Investigating the Use of Compliments and Compliment Responses in Persian : Effect of Educational Background

The present study aimed at investigating the effect of the social variable of education on the use of compliments and compliment responses in Persian. To this end, a Discourse Completion Task (DCT) was administered to 200 native Persian speakers from different educational backgrounds. In general, the results revealed that participants tended to use explicit unbound semantic formula as well as non-compliment strategies to give compliments the most. However, they used future reference, contrast, request, and 'other' strategies the least. Furthermore, they followed accept, reject, and evade trend when replying to compliments. Surprisingly, the most common subcategory of compliment response strategy used by participants was downgrade. Return and appreciation tokens were the second and third most frequently used strategies. However, they never used reassignment and topic shift to respond to compliments. In particular, the results suggested the effect of education on determining compliments and compliment responses patterns. While lower educated people preferred non-compliment strategies, higher educated people preferred explicit semantic formula strategies to give compliments the most. In replying to compliments, downgrade occurred most frequently across different educational levels except PhD/MD level. PhD/MD holders used appreciation token the most. The second most frequently used compliment response strategy by all educational levels was return. However, despite minor differences, no marked difference was found among educational levels regarding the least frequent compliments and compliment responses. The findings can provide valuable insight into the cultural and socio-cultural factors affecting the way people compliment, perceive the compliments, and respond to the compliments made on them.


Introduction
For a long time, pragmatic features of language had been ignored in research; however, the inability of learners to handle different situations on the one hand, and the emergence of Hyme's (1971) communicative competence in which pragmatic competence was seen as an essential part of language competence on the other hand, gave importance to the subject of pragmatics in language research (Mohammad-Bagheri, 2015;Razmjoo, Barabadi, & Arfa, 2013).Pragmatic competence is the competence which enables speakers to behave and talk appropriately in different contexts.According to Barron (2003), pragmatic competence can be seen as knowledge of the linguistic resources available in a given language for realizing particular illocutions, knowledge of the sequential aspects of speech acts (SAs) and finally, knowledge of the appropriate contextual use of the particular languages' linguistic resources.First language (L1) speakers acquire pragmatic competence during language acquisition and use pragmatic rules unconsciously when interacting with others; however, second language (L2) learners differ significantly in their performance even if they have access to good amount of pragmatic input (Bardovi-Harlig, 2001).Researchers, like Bardovi-Harlig (2001), Eslami-Rasekh, Eslami-Rasekh, and Fatahi (2004), and Kasper (1997), found that pragmatic competence could not be expected to be developed along with the grammatical competence and there are many linguistically competent learners who do not have comparable pragmatic competence.Thus, Mohammad-Bagheri (2015) suggests that L2 learners need to be provided with tools to acquire pragmatic competence.Among different aspects of pragmatics, SAs such as apologizing, requesting, and complimenting have taken the majority of researchers' attention, and among different types of SAs, compliments (Cs) and compliment responses (CRs) require a great deal of pragmatic insight since they are loaded with cultural and socio-cultural factors and thus they have been of interest for many researchers (Mohammad-Bagheri, 2015).Cheng (2003) believes that research on the SA of complementing can provide useful information about the rules of language use in a speech community, the value system of individual speakers, and the context of culture and situation.These aspects of pragmatics appear to contribute significantly to construction of everyday communication.The performance of SAs depends on sociocultural concerned with instruction of L2 pragmatic features (see, for example, Wolfson, 1981) have found that there is much variation with regard to both pragma-linguistic and socio-pragmatic aspects of L2 and L1 so that learners have to focus their attention on these differences.Thus, such studies are required to be done to describe and compare the SAs of various languages across social and cultural variables.These would help increase the understanding of the norms of language use in other cultures and would help reduce instances of serious miscommunication which might occur in inter-cultural communication situations.Although a plethora of research studies have been conducted so far on Cs and CRs, only a few have worked on the relationship between different strategies of Cs and CRs used by native speakers across social variables in Persian.Majority of studies have attempted to compare different ways of complimenting and responding to Cs across different languages, communities, and cultures (see, for example, Jin-pei, 2013;Sadeghi & Zarei, 2013), or others have taken other factors such as gender difference in Cs and CRs in English into consideration (see, for example, Heidari et al., 2009;Mohammad-Bagheri, 2015).However, the present study is an attempt to investigate the relatively unexplored effect of the social variable of education on C and CR strategies employed by native speakers in Persian since there seems to be a gap in literature regarding this issue.The authors hope that the findings of this study can have a contribution to the interlanguage pragmatic competence of both Iranian learners of English and those who want to learn PSL/PFL.They can also help foreigners who wish to interact with Persian speakers understand factors motivating patterns of Cs and CRs in Persian and avoid misinterpretations especially in high-stake cases like in political domains.Taking the above-discussed issues into consideration, the study thus tries to find out what major categories of Cs and CRs are used by native speakers of Persian and whether or not these features are influenced by their educational levels.

Participants
Considering the social variable investigated in the study, 200 Persian native speakers were selected from different educational backgrounds (40 under high school diploma, 40 high school diploma, 40 BA or BS, 40 MA or MS, and 40 PhD or MD holders).It has to be noted that 100 of the participants were male and 100 were female from various age groups, occupations, and cities of Iran including Tehran, Isfahan, Mashhad, Shiraz, Qom, Zanjan, Qazvin, and Urmia in order that they could be as representative of the target population, Iranian people, as possible.The sites from which participants were chosen were as diverse as possible, as follows: parks, streets, academic settings, family gatherings, markets, e-mails, and social networks.

Instruments
Many studies on pragmatics have used Discourse Completion Tasks (DCTs) to elicit data from the participants (Mackey & Gass, 2005).It is worth noting that data collection methods employed by researchers influence the results.A number of studies (e.g., Golato, 2003;Yuan, 2001) have compared data obtained from DCTs and those from other methods, such as recording spontaneous conversation, and found both similarities and differences.Data obtained from DCTs, however, do not always correspond to natural data (Aston, 1995;Hartford & Bardovi-Harlig, 1992).Similarly, Golato (2003) argues that DCTs are in a crucial sense metapragmatic in that they explicitly require participants not to conversationally interact, but to articulate what they believe would be situationally appropriate responses within possible, yet imaginary, interactional settings.Golato also holds that the DCT is a valid instrument for measuring not pragmatic action, but symbolic action (2003, p. 92).Despite their limitations, DCTs have been widely used as controlled elicitation tools to collect data in the fields of pragmatics mainly because (1) they allow researchers to control for certain variables (e.g., age of respondents), (2) they also allow them to quickly gather large amounts of data and to statistically compare responses from various speakers (Golato, 2003).In addition, their simplicity of use and high degree of control over variables lead to easy replicability (Mackey & Gass, 2005;Yuan, 2002).Taking DCTs' advantages and disadvantages into consideration, the present study employed a DCT developed by Yuan (2002) and translated into Persian to elicit data from the participants.The current Persian DCT was checked, modified, and verified for their authenticity of language by several native speakers of Persian.It was a written questionnaire starting by asking demographic questions regarding the participants' gender, age, and educational levels.It went on with a number of hypothetical situations in which participants were required to write in the space provided, what they would say in real life if similar situations happen to them.The DCT questionnaire had two parts.Part one was designed to explore the possible C strategies used by the participants.As Table 1 indicates, eight topics or scenarios which aimed to reveal positive characteristics of the complimentees (e.g., appearance, possession, kindness, and ability) were described to the participants, so they could have a clear picture of what the topic is and what the relationship between the speakers is.The participants were asked to play the role of the complimenters and give Cs.

Data Collection Procedures
The procedures followed in this study were divided into two main stages.First, based on the social variable considered in the study, 200 Persian native speakers living in different cities of Iran were selected.In the second stage, the Persian DCT was administered to the participants who were given adequate time to complete the questionnaires at their own pace.The collected date were organized and prepared for further analyses.

Part One
The present study employed a framework proposed by Yuan (2002) to analyze different C patterns used by the participants in part one.The participants' responses to the situations provided in part one were divided into three types: Compliment, Non-compliment and Opt out (see Table 3).Yuan divided semantic formulas for Cs into two types: unbound semantic formulas and bound semantic formulas.Unbound semantic formulas refer to those expressions that can act independently as Cs, whereas bound semantic formulas refer to those ones that cannot be considered as Cs by themselves but must be attached to or co-occur with one of the unbound semantic formulas to be interpreted as a C. In the context of fixing a laptop, for example, a response from one can be "Where did you learn to fix laptops?"This response is not regarded as a valid C because it is more of a question seeking for information than a C. It must co-occur with an unbound semantic formula (e.g., You're really good at fixing laptops) to be interpreted as part of a C. Unbound semantic formulas can be also divided into two sub-types: explicit and implicit Cs.Explicit Cs refer to those expressions that carry at least one positive semantic value and are realized by a small set of conventional formulae outside of context.Implicit Cs are those in which the positive value of an expression can be inferred from what is said in a particular situation (Herbert, 1997).Explanation, information question, future reference, contrast, advice and request are examples of bound semantic formulas.Explanation alone cannot be regarded as C.But it co-occurs with an explicit semantic formula.Information question is used in addition to the explicit semantic formula by the complimenter to ask more information about the complimentee's qualification.A complimenter can use future reference to note that the complimentee will have a great future due to his/her good quality.Contrast refers to the case that the complimenter compares or contrasts the qualification of the complementee to that of another one.Advice refers to the case that besides giving a C, the complimenter also gives advice to the complimentee.In some contexts, the complimenter makes some requests to the complimentee which is referred to the request C strategy.
On the other hand, non-compliment refers to responses that cannot be seen as Cs, be it either mere expression of thanks, or bound semantic formula occurring on their own, or replies that do not carry any positive meanings.Finally, opt out refers to the cases where the participants indicate that "I would not say anything" when a C is expected in that situation (Jin-pei, 2013).The present study also used a framework proposed by Holmes (1986) to analyze different CR patterns used by the participants in part two.Holmes divided CRs into three major types: Accept, Reject, and Evade (see Table 4).The first type of responses includes appreciation token, return, and upgrade.Appreciation token refers to verbal or non-verbal signs that a C has been noticed and accepted.Return refers to the case that the complimentee returns the Cs to the complimenter.Upgrade refers to the case that complimentee accepts the C and thinks that the complimenter undercompliments him/her or the complimentary force is insufficient.On the other hand, downgrade or scale down and disagreement are types of rejections.Downgrade refers to the case that the complimentee disagrees with the complimentary force, pointing to some flaw in the object.Disagreement refers to the case that the complimentee does not agree with the Cs.Finally, evade strategies are explanation, reassignment, request interpretation (offer), topic shift, and reassurance.Explanation refers to the case that complimentee offers a comment on how he/she does something.
When the complimentee uses reassignment, he/she transfers the credit to another person.Request interpretation refers to the case that the complimentee interprets the C he/she receives as a request.Topic shift is used when the complimentee is not limited to the CR and he/she initiates a new topic.Finally, reassurance refers to the case that the complimentee is asking confirmation from the complimenter that the C is directed to her/him.

Part one: Compliments (Cs)
In this part, different C strategies used by the participants in different situations are explained.To address the major C types in Persian, 1598 Cs were elicited from 200 Persian native speakers.As indicated in Table 5, participants tended to use explicit unbound semantic formula to give C the most, accounting for 35.16% of the total responses elicited through DCT.Participants used direct Cs to create and reinforce solidarity between the addressees and themselves.This finding supports Yuan's (2002) as well as Jin-pei's (2013) argument that people tend to make direct and positive statements when they pay Cs.Some participants, for example, employed the following C to explicitly value the complimentees' kindness (in all the examples, the italics are Persian and the non-italics are their English equivalents):

Che ghalbe mehraboni dari!
What a kind heart you have!
On the other side, as mentioned earlier, the type of data collection methods employed in research influences the results.
That is, elicited date may differ from naturally occurring speech.It raises a question about the extent to which the elicited data can represent participants' pragmatic competence.The DCT used in the present study might favor the production of Cs which were direct explicit formulas.But if another researcher employs other methods to collect data (e.g., recording spontaneous speech, or role plays), results will probably change.
The second most frequently used strategy was non-compliments which accounts for 27.72% of the total responses.The relatively high frequency of non-compliments, i.e., not paying Cs when the situation calls for them, may be due to the fact that the respondents were not able to distinguish expression of thanks from expression of Cs because borders between the two expressions are blurred.These findings corroborate those of Yuan (2002) and Jin-pei (2013), who found that the most frequently used C strategies by the participants were explicit semantic formula and noncompliments.The following example shows that participants tended to thank rather than compliment when Cs were expected or they failed to distinguish between thanking and complimenting: Additional reason for high percentage of non-compliments is that the point is not only that respondents fail to distinguish between Cs and other expressions, but that in real situations there might be an intrinsic ambiguity or that the respondents intend to use an avoidance strategy.Moreover, in some cases, thanking seems to be more common than complimenting.For example, it is common for people to thank the interlocutor because he/she has been listening to them for a long time than to compliment his/her kindness.
However, the participants used other strategies, including offer, prayer, and blame, future reference and contrast strategies the least which amount to 0.43%, 1.37%, and 1.37% of the total tokens, respectively.These findings also lend credence to Yuan's (2002) and Jin-pei's ( 2013), who reported that other strategies such as joke, contrast, blame, and offer, future reference, contrast, advice, and request were among the least popular C strategies.In fact, they were less preferable because they are more likely to act both as positive politeness strategies consolidating solidarity between people as well as a threat to the negative face than other strategies, especially explicit ones.Use of ironic or ambiguous sentences as Cs, for example, could simply lead to misunderstanding.The study also attempted to investigate whether educational level affects the type of C. To this end, 1598 Cs were taken from Persian native speakers of five different educational levels.As illustrated in Table 6, educational background affected the type of C strategies among the participants.Lower educated people (i.e. the under-diploma and the diploma groups) tended to use non-compliment strategies the most, amounting to 30.85% and 31.63%,respectively, whereas higher educated people (i.e. the BA/BS, MA/MS, and PhD/MD holders) tended to employ explicit semantic formula strategies the most, making up 38.06%, 42.90%, and 39.54% of the total responses, respectively.
Lower educated people may fail to distinguish between thanking and complementing and they also find other ways to express solidarity like asking questions instead of complimenting.Some participants, as an example, asked the following question to get more information about the topic of the Cs and/or establish solidarity:

Az koja gereftish ino?
Where did you get it?
This strategy may be the result of intimacy between the complimenter and complimentee which allows the complimenter to use information question rather than complimenting.It also reveals that lower educated people show their curiosity more explicitly than their counterparts since this behavior may seem less prestigious among the latter.Higher educated people, at the same time, usually give explicit Cs to express politeness and solidarity.They usually tend to explicitly compliment since they consider explicit Cs more polite than other strategies.This finding is in line with that of Jin-pei (2013) revealing that the most commonly used C strategy is explicit semantic formula.
The second most frequently used C strategy by the under-diploma and the diploma groups was explicit semantic formula Cs, accounting for 27.42% and 29.50% of the total responses, respectively.The second most frequently used C strategy by the BA/BS, MA/MS, and PhD/MD holders was also non-compliment responses, accounting for 22.65%, 22.83%, and 30.39%, respectively.
However, the participants who did not have diploma rarely used other strategies (0.85%), and future reference (1.55%).
Finally, the least frequently used C strategies by the PhD/MD holders were other strategies (0.32%), contrast strategy (0.98%), and future reference (0.98%).This finding gives support to that of Golato (2002), Yousefvand (2010), Yousefvand (2012), Razmjoo et al. (2013), and Sadeghi and Zarei (2013), who found that the accept category occurred most frequently in the C exchanges.Golato (2002), for example, found that Germans prefer to accept Cs by saying 'Yes' when receiving Cs.Accordingly, such finding is in conflict with that of Sharifian (2005), who revealed that speakers of Persian largely tended to reject Cs rather than accepting them.This finding is also in contrast with those studies that claim that Asian people tend to reject Cs, or lessen embarrassment and tensions between interlocutors (Chen, 1993).Such observed discrepancy may be attributed to the fact that they might look at the subcategories of CRs rather than major categories.Regarding major categories of CRs, participants used accept strategies to respond to Cs the most; however, the most common subcategory of CR strategy used by them was downgrade which accounts for nearly 25.76%.It was found that Persian native speakers often tried to downgrade themselves which is rooted in their modesty and considered as an important component of Persian culture.This lends credence to Pomerantz's (1978) finding that complimentees have to do two challenging tasks simultaneously: they must agree with the Cs given by the complimenter and avoid self-praise.It is also in harmony with the modesty and agreement maxim principles proposed by Leech (1983).In fact, he proposes six conversational maxims in relation to his politeness principle: tact maxim minimizing cost to hearer and maximizing benefit to hearer; generosity maxim minimizing benefit to speaker, maximizing cost to speaker; approbation maxim minimizing dispraise of hearer and maximizing praise to hearer; modesty maxim minimizing praise to speaker, maximizing dispraise of speaker; agreement maxim minimizing disagreement between speaker and hearer and maximizing agreement between them; and sympathy maxim minimizing antipathy between speaker and hearer and maximizing sympathy between them.
In a similar vein, Sharifian (2005) refers to this modesty maxim as the Persian cultural schema of shekasteh-nafsi 'modesty' which encourages the speakers to downplay their talents, skills, achievements, etc. while praising a similar trait in their interlocutors.Participants, for example, employed the following expression to downgrade their good qualities as a means of expressing their modesty and as a result their politeness: Such finding is different from that of Yuan (2002), who found that Kumming Chinese native speakers employed explanation the most (i.e., 43.28%).This difference can be due to the fact that, as Pomerantz (1978) argues, complimentees must accept the Cs and avoid self-praise simultaneously.While Kumming Chinese speakers evade the Cs by means of explanation to avoid self-praise, Persian speakers downplay their good qualities just due to the Persian cultural schema of shekasteh-nafsi to achieve the same goal.At the same time, both CR strategies may imply acceptance of the Cs even though they do not necessarily express so overtly.In addition, Chinese people are more ready to accept Cs than Iranian people although they prefer to do so implicitly with explanation.
Return and appreciation token were the second and third most frequently used strategies by Persian native speakers, accounting for 23.54% and 19.11%, respectively.Concerning the frequency of different response types in Persian, it can be argued that the accept category comes at the top of CRs with return and appreciation token with high frequency.The main characteristic of this type of CR is that at the same time that the complimentee accepts the illocutionary force of C, he or she tries to accompany this type of response with one or two forms of formulaic expressions in order to avoid selfpraise.In fact, Persian native speakers usually use multiple expressions in response to the C.For example, they initially start with expressions of gratitude (Merci, Thanks) followed by the formulaic expression of return (Lotf darid, Kind of you) and go on by denying the praise (Intori nist, It isn't like this).Relevant literature suggests that speakers of other languages including English also used their own specific formulaic expressions in responding to the Cs; however, the formulaic expressions that were used by Persian speakers were different.These types of responses to Cs could be assumed to come from Persian speakers' culture that allowed the addressees to avoid acceptance of the Cs, which were assumed as self-praises.Moreover, the function of these expressions in Persian speakers' culture, as Sharifian (2005) argues, was to decrease embarrassment and tension between complimenters and complimentees.If the function of the Cs was to make the hearers feel good, the function of the responses other than acceptance would be the same.As indicated in example 5, speaker A complimented on the speaker B's appearance and speaker B first accepted the C but then she returned it to the speaker A to express shekasteh-nafsi, politeness, and respect and to avoid self-praises.
In addition, the popularity of return strategy in responding to Cs may indicate a tendency among Iranians to make offer when they are given a C on their possessions which can be explicated in light of the concept of Ta'arof in Iranian culture.Ta'arof is a part of Iranian culture which indicates politeness, humility, respect, standing on ceremony, as well as hospitality (Sahragard, 2004).This finding corroborates that of Razmjoo et al (2013), who found that making offer was used by Iranians remarkably.However, it contrasts with Yuan's (2002) finding, who reported that making offer was scarcely used by native speakers of Kumming Chinese (i.e., 1.52%).In fact, the concept of Ta'arof in Iranian culture plays a crucial role in making such difference between two languages regarding CR patterns.As seen in example 6, speaker A complimented on the speaker B's possession and speaker B accepted the C and then she made offer.

A: Laptopet kheili zibao aalie!
Your laptop is really beautiful and perfect!B: Merci.Ghabele shomaro nadare!Thanks.It's not worthy of you!However, the participants never used reassignment (0.00%) and topic shift (0.00%) to respond to Cs.Such finding gives support to that of Razmjoo et al. (2013), who reported that reassignment only accounted for 2.2% of all CRs used by Persian speakers.It also confirms Yuan's (2002) finding that people rarely used reassignment (i.e., 1.24%) to reply to the Cs made on them.Reassignment and topic shift are subcategories of evade which show no sign of modesty, respect, politeness, and Ta'arof.That would be only one reason why Iranians have no tendency to employ them.However, who gives Cs to whom, the topic of Cs, when, where, why, and how Cs are made may play a crucial role in determining whether they are evaded or not.For the sake of brevity, the authors will not repeat these C strategies in the following since they were never used by all participants.
The second and third least frequently used CRs by the participants were also other strategies, including silence, smile, blame, insult, or thanking God, and reassurance strategy, accounting for 1.10% and 1.36%, respectively.This finding corroborates that of Razmjoo et al (2013), who found that no respondent remained silent after receiving a compliment.
New categories of CR strategies emerged in this situation which have been called other strategies.This might be the result of misunderstanding from the C receiver side who misinterprets the C as a negative act rather than a positive one.
In other words, the C receiver might had interpreted the C as a sign of envy, tease, sarcasm, or things alike rather than a true C and this had led him to respond with being silent, smiling, blaming, or insulting.The following example clearly illustrates that speaker B misinterpreted the C given by speaker A as sign of tease and hence the C acted in a wrong way rather than a positive way.Another strategy seen in Persian speakers' responses was thanking God.Due to their strong ties with their religion, Persian speakers showed their faith in God deeply embedded within their SAs.Some of the CRs were in the form of a small prayer to the effect that the speakers be blessed from God, and get whatever they wanted with the help of God.These patterns might be linked to such larger aspects of socio-cultural organization as religion.At the same time, these CRs may be motivated by the Persian cultural schema of shekasteh-nafsi in which Persian speakers downplay their positive qualities and attributed them to God or even to chance and luck to express more modesty and politeness.The effect of these factor on the CRs can be simply seen in the following examples: Thank God, it isn't bad.
On the other hand, after other strategies, disagreement and reassurance strategies were used the least which may be due to the fact that these strategies can be a sign of confidence lack among participants.This finding is not in harmony with Yousefvand's (2012), revealing that after appreciation tokens and formulaic expressions, the third most frequently used subcategory by participants was reassurance.This discrepancy might be related to the participants of these studies.The participants of Yousefvand's (2012) study were EFL students with Persian as their L1 background.Hence, pragmatic transfer from English into Persian might have influenced their responses in Persian.Some participants, for example, used CRs like: "Really?" or "Are you sure?" which are common in English and categorized as reassurance strategies.
Whereas the participants of the present study were only Persian native speakers and they did not know necessarily other languages.In fact, even if Persian complimentees get surprised at the Cs made on them, they rarely use reassurance to express their surprise.When they do not feel that their qualities are as good as what the complimenter thinks, they use gratitude (e.g., Thank you!) and then they use rejection through downgrading or disagreement rather than they use reassurance.Finally, the last area of investigation aimed to see whether educational level affects the type of CR.To answer this question, 1172 CRs were elicited through DCT administered to 200 Persian native speakers of five different educational levels.As indicated in Table 8, educational background affected the type of CR strategies among the participants.In responding to Cs, downgrade occurred most frequently across different educational levels except PhD/MD level.Lower educated participants, the under-diploma and the diploma groups and higher educated participants, the BA/BS, and MA/MS groups, tended to use downgrade the most, accounting for 25.71%, 25.71%, 25.00%, and 26.19%, respectively.However, PhD/MD holders used appreciation token the most, accounting for 25.25%.
The findings are in agreement with those of Allami and Montazeri (2011), who found that participants from different educational background equally attempted to accept the Cs and avoid self-praise at the same time.It leads the CRs to downgrade categories.The majority of CR strategies used by higher educated participants, however, were appreciation token which may be the sign of confidence that such participant could feel when encountered with a C giver.They are more likely to accept Cs, and they like to be complimented.In addition, they want to avoid threatening the positive face of those people who compliment them (Razmjoo et al, 2013).They see the disagreement response pattern as an inappropriate option.Another reason for high percentage of appreciation token (e.g., 'Thank you') is related to the data collection method.The nature of DCT questionnaire could oblige speakers to write a response even when they would be normally quiet in spontaneous conversation.
As discussed earlier, the relatively high frequency of downgrade and acceptance is due to the fact that Persian complimentees use multiple CR expressions: they firstly use gratitude, the most common form of acceptance, to accept the credit proposed in the C without obviously accepting the content communicated by the C; they secondly downgrade the C to avoid self-praise.In fact, Iranian complimentees used downgrade the most to express their modesty through the strategy of shekasteh-nafsi , humbling oneself or modesty, which can be considered as one of the culturally-specific Iranian features.This finding confirms the role of modesty in Persian speakers' society.Holmes (1986) found that New Zealand English speakers rarely reject a C. Persian native speakers are also expected to agree with the complimenter but there is still strong pressure on how they can accept the C without seeming to praise themselves.As a result, they reject the C through the use of downgrade to show their modesty and politeness.In Example 10, the complimentee accepted the C but then she used downgrade strategy to show shekasteh-nafsi.The second most frequently used CR strategy by all educational levels was return, the under-diploma group (22.77%), the diploma group (23.26%), the BA/BS group (22.30%), the MA/MS group (25.71%), and PhD/MD group (24.24%).
However, after reassignment and topic shift, the first three groups used other strategies to respond to Cs the least, the under-diploma group (0.77%), the diploma group (0.40%), and the BA/BS group (0.76%).The least frequently used C strategy by MA/MS and PhD/MD holders was also reassurance, accounting for 0.95%, and 0.50%, respectively.Moreover, the second least popular strategy for the under-diploma group (1.93%), the diploma group (1.63%), and the BA/BS group (1.53%) was reassurance.The second least frequently used strategy by MA/MS and PhD/MD holders was disagreement which amounts to 1.42% and 1.51% of the total responses, respectively.Other strategies were also used as much as disagreement by PhD holders (about 1.51%).
In other words, Persian speakers rarely disagreed with the Cs in a simple way, but they tended to make a comment to show their modesty (see example 11).Since the absence of CRs often leads to situations that threaten the positive face (the need to be approved) of complimenters, it is important for complimentees responding to the Cs to minimize the sense of threat.No one remained silent after receiving a compliment; however, this does not mean that the participant did not reject Cs; rather they resorted to verbal responses such as downgrade and question.In Iranian culture, remaining silent as a sign of rejecting a C may not be considered a polite way of responding, although remaining silent along with a facial expression such as smiling as a way of accepting the C is not far from expectation (Razmjoo et al, 2013).
No way.I am not handsome at all.
Generally, such results support those of other studies (e.g., Allami & Montazeri, 2011) suggesting that educational levels has a role in determining CR patterns.

Conclusion
The study was an attempt to investigate general patterns of Cs and CRs in Persian and the possible effect of educational background on them.It was found that native Persian speakers used explicit unbound semantic formula and noncompliment strategies to give C the most.In responding to Cs, they employed downgrade, return, and appreciation token strategies the most.Taking their educational levels into consideration, the study found that lower educated people tended to use non-compliment strategies the most, whereas higher educated people tended to use explicit semantic formula strategies the most.In replying to Cs, downgrade occurred most frequently across different educational levels except PhD/MD level.PhD/MD holders used appreciation token the most.The second most frequently used CR strategy by all educational levels was return.
As seen above, Persian speakers' Cs and CRs needed to be interpreted in the Persian language and culture context.One must bear in mind that the speech event of Cs and CRs could not be interpreted apart from social and cultural context because they relied on shared beliefs and values of the speech community coded into communicative patterns (Yousefvand, 2010(Yousefvand, , 2012)).Similarly, Haliday (2003) states that this pragmatic variation among languages proves that language is not an object devoid of social and cultural elements, but it is a process which is inseparable from the rest of human social life.The present study also confirmed that language and culture were closely intertwined.
The findings of this study will have some implications for cultural studies, foreigners and tourists who interact with Persian speakers, PSL/PFL teachers, materials developers, and researchers.The findings can provide valuable insight into the cultural and socio-cultural factors which affect the way people compliment, perceive Cs, and respond to the Cs made on them in Persian.Such results can also unravel the cultural norms which dominate any society.Particularly they can help foreigners to communicate with Persian speakers appropriately, especially in political domains.No one can deny that misinterpretations in political domains will have serious effects on the countries' relationships.Cs and CRs in Persian are usually employed to have a positive effect on interpersonal relations.Thus, both of them need to be handled appropriately for the outcome actually to be positive.In particular, Persian speakers' CRs may be motivated by the role of modesty and agreement maxims, proposed by Leech (1983), the Persian cultural schema of shekasteh-nafsi, proposed by sharifian ( 2005), and the Persian cultural schema of ta'arof, proposed by Sahragard (2004).They may disagree with and downgrade Cs rather than accept them which reflects the role of shekasteh-nafsi or modesty and politeness, existing in Iranian culture, in determining CR patterns and strengthening solidarity whereas in other cultures disagreeing with and downgrading Cs may be seen unfriendly.Persian complimentees also have tendency to make offer when they are given a C on their possessions even on their appearance and to return positive qualities to the complimenter or others rather than to accept them.This can be explained through the role of Ta'arof, another concept in Iranian culture, in determining CR patterns.They, however, also tend to agree with their interlocutors to establish and maintain solidarity and respect which reflects the role of agreement maxim.Furthermore, although Persian speakers tend to make direct and positive statements when they decide to pay Cs, they sometimes use non-compliments instead of explicit Cs.It may be attributed to the fact that the boundaries between the SA of gratitude and that of C are blurred.It has to be noted that although non-compliments (thanking) are an appropriate way to value positive qualities in Persian, they are not seen as Cs and consequently they may not have such a positive effect.On the whole, if foreigners are consciously aware of the pragma-linguistic and socio-pragmatic similarities and differences between their native languages and Persian, negative effects of transfer will most probably be inhibited.
Moreover, in L2 teaching, SAs and other pragmatic features of language should be high on the agenda.Both teachers and material developers are strongly recommended to pay more attention to these aspects of language.Learners need to become knowledgeable not only in the rules of grammar but also in the sociolinguistic rules of language use.
The results can be also helpful for those teachers whose students are PSL/PFL learners.They help them to find effective ways to promote sociolinguistic competence.They can use them in their materials and syllabi to provide the learners with the necessary tools to make the appropriate pragma-linguistic and socio-pragmatic decisions in the target language.They should also explain the similarities and differences among languages regarding their patterns of complementing.Moreover, researchers can benefit from the findings and conduct further research to shed light on other similar sociolinguistic aspects of languages.
However, the present study was limited in the size of the participants, and in their variation regarding social variables.Further research can be conducted at a larger scale and with a variety of participants from different social variables.
Besides, the tool of data collection can be enhanced by incorporating other ways like direct and indirect observation since the study used only a DCT to collect data.As Kasper and Dahl (1991) opine that some of the most successful studies in literature have employed combined methods of data collection.

Table 1 .
Topics of Cs in the DCT Questionnaire the DCT aimed at eliciting the possible CR strategies used by the participants.As Table2illustrates, four scenarios, in which the informants were the recipients of Cs, were designed.The participants were asked to play the role of the complimentees and respond to Cs they received.

Table 3 .
Yuan's (2002)Framework of C Strategies Advice You'd better open a laptop store.Request Can you check my cell phone too?Non-compliment Your friend listened to your problems.Sorry to take your time.Opt-out Your neighbor is wearing a new shirt.I would say nothing.

Table 4 .
Holmes' (1988)Framework of CR Strategies RejectDowngradeYour friend says: "You speak English very well".I can speak English but not very well.DisagreementNo, I still have a lot to learn.Evade ExplanationYour friend says: "You have a nice laptop".It took me a lot of time to choose the nice one.Reassignment My father gave it to me.

Table 5 .
Distribution of Cs across the Participants

Table 6 .
Distribution of Cs across Educational Levels 3.2 Part two: Compliment Responses (CRs)In this part, different CR strategies employed by the participants in different situations are explained.The study elicited 1172 CRs from 200 Persian native speakers to address this issue.According to Table7, Persian native speakers employed accept strategy to respond to Cs the most, accounting for 48.45% of the total responses.Whereas, the least frequently used CR strategy by the participants was evade strategies, accounting for 23.01% of the total tokens.They also used reject strategies (28.4%) more than evade ones.It was found that the overall pattern of complimenting was accept, reject, and evade.This simply indicates that Iranians are more likely to accept rather than reject a C.

Table 7 .
Distribution of CRs across the Participants

Table 8 .
Distribution of CRs across Educational Levels