Electronic Complaints: An Empirical Study of Negative Reviews from Amazon.com Users Advances in Language and Literary Studies

Unprecedented advances have been seen in E-commerce with the spread of digital commerce and customer relations on commercial websites, such as Amazon. As a result, investigation of this type of communication has opened up new horizons for discourse analysts. This study aims to identify the complaint strategies used by customers and the reasons behind them. With this in mind, the researchers formed a corpus of the most helpful negative reviews posted on Amazon. Similar to previous studies on complaints in spoken communication, the present analysis investigated the different complaint strategies speakers used to formulate their complaints. Additionally, politeness strategies and face concerns were examined in these complaints. The results indicated that customers used eight different complaint strategies, the most frequent being showing their disappointment, anger and annoyance while making their complaints. The reasons for their complaints were delivery problems and unhelpful customer services behaviours.


INTRODUCTION
The digital revolution has influenced many sectors in the world of trade. It has quite literally reshaped our understanding of commerce. Its main impact has been felt through websites and social media. According to the 2021 eMarketer report (backlinko.com/social-media-users), the number of social media users has grown to 4.48 billion users. Almost one third of this population use social media regularly (Grewal et al 2021; Sotiriadis, 2017). This is partly because social media platforms have become key players in the world of commerce. Users have found a global arena where they can communicate with other users from different countries. An additional factor in this growth is the overwhelming spread of the English language which is the lingua franca of the modern world, a language upon which online communication and commerce grow (Seidlhofer 2002). Websites are mostly published in English, which makes interaction on a global scale possible.
One important aspect of online shopping is that customers can post their comments about goods and services. Some examples of such platforms are TripAdvisor, eBay, and Amazon. Not surprisingly, the most commonly used language in these comments is yet again English. What custom-ers say about goods and services has always been a valuable source of information for retailers and discourse analysts (Vasquez, 2011). Consequently, both researchers from sectors such as hospitality (Sotiriadis, 2017) and discourse analysts wish to investigate customer reviews in depth (Ho, 2017, Karatepe (in press), Van Herck 2022, Vásquez 2011.
During the last two decades, different retail sectors have adapted discourse analysis to survey customer satisfaction. While the present study has benefited from their findings, it mainly focuses on how customers express their thoughts and emotions in their online complaints and so basically draws from discourse analysis. It should be noted that it is mainly based on the Speech Act Theory and Brown & Levinson's Politeness Theory (Yule, 1996).
As buyers scan online for item data and to assess item options, they are frequently exposed to numerous item comments from different customers. The presence of customer reviews on a website has been shown to improve customer perception of the usefulness and social presence of the website (Kumar and Benbasat, 2006) Van Herck et al 2021;. E-commerce companies have widely used review "helpfulness" as the main method of measuring how consumers assess a comment. For instance, after each customer review, Amazon.com asks, "Was this review helpful to you?" Amazon places this helpfulness information (information such as "26 of 31 individuals found the accompanying review helpful") close to the review and positions the most helpful reviews more prominently on the item's data page. Customers can also classify reviews by their level of helpfulness (Ho, 2017;Mudambi, 2010).
In the light of previous studies, researchers decided to collect customers' reviews posted on Amazon.com and to analyse them according to complaint strategies used and the reasons behind the complaints. It has been argued that platforms such as TripAdvisor and eBay form their own community and discourse (Vasquez, 2014). While some studies have been carried out on reviews posted on TripAdvisor and eBay, reviews posted on Amazon.com have not yet been investigated. Hence, the present study aims to fill in this gap. These reviews will be investigated to answer the following research questions: 1. Does the reason for complaint influence comment writers' linguistic choices? 2. Does it make a difference whether they have one or two reasons for their online complaint? 3. How do customers express negative politeness strategies?

LITERATURE REVIEW
Previous studies have looked into the use of language in online reviews.  put forward that 'It is important to study systematically the language that is characteristics of this genre' (p. 1). In one study, Vasquez (2011) investigated negative hotel reviews. Her study revealed a typical generic structure for negative hotel reviews. Karatepe (2021 'in press') looked at how the restrictive adverb 'only' reflected writers' stance in negative hotel reviews. Furthermore, Meinl (2017) wrote her Ph.D. thesis on the politeness strategies of British and German users on eBay. Based on her findings, she concluded that; "So all in all, individuals from the whole eBay people group appear to have created specific standards on this discussion concerning the utilization of complaint strategies, alteration, and dialect rightness, which conceal other potential social contrasts as to complaining behavior" (p. 225). It is interesting that e-commerce is leading the way in surpassing cultural differences. In the process of interaction, idiosyncratic differences and cultural differences may influence us and we make certain decisions to choose particular communication strategies. We do this in order to protect our social identity, which is referred to as 'face'.

Face as a Social Commodity
Goffman (1967) regards our social face as a commodity. We need to take care of our social face. It should be protected or lost and can it also be saved. Brown & Levinson (1987) borrowed the concept of face. 'They proposed that we want to protect our face against possible damage when we interact with each other.' (Meyerhoff, 2011: p. 88). Brown & Levinson (1987) distinguished two levels of face within the context of a person's fundamental needs as a social being (negative face) and his/her desires (positive face). In order to protect one's own face and that of the addressee, they have to take care of both levels (Brown & Levinson, 1987). In this context, when we make a complaint, we put our face in danger. The addressee can take our complaint as a face threatening act (FTA). For this reason, performing certain speech acts requires extra caution (Yule, 1996).

Complaint as a Speech Act
Speech acts are analysed in terms of their force on three levels. The first one is 'Locutionary force'. It is the basic meaning of the speech and consists of the literal meaning of the words in the speech act. For example, in a warm and stuffy room, one might mildly complain as follows: 'Oh, it is really warm here.' On the surface, this can be taken as a comment on the condition of the room. This is referred to as 'Locutionary force'. This utterance could also be both a complaint and a request aimed at another person in the room, more specifically, a person who is sitting near to a window. If this is the case, it is referred to as the 'Illocutionary force' with the effect of making a request. The third level is called 'Perlocutionary force' and is related to the impact of the illocutionary force. That is, the person who made the comment expects the person sitting near to the window to open the window (Yule, 1996).
When complaints are put into words in an indirect way, there is a possibility that their illocutionary force will not be interpreted. When they are too direct, the complainer may regarded as being impolite. Therefore, it is important to find the most acceptable way of making a complaint.
The way in which people from different cultural backgrounds express their complaints can vary. For example, Karatepe (2021) investigated complaints in complaint letters from university students to the faculty registrar in Turkish. Students asked the registrar to correct their grade which had been mistakenly recorded incorrectly. She found that students tended to use impersonal and indirect methods to express their complaint. Karatepe (2021) concluded that students' choices here may have been influenced by the asymmetrical relationship between themselves and the registrar. In the Turkish context, when you refer to a person in high office, you should be respectful and considerate. Therefore, students preferred to use indirect and impersonal language forms to avoid projecting an accusatory tone, even though the situation was the registrar's fault. They did this in order to save their own face and the other person's face (Brown & Levinson, 1987). Murphy & Neu (1996) studied complaint strategies used by native speakers of American English. First of all, they found that their informants tended to use the pronoun 'we' to indicate that both complainee and complainer are to be blamed. The second major finding was that they asked questions to receive advice and to ask permission to explain their purpose. The other main finding was related to the depersonalisation of the problem and the deliberate diversion of the attention from the individuals in question to the inanimate aspects of the problem. In this way, they avoided blaming anybody. Murphy & Neu (1996) also reported that their native speaker informants were inclined to use mitigating features to soften their complaint. These were 'may be', 'just', 'a little', 'kind of', 'perhaps', 'really', 'you know', 'I don't know'. (p. 204). Finally, these informants preferred to indicate that part of the responsibility of the blame was on themselves. Researchers argued that native speakers of American English used these strategies to facilitate communication between the complainer and the complainee. These findings help us understand how complaints are performed in face-to-face interactions. However, online communication has other factors by default. The current study would also like to investigate to what extent these findings can be applicable in this new way of communication.
Comments on eBay or any other online platform provide a very different context where every single individual brings along their cultural background and understanding of politeness and etiquette -a system which is very conveniently called Netiquette (see also Meinl 2017).
For this reason, investigating communications between individuals from different backgrounds will reveal interesting results.
Communication online has the advantage of anonymity. As some people feel comfortable in anonymity, they tend to use this advantage to its utmost levels in order to insult others. In real life, one would rather avoid this to protect their social face. People are far more cautious while interacting face to face, otherwise, there might be some consequences and that would not be a pleasant experience.
According to Searle's taxonomy (1975), complaints are categorised as expressives since speakers express their sentiments and demeanours in specific situations. Their mental state is that of inconvenience and outrage, which they might possibly make reference to in the propositional content.
When focusing on complaints, one must distinguish between direct and indirect complaints. While indirect complaints (also referred to as 'griping' and 'grumbling'), have the capacity to set up affinity and solidarity between a speaker and listener, a speaker making an immediate protestation communicates his/her objection to a past or continuous activity which does not acclimate with the speaker's desires and interests. The outcomes of this activity are at the expense of the speaker, who considers the listener at any rate to be somewhat in charge of or fit for curing the apparent offense. (Meinl, 2017).
The present study will focus on a set of complaints presented in Amazon data. This category includes both direct and indirect complaints.

The Data of the Present Study
This study comments on the data i.e., electronic complaints which have been downloaded from Amazon. Detailed information is provided below. (See Figure 1)

Data Collection Procedures
The data was collected manually from the feedback forum of the online auction house Amazon, specifically https://www. consumeraffairs.com/online/amazon.html. The corpus included the content of reviews posted from 1 st January 2017 until 1 st December 2017.
Researchers listed the one-star comments and grouped together the most helpful (those comments with helpfulness scores between 13 and 28). Numerical star evaluations for online client reviews ordinarily range from one to five stars. A low evaluation (one star) demonstrates an extremely negative perspective of the item, a high evaluation (five stars) reflects an exceedingly positive perspective of the item, and a three-star rating reflects a moderate view. The star appraisals give an indication of demeanour limit, that is, the deviation from the midpoint of a state of mind scale (Krosnick et al. 1993). One issue with review extremity is how the helpfulness of a review with an outrageous rating of one or five contrasts to that of a survey and a moderate rating of three. That is the reason to analyse the most helpful reviews.

DATA ANALYSIS
Here the focus turns to the analysis of the collected data. Specifically, the use of complaint strategies and the reasons behind the complaints. There are eight complaint strategies, and each review was analysed according to these strategies.

Strategy 1: Expression of disappointment:
When expressing disappointment, the customer in example 1 avoided conflict. S/he did not attribute blame to anybody. Instead, they expressed their disappointment by focusing on his/her own feelings. However, this customer did not refrain from sending this post which included the phrase 'big disappointment' and the word 'fiasco'. Here, we can see that this person chose to use an impersonalised way of expressing their compliant, which agreed with one of the findings in Murphy & Neu (1996). When we look at this from Brown & Levinson's face concept perspective, the customer attempted to protect both his/her own face and the addressee's face. Example: • ….it's turning out to be one big disappointment. The latest fiasco is their promised 2-day delivery….

Strategy 2: Expression of anger or annoyance:
In the case of Strategy 2, the complainer avoided mentioning the name of the party to be blamed. Instead, s/he expressed his/her anger and annoyance about the situation by using powerful vocabulary such as 'horrible, pathetic, suffer, mismanagement and never'. By using netspeak jargon 'do Amazon', this customer seemed to show that s/he was not new to shopping online. Example: • Horrible&patheticapproach -why do we as customer have to suffer on their mismanagement. Never do Amazon again!!!

Strategy 3: Explicit complaint
In an explicit complaint, the complainer gives the precise name of the offensive act and its actors. In our example, s/he however, neither explicitly judged the addressee's behaviour nor sought a payoff. S/he appeared to report what had happened by mentioning the name of the company. This report confirmed that s/he had done his/her duty, which was a face threatening act though the company seems to have failed to do its duty without attempting any obvious attack. Nonetheless, the post was an attack in disguise. Example: • Well, there was no signature although it was required. I had a missing package and informed Amazon.

Strategy 4: Negative judgement
In using Strategy 4, the customer explicitly stated that an offensive act had happened, and the company was responsible.

Strategy 5: Drawing one's own conclusion
After having a negative experience, the customer used Strategy 5, s/he made a decision about the future. That is, s/he decided not to shop on this website anymore. This is something that no retail company would like to hear. Actually, for them, this is a kind of threat. The customer declared that s/ he did his/her part by paying the money by identifying the name of the company (I paid the money to Amazon), but s/ he claimed that the company was not cooperative (they ran away when things go bad.), which obviously disappointed him/her. Example: • I paid the money to Amazon -they are happy to take your money, but they run away when things go bad. I have spent thousands of dollars with them but they will never get another penny from me.

Strategy 6: Warning others
The customer who used Strategy 6 explicitly warned the whole Amazon community by imposing his/her judgement on them by using capital letters, which is a substitute for shouting in digital text. That is, s/he wanted to make an announcement. Again, this is probably something that no commercial company wishes to see or hear. Moreover, word of the mouth apparently works very well in both the digital and non-digital world (Ho 2017;Ruytenbeek et al 2021;Vasquez 2014). It seems the customer made an attack on the company's face as s/he took their seemingly indifferent attitude as a face threatening act. Example: • DO NOT USE AMAZON HOME SERVICES. They are more concerned with covering themselves than helping you.

Strategy 7: Threat:
A threat is the last thing any commercial company would like to hear. It threatens their reputation. In fact, any kind of threat may eventually also threaten the customer. Although it is like a double-edged sword, some people may wish resort to this level of complaint. That is, the conflict may escalate to the level of court judges and lawyers. However, while the threat was indirectly expressed, the customer seemed to have faith that if the case went to court s/he would get what s/he was entitled to. The fact that this customer did not openly write that s/he would sue Amazon was interesting. It showed that s/he still wanted to be discrete. Another striking point was the use of the pronoun 'we', which indicated that there are many customers, and the person was not alone. Even the use of inclusive 'we' alone could be taken as an apparent threat. In addition to this, the sarcastic expression 'That's fine.' and exclamation marks appeared to complete the face threat in the message. Example: • "That's fine. We are covered legally!!"

Strategy 8: Insult:
In using insults, complainers explicitly condemn the accused personally, whereby they often call them demeaning names. Calling people names is regarded as extremely impolite. What is more, it was reinforced with other adjectives 'disgraceful'. The customer sounded extremely angry. Example: • "Totally disgraceful. Amazon& rugged outfitters are lying cheats".

Number of Occurrences of Reasons for Complaints on Amazon.com
There are many different reasons for complaining on Amazon. As illustrated in Figure 2, customers mostly suffer from delivery problems and bad customer service. Many customers complain about shipping problems, late shipping, and misdelivery problems. For example: 'I came home to find dissatisfied with customer service. The main reason behind customer service causing dissatisfaction is the inability of service agents to speak in English. As customers cannot make themselves understood by the service staff, no one can help them to solve their problems. They become fed up with the situation and express this frustration readily. For example: …When you call you always get folks in other countries where English is not their first language, and they often do not understand details of what you are trying to convey to them, but they are quick to pretend like they do… After these two, no refund was given, broken or not working items and the item has not been received are the next most common reasons for complaints. Rudeness of the seller, different item, and postage was too high make up the rest of the reasons.

DISCUSSION AND CONCLUSION
The modern world offers several opportunities to facilitate modern people's lives. One of them is online shopping. In the last decades, online trading has become increasingly popular and widespread. With this popularity, the number of online communities has also been increasing, the most popular of these being e-bay, yelp, Netflix, and Amazon. These communities provide service to millions of people and receive online feedback, both positive and negative. There has been a certain amount of research conducted on online reviews of e-commerce companies. In the global context, Zhu and Zhang (2006), Clemons (2006), Ghose and Ipeirotis (2006), Mudambi and Schuff (2010) are the most prominent examples, while in the local context Altun (2019) analysed the positive reviews of Amazon.com users. The present study was inspired by these studies and the researcher looked at the most helpful negative reviews on Amazon.com in consideration of these three research questions: (1) Does the reason for complaint influence comment writers' linguistic choices? (2) Does it make a difference whether they have one or two reasons for their online complaint? (3) How do customers express negative politeness strategies? According to our analysis, Amazon users chose to state their complaint in a civil and helpful manner. However, if their problem was not solved in a reasonable way, they expressed their feelings impolitely without hesitation. In addition, if their reasons for complaint piled up, their linguistic choice changed in a way that addressees may not have wished to hear or read. two packages opened and empty. The items had been stolen from my front door. Amazon's negligent driver did not follow instructions clearly labelled on my front door to deliver next door……. I have a been a customer for 10 years and purchase a lot from them. No more!!!'. In this example, the customer complained about the misdelivery of his/her package and because of this unpleasant situation, s/he says that s/ he is going to stop purchasing things on Amazon. Moreover, when customers call customer services to resolve problems, the agent generally neglects the customer or says something irrelevant. In some cases, the customers are fed up with the inadequacy of service workers in speaking English. For example; 'I am so sick and tired when I call customer service for Amazon I get someone out of the country. It is very annoying that they can hardly speak English and all I hear is UMMM. IT'S RIDICULOUS! I guess I will have to stop shopping Amazon for this very reason and I am not the only one. I have several people complaining about it.' Customer services' inability in speaking English makes customers consider stopping purchasing goods on Amazon. Most likely, this would be the last thing any online trading company would like to see, hear or experience.

RESULTS
In this study, the researchers manually downloaded content from Amazon to try to identify the complaint strategies used by Amazon.com users and the reasons behind them. Among thousands of reviews, the most helpful negative one-star rating reviews were collected and analysed. According to these analysis results, Amazon users chose to express their complaints politely rather than insult the addressee. The users employed 'expression of disappointment', 'explicit complain' and 'double complaints' as the most preferable strategy to state their problems. While expressing their concerns, most of them chose to do so in a civil manner but some of them didn't hesitate to use offensive or threatening phrases such as 'Why the ** you are fooling the customer then.', 'Amazon is the worst LIAR and customer SERVICE EVER' or "That's fine. We are covered legally"!! However, when it comes to the reasons behind these complaints, two reasons come to the fore: delivery time and customer service. Consumers suffer from no delivery, late deliveries, and misdeliveries. In such circumstances, they have severe issues with finding an addressee. This leads the users to consider how bad the customer service is. This is not the only reason for users to be Consequently, negative reviews on e-commerce platforms may not only be a good source for researchers to explore the linguistic changes in online communities but they may also present an opportunity for the companies to understand their customers' expectations.
The analysis of reviews revealed that customers do not hesitate to perform face threatening acts and using negative politeness strategies. They may even go further and express their anger by using degrading words and expressions (disgraceful), insults (stupid) and threats. Sometimes these threats can be disguised, as in the example 'We are covered legally.' The results overlap those of Meinl (2017). As Meinl indicated at the end of her study, it seems that there is a tendency towards the creation of a kind of netspeak where cultural differences are transcended and netizens complain, are impolite, insult and make threats by following similar strategies and norms.