Males and females ’ complimenting behaviour on the celebrities ’ I nstagram comments

Article history: Received: Jan 22, 2021 Revised: Sept 29, 2021 Accepted: Nov 11, 2021 A compliment is a sincerely positive comment, remark, utterance, or expression about the speaker’s intention to have a good relationship with others by delivering positive or good values toward someone else. This research is purposed to analyze complimenting behavior by males and females found on celebrities’ Instagram comments. This research employed descriptive qualitative research through a documentaries method. 250 comments from 10 most-followed celebrities’ (five males and five females) comment columns were used to identify the complimenting behavior in the form of words, phrases, or clauses containing compliments. This revealed that the differences between males and females in complimenting behavior showed who complimented, to whom they complimented, the compliment types, and the compliment forms. Females complimented more than males did. Besides, complimenting celebrities is commonly directed at the same sex rather than across sex. However, both males and females were more likely to use direct compliments and less indirect compliments. Compliments by both males and females also showed that direct compliment form 8 (ADJ NP!) is dominantly used. The dominance of form 8 occurrence indicated that shorter form of compliment seemed suitable and reasonable for a virtual interaction like Instagram. From this research analysis, it can be concluded that some different contexts of communication require other compliment forms. Therefore, adding a sub-category and additional category from the original compliment forms should be considered.

by a speaker within an utterance, such as stating, asking, requesting, advising, warning, persuading, considering the content of the message, the intention of the speaker, and the effect of the listener. A compliment is regarded as an expressive speech act which expresses the speaker's psychological state of attitudes and emotions, one of which is compliment.
The preliminary complimenting behavior studies in 1970s and 1980s have been influential in current complimenting behavior studies. Regarding to the compliment form, Ye (in Miranda & Hamzah, 2018)) states that compliment is divided into indirect compliment and direct compliment. Indirect compliment refers to unstructured or other creative patterns of compliments. Meanwhile, direct compliment refers to a strictly patterns of compliments. Furthermore, regarding to the direct compliment, the results of Manes & Wolfson (1981) revealed that there were nine formulaic patterns found in American English compliments in face-to face interaction. They are: (1) NP {is/looks} (really) ADJ (e.g.: Your earrings look gorgeous), (2) I (really) {like/love} NP (e.g.: I love your clutch), (3) PRO is (really) (a) ADJ NP (e.g.: That is a nice pose), (4) You V (a) (really) ADJ NP (e.g.: You did a good job), (5) You V (NP) (really) ADV (e.g.: You played guitar really well), (6) You have (a) (really) ADJ NP (e.g.: You have a beautiful daughter), (7) What (a) ADJ NP! (e.g.: What a beautiful view), (8) ADJ NP! (e.g. Nice picture), and (9) Isn't NP ADJ? (e.g.: Isn't that dress pretty?).
However, Manes & Wolfson (1981) examined compliments in face-toface interaction. Nowadays, the globalization era influences other linguists' interests to analyze complimenting behavior in virtual interaction which results different formulaic patterns from the face-to face interaction. Utilizing Manes & Wolfson (1981) nine formulaic patterns, Placencia & Lower (2013) only found patterns 1-4 and 7-8 in Facebook corpus covering American English interaction. Furthermore, Placencia & Lower's (2013) study introduced subcategories in elliptical forms from the original category of Manes & Wolfson (1981). These authors reasoned that virtual interaction shows a greater number in elliptical subcategories. An elliptical form is also regarded as the informal form of compliment in a virtual written medium.
Additionally, some researchers have examined the complimenting behavior studies by sex-based perspective. Manes & Wolfson (1981) and Holmes (1986) revealed that females gave and received more compliments than males. Regarding to whom the compliments are directed, Bolton (1994) states that both males and females directed compliments to the same sex more frequently than to across sex. On the contrary, Miranda & Hamzah (2018) found that both males and females tended to compliment across sex than the same sex. Regarding this finding, they argue that females nowadays have open-minded, thoughtful, and independent speech style in expressing their thoughts and ideas, especially in complimenting. Similarly, Lakoff (1973) states that no matter what status the females are, females speak a subtle and powerful way of disregarding their subordinate position. The differences of males and females in complimenting, according to Khan & Rustam (2017), might happen due to the universalities of the speakers which reflect to the heterogeneity of the speakers' sociolinguistic behavior.
Since some preliminary studies have tried to analyze complimenting behavior between directed at relatives within particular speech community, this research was aimed at complementing the previous research regarding complimenting behavior from sex-based perspective found on the world's most-followed celebrities on Instagram. We consider that analyzing the compliments from the celebrities' Instagram comment columns would result in providing another evidence of the differences between males and females in using language, especially in complimenting. Furthermore, research on complimenting behavior in virtual interaction; especially on Instagram is still lack of attention since research on complimenting behavior in face-to-face interaction gains a much interests of many researchers This research employed the compliments data obtained from the compliments directed to the world's most-followed celebrities on Instagram. Complimenting behavior directed at the most-followed celebrities on Instagram might appear differently since the celebrities are followed by people from different sexes, ages, first language, ethnics, cultures, social stratification, jobs, and education levels. Moreover, the celebrities were commented by hundreds, thousands, to millions of people, some of whom have family relationships, friendships, coworkers, acquaintances, fans, or are just strangers to the celebrity. Followed by millions of Instagram users in the world, many people compliment these celebrities in various ways, such as about their physical appearance, the clothes they wear, the things around them, their attitudes or personalities, and other topics that cannot even be identified.
In identifying the compliment behavior on celebrities' Instagram comment columns by sex-based perspective, we take into account the occurrences of who compliment, to whom the compliments were directed, the types of compliments being employed (direct compliment and indirect compliment), and the formulaic patterns of compliment being employed. The research concludes that female complimented more than males. This conclusion is supported by Placencia & Lower (2013) and Sciriha (2003) studies who also found that females gave more compliments than males. However, compliments on Instagram community commonly happen between the same sexes despite of across sex. In directing the compliments to the same sex, a direct compliment was frequently used than indirect compliment. Regarding to the direct compliment formulaic patterns, this research presented a high occurrence of pattern 8 (ADJ NP) along with its sub-category by both sexes.
The results of this research have led conclusions that complimenting behavior between different sexes differs according to who compliments, to whom they complimented, the compliment forms, and the direct compliment formulaic patterns. Complimenting behavior on virtual interaction of Instagram seems to be similar to complimenting behavior in face-to face interaction, but appear to be more concise and simpler yet more creative than in face-to face interaction. Therefore, this research suggests adding additional categories along with their sub-categories for further study of complimenting behavior on an internet-based communication.

METHOD Subject
The subjects of this study were obtained from the world's most-followed male celebrities and female celebrities on Instagram, who have different nationalities (Portuguese, Spanish, Canadian, and American). Meanwhile, the data used in this research were comments in the form of English words, phrases, clauses, and sentences containing compliments in the comment columns on the celebrities' Instagram accounts. The data sources of this study were purposively chosen from the world's most-followed celebrities listed in Tracalysitics (2020). In identifying the complimenting behavior by sex-based perspective, the researcher limited the world's most-followed celebrities into five world's most-followed male celebrities and five world's most-followed female celebrities on Instagram. The limitation of these celebrities was purposed to equalize the number of celebrities in relation to analysee the complimenting behavior by sex-based perspective. Since these celebrities received huge comments, the compliment data were collected from the five top comments containing compliments from any public accounts (non-private account) on the celebrities' five latest photos. Public accounts (non-private account) were selected to ease the researcher in identifying the complementors' sex. Besides, the photos were also limited into five photos showing the celebrities' portraits uploaded before November 30 th , 2020. Therefore, the total comments containing compliments were 250 data.

Instruments
The main instrument of this study was the researcher herself through reading, identifying, and processing the data of compliments by different sexes. Since this study was limited to describe complimenting behavior by sex-based perspective, including who directed compliments, to whom the compliments were directed, the compliment forms, and the direct compliment formulaic patterns, a datasheet was used as the secondary instrument. The datasheet is presented in Appendix 1.

Data analysis procedures
In analyzing the data, compliments found on males' Instagram comment columns were separated from compliments found on females' Instagram comment columns. Each compliment on different sexes' Instagram accounts was analyzed using Ye's (in Miranda & Hamzah, 2018) theory of compliment form (direct compliment and indirect compliment form), and Manes & Wolfson's (1981) theory of nine direct compliment formulaic patterns. Furthermore, the data are presented in percentages in the form of tables and figures, and analyzed descriptively.

FINDINGS Who directed compliments?
Complimenting behavior on Instagram by sex-based research was considered by identifying the compliment occurrences directed by males and females. As one can see, Table 1 presents the frequency of sex-based distribution in directing compliments on the celebrities' posted photos on Instagram.  Table 1 shows most of the female Instagram users prefer to compliment more frequently than males. The compliments obtained from the male users account for 47.2%, while the compliments obtained from the female users account for 52.8%. However, it can be seen from Table 1 that there is no significant difference between male and female in directing compliment.

To whom the compliments were directed
This study is also purposed to identify the complimenting behavior directed at the celebrities' Instagram comment columns from a sex-based perspective seen from to whom the compliments are directed. Figure 1 presents male and female differences in directing compliments to the different sexes.  Figure 1 presents that both males and females directed the majority compliments to the same sex while directed fewer compliments across sex. As one can see, Figure 1 presents variation in terms of to whom the compliments are directed, from males to males, males to females, females to females, or females to males. However, still, Figure 1 presents there are no significant differences in directing compliments, both to the same sex and across sex.

The frequency of compliment forms
Another purpose of this research is to identify the complimenting behavior directed at the celebrities' Instagram comment columns from a sex-based perspective seen from the frequency of the compliment forms. According to Ye (in Miranda & Hamzah, 2018) compliment form is divided into direct compliment, indirect compliment, and non-compliment. However, this research only focuses on direct and indirect compliment to identify the complimenting behavior by a sex-based perspective. As one can see, Table 2 shows variation in terms of the compliment forms in reference to whom the compliments are directed.  Table 2 shows insignificant difference between the use of direct and indirect compliment. However, we identified that both males and females directed more direct compliments than indirect compliment. Moreover, Table   Male  3 indicates significant difference of each compliment form in reference to whom the compliments are directed. It can be seen that Instagram users, both males and females, tend to give direct compliment and indirect compliment to the same sex than across sex. Furthermore, the discussions of the direct and indirect compliment forms are presented in the following sub-findings.

Direct compliments
Regarding to the syntactic aspects of complimenting behavior in American English by Manes & Wolfson's (1981), they found nine formulaic compliment forms. They are: (1) NP {is/looks} (really) ADJ, (2) I (really) {like/love} NP, (3) PRO is (really) (a) ADJ NP, (4) You V (a) (really) ADJ NP, (5) You V (NP) (really) ADV, (6) You have (a) (really) ADJ NP, (7) What (a) ADJ NP!, (8) ADJ NP!, and (9) Isn't NP ADJ? All these formulaic compliment forms occurred more than two times in Manes & Wolfson's (1981) data; compliment form 1, 2, and 3 are dominantly occur than the rests. Meanwhile, the use of nine formulaic compliment forms in this research is presented in Figure 2. Figure 2 shows that the direct compliment form 8 is far most frequently used by both sexes. Specifically, this research revealed that males are more likely to use the direct compliment form 8, following compliment form 3, compliment form 2, compliment form 1, and compliment form 5, and the compliment forms. On the other hand, females mostly used the direct compliment form 8, following compliment form 2, compliment form 1, compliment form 5, and other compliment forms. Furthermore, direct compliment form 8 is a shorter form than direct compliment 1. On the other side, Figure 2 indicates that males used five out of nine direct compliment forms, in which the direct compliment form of 4, 6, 7, and 9 did not occur. Meanwhile, females used four out of nine direct compliment forms, in which the direct compliment form 3, 4, 6, 7, and 9 did not occur.

Indirect compliment form
The findings on indirect compliments indicate the creativity of Instagram users in directing compliments. Generally, indirect compliments between virtual interactions and indirect compliments in face-to-face interactions have no significant difference, and neither has the indirect compliment between males and females. However, the indirect compliments data obtained in this research seem to have patterns and topics that cannot be specifically described. Therefore, it is difficult to see sex perspective in indirect compliments. If the analysis of direct compliments can be seen from the use of the formulaic forms, the analysis of indirect compliments seems quite difficult to determine certain criteria.

DISCUSSION
This study is purposed to identify the complimenting behavior directed at celebrities' Instagram comment columns from a sex-based perspective seen from who directed the compliments. Each photo in the selected celebrities' photo was examined and a tally was taken from compliments made by other users. This research has obtained 250 compliments from 25 selected photos of 10 selected celebrities' Instagram accounts.
Regarding who directed compliments, this research found that female tend to give more compliments than males. This finding is similar to Sari (2013) who claims that females gave more compliments than males. These findings are reinforced by the results of Placencia & Lower's (2013) and Sciriha (2003) studies which also found that females tended to give more compliments than males. In other words, the high frequency of females' compliments supports the findings of the preliminary studies (Wolfson, 1983;Lee, 2009;Sifianou, 2001;Qanbar, 2012). According to Al-Rousan, Awal, & Salehuddin (2014), females assume compliments as a positive device in building positive relationship while males assume compliments as an embarrassing or a face threatening act. The difference of sex-based distributions of complimenting behavior might happen due to the different understanding of compliment function between males and females.
The reasons to sex-based variations in complimenting behavior are the socio structural aspects, the interpersonal aspect, and the individual aspect. Regarding to these aspects, Zhao (2013) found out that females compliment more because they recognize that complimenting others is regarded as a positive speech act. Specifically, according to Sciriha (2003), they complimented because they also wanted to be complimented. Hence, they pay compliments as much as possible as they wanted the pay back of the compliments. In contrary, male compliment less because they recognize that complimenting others is regarded as a face-threatening act. That is to say, females' compliments function to show their politeness and positive side to build up positive relationship with others. Nevertheless, the finding of this research is contrast with Widyanita and Pasaribu (2019) who found that males gave more compliment than females. Their findings differ as the impact of the narration which governs the characters' personality. Another finding is also found out by (Rees-Miller (2011) that males gave more compliments than females.
However, this research found that the males and females did not present a significant difference in directing compliments. This result might happen in reference to the influence of the technology use in this sophisticated era. Research by Manes & Wolfson (1981), for example, shows significant differences between male and female of American English in directing compliments in face-to-face interaction. On the other side, research by Miranda & Hamzah (2018) revealed insignificant differences between male and female in directing compliments on Facebook. Miranda & Hamzah (2018) claim that the more advanced an era, the less difference between male and female in using language. Furthermore, Either Wolfson (1983) or Placencia & Lower (2013) conclude that females appear to receive more compliment than males. Nevertheless, this research does not consider males and females' differences in receiving compliments because we had balanced the number of complimentee to 5 male celebrities and 5 female celebrities, and limiting 25 compliments on each of these celebrity accounts.
Regarding to whom the compliments were directed, this research seems to be similar with Holmes (1986) and Bolton (1994) who also found that both males and females complimented the same sex more frequently than across sex. In contrary, Sciriha (2003) and Miranda & Hamzah's (2018) studies found that both males and females tend to give more compliments to the opposite sex. They believe that this result might happen due to the phenomenon of sex equality so that people may express their compliments to others without considering the sex of others. They argue that females nowadays have open-minded, thoughtful, and independents speech style in expressing their thoughts and ideas, especially in complimenting.
In this research, the compliments directed by males and females to the same sex of the celebrities might be influenced by the special relationship between top users' and the celebrities. Some users are probably the male celebrities' brothers, boy-friends, male relatives, male co-workers, male acquaintances, male fans, or male strangers. Similarly, some users are probably the female celebrities' sisters, girl-friends, female relatives, female co-workers, female acquaintances, female fans, or female strangers. For some of the celebrities' fans, they might keep following every update of the celebrities they admire or adore. Therefore, it is very common that this type of users always appears in the top comments that fill the comment column space with their compliments or praises of their idols. We can also assume that the users admire, crave, and make the celebrities they admire become their role model. For example, a female celebrity may become a role model of ordinary females who commented on their photos. Vice versa, a famous male athlete may become a role model of other common males who commented on their photos. For some of the celebrities' strangers, they might not always keep following the celebrities, but the compliments they direct might be influenced by the users' interest in the celebrities' job as actors/actresses, musicians, athletes, models, and the like. That is to say, the celebrities attract the strangers to compliment because they also work in the same area as the celebrities do.
The complimenting behavior by a sex-based perspective can also be seen from the use of compliment forms, including direct and indirect compliment forms. This research found that both males and females used direct compliment form than indirect compliment form. This result is in line with Sari's (2013) study which also found that both males and females pay more direct compliments. According to Purnomo (2015), the use of direct compliments reflects closeness signs, affiliation and solidarity. Additionally, they revealed that speakers tend to use direct compliments because they did not intend to make small talk so the speakers compliment to the point or directly. However, this research found that both males and females tend to give direct compliment and indirect compliment to the same sex than across sex. This finding is in contrast with Miranda & Hamzah's (2018) study which revealed that in directing compliments across sex, males are more likely to use direct compliments, while females are more likely to use indirect compliments. The same results also appear in directing compliments to the same sex, in which males are more likely to use direct compliments, while females are more likely to use indirect compliments. The finding of the present research is possibly caused since it is directly proportional to the number of compliments directed to the same sex as compared to across sex.
Specifically, to the direct compliment form, this research found that direct compliment forms on Instagram are indicated by the positive semantic load utilizing adjectives such as "amazing" and "beautiful" or verbs such as like and love. Nevertheless, this research identified another positive semantic load by adjectives, such as stunning, incredible, beautiful, amazing, perfect, awesome, great, the best, unbelievable, cute, pretty, sick, funky, kind, interesting, dope, gorgeous, okay, happy, and fantastic. Moreover, compliments containing the positive semantic load may also be reinforced with intensifiers, such as so, very, completely, always, absolutely, and truly. Moreover, this research found that the distributions of direct compliment forms vary.
These different findings are due to people from different backgrounds who communicate across linguistic and cultural boundaries. Since the Instagram covers people from different sexes, ages, nationalities, ethnics, cultures, social stratification, jobs, and educational levels, the syntactic forms of compliment are also varied. For example, using shorter compliments such as compliment form 8 might be worth for some cultures, but some cultures might regard it as ignorant compliments. Moreover, these differences might also happen since the communication medium is a crucial factor that affected the way complimenters complimented the linguistic boundaries. This statement is supported by Indah & Rifana (2017) who state that the informal online characteristic of interaction on online platforms, especially on Instagram, causes the tendency of using shorter or abbreviated comments by the Instagram users. They add that visual contents presented in Instagram in the form of photos and videos only attract Instagram users' short comments.
However, Mambo (2019) states that the distinctions of the formulaic compliment forms are possibly affected by the focus of the study (sex, age, nationality, ethnic, culture, social stratification, job, or educational level). Al-Rousan et al. (2014) for example, argue that the findings of the previous studies point out that compliments are syntactically formulaic in many languages. Regarding to the formulaic compliment forms, we found 14 additional compliment forms in the data. We decided to take additional compliment forms since some data could not be classified into the current nine formulaic compliment forms. The additional forms are the elliptical form of the original form and another form that has not been included in the nine formulaic compliment forms by Manes & Wolfson (1981). Along with the globalization era, according to Placencia & Lower (2013), abbreviation or elliptic is affected by communication media and the informal context of communication. The invention of these other direct compliment forms is also affected by the differences between the compliment forms proposed by Manes & Wolfson (1981) and the target languages analyzed in some studies. The highlight of additional compliment forms in this research is shown in Table 3.  Table 3 shows that according to this research data, some forms have suggestions for adding sub-categories, some forms have suggestions for changing their forms from their original form, and some new forms. Forms that do not have suggestions for adding sub-categories and changing their forms from their original form are forms 4, 6, 7, and 9 because data are not found in those four forms. Additionally, the forms that have suggestions for adding sub-categories are forms 2, 5, and 8. Meanwhile, the forms that suggestion for changing their forms from their original form is form 1. On the other hand, there are 3 new forms along with their sub-categories.
In form 1, we found that the proposed changes to its original form are: (1) changing is/ looks to a LV (linking verb), and (2) changing really to ADV (adverb). These changes are purposed to other compliments that contain linking verbs apart from is/looks, and adverbs apart from really. In form 2, there are four suggestions for adding sub-categories. The first sub-category is a suggestion of other forms of the original form, including (1) changing I to SubPRO (subject pronoun), (2) changing really to ADV (adverb), (3) changing like/love to a positive verb of compliments, and (4) adding N (noun) other than N.P. (noun phrase). The second subcategory is the elliptical form from the first subcategory without involving N.P. (noun phrase). The third subcategory is the elliptical from the original form without involving I/SubPRO, really/ADV, and N.P. The fourth sub-category is the elliptical form of the original form which does not involve I/SubPRO, really/ADV, but adds N (noun) other than N.P. (noun phrase). The suggestion to add this sub-category is needed so that other compliments that contain subject pronouns apart from I, verbs apart from like/love, and adverbs apart from really can be categorized as form 2.
In form 3, we propose the changing form from its original form, including (1) changing PRO (pronoun) to SubPRO (subject pronoun), (2) changing is to AUX (auxiliary), (3) changing really to ADV (adverb), and (4) adding N (noun) apart from N.P. (noun phrase). These changes are needed because pronouns have several categories. In the context of compliment forms, only two types of pronouns can be specified, namely subject pronouns and object pronouns. In form 3, the appropriate type of pronoun is a subject pronoun. Additionally, these changes are also needed so that other compliments containing auxiliary verbs apart from is and adverb apart from really can also be categorized as form 3.
In form 5, there is no suggestion to change the form from its original form because no such data was found. However, from the data of this research, it can be suggested to add subcategories which are the elliptical form from the original form without involving You, V (verb), NP (noun phrase), and really. Similar to form 5, form 8 also does not have a proposed change of form from its original form because no such data was found. However, from this research data, it can be suggested to add sub-categories which are the elliptical form from the original form without involving NP (noun phrases). The proposed sub-categories are needed so that other compliments containing only an adverb can be categorized as form 5, and other compliments containing only an adjective can be categorized as form 8.
Meanwhile, the results of this research led to 3 new form suggestions that need to be considered as new forms of compliments in virtual interaction. These 3 proposed new forms, probably, can also be used to analyze compliments in face-to-face interactions. These three proposed new forms also include sub-categories which are the elliptical form of the proposed new form, or other forms of the proposed new form. However, it should be noted that the proposed form found in this research is limited to this research data only. To validate the newly discovered forms, it is necessary to analyze data from various other sources, both face-to-face interaction, and virtual interaction. Additionally, there is one other thing related to the new form of compliment that has been found, which may still be considered in writing it in the formulaic form. It is an adverb in compliments that is meant to be limited to the adverb of degree, adverb of frequency, and adverb of manner. However, other considerations are still needed in writing the adverb term formula.
Besides, regarding to indirect compliment form, this research found that the unavailable criteria for analyzing indirect compliment form cause difficulties since the creativity of Instagram users in directing compliments seem to have patterns and topics that cannot be specifically described. Generally, because Instagram includes various users from various languages and cultures, the patterns used in complimenting also differ. For example, in the researcher's first language, namely Bengkulu Malay, the pattern used in compliment tends to begin with an expression of admiration, and then proceed with compliments. In other languages, however, it may be quite common to compliment by only using admiration expressions. Besides, perhaps in other languages, it may be very common to compliment something by relating any other than that which is compliments. In research data, Instagram users used patterns that tend to be different. For example, compliments in the pattern of greetings, questions, or admirations. Additionally, the different language of the Instagram users resulted in different patterns of English compliments. When involving patterns of English compliments, variations appeared creatively that were formed accidentally.
The topics of indirect compliment in this research tend to cover topics related to things that are contained in photos uploaded by celebrities. For example, compliments about celebrities' appearances (face or body shape), about certain celebrities' abilities or skills (soccer skills, acting skills, singing skills, etc.), about the celebrities' personalities or characteristics (celebrities' attention about charity, the celebrities' condolences or concerns for other people's misfortunes, etc.), about things related to the celebrities' possessions (celebrities' clothes or accessories, song albums, children, pets, etc.), and other compliment topics. However, indirect compliments also often involve topics of compliment that are not related to the things contained in the photos uploaded by celebrities. Some of the users leaving indirect compliment by saying their affective feelings (It makes my heart smile and Thankful for you), employing a noun or a phrase reflecting the celebrities they compliment on (goat, king, boss, goals, Gods of football, God on the pitch, babies, beauty, queen, Mami, and Santa baby).
Additionally, this research found that the phenomenon of complimenting behavior on celebrities' Instagram comments happen because the celebrities might become the complimenters' role models in the celebrities' working area. Besides, the complimenters might try to build a relationship with celebrities (as a fan or a stranger) or maintain the current relationship (as siblings, nieces, best friends, co-workers, and acquaintances). Instagram users were more likely to direct compliments freely and pay no attention to sex boundaries; there are only slight differences in complimenting behavior, including who directed compliments, to whom the compliments were directed, and the compliments forms. This insignificant difference is the influence of sex equality in mastering technology which fades males and females' differences in various aspects, especially in the language use of compliments.
Based on the discussion above, this research has limitations that are considered for the discussion section. First, this research was limited to how the complimenting behavior is viewed from one of the external factors of language, namely the speakers' sex factor. The word "sex" is used to distinguish humans based on their reproductive organs and functions; sex is categorized into males and females (Torgrimson & Minson, 2005). Meanwhile, the word "gender," as another term that is very commonly misunderstood, is a term used to distinguish humans based on their self-representation, whether as male or female, or social responses to one's self-representation (Torgrimson & Minson, 2005). Because this study looks at the differences in complimenting behavior between men and women, this research is linguistic research through sex-based research. Therefore, the term "sex" is more appropriate to use in this study than the term "gender." Second, this research data is limited to 250 compliments found in the comments column of 10 celebrities that contain compliments. By limiting the number of celebrities and the number of compliments collected from each celebrity account, the same number is expected to avoid the tendency of the data in one of the speech groups (male or female). Third, this study only analyzes the complimenting behavior of men and women based on who is praising, to whom the compliment is addressed, and how the compliment forms were used. Fourth, in analyzing the compliment form, the data of this study are limited by indicators of the form of compliments proposed by Ye (in Miranda & Hamzah, 2018), including direct compliment form and indirect compliment form. Fifth, the direct compliment form data is limited to the nine direct compliment patterns indicator proposed by Manes & Wolfson (1981).

CONCLUSION
The research of complimenting behavior on celebrities' Instagram comment by sex-based perspective revealed that females gave more compliments than males. Females were more likely to address compliments at the same sex than across sex. Both males and females used direct compliment forms than indirect compliment forms. Specifically, the direct compliment 8 (ADJ NP!) was dominantly used by both sexes which reflected that both sexes were likely to use simple, short, and concise compliments.
Regarding the direct compliment form, adding a sub-category and additional category of direct compliment forms is necessary for classifying forms of direct compliment originating from internet-based communication media. The new compliment forms are needed because virtual interaction's time now shows differences to face-to-face interaction. Virtual interaction is commonly restricted to time and space so that people tend to use shorter or any other way to compliment others. In another case, different compliment forms are influence by culture differences.
Regarding the indirect compliments, we found that sex-based perspective in the indirect compliment cannot be explained specifically because there is no specific criterion to characterize the use of indirect compliments by different sexes. The indirect compliments could only be determined with compliments outside of the direct compliments which have formulaic patterns. Therefore, we expect further research could contribute in providing certain criteria of an indirect compliment is needed to classify certain categories or subcategories in indirect compliments, both in face-toface and virtual interactions. Moreover, more research is also still required to distinguish the complimenting behavior between virtual interaction and face to face interaction within a certain social group/community. Note : (*) : represent the use of emoticons, such as smiley faces or hearts, conveying or supporting positive meanings in compliments.