(Im)Politeness and Offence in Greek Food Blogs

Abstract

Digital communication has been discussed as the locus of impoliteness and conflict par excellence. The aim of this paper is to examine impoliteness in a context of digital communication, that of food blogs, where impoliteness seems to be rather rare. The dataset consists of 2660 comments from 11 Greek food blogs. The data are analysed with the aid of strategies proposed in impoliteness research. The study aims to examine the frequency of offensive behaviour and to identify the issues that may trigger it. Furthermore, it purports to delve into the types and strategies of impoliteness used, and into interactants’ responses to offence. Data analysis showed that impolite behaviour is rare in this specific context and that it is triggered by issues related to features of good recipes and healthy eating practices, among others. It was also found that offence is usually mitigated through politeness strategies. Finally, several cases of offence were found to be disregarded by interactants, while others were resolved amicably. A tendency emerges in Greek food blogs towards the avoidance of impoliteness and the cultivation of relationships of closeness and solidarity.

1. Introduction

Food blogging has recently surged in popularity, even surpassing travel and fashion (Mainolfi et al., 2022). Despite this rise and the fact that some food bloggers engage in influencer marketing, there is limited research on food blogging (Mainolfi et al., 2022, p. 431), particularly regarding how language is used. Addressing this gap, the present study focuses on the discursive practices of food bloggers and commenters in relation to offensive behaviour identified in Greek food blogs. In so doing, the study also responds to Haugh and Sinkeviciute’s (2019, p. 209) call for examining offence and conflict in a range of different settings, including everyday, interpersonal settings. This study expands on previous work on Greek food blogs (Tzanne, 2022) using a larger and slightly differentiated dataset (see Section 2.1). It delves into the discourse employed in this context by focusing on the types and strategies of impoliteness used and the development of instances of offence in a digital context where impoliteness, though rare, deserves exploration. The study purports to contribute to the investigation of food blogs as a genre (Swales, 1990, p. 58), that is, a class of communicative events with specific purposes that shape the structure of the discourse and affect and constrain content and style.

1.1. Food Blogs

Food blogging is “the practice of publishing food-related posts on a blog” (Lee et al., 2014, p. 228) through which experts and non-experts exchange information about food (Cesiri, 2017; Lofgren, 2013). By engaging in food blogging, expert chefs and amateur cooks make their cooking practices known to a wide audience of interested users, with their interaction being characterised by collaborative creativity (O’Hara et al., 2012; Senyei, 2011).

Food blogs have been defined as a written, asynchronous genre of computer-mediated communication (CMC) which revolves around buying, preparing, photographing, consuming and evaluating food in all varieties and contexts (Diemer & Frobenius, 2013, pp. 53–54). In other words, food blogs are “online forums dedicated to the preparation and evaluation of food in a lifestyle context” (Diemer et al., 2014, p. 8), and have been characterised as “a valuable source of information and knowledge regarding food” (Mainolfi et al., 2022, p. 432; see also Lee et al., 2014).

Posting recipes and comments on food blogs often mirrors face-to-face interactions between friends, fostering involvement and contributing to the creation of cohesive ties among interactants. In this sense, food blogs may be viewed as an online community of practice (see, e.g., King, 2019; Kopaczyk-McPherson & Jucker, 2025), with ‘community of practice’ being defined as “an aggregate of people who come together around mutual engagement in some common endeavour” (Eckert & McConnell-Ginet, 1992, p. 464). In this community, some participants are central to the group, while others are non-regular, peripheral members in terms of frequency of participation and affiliation with other members (Mills, 2003; Myers, 2010). As will be argued in Section 3.4, being a core or a peripheral member seems to have implications for the way offensive behaviour is introduced, responded to and develops in the context examined.

1.2. Impoliteness in Digital Communication

Impoliteness can be viewed as the umbrella term, under which can be subsumed types of negatively evaluated behaviour ranging from mild disagreements to situations of conflict. The concept of ‘offence’ is discussed in the relevant literature as being closely related to the concepts of impoliteness and conflict (Culpeper, 2011; Haugh & Sinkeviciute, 2019). According to Culpeper (2011, p. 22), impolite behaviour always causes or gives offence. In Culpeper’s (2011, p. 23) terms, negative and impolite behaviours “always have or are presumed to have emotional consequences for at least one participant, that is, they cause or are presumed to cause offence”.

Digital communication is often claimed to be characterised by the ubiquitous presence of impoliteness, aggression and conflictual behaviour (Antoci et al., 2019; Bou-Franch & Garcés-Conejos Blitvich, 2014; Graham, 2019; Sifianou, 2019). The present study aims to examine whether a similar claim can be made in relation to the specific digital context of food blogs.

1.3. Impoliteness in Food Blogs

Offence and impoliteness in food blogs appear to be largely unexplored. The few studies on this topic examine offensive comments in French and Italian (Thaler-Mannheim, 2014) and Greek food blogs (Tzanne, 2022).

Thaler-Mannheim (2014) examines negative evaluation in 700 comments from two food blogs, one French and one Italian, that expressed disapproval, criticism and disagreement. Analysing her data by using Brown and Levinson’s (1978/1987) politeness strategies, Thaler-Mannheim (2014, pp. 282–284) finds that these comments were usually expressed in an indirect way and that most involved positive politeness strategies (e.g., ‘giving agreement before disagreeing’) used as face-saving strategies. Thaler-Mannheim (2014) does not address the impoliteness strategies used in the offensive comments she examines, and limits her work to the presentation of face-saving strategies used to cope with the threat. In order to address impoliteness in food blogs in a more thorough way, the present study considers both politeness and impoliteness strategies that are involved in the expression and mitigation of negative evaluation.

In another study focusing on the comments section of Greek food blogs, Tzanne (2022) uses taxonomies of politeness (Brown & Levinson, 1978/1987) and impoliteness (Culpeper, 1996, 2011; Garcés-Conejos Blitvich, 2010a, 2018) to analyse the strategies participants employ to express evaluation in 10 Greek blogs. Tzanne (2022) finds that only a small minority of the comments (2.7%) involved impoliteness strategies, while most of the comments involved positive politeness strategies. The study concludes that, when viewed against a backdrop of politeness, the identified instances of impoliteness become foregrounded and deserve our full interpretative attention as important indications of the discourse used in the communicative genre of food blogs.

1.4. Aims of the Study—Principal Conclusions

The present study expands and further elaborates on an earlier study (Tzanne, 2022) on negative relational work in Greek food blogs. Based on a larger dataset, this study deals with instances of offence and impoliteness by focusing on the types (on-record or off-record) and the strategies (positive or negative) of impoliteness used by Greek bloggers and commenters. Also working towards a comprehensive examination of impoliteness in the discourse of food blog comments, the study investigates whether findings from Tzanne (2022) can be corroborated, which could thus provide evidence concerning the discursive features of Greek food blogs, in particular, but also of food blogs as a genre, in general.

A first finding is that offensive behaviour is rather rare (3.8%) in Greek food blogs. Another finding of the present study is that several cases (59.9%) of offensive behaviour are left unaddressed, while only a few are responded to with further offence. When viewed together, these findings suggest that, in the communicative context of food blogs, there is a tendency to minimise or avoid aggression. Additionally, instances of impolite behaviour in the data are found to combine impoliteness with politeness strategies. For example, the impoliteness strategy “Associate other with a negative aspect” may combine with the politeness strategy “Hedge”, probably in an attempt to mitigate the impact of the offence (see Section 3.3). Finally, it is concluded that both food bloggers and commenters seem to be concerned with the protection of their positive face and the maintenance of relationships of closeness and solidarity. This may relate to traits associated with Greek culture, such as the tendency to use positive politeness strategies in their encounters (Makri-Tsilipakou, 2001; Sifianou, 1992). Preoccupation with the protection of positive face and maintenance of relationships of closeness could also be viewed as features pertaining to food blogs as a communicative genre.

2. Materials and Methods

This section presents the process of collecting data and the taxonomies of impoliteness used for the analysis of these data.

2.1. Data Collection

The comments examined in this study were collected from blogs run by amateur cooks, not professional chefs. More precisely, from the information they post on their blogs and other social media accounts, it transpires that these bloggers are not chefs or cookbook authors, nor do they own a restaurant or run a television cooking show. Rather, they are individuals who have discovered an online space in which to share their passion for food and to exchange ideas and experiences about food.

The decision to focus exclusively on comments from amateur, non-professional blogs is based on the assumption that feedback on recipes by well-known professional chefs might be “contaminated” by the commenters’ admiration for the chefs’ public image. Another important reason for focusing exclusively on amateur food blogs is that, in a context where there are no set institutional roles (the chef and her/his fans/followers), the distribution of power is not predetermined and relationships are more likely to be symmetrical, the study of discourse may offer valuable insights into the role of (im)polite comments in the dynamic construction and management of interpersonal relationships.

In order to ensure that the selected blogs were relevant to food blogging, the keywords συνταγές ‘recipes’ and λιχουδιές ‘delicacies’ were used in the search. The search yielded 24 food blog sites containing these keywords in their name, as in Oι λιχουδιές της Κάτιας ‘Katia’s delicacies’ (names of blogs and participants in this study have been replaced by pseudonyms). Most of these sites mentioned a woman’s first name (as in the above site) or a female family relation (e.g., Oι συνταγές της γιαγιάς ‘Grandmother’s recipes’), which can be viewed as creating a sense of familiarity, closeness and intimacy between bloggers and blog visitors. From the initial set of blogs, blogs that concerned printed publications or blogs run by professionals (e.g., Συνταγές της AΒ ‘Recipes by AΒ’, with AB being a well-known Greek chef) were excluded. The final list of food blogs used in this study comprises 11 food blogs, all run by female amateur bloggers. The present dataset differs from that used in Tzanne (2022) as two of the blogs previously analysed became unavailable and were removed from the corpus, while another three new blogs were added to the dataset, all identified in the way described above.

The large size of some blogs (e.g., a blog with 72 categories of dishes) and the huge number of recipes in some others (e.g., a blog with 2873 recipes) made it imperative to limit the number of comments to a manageable size and, at the same time, to include comments from all categories of recipes. To this end, comments from the first ten recipes of all categories of the larger blogs were collected, while comments from the smaller blogs were included in the dataset in their entirety. Comments were numbered according to the order in which they appeared in the blog. The final dataset for the present study consists of 2660 comments, which were posted between 2009 and 2023 and collected between October 2019 and November 2024. In the study, comments appear with their original spelling, grammar and punctuation.

Although the comments analysed are publicly accessible, I concur with Bou-Franch and Garcés-Conejos Blitvich (2018, p. 140) that this does not imply that participants lack expectations of privacy (see also Zimmer, 2010). Accordingly, and in line with the ethical guidelines for Internet research outlined by franzke et al. (2020), measures were taken to safeguard the privacy of all individuals involved in the corpus. Specifically, any names of people, places or dishes mentioned in the study have been replaced by pseudonyms.

Identification of offensive comments in the data was a difficult task, especially in cases where there was no discursive evidence to suggest that offence had been taken. In such cases, I engaged in interpretative work as a ‘legitimate analyst’ (see Holmes, 2005), based on co-text and/or elements indicating the presence of strategies from the relevant taxonomies of impoliteness that have been proposed, modified and refined by Culpeper (1996, 2011) and Garcés-Conejos Blitvich (2010a, 2018). The final dataset comprised 101 offensive comments involving 292 impoliteness strategies.

After the identification of offensive comments, a number of categories were established for analysis: (1) frequency of occurrence, (2) triggering event (issue of contention emerging from the relevant comments), (3) type of impoliteness (on-record and off-record impoliteness), (4) positive and negative impoliteness strategies, (5) politeness strategies (generally employed as mitigation strategies), and (6) ways of responding to offence. The identified offensive comments were then coded accordingly in an Excel file, which enabled the quantitative and qualitative analysis of the data through various searches and sorting operations.

2.2. Taxonomies of Impoliteness for Data Analysis

In his first attempt at the examination of impoliteness, Culpeper (1996, p. 355) approached impoliteness by building a framework for it in relation to Brown and Levinson’s (1978/1987) politeness strategies. Culpeper (1996, p. 356) presents five impoliteness superstrategies which he sees as the direct opposites of Brown and Levinson’s (1978/1987) politeness superstrategies and claims that “they are a means of attacking face”. The impoliteness superstrategies Culpeper (1996, pp. 356–357) proposes are:

(1)

Bald on-record impoliteness—the FTA is performed in a direct, clear, unambiguous and concise way in circumstances where face is not irrelevant or minimised.

(2)

Positive impoliteness—the use of strategies designed to damage the addressee’s positive face wants.

(3)

Negative impoliteness—the use of strategies designed to damage the addressee’s negative face wants.

(4)

Sarcasm or mock politeness—the FTA is performed with the use of politeness strategies that are obviously insincere, and thus remain surface realisations. […]

(5)

Withhold politeness—the absence of politeness work where it would be expected.

Proceeding with the output strategies that aim to satisfy the strategic ends of superstrategies (2) and (3), Culpeper (1996, pp. 357–358) suggests a list that involves output strategies for positive and negative impoliteness. This list has formed the basis of the taxonomy of impoliteness presented in Table 1.

Table 1. Taxonomy of impoliteness strategies.

On Record Impoliteness (ON-IMP)	Off Record Impoliteness (OR-IMP)
Positive impoliteness (PIMP): ignore/snub the other;	Implicated impoliteness
exclude other from activity (EOA); dissociate	(IP)
from other (DFO); be disinterested, unconcerned,
unsympathetic (DUU); use inappropriate identity
markers (IIM); use obscure secretive language
(OSL); make the other feel uncomfortable (MOFU);
seek disagreement (SD); use taboo words (TW);
call the other names (CON)
Negative impoliteness (NIMP); frighten (FR);	Sarcasm (SRC)
condescend, scorn, ridicule (CSR); invade the
other’s space (IOS); explicitly associate other with
a negative aspect (ANA); put the other indebtedness
on record (PIR); hinder or block the other, either	Withhold politeness (WP)
linguistically or physically (BO)

Finally, in another study, Culpeper (2011, p. 134) distinguishes between conventionalised and non-conventionalised (implicational) impoliteness, with the former category involving, among others, insults, pointed criticisms, dismissals, silencers and threats “to which somebody, typically the target, displayed evidence that they took the utterance as impolite”. In the present study, the interpretation of comments that were not responded to as offensive by any of the participants involved was greatly assisted by the occurrence of certain conventionalised impoliteness formulae that led to the evaluation of such comments as ‘potentially offensive’ (see Section 3.2).

The other taxonomy of impoliteness strategies used in this study is Garcés-Conejos Blitvich’s (2010a, 2018). According to Garcés-Conejos Blitvich (2010a, p. 71), this taxonomy is a modified and refined version of the relevant work by Bousfield (2007, 2008), Kienpointner (2008) and Garcés-Conejos Blitvich (2010b). Following Culpeper’s (1996, 2005) taxonomy, she distinguishes between on-record and off-record impoliteness, with the former category including positive and negative impoliteness strategies that threaten participants’ positive and negative face, respectively. Concerning the broad category of off-record impoliteness, Garcés-Conejos Blitvich (2010a, p. 71) distinguishes between:

“implicated impoliteness” (cases where the implicated meaning could correspond to any of the myriad of impolite meanings realised on-record by the strategies listed in the taxonomy), “sarcasm” (cases where the use of politeness is obviously insincere) and “withhold politeness” (cases where politeness is absent where it should be expected or mandatory).

The taxonomy of the types and strategies of impoliteness proposed by Garcés-Conejos Blitvich (2010a, p. 72) is presented in Table 1 below.

The strong point of this taxonomy and the reason it is used in the present study is that the modifications proposed are data-driven, that is, they are modifications that have suggested themselves from the analysis of a corpus of authentic data, namely comments posted in response to YouTube videos (see Garcés-Conejos Blitvich, 2010b).

3. Results

In this section, the main findings of the study are presented. The section begins with the frequency of offensive comments and the issues of contention that trigger them (Section 3.1) and continues with the strategies and types of impoliteness (direct or implicated) these comments involve (Section 3.2). Section 3.3 presents the mitigation strategies participants employ. Finally, the main trajectories that offensive comments may take are presented (Section 3.4).

3.1. Frequency of Occurrence and Issues of Contention

In a dataset of 2660 comments, only 101 cases of offensive behaviour (potential and actual) were identified. This low frequency (3.8%) is similar to that identified in previous studies (Tzanne, 2019, 2022) and could be a characteristic of the discourse of comments in (Greek) food blogs.

Concerning the issues of contention that trigger impoliteness, it was found that the most common trigger, leading to 61% of the impoliteness strategies used, related to negative features of recipes (e.g., inaccurate, incomplete, unclear) that left participants confused and disappointed, especially if they had put a lot of effort into executing the recipe. The frequency of this trigger is not surprising given the topic of the blogs examined; what is worth pointing out, though, is that its percentage in the off-record subset is much higher (70.2%) than that in the on-record subset (52.3%), which suggests that such a trigger is more likely to lead to a comment involving indirect, mitigated face threat. This may probably relate to the commenters’ perception of the trigger as an issue that is highly likely to cause offence in the specific context.

The second most common trigger in the dataset involved issues falling into the category of ‘healthy eating and cooking’ (26%), with interactants arguing about the value of, for example, stock cubes, or the hygiene risk one takes when consuming street food. The latter was also found to lead to the most heated argument and case of conflict in the data, probably because it also involved important local business interests (due to space limitations, this case is not presented here but can be found in Tzanne, 2022, p. 238ff.). Issues of ‘healthy cooking and eating’ were found to trigger offence more often in the on-record (32.5%) than in the off-record (19.1%) subset, probably because such issues are less likely to cause offence.

Other issues of contention were found to relate to the (mis)interpretation of comments made by bloggers (5.1%) or to plagiarism (3%), which concerned cases of uploading a recipe without mentioning its original source (e.g., a cookbook). Less common triggers were matters of taste (2.4%), appropriate use of the Greek language (1.4%), and fasting for religious purposes (1%). These percentages indicate that, for Greek food bloggers and commenters, successful communication and accurate reference to source texts are more important than, for example, individual preferences and taste.

3.2. Types and Strategies of Impoliteness

Almost half (48.3%) of the instances of impoliteness identified in the data are cases of on-record impoliteness, while the rest (51.7%) are cases of off-record impoliteness where the offensive comment is expressed mainly through implicated impoliteness strategies.

Example (1) involves both on-record and off-record impoliteness interpreted as such by one of the participants (‘actual offence’) and by the analyst (‘potential offence’).

(1)	(Blog 2, comments 507–508)
	Tina 21 Ιουνίου 2013—7:59 μ.μ.
	Λείπει το γάλα!!!!!!!!!!!!!!!!!!!! H συνταγή χρειάζεται και γάλα!!!
		Γεωργία 25 Ιουνίου 2013—9:44π.μ.
		Aν δοκιμάσουμε μια συνταγή και δεν μας πετύχει, ενώ έχει πετύχει σε άλλους, δεν σημαίνει ότι έχει λάθος η συνταγή!!! Μάλλον αλλού είναι το λάθος! Ευχαριστώ πολύ!
	Tina 21 June 2013—7:59 p.m.
	Milk is missing!!!!!!!!!!!!!!!!!!!! The recipe needs milk, too!!!
		Georgia 25 June 2013—9:44a.m.
		If we try a recipe and it doesn’t work, while it has worked for others, it doesn’t mean that there is a mistake in the recipe!!! The mistake is probably elsewhere!!! Thanks very much!

In this example, Georgia, the blogger, seems to have been actually offended by a comment which associates her recipe with a negative aspect (ANA), that of being incomplete (Λείπει το γάλα ‘Milk is missing’). The repetition of the strategy (‘Milk is missing!!!!!!!!!!!!!!!!!!!!’—‘The recipe needs milk, too!!!’) and the marked use of punctuation (exclamation marks) may have contributed to the negative force of the comment that appears to have irritated the blogger. In her response to Tina’s comment, the blogger begins by clearly expressing her disagreement (“Seek disagreement (SD)”) with the commenter (‘it doesn’t mean that there is a mistake in the recipe!!!’). Again, in this case, the use of punctuation (exclamation marks) seems to further intensify the disagreement as well as provide evidence for the blogger’s irritation as a result of the offence she has taken. Moreover, the blogger associates the commenter with a negative aspect (ANA), but only indirectly, when she implies that it was the commenter who made a mistake while executing the recipe (‘The mistake is probably elsewhere!!!’), which further implies that the commenter is not a competent cook.

Blogger’s response to Tina’s comment can be seen as evidence of her taking offence at the comment, which marks this case as one of ‘actual offence’ (see Tzanne, 2022, p. 185ff.). On the other hand, this same response constitutes a case of ‘potential offence’ (see Tzanne, 2022, p. 190ff.), as it is an offence that arguably emerges from the above analysis but is not based on discursive evidence, since it was left unanswered by Tina. Therefore, it can only be suggested, but not affirmed, that Tina was offended, too.

3.2.1. On-Record Impoliteness

This category involves strategies with the aid of which the threat is clearly addressed to participants’ positive and/or negative face. The most commonly used on-record impoliteness strategy is “Associate other with a negative aspect (ANA)” (37%). In such instances, it is almost always a recipe or a dish that is characterised negatively (Ελλιπής η συνταγή! ‘The recipe is incomplete!’) and only very rarely (2%) the blogger herself (θεωρώ τη συμπεριφορά σου προσβλητική και χαμηλού επιπέδου ‘I consider your behaviour to be offensive and of low level’). This shows that participants tend to use this strategy with inanimate rather than animate entities, probably because it is less face-threatening to attribute a negative feature to the product (the recipe) than to the face of the producer (blogger) directly.

The second most common strategy used in on-record impoliteness is “Condescend, scorn, ridicule (CSR)” (25.8%). In these instances, the strategy involves the blogger herself in a direct and personal way (χρησιμοποιείς λεξιλόγιο που δεν σου ταιριάζει ‘you use vocabulary that doesn’t become you’). In the whole dataset, the direct use of this strategy is rather limited (13.3%), which suggests a tendency to avoid threatening the blogger’s face directly.

Another common strategy of on-record impoliteness in the data is “Seek disagreement (SD)” (22.5%). In such cases, the commenter expresses clear disagreement with the ingredients of a recipe (θέλει κρασί, όχι κονιάκ ‘it needs wine, not brandy’) or the blogger’s presentation of it (αυτό δεν κάνει για άτομα με υπέρταση ‘this [dish] is not for people with high blood pressure’). Other on-record impoliteness strategies identified are “Dissociate from other (DFO)” (5.3%), “Call the other names (CON)” (2.7%), “Frighten (FR)” (2.7%), “Exclude other from activity (EOA)” (2%), and “Use inappropriate identity markers (IIM)” (1.3%). It is worth noting that, in cases such as δεν ξέρεις τι λες ‘you don’t know what you’re talking about’, the commenter expresses both disagreement (“Seek disagreement (SD)”) and scorn (“Condescend, scorn, ridicule (CSR)”) for the blogger. In general, CSR has been found to combine with other impoliteness strategies as it may be seen as an accompanying force, a result and a consequence of the use of strategies such as “call the other names (CON)” (είσαι τρελλή ‘you are crazy’) or “use inappropriate identity markers (IIM)” (εσύ με τα περί τροφικής δηλητηρίασης ‘you with this food poisoning thing’).

On the basis of the identified strategies, it was found that, in the majority (67.6%) of the cases of on-record impoliteness, participants used strategies of negative impoliteness, while the percentage of positive impoliteness strategies used was much lower (32.5%). This shows that, in Greek food blogs, participants’ positive face was threatened by offensive comments less frequently than their negative face, which indicates a tendency to protect positive face more.

3.2.2. Off-Record Impoliteness

Off-record utterances are, in essence, indirect uses of language (Brown & Levinson, 1978/1987, p. 211). In this sense, the instances of off-record impoliteness identified in Greek food blogs involve strategies with which offence is expressed indirectly. An example of off-record impoliteness in the data is found in the question Μήπως έχεις κάνει κάποιο λάθος; ‘Have you perhaps made a mistake?’, where a request for information is used to express indirect criticism or accusation (Έχεις κάνει κάποιο λάθος ‘You’ve made a mistake’).

The strategies that fall into this category are mainly implicated positive or negative impoliteness strategies, that is, “cases where the implicated meaning could correspond to any of the myriad of impolite meanings realised on-record by the strategies listed in the taxonomy” (Garcés-Conejos Blitvich, 2010a, p. 71). Almost all (94.4%) of the instances of off-record impoliteness involve the implicated strategies “Condescend, scorn, ridicule (CSR)” (51.8%), “Associate other with a negative aspect (ANA)” (31.2%) and “Seek disagreement (SD)” (10.7%). This shows that fewer categories of impoliteness strategies are involved in off-record impoliteness, which is reasonable, given the fact that some strategies (for example, CON and IIM) can only be addressed to someone directly (on-record impoliteness) and not indirectly (off-record impoliteness). Another observation that can be made in relation to these percentages is that the overwhelming majority (88.7%) of the implicated strategies used concerns participants’ negative face (“Condescend, scorn, ridicule (CSR)”, “Associate other with a negative aspect (ANA)”, “Frighten (FR)”) and only a small percentage concerns their positive face (“Seek disagreement (SD)”, 11.3%).

Similar to what was found in relation to the explicit association of others with a negative aspect (on-record ANA), implicated ANA is mainly related to recipes and not participants (bloggers or commenters). For example, the commenter who posts δοκιμάστε και την εύκολη δική μου συνταγή ‘try my own easy recipe, too’ implies that the posted recipe on which they comment is a difficult one, with ‘difficult’ being considered as a negative feature of recipes in general. Furthermore, implicated “Condescend, scorn, ridicule (CSR)” is often found to result from the use of “Associate other with a negative aspect (ANA)”, as in τα υλικά είναι σίγουρα γραμμένα όλα; ‘are all ingredients definitely written down?’; in this comment, ANA is realised indirectly in the form of a question with which the commenter challenges the accuracy of the list of ingredients and implies that the recipe is incomplete. This, in turn, implies that the blogger who posted the recipe is unreliable and at fault here, which indicates the use of implicated CSR in the same comment. It is interesting to note that, in this comment, the use of the negative impoliteness strategy ANA (“Associate other with a negative aspect”) seems to threaten the positive face of the blogger by implying that she is unreliable. In fact, as Lorenzo-Dus et al. (2011) have argued, this strategy is very close to the positive impoliteness strategy CON (“Call the other names”) and should be subsumed under the same superstrategy, that of “Positive impoliteness”. However, as this re-classification has not been followed through by the same authors (see Bou-Franch & Garcés-Conejos Blitvich, 2014; Garcés-Conejos Blitvich, 2018), the present study remains faithful to the original classification, as modified and presented in Garcés-Conejos Blitvich (2010a).

Finally, it was found that the implicated “Seek disagreement (SD)” and/or “Condescend, scorn, ridicule (CSR)” are often combined with “Associate other with a negative aspect (ANA)” (direct or implicated), as in η χρήση του κύβου είναι απαράδεκτη ‘the use of stock cubes is unacceptable’. In this case, the on-record ANA, which associates a cooking practice (‘the use of stock cubes’) with a negative feature (‘unacceptable’), may also imply that the commenter disagrees with the specific cooking practice (“Seek disagreement (SD”) and scorns those who use stock cubes (“Condescend, scorn, ridicule (CSR)”).

In addition to implicated impoliteness strategies, off-record impoliteness involves an insignificant number of cases of “Sarcasm (SRC)” (5%) and “Withhold politeness (WP)” (0.6%). The former involves cases of insincere use of politeness, such as Ευχαριστώ για το ενδιαφέρον ‘Thank you for the interest’ in a case where the commenter complains to other participants that they are indifferent to her problem with the execution of a recipe, or where the commenter calls a dish which, to them, is suspect for food poisoning αυτό το υπέροχο έδεσμα ‘this wonderful delicacy’. As for withholding politeness, there is evidence in the data that the absence of some expected behaviour constitutes impoliteness; in particular, one commenter accuses another commenter of failing to justify her evaluation of a recipe and concludes with λίγη ευγένεια δεν βλάπτει ‘a little politeness does no harm’, which implies that the absence of the expected justification was deemed to be impolite. Interestingly, this last comment can be seen as an instance of potential offence itself, threatening the first commenter’s face with the use of implicated “Associate other with a negative aspect (ANA)” (you have been impolite) and “Condescend, scorn, ridicule (CSR)” (I have contempt for you because of that).

In the dataset examined, off-record impoliteness was often employed in combination with positive and negative politeness strategies that are presented in Section 3.3.

3.3. Mitigating Offence by Using Positive and Negative Politeness Strategies

When examining impoliteness in Greek food blogs, an interesting finding is that the clear majority of these cases (66%) are instances of off-record impoliteness (see Section 3.2.2), but also cases that combine (off-record) impoliteness with positive and/or negative politeness strategies (Brown & Levinson, 1978/1987).

According to Brown and Levinson (1978/1987, p. 60), face is constantly under attack in interaction by Face-Threatening Acts (FTAs); nevertheless, speakers have various strategies at their disposal in order to reduce the threat directed at their own face and at the face of their interlocutor(s) (Brown & Levinson, 1978/1987, p. 68ff.). Among these strategies, positive politeness strategies constitute actions that redress the face threat directed at participants’ positive face (Brown & Levinson, 1978/1987, pp. 101–102) and negative politeness strategies are actions that redress the threat directed at participants’ negative face (Brown & Levinson, 1978/1987, pp. 129–131).

In Greek food blogs, it was found that positive politeness strategies were employed as frequently as negative politeness strategies in combination with some impoliteness strategy. The most frequently occurring positive politeness strategy was praise, an output of “Notice, attend to H (his interests, wants, needs, good” (Brown & Levinson, 1978/1987, pp. 102–103), which satisfies the blogger’s need for approval and appreciation of their recipe (see Tzanne, 2022, p. 87). In the data, praise usually preceded the offensive remark (ωραία συνταγή αλλά βγάζει πολύ μικρή ποσότητα ‘it’s a nice recipe, but it makes a very small quantity’), but also followed it in some cases (παρόλο που χρειάστηκε να σουρώσω το λάδι, έγινε άπαιχτο ‘even though I had to strain the olive oil, it turned out great’).

Another positive politeness strategy used was “Use of in-group identity markers” (Brown & Levinson, 1978/1987, p. 102), as in ας είμαστε λίγο πιο γλυκούληδες βρε παιδιά ‘let’s be a bit sweeter vre guys’, with the particle ‘vre’ (used in informal settings and denoting familiarity and closeness between participants) and the term of address ‘guys’ being the in-group markers used.

The last positive politeness strategy as a mitigator is “Give reasons” (Brown & Levinson, 1978/1987, p. 102), as in δυστυχώς δεν τη φάγαμε γιατί είναι πολύ διαφορετική από αυτήν που φτιάχνω ‘unfortunately, we didn’t eat it because it is very different from the one I make’. In this case, the commenter tries to justify their dislike and rejection of the dish on the grounds of it being simply different from what the commenter’s family is used to. This dissociates the dish from any negative feature, which may help reduce the face threat involved in the rejection and protect the face of both the commenter and blogger in the exchange.

Positive politeness strategies were found in cases where the positive face is primarily threatened, as in cases of “Seek disagreement (SD)”, but also in cases where the negative face of the blogger is also threatened. Moreover, positive politeness was used much more frequently in on-record impoliteness than in off-record impoliteness, which indicates that participants considered it essential for redressing direct face threat.

In addition to positive politeness strategies, negative politeness strategies also co-occurred with impoliteness strategies in offensive comments. The strategy most commonly used in the data was “Question, hedge” (Brown & Levinson, 1978/1987, p. 131) as in Μήπως έχεις κάνει κάποιο λάθος; ‘Have you perhaps made a mistake?’. The other two negative politeness strategies were “state the FTA as a general rule” (Brown & Levinson, 1978/1987, p. 131), as in όταν ένα από τα υλικά που χρησιμοποιούνται δεν αναφέρεται στη λίστα την συνταγή την έχει γράψει άσχετος ‘when one of the ingredients to be used is not on the list the recipe has been written by someone who has no idea what they’re doing’ and “impersonalise” (Brown & Levinson, 1978/1987, p. 131), as in έτσι όπως έχει γραφεί η απάντηση δείχνει εμπαιγμό ‘the way the answer is written shows mockery’.

Finally, it was found that, in a few cases of on-record impoliteness, positive and negative strategies were combined. An example of combined strategies is μου φαίνεται ότι κάτι του λείπει στη γεύση, βέβαια είναι ωραίο κέικ σοκολάτας ‘it seems to me that something is missing in the taste, it is of course a nice chocolate cake’. In this case, the negative politeness strategy ‘hedge’ (μου φαίνεται ‘it seems to me’), combines with the positive politeness strategy of praise (βέβαια είναι ωραίο κέικ σοκολάτας ‘it is of course a nice chocolate cake’), probably in an attempt to redress the threat of the direct “Associate other with a negative aspect (ANA)” strategy used (‘something is missing in the taste’, meaning that the cake is rather tasteless).

3.4. (Non-)Responding to Offence

Instances of offensive behaviour in Greek food blogs are found to develop in various ways. First, the majority of these instances (59.9%) are left unaddressed, which renders them instances of ‘potential offence’ (see the analysis of the Example in Section 3.2). In such cases, there is no discursive evidence for the occurrence of offence and the analyst has to rely on their own interpretation for their identification.

From the instances of offensive behaviour that have been met with a response, only a few cases (7.9%) were responded to in an equally (or more) offensive manner. This way of acknowledging and dealing with offence is found to relate mainly to instances of on-record, unmitigated impoliteness. On the contrary, no intensified aggression or conflict was identified among cases of off-record impoliteness.

Aggression and conflict were also found to arise mainly among non-regular and anonymous participants. Furthermore, some aggressive comments involved participants who perceived themselves to be more powerful than others on account of their (claimed) expert knowledge about a dish.

As for the rest of the cases of offence that were met with a response, bloggers and commenters were found to deal with the offence by adopting an apologetic or reconciliatory tone, which ended the negative episode and, more often than not, led to the creation of a positive atmosphere among interactants (see also Tzanne, 2022, pp. 224, 229ff.).

4. Discussion

In this section, the findings of the study are interpreted and discussed in terms of the threat directed at participants’ face and of the relationships cultivated among them. It is argued that the types of impoliteness and the strategies used suggest that participants tend to mitigate face threat in the encounter in various ways and, in general, avoid using direct and conventionalised impoliteness strategies that constitute a direct blow to the face of others.

The first finding, which confirms similar findings from previous studies on English (Cesiri, 2020) and Greek food blogs (Tzanne, 2019, 2022), is that the percentage of offensive behaviour in the dataset is very low (3.8%). Moreover, only a small minority (2%) of such instances involve conflict. These percentages strongly suggest that, in this specific digital context, aggression and conflict are avoided.

Concerning the issues of contention that triggered offence, it was found that most of the cases of offensive behaviour concerned problematic recipes (incomplete or inaccurate) or practices related to (un)healthy cooking and eating. In the case of problematic recipes, implicated impoliteness strategies were generally used, while in the case of (un)healthy cooking and eating, on-record impoliteness strategies were more frequent. This probably relates to participants’ perception of the former case as more likely to cause offence.

An unexpected finding was that a simple cooking ingredient, such as stock cubes or the practice of consuming street food, could trigger on-record, unmitigated impoliteness and, eventually, highly intensified aggression. In these (few) cases, the on-record impoliteness strategies used levelled direct attacks at participants’ positive face, which undermined interpersonal relationships too. Intensified aggression could here be explained in terms of who participated in these episodes. More specifically, it was found that all encounters of intensified aggression involved only non-regular participants, that is, commenters who appeared only once in the comments section of a blog, and/or anonymous participants. This finding does not lend support to Ogiermann’s (2019) claim that there is no straightforward relationship between the degree of intimacy between participants and the possibility of conflict arising in the encounter. Non-regular presence and anonymity may have led to reduced inhibition, which has been claimed to lead to conflict in digital settings (see, e.g., Bou-Franch & Garcés-Conejos Blitvich, 2014; Graham, 2019; Lorenzo-Dus et al., 2011).

As already stated, approximately half of the cases of offensive behaviour in Greek food blogs are instances of on-record impoliteness. A careful look at these cases reveals that several (28.5%) are mitigated through the use of positive and/or negative politeness strategies. If these mitigated cases are added to the many instances of off-record implicated impoliteness identified, it could be argued that, in the biggest part of the dataset, interactants seem to make an attempt to reduce the risk of the offence and protect the face of all participants involved. The use of mitigation strategies in food blog comments has also been reported in previous studies (Cesiri, 2020; Thaler-Mannheim, 2014) and has important implications for the type of relationships participants tend to cultivate in the digital context. Attempting to reduce the impact of offensive comments is probably done in order to preserve the atmosphere of closeness, solidarity and ingroupness usually established in Greek food blogs where comments involving positive politeness strategies prevail (Tzanne, 2022).

An interesting category of off-record impoliteness involves the use of questions (τα υλικά είναι σίγουρα γραμμένα όλα; ‘are all ingredients definitely written down?’/Μήπως έχεις κάνει κάποιο λάθος; ‘Have you perhaps made a mistake?’) functioning as indirect criticism or accusation (‘you haven’t written down all the ingredients’/‘you’ve made a mistake’). In these cases, the commenter appears to be asking a question that indirectly associates the blogger with a negative aspect (you are not careful, you are unreliable). The form of such implicated impoliteness strategies probably relates to the commenter’s attempt to protect their face by building a base for deniability that is “the speaker’s perceived ability to actually deny a potential message or inference somehow associated with her [sic] utterance in interaction” (Sternau et al., 2017, p. 102). In the specific example, the commenter can protect their own face by claiming that they only asked a simple question and did not criticise the blogger. At the same time, the indirect force of the utterance may allow the blogger to protect their face, too, by ignoring the point of criticism and treating the question as a simple request for information.

In the vast majority (75.7%), the on-record and off-record impoliteness strategies used in Greek food blogs were found to concern primarily participants’ negative face. By contrast, participants’ positive face was the target of offence in noticeably fewer cases (21.6%). This could be viewed as an attempt to avoid levelling attacks at this aspect of face, probably because of the importance Greek interactants attribute to it, given their orientation to positive politeness (see Sifianou, 1992). It was also found that cases of on-record impoliteness were mostly mitigated through the use of positive politeness strategies. In particular, positive politeness strategies were found to combine with on-record positive impoliteness strategies such as “Seek disagreement (SD)”, but also with implicated negative impoliteness strategies such as “Condescend, scorn, ridicule (CSR)”. It is reasonable to argue that these findings reveal the need of participants to prioritise the protection of everyone’s positive face from the threat posed by offensive comments.

Realisations of positive politeness are, according to Brown and Levinson (1978/1987, p. 101), “in many respects simply representative of the normal linguistic behaviour between intimates, where interest and approval of each other’s personality, presuppositions indicating shared wants and shared knowledge, implicit claims to reciprocity of obligations or to reflexivity of wants, etc. are routinely exchanged”. The fact that such linguistic realisations are found in Greek food blogs as mitigators of offence may relate to the attested preference of Greek speakers for cultivating informal relationships of intimacy and solidarity (Makri-Tsilipakou, 2001; Sifianou, 1992; Tzanne, 2001).

In addition to positive politeness strategies, negative politeness strategies were also used together with impoliteness strategies in offensive comments, probably in an attempt to mitigate threat in instances of either positive or negative impoliteness. In this sense, negative politeness seemed to be an appropriate mitigator for threat directed at either the positive or the negative face of the participants. This probably relates to the nature of the specific strategies used (“state the FTA as a general rule”, “impersonalise”) in the dataset, which renders them suitable for all categories and types of impoliteness.

As far as the response to offence is concerned, a first important finding is that the majority (59.9%) of the instances of impoliteness in the data were left unanswered. This may be due to the asynchronous nature of digital communication; for example, a commenter may feel it is not worth replying to an offensive comment posted a week earlier. Lack of response to offensive comments could also be related to the human factor, for example, when, for their own reasons, bloggers may simply stop posting or responding to comments.

Aside from these possibilities, it is also likely that bloggers and commenters may not respond to offensive comments because they do not want to prolong a difficult and perhaps embarrassing (for them) situation. In other words, unaddressed offence may relate to participants’ wish to protect their face against any additional face threat that could be generated if the initial offence were addressed. Overall, unaddressed offence may enable interlocutors to avoid exacerbating the negative atmosphere and to maintain positive relationships with other interactants (Cesiri, 2017; Tzanne, 2022). Finally, at a more general level, leaving offensive comments unaddressed may be associated with the communicative purposes and the kind of interpersonal relationships that are deemed relevant and appropriate in food blogs. In other words, lack of response to offence can be understood as a form of relational work through which interactants orient to and reproduce locally salient norms of appropriate behaviour, thus foregrounding social goals such as the minimization of aggression and conflict and the maintenance of civility in this specific context.

Concerning response to offence, it was found that only a small percentage (7.9%) of offensive comments was responded to in the same tone. These comments were instances of on-record, unmitigated impoliteness, which probably caused enough offence to call for some kind of retaliation as a consequence. On the other hand, no comment involving off-record impoliteness was responded to in an aggressive or conflictual manner, which indicates that such comments are not deemed to be as offensive as those involving on-record impoliteness. Regardless of the gravity of the offence, another possible explanation for the lack of intensified aggression or conflict among cases of implicated impoliteness could be the indirect way in which off-record impoliteness is expressed, which allows people to ignore it and avoid escalation. In general terms, the fact that interactants responded to offence with further offence in only a few cases (7.9%) strongly suggests their unwillingness to engage in conflict and lends further support to the view that the specific context is one where impoliteness is generally avoided.

The main findings of this study (low frequency of intensified aggression, extensive use of positive politeness strategies and limited use of positive impoliteness strategies, and a high percentage of unaddressed offence) can be associated with Greek participants’ attested preference for closeness, solidarity and in-group relationships. However, relating these findings to cultural traits seems to capture only part of the picture, as similar findings from other studies on the discourse of (non-Greek) food blogs (e.g., the use of politeness strategies as mitigators in French and Italian food blogs, Thaler-Mannheim, 2014) suggest that avoidance of aggression and prioritisation of positive face needs could also be perceived as features of food blogs as a communicative genre. Studying types and strategies of (im)politeness in food blogs from various cultural backgrounds is therefore an interesting research prospect in that it may lead to the identification of genre-related discursive practices across datasets, but it may also reveal important cross-cultural differences that can deepen our understanding of, and enhance our knowledge about, the discourse of food blogs.