2001; Mack and Sykes 2009;Maíz-Arévalo 2010, Maíz-Arévalo 2012Siebold 2006). In building solidarity, Peninsular Spanish speakers understand that a sign of gratitude may be interpreted as insufficient and perhaps insincere due to its formulaic nature. As noted, our Spanish respondents mentioned that they question the sincerity of the compliment, which may explain why a simple thank you in Spanish is not enough. Commenting on the nature of the complimented object shows that the listener paid careful attention to the compliment and subsequently, to the complimenter, thus, the preferred use of showing agreement when accepting a compliment. The same can be said about the use of detailed informative comments about the contextual factors surrounding the complimenting event or scaling down the object of the compliment. Peninsular Spanish speakers show affiliation with the complimenter by removing any self-praise, and placing both the complimenter and the complimentee at the same level. In contrast, American English speakers balance positive and negative politeness in using strategies that irrefutably show acceptance of the compliment, such as a sign of gratitude, and in addition, show self-praise avoidance by returning the compliment. The frequency of compliments in American English may also help explain why returning a compliment is so common since in this mitigating act the complimenter now becomes the recipient of another compliment.

Strategy use across social contexts.
The scenarios included in the DCT reflect a combination of contextual variables often employed in pragmatics research (i.e. familiarity, power). However, these variables did not uncover clear patterns when examining our data. A closer look at four strategies, on the other hand, led to noteworthy cross-linguistic and sociolinguistic complexities.

Returning the compliment.
When the complimented object was due to someone's skills, our English informants typically chose to return the compliment. In the Tennis (S4), Essay (S5), and Computer (S8) scenarios, returning the compliment (2.3 and 2.3 a) was the most preferred CR among the English group (S4: 29.9%, S5: 43.1%, S8: 31.5%). In the Tennis and Computer contexts, the coach's and the professor's teaching skills are partly responsible for the success in the complimentee's behavior, which explains the need to praise the complimenter. However, it is in the Essay context, where one friend compliments another on a well written paper, where our English participants displayed the highest number of returns (27). The Spanish group also returned the compliment in these scenarios (28) but not as frequently as their English counterpart (S4: 10.1%, S5: 10.1%, S8:13.5%).
It is the nature of the relationship and the cultural context that determines the importance of returning the compliment among our American English respondents. For example, in the Essay scenario, the often close relationship between classmates in the American academic setting, where group projects, peer-assessment, and active class participation are the norm rather than the exception, may explain the need to ensure that the complimenter is equally praised. Despite changes to ensure common standards imposed by the signed Bologna treaty (Bologna Declaration 1999) with other European countries, the Spanish academic setting is still quite traditional and peer-collaboration among students is not as common as in the United States. In addition, in Spain many universities are located in big cities and many students live with their families. This means that friendships within the university are not as easily developed as in the case of students within American campuses. The more distant relationship between classmates in the Spanish setting and the lack of experience sharing personal items, such as an essay, may explain the low frequency of compliment returns in the Essay scenario among the Spanish participants (2.3:10.1%). In a highly competitive and individualized culture as the American English one, the acceptance of one being superior to another should be the norm; but it is also true that the American dream where everyone has the same potential to achieve success is highly valued. In returning a compliment, our English respondents' desire to acknowledge the complimenter's positive face by praising his/her skills in writing a paper prevails over the need to accept the compliment. In addition, as mentioned, Peninsular Spanish speakers do not seem to compliment as often as their American English counterparts. Responding to a compliment with another compliment may seem redundant in Peninsular Spanish, especially when the object of the compliment is someone's skills or efforts. However, when the object of the compliment is on appearance, as in the case of the Outfit scenario (S6), Spanish and English speakers agree on the importance of deflecting the illocutionary force of the compliment by returning the compliment (ENG: 27.1%, SPA: 31.1%)
We can think of two possible explanations for the Spanish respondents' preference for the scaling down sub-strategy, instead of the compliment return, as in the case of the English informants. In the first place, the greater skepticism among Spanish informants in the sincerity of the compliment makes them avoid a response which consists precisely in performing this very same act. Another explanation could be a more hierarchical perspective of social relations on the part of Spanish respondents, which places less value on the individual and his/her personal achievements, especially when these achievements are of an ordinary type. Contrary to American English culture, Spain's political and historical picture reveals a country that spent 40 years under an oppressive regime where only some selected individuals could excel. By many standards, Spain is still a young democracy and it was not until its entry into the European community that Spaniards started to feel valued. It is this cultural context that may explain the perceived need to remove any attention from one's positive outcomes and skills. Nonetheless, avoiding self-praise by diminishing the importance of the complimented object is a strategy that threatens the negative face of the complimenter, who may feel pressured to reiterate and increase the force of the compliment.

Expressing humor
According to Lorenzo-Dus (2001), compliments can be interpreted as an appreciation gesture or as teasing behavior. When a compliment is perceived as a sign of gratitude, a humorous CR serves to avoid self-praise, and instead, increase solidarity among members. On the other hand, when a compliment is interpreted as a teasing act, the humorous response is intended as a defense mechanism to negotiate power against the teasing behavior (Lorenzo-Dus, 2001: 116-117).
Overall our English participants offered more humorous responses (2.6 + 2.6a: 8.1 %) than the Spanish group (5.5%) ( Table 1). The Hair scenario triggered the highest number of humorous responses by the English group (S1: 17.2%) and the second highest by the Spanish group (10.2%), which was only slightly preceded by the Eyes context (S9: 10.4%). In responding to a compliment on a new haircut (S1), both language groups offered teasing responses loaded with comments about the perceived old age of the complimentee. Often the comments appeared in the form of a question (31, 32).
(31) Thanks, but does it really?? I don't want to look younger; I already look like I'm 16! (S1) (32) ¿Me estás llamando vieja? ¡Si sólo tengo 20 años! (S1) "Are you saying I am old? I am only 20 years old!" The interaction between friends allowed for this friendly and teasing exchange mostly triggered by the young age of both participants. Respondents in our study were all university students in their twenties, who most likely want to come across as older rather than younger. In responding to a compliment that focuses on the youthful appearance of the complimentee ("That hair cut makes you look great. It makes you look younger!"), our young informants resorted to humorous references to their age. Teasing and mocking have the potential of being regarded as aggressive behavior because they are non-inclusive and can be a means of social control. However, among friends, teasing is rarely interpreted as insulting or impolite (Shardokova, 2012).
In examining the humorous CRs in the other two scenarios between friends, Essay (S5) and Conference (S7), we see that whereas the English group relied on the use of ironic upgrades (33) and on-record strategies challenging and attacking the other (34), the Spanish group preferred to make funny remarks about themselves and their skills (35). The asymmetrical relationship between interlocutors influenced humorous CRs differently for both English and Spanish. In the Cook scenario (S3), where a student compliments the teacher on his/her cooking abilities, our English respondents relied on the use of humor (16.3%) to mitigate the compliment, offering jocular remarks to remove attention from the complimenter (36). Our Spanish informants, on the other hand, preferred to respond using ironic upgrades (9.8%; 37). The strategy to ironically upgrade the compliment was also present in the Eyes scenario (S9) where an employee compliments the boss' eyes, especially among the Spanish group (10.4%, 38) but not so much by the English group (3%).
The use of humor by the two groups reveals an intricate behavioral picture that we have barely explored here. The data point to some culture-specific humor styles. The English responses are consistent with a view of culture which fosters a friendly but sometimes condescending attitude, in which challenging and personal attacks are acceptable especially among friends. In addition, the use of off-record comments that draw the attention away from the compliment itself allows complimentees to mitigate the illocutionary force by redirecting attention to the positive face of the complimenter. On the other hand, Peninsular Spanish speakers enjoy making fun of themselves and/or ironically upgrading the compliment. However, we should not forget that humor is a multi-faceted behavior which has a complex and varied array of linguistic devices, often difficult to decipher, especially in a written questionnaire. Humor is also multifunctional indicating solidarity, power negotiation, entertainment, aggression, identity, etc. Therefore, these conclusions are only exploratory pending further research on the use and interpretation of humorous CRs in both Spanish and English.

Disagreeing with and refusing a compliment
As indicated in the methodology section, disagreeing with and refusing a compliment are clear ways to reject a compliment and thus, both appear under the Non-Acceptance category. For comparative purposes, we will focus our analysis on disagreement (3.1 and 3.1a) and refusal (3.2 and 3.2a) combined. As evidenced in other studies (Lorenzo-Dus 2001;Maíz-Arévalo 2010), Peninsular Spanish speakers seem more at ease displaying their feelings, even when such an expression implies an attack on the addressee's negative face, as in the case of compliment rejection. Our study confirms these results. Although verbalizing rejection is the least preferred CR, our Spanish participants opposed the complimenter more frequently and across more contexts (3.1+ 3.2 = 3.9 %) than the English respondents (3.1+ 3.2 = 2.6%) ( Table 1).
A qualitative analysis of the data reveals that Spanish rejections were quite formulaic, including expressions like ¡Qué va! ("No way!") or ¡Qué dices! ("What are you talking about!"). Maíz-Arévalo (2010) also found similar expressions in her naturalistic data, and in examining complimenting sequences, she also points out that these rejections were well accepted, as interlocutors did not show any sign of embarrassment. In addition, in our study, as already seen in other types of CRs, the Spanish group expressed their feelings in greater depth by adding informative comments and/or rebukes (39). American English responses, on the other hand, tended to be quite concise (40).
(40) I don't like it. (S1) Responses to compliments on personal appearance resulted in a high number of disagreements and refusals among both language groups. The verbalization of these rejections in the Hair scenario (S1) was quite similar for both Spanish and English, but in the Eyes context (S9), differences in response type were evident. The Eyes scenario revealed that our American English respondents explicitly refused the compliment (41) or interpreted it as a flirtatious request (42). Conversely, the Spanish responses often reflected an interpretation of the compliment as a way to please someone for their own gain (43). The employer-employee relationship in the Eyes scenario appears to be understood differently by American English and Peninsular Spanish cultures. In the American English culture, the assumption that an employee may compliment his/her boss on a personal attribute for personal gain is not common, perhaps due to the fear of sexual harassment. The Peninsular Spanish culture may place greater emphasis on 'connections' and thus, in the work environment one may look to please people in higher positions in the hope that they may help him/her advance professionally. This explains why Spanish respondents did not react to the content of the compliment and its possible hidden flirtatious intentions but instead, responded to the complimenter´s attempt to find favoritism.
In disagreeing with the complimenter, the power relationship was also interpreted differently in the Eyes context. Our American English participants did not show disagreement with the complimenter, whereas the Spanish speakers felt comfortable explicitly disagreeing with the complimenter, who was in a lower social position. The atypical nature of the interaction prompted our English respondents, unsure of what to say, to make the interaction as short as possible. Spanish speakers, on the other hand, expressed contrary views explicitly and concisely (44).

Conclusions
The aim of this paper has been to compare and contrast CRs in American English and Peninsular Spanish. Our data has revealed cross-linguistic differences and similarities. American English speakers claim to use compliments in their everyday communication with more frequency than their Peninsular Spanish counterparts, which may be consistent with the presence of respondents in this latter group who showed a lack of trust on the sincerity of the compliment.
American English CRs are more predictable and systematic than Peninsular Spanish CRs. Saying Thanks or Thank you, alone or in combination with other response types, is a frequently used response type in American English. In order to mitigate the illocutionary force of the compliment, American English speakers like to return the compliment, especially in scenarios where the complimented action is attributed to someone's skills. Peninsular Spanish CRs, on the other hand, are more variable and expressive. In accepting a compliment, speakers directly refer to the complimented object and in avoiding self-praise, speakers like to explain contextual circumstances and/or scale down the force of the compliment.
Humorous CRs were similarly favored by Spanish and English respondents, as they represent the third and the fourth most preferred self-praise avoidance sub-strategy for both language groups respectively. However, humor was expressed differently by English and Spanish speakers. In American English, speakers tease by offering challenging and personal attacks towards the complimenter. In Peninsular Spanish, speakers laugh at themselves or exaggeratedly upgrade the compliment.
Finally, disagreeing with and refusing a compliment are definitely nonpreferred CRs in both groups of respondents. However, Peninsular Spanish speakers are more explicit in their verbalization and/or justification of the rejection, even if this implies an attack on the addressee's face. In addition, Peninsular Spanish speakers do not show reluctance to disagree with a complimenter in a powerful position, whereas American English speakers avoid disagreement in these conditions. All in all, the degree of familiarity with compliments in everyday use in English and Spanish explains not only the more formulaic and less diverse nature of CRs of American English speakers, but also the Peninsular Spanish respondents' perception of complimenters' lack of sincerity, their lower preference for returning the compliment to the complimenter as well as the type of irony and humor directed to the complimenter. The positive politeness nature of the Peninsular Spanish culture may justify the need to offer longer CRs filled with explanations and justifications in an effort to establish a solidarity relationship with the complimenter. American English speakers, however, are used to giving and receiving compliments as part of their everyday interaction regardless of whether they are sincere or not. Therefore, their brief acceptance of the compliment by saying thank you, and perhaps follow it up with praise of the complimenter are acceptable ways to express solidarity and to respect individual freedom.
Our conclusions come with some limitations. On the one hand, the DCT simply shows the explicit pragmatic knowledge displayed in participants' responses. In addition, the non-interactive nature of the DCT does not allow for a full analysis of CRs within the speech event of complimenting. Our results should therefore be verified by an analysis of natural data. In addition, gender has been found to be a significant factor in the act of complimenting and responding to a compliment (Herbert 1990;Holmes 1988;Lorenzo-Dus 2001;Rees-Miller 2011). Due to data collection constraints, we did not control for gender in our study. Finally, our coding scheme, although based on previous studies, reflected the responses obtained in the data. Hence, comparisons with other studies should consider how other researchers define and identify strategies in the data. *The presence of an appreciation token such as thanks or thank you is indicated by 'a'