Slurring individuals

ABSTRACT

This paper explores the derogatory uses of nicknames within closely-knit social settings such as villages, households, and schools. By examining ethnographic and psychological data on nicknaming practices, this paper contends that pejorative nicknames and slurs share structural and functional attributes. On the one hand, pejorative nicknames and slurs can elicit deep offence regardless of the speaker’s intentions or whether they occur within speech reports. On the other, pejorative nicknames can contribute to creating and reinforcing unjust intra-group hierarchies, hence mirroring the role of slurs within a smaller social scale. To explain these shared attributes, this paper argues (i) that both forms of verbal aggression index multiple dimensional qualities such as ‘negative valence’, ‘neutral arousal’, and ‘high dominance’ rather than discrete emotional categories such as ‘contempt’ or ‘anger’, and (ii) that the expression of high dominance in social interactions underlies their capacity to offend. Then, it translates this hypothesis into a Bayesian model of sociolinguistic variation inspired by Heather Burnett’s work on identity construction, thus integrating pragmatic reasoning into a psychologically informed framework for interpreting emotional cues. By studying both phenomena in tandem, this paper shows how understanding nicknaming dynamics within smaller speech-act communities can contribute to our comprehension of the derogatory impact of slurs in more intricate social contexts.

KEYWORDS:

• slurs
• nicknames
• expressives
• name-calling
• indexical field
• sociolinguistic variation

Derogatory, generic names for ethnic groups in historical American English, or in the language of any plural society, serve social uses of informal social control in the speech community similar to those that personal nicknames serve in smaller groups. (Allen 1988, 217)

1. Introduction

Users of any language know that verbal forms of aggression can be as varied as non-verbal ones. Within the realm of disparaging expressions, slurs are standardly considered the category with the most pernicious impact. Slurs are defined as expressions targeting individuals singled out based on socially relevant attributes such as race (e.g. ‘Spic’), religion (e.g. ‘Kike’), or sexual orientation (e.g. ‘faggot’). When employed as weapons, they overtly convey the speaker’s derogatory attitudes towards those affiliated with these categories. However, unlike other types of derogatory expressions, slurs aptly transform into tools for silencing, oppressing, and ultimately dehumanising members of social groups. Hence, their utterance provokes offence in those who find unjust forms of inter-group dominance relations detrimental to society.

Slurs’ offensiveness has often been investigated by juxtaposing slurs with other forms of verbal aggression. Hom (2012), for instance, asserts that both slurs and adjectives like ‘damned’ or ‘fucked’ derive derogatory truth-conditional meanings externally from social institutions. Jeshion (2013), in turn, argues that slurs conventionally express a speaker’s affective states in a manner analogous to their neutral counterparts when modified by expletive adjectives (e.g. ‘Spic’ vs ‘fucking Hispanic’). Bolinger (2017), in contrast, asserts that slurs’ offensiveness, although more severe, ultimately arises through pragmatic mechanisms similar to those governing impolite and rude speech. Similarly, Nunberg (2018) contends that slurs’ offensiveness derives from conversational mechanisms resembling those underpinning expressions that disparage individuals for their occupations, behaviours, or political orientation.

However, the relationship between slurs and pejorative nicknames has remained largely unexplored. Pejorative nicknames, like other types of nicknames (e.g. pet names), are expressions that modify or substitute individuals’ given names. Hence, they are intricately interwoven into an individual’s history, social interactions, and beliefs within close-knit communities across familiar or professional spheres. Despite their transient nature, anthropologists and social psychologists have meticulously collected data regarding nicknaming practices across various cultures and settings, employing interviews, participant observation, and data analysis methodologies. This body of research provides an empirically grounded outlook on this form of verbal pejoration, thus facilitating a meaningful comparison to slurs.

Upon synthesising this data, this paper argues that slurs and pejorative nicknames possess analogous offence-generation patterns and, what is more, the same discursive function within the speech networks where they circulate. On the one hand, both slurs and pejorative nicknames can elicit deep offence regardless of the speaker’s intentions or whether the nickname is merely reported. Moreover, both can be ‘reclaimed’, i.e. be the subject of processes whereby targets opt to embrace their pernicious sobriquets. On the other hand, in some circumstances, such as in the context of high-school bullying, pejorative nicknames can convey that targets are inferior or unworthy of respect, thereby inflicting enduring and profoundly negative psychological harm upon their targets. In these cases, nicknames effectively function as slurring terms, although reinforcing intra-group, rather than inter-group, dominance relations.

Building upon these premises, the paper develops a unified theory. This theory posits both forms of verbal aggression as indexing multiple dimensional qualities such as ‘negative valence’, ‘neutral arousal’, and ‘high dominance’ rather than single discrete emotional categories such as ‘contempt’ or ‘anger’ and contends that their capacity to express speaker’s high dominance states within social interactions is at the core of their capacity to warrant moral offence. Moreover, the paper translates this hypothesis into a Bayesian model of sociolinguistic variation, drawing inspiration from Heather Burnett’s (2017, 2019) pioneering work on identity construction. This new approach not only accounts for slurs and pejorative nicknames’ uniform offensive profile but also situates pragmatic reasoning within a psychologically grounded framework for interpreting emotional signals.

2. Nicknames, a multiplex phenomenon

Ethnographic examinations of nicknames have predominantly focused on rural settlements across the globe. In these, the small number of surnames, combined with rigid naming requirements, leads to the replication of official names within the same population (Brandes 1975; Glazier 1987). Dorian observes, for instance, that in Gaelic communities, ‘official names were virtually nonfunctional (…) for the simple reason that too many people had the same name’ (1970, 305). Hence, at first sight, nicknaming practices are perceived as emerging for identification purposes. However, it quickly becomes apparent upon closer observation that nicknames can assume various non-practical functions.

Following Pitt-Rivers (1971), numerous anthropologists regard nicknames as a means of upholding social control within a community (Gilmore 1982; Holland 1990). By subjecting socially undesirable traits or behaviours to ridicule, nicknames serve to ‘educate’ about appropriate social behaviour norms. This function transcends cultural boundaries. For instance, Mashiri (2002) notes that the Shona people primarily rely on nicknaming practices to delineate ‘inappropriate or excessive behaviour, uphold cultural ideals and politely rebuking deviant behaviour or personalities.’ (34). Collier and Bricker (1970) observe that Zinacantecos deride those whose appearance or behaviour deviates from the accepted norms. Moreover, Glazier (1987) observes that, in an immigrant Jewish community, nicknames remind individuals that, regardless of their wealth or success, any attempt ‘to put on airs’ will ultimately prove futile among peers.

However, nicknaming practices abound in contradictions. They can connote the speaker’s closeness to the subject or manifest deep-seated animosity (Tait 2006). On the one hand, nicknames can operate as an intimate mode of address within a close-knit community, unlike the more distant, official nomenclature (Lele 2009). Skipper (1986), for example, found that highly derogatory nicknaming among coalminers serves as a mechanism for solidifying one’s affiliation with a group, thereby enhancing ‘solidarity of the work group so necessary for both production and survival in the coal mines.’ (1986, 145). On the other hand, nicknames are wielded as weapons strategically employed in specific situations to steer individuals back towards conformity or, if they already adhere to societal norms, to ensure their continued compliance (Antoun 1968). As Adams (2009) points out, nicknames carry a ‘verdictive force’. By bestowing a pejorative nickname, the coiner asserts their prerogative to appraise the target negatively and rank them low within social interactions. It is this phenomenon that we shall explore further now.

2.1. Nicknames that derogate individuals

2.1.1. Dorian (1970)

In her study of a Gaelic community, Dorian (1970) reports that nicknames (or ‘by-names’) can be broadly categorised as ‘descriptive’ or ‘nonsensical’. The former category encompasses nicknames with discernible meanings or etymological origins. These may allude to distinctive attributes of the individual, such as their physical appearance, personality traits, or behavioural idiosyncrasy. Among these, some are overtly derogatory and thus ‘painfully pointed’. An effeminate man, for example, was branded as ‘Johnnie Lassie’, and an individual who has experienced bedwetting in childhood was christened ‘Spootie lsputil’. In contrast, nonsensical nicknames appear to have either no lexical content or no apparent connection with the individual to whom it is ascribed. Terms such as ‘Fildy’ or ‘Nogie’ fall within this classification.¹

However, the offensiveness of nicknames is often orthogonal to their descriptive connotations. Complimentary nicknames can harbour derisive undertones, particularly those aimed at humbling the individuals for their pretensions to lordly bearings or unique talents (e.g. ‘The King’, ‘The Laird’, ‘The Bard’, etc.). Similarly, nonsensical nicknames, lacking any descriptive connotations, may or may not be offensive. Whereas ‘Fildy’ and ‘Nogie’ are reported to offend, ‘Dodgey’ and ‘Bebban’ are not. Therefore, the appropriate use of nicknames requires deep knowledge of communities’ social structure. As aptly articulated by the author:

Learning to use the bynames is a problem of establishing the offensiveness or inoffensiveness of each byname, a quality that is not always inherent in the by-name but may vary with the identity of the speaker and the designee. That is, for each by-name of (for example) the nonsense variety, one must ask “(in)offensive to whom?” and “(in) offensive from whom?” (306)

In sum, nicknames’ offence can vary as follows:

• Target variation: Targets may or may not take offence when reminded of a particular attribute (e.g. red hair, a lame leg, etc.). Having a father who worked as a cobbler might be a source of pride to one individual but a source of shame to another.

• User variation: Whereas some individuals, often close friends, can use nicknames with impunity, the same usage from a younger or unfamiliar person will likely be met with resentment. Informants who supplied the by-name ‘Cut’ for another villager, for example, noted that they would not address him by that name but that individuals closer in age and friendship can.

• Context variation: Additionally, those interviewed assert that a nickname's offensiveness is neutralised in contexts where the user and the target are close acquaintances and when the appropriate intonation is used.

Additionally, the author describes the speaker’s ‘lack of innocence’ when using pejorative nicknames. The speaker’s noticeable ignorance of the offensiveness of a nickname does not block its harmful impact. English speakers, outsiders to the Gaelic community, often find themselves in difficulties when they try to imitate the linguistic customs of their Gaelic-speaking acquaintances. Some ‘use all the by-names indiscriminately as the functional equivalents of official names or nicknames until they are brought up sharply by anger or laughter, depending on whether the misused by-name is offensive or inoffensive’ (316). The absence of malice does not suffice as grounds for exculpation, a norm that applies even to the author herself:

I had heard one Golspie man called nothing but Nogie (…). I had no idea, until I referred to him that way in what turned out to be his brother-in-law’s house, that he, and on his behalf his family and close friends, objected to the by-name. (305)

2.1.2. Gilmore (1982)

Like Dorian (1970), Gilmore distinguishes between descriptive and nonsensical nicknames. In his study of an Andalusian village pseudonymised as ‘Fuenmayor’, the former category often comprises labels that ridicule the target’s snobbery, such as ‘Pepe “el de la Clase Media”’ (‘Joe Middle-Class’) and gender-reversing insults such as ‘Maripepa Caramelo’ (‘Little Miss Caramel’), used in reference to a man with a high-pitched voice. Nonsensical nicknames, in contrast, can have more vigorous effects than their descriptive counterparts. For instance, an individual christened with the nonsensical nickname ‘Matruco’ achieved notoriety in the town for his furious and unrestrained displays of temper upon hearing the word.

Additionally, nicknames can also be categorised based on whether they arise as distortions of given names or not: an individual with the middle name ‘Hermogenes’ is called ‘Er Mogenes’ (‘Er’ being Andalusian for ‘El’) and ‘Los del Catalino’ are descendants of a woman named ‘Catalina’, considered an unusual name in Fuenmayor. In all these instances, Gilmore observes, ‘meaning’ is irrelevant. Even when the descriptive element of the nickname denotes a neutral attribute, such as the target’s birthplace (e.g. ‘El Sevillano’), they can be viewed as deeply derogatory due to the generalised impression that it ‘cheapens’ the individual, overshadowing all other aspects of their identity.

Subsequently, Gilmore emphasises the ambivalent value of nicknames, whose utterances simultaneously constitute ‘a joke and a slight’ (693). The mere articulation of a nickname engenders satisfaction to everyone present but the target himself, who becomes the object of public amusement. Simultaneously, the utterance inescapably wounds the recipient. In this context, however, nicknaming signifies more than ‘simple rejection’ by the community. Instead, it threatens an individual’s sense of control, an attempt to subjugate them by conveying that it is not them, but the community, to decide for their public image. Among men, nicknames are perceived as assaults, a ‘kind of symbolic castration’ (697). According to Gilmore, this may explain the intense, often futile, outbursts many people exhibit in response to hearing them.² Interviewees declared that their use of nicknames is enhanced by the knowledge that the nickname provokes a strong reaction in the bearer.³

2.1.3. De Klerk and Bosch (1996)

Numerous ethnological studies are set in small-scale peasant villages, thus raising questions about the extension of their dynamics to different contexts. As Brandes (1975) posits, for nicknaming practices to effectively exert their social control function, an agreement among members of the community regarding what constitutes right or wrong (or ‘moral unity’) is essential. Consequently, integrating nicknaming practices into the fragmented design of urban life may appear doubtful. Nevertheless, evidence shows that nicknaming practices manifest across various urban and rural settings insofar as social relations within them are ‘primary, face-to-face, and multifaceted’ (Glazier 1987, 84).

De Klerk and Bosch’s (1996) study of nicknames among South-African adolescents stands as representative. Like Gilmore (1982), the authors distinguish among nicknames that are transformations of an individual personal name (e.g. ‘Furry’: ‘my name is Jennifer, which became furry’) or replacements thereof (e.g. ‘Sput’: ‘he was born when Sputnik was in the news a lot’). Of particular relevance, the authors emphasise the usage of gendered nicknames, which are applied based on prevailing social norms or ideologies. Within this category, one finds ‘objectifying’ nicknames such as ‘Sexy Ankles’. These labels, by drawing attention to specific attributes of the bearer, influence perception and foster false expectations grounded in gender-role stereotypes. This suggests that nicknaming practices among schoolchildren can effectively serve as ‘instruments of the social control of personal appearance and personality projection’ (Starks, Leech, and Willoughby 2012, 136).

In De Klerk and Bosch (1996)’s study, interviews also unveiled that grounds for disliking a nickname do not reduce to a single factor but are variegated. The reasons mentioned by interviewees included the perception that the name was derogatory (e.g. ‘Ndludlu’: ‘When I was young I was fat, but I do not like it because it doesn't sound good’; ‘Gofor’: ‘my older brother always tells me to get things for him … makes me sound like his slave’), a distaste for the nickname based on ‘social’ reasons, such as it no longer aligning with the individual’s identity (e.g. ‘Tomqi’: ‘I am no longer small’), embarrassment (e.g. ‘Toesie’: ‘I get shy if others hear it’), or a belief that the name is overly childish (e.g. ‘Kosie’: ‘dis nie goed vir my manlike ego nie’ [it is not good for my male ego]).

Finally, the study found that interviewees frequently found themselves uncertain about their stance on when they approved or disapproved of nickname usage, hence indicating their awareness of the strong contextual constraints they impose (‘it depended on who used it and in what context’, ‘when I first met him his friends used it and as the friendship progressed I started using it as well’). Lastly, the authors observe that certain nicknames are not disliked despite their offensive connotations. This tolerance stemmed from the playful or affectionate effects inherent in their use by particular people (e.g. ‘Pong’: ‘from the rhyme Inky pinky ponky, daddy bought a donkey – only my brother uses it’). Astonishingly, some individuals actively embrace and propagate their pejorative nicknames. Despite the harmful intent behind these nicknames, their uniqueness was perceived by their bearers as enhancing popularity (e.g. ‘Mandoza’: ‘It's nice to be called differently sometimes’).

2.2. Nicknames, name-calling, and bullying

The investigations summarised have primarily centred on practices within specific communities and relatively circumscribed timeframes. In these communities, nicknaming emerges as a tool to target any individual, including dominant community members. Consequently, ascertaining the longitudinal impact of nicknaming practices becomes challenging, leaving it unclear whether and to what extent their social and psychological effects may diverge from those of other community social control practices such as gossiping or public censure.

Nonetheless, nicknames’ longitudinal effects become more evident within the realm of ‘name-calling’, one of the most pervasive forms of bullying. Following Olweus’ framework (1993), bullying constitutes intentional acts of aggression that are carried out repeatedly and involve an imbalance of power, either actual or perceived, between the victim and aggressor. This definition encompasses various forms of abuse, including verbal bullying (teasing and name-calling) and physical bullying (physical threats or harm). Among these, name-calling is the most frequently reported (Peterson and Ray 2006), particularly among junior/middle and secondary school children (Whitney and Smith 1993). Notably, verbal forms of aggression often escalate into physical forms as well (Boulton, Bucci, and Hawker 1999) and constitute a significant risk factor for poor physical and mental health (Espelage, Rao, and De La Rue 2013).

Thus, within inter-individual dominance relations, imposing undesired labels constitutes verbal aggression capable of harming the target’s psychological and social well-being. It may be argued, however, that name-calling and nicknaming constitute different forms of verbal aggression. Even though nicknaming and name-calling have been standardly considered a single phenomenon within the literature on bullying, an individual may be subjected to different labels (e.g. ‘effeminate’, ‘skinny’, ‘stupid’, etc.) without necessarily being associated with a unique nickname (Starks, Leech, and Willoughby 2012). Therefore, findings regarding the enduring detrimental effects of name-calling may not necessarily be attributed to nicknaming practices.

However, calling an individual with a label F (as in vocative constructions such as ‘Hey, F!’) and entirely replacing a target’s name with F (as in ‘F is reading’) constitute two phases within the same overarching process. Crozier and Skliopidou’s (2002) research on the long-term effects of being called names during school years is relevant in this context. The authors conducted surveys seeking adults’ current views on name-calling experiences. Most rated the names they had received as moderately, very, extremely, or quite hurtful. The reported age of onset ranged from 4 to 18 years, with a mean of 11.30 years, and its persistence could go from less than one year to between 2 and 6 years (71%) to at least 10 years (22%). Notably, the same expressions tended to be employed by several people rather than a sole person. Among those who also endured physical bullying, half said it was connected to name-calling. All these results were more pronounced for the ‘most hurt’ category, which reported an earlier mean of onset, longer duration, higher incidence of physical bullying, and more significant long-term effects on their personality than the ‘moderately hurt’. Therefore, in name-calling practices, the most hurtful names often operate like nicknames rather than ephemeral expressions only employed in the heat of the moment.

2.3. Summary of the data

As observed, nicknames mainly serve as tools of social control, reinforcing established norms and power differentials. They can be categorised along various dimensions:

• Descriptive or non-sense: Many nicknames incorporate descriptions, while others remain enigmatic.

• Derivations or replacements: Nicknames may derive from the target’s official name or replace it altogether.

• Normative or not: Some nicknames exert their social control function by directly appealing to norms regarding gender (e.g. sex-reversal insults), while others do not.

It is worth highlighting that no single factor about a nickname is sufficient to make it offensive. Nicknames can be descriptively positive, neutral, or non-sensical yet retain the capacity to inflict offence. Moreover, their degree of offence is modulated by multiple contextual factors, including the speaker's identity, the situation in which the term is employed, etc. Nicknames can offend whether they are used or merely mentioned, whether in a scientific context or joyful conversations. Crucially, they offend regardless of whether the speaker knew the term’s offensiveness. However, nicknames can simultaneously enhance affiliation in two ways: (i) among friends, where their use is accepted as part of ‘mock aggression’, and (ii) among bigots, who assert a common distance from the target, thus promoting cohesion and trust. Lastly, some nicknames are accepted with pride despite their negative connotations, mostly when the target wants to underscore their singularity.

3. Comparing slurs and pejorative nicknames

3.1. Slurs: definition, functions, and offence-generation profile

Slurs encompass derogatory expressions aimed at specific groups singled out by collective affiliations such as ethnicity or religion. Thus, unlike expressions such as ‘stupid’, slurs target individuals as members of a particular group, not as individuals singled out by their appearance, behaviours, or quirks. However, the distinction between derogating individuals by their group membership versus their individuality is not clear enough. To wit, any individual feature may be shared by various individuals that could be bundled up in an identifiable social group (Diaz-Legaspe 2020; Jeshion 2013). As a result, for numerous derogatory terms, there is no universal agreement regarding their classification as slurs or personal insults (e.g. ‘fatso’, ‘whore’, ‘retard’, etc.).

Instead, what distinguishes slurs is the ‘particular moral or political tenor’ of the offence they inflict (Nunberg 2018, 239). In contrast to other forms of verbal aggression, slurs possess the capacity to ‘dehumanise’ or imply that the target is inferior, thus representing a kind of verbal thoughtcrime. As Jeshion (2017) points out, to dehumanise, slurs need not convey that their targets are subhuman, but ‘only’ that they are beneath the rest as persons. Given their moral implications, it is irrelevant whether a slur’s utterance elicits offence in the audience. As Bolinger (2017) emphasises, an utterance may warrant but fail to generate offence (owing to circumstances such as the absence of an audience or the audience sharing the same derogatory attitude). What is relevant is that, in such situations, audience members, including the target, are ‘morally justified’ or entitled to act offended regardless of how ‘hard-skinned’ or resilient they are (Diaz-Legaspe 2020).

Nonetheless, an exclusive focus on slurs’ offensiveness obscures the fact that slurs are mainly employed in casual or humorous exchanges among bigots. In those contexts, the speaker aims to provide their peers pleasure and gratification rather than subject their targets to humiliation (Camp 2013; Nunberg 2018). Indeed, slurs foster a sense of camaraderie and shared sentiment of superiority, aiming to revel in the schoolyard-style naughtiness of using forbidden words and underscore the group's normative values. However, note that both phenomena are not mutually exclusive but can be analysed as two sides of the same coin. In such situations, humour serves as a veil for underlying contempt, with the experienced amusement deriving precisely from the harm inflicted upon the target.

That being acknowledged, what are slurs’ distinctive characteristics?

• Autonomy: Except in specific contexts, a slur's offensiveness remains ‘independent’ of the beliefs, attitudes, and intentions of individual speakers (Hom 2008). As McCready and Davis (2017) point out, slurring utterances act as ‘invocations’ that produce effects that go beyond, and in many cases despite, whatever intentions the utterer may have had or is inferred to have had.

• Projection: The offensiveness of slurs ‘projects’ out of various forms of embedding, including negations or disjunctions (Bolinger 2017). Saying ‘There are no Spics living here’ inflicts offence towards Latino Americans even though the slur occurs under the syntactic scope of negation. Even though projection may be construed as presupposing a semantic view of slurs, where their derogatory ‘content’ projects, slurs’ objectionable status also pertains to the mere presence of their tokens. For example, a slur occurring within quotation marks can offend (Anderson and Lepore 2013).

• Variability: Some slurs are more insulting than others, exhibiting varying degrees of offensiveness (Camp 2013; Hom 2008; Jeshion 2013, among others). These differences can be classified into different types (Popa-Wyatt and Wyatt 2018): ‘word-variation’, which can be ‘inter-group’ (‘Nigger’ is perceived as more offensive than ‘Chink’) or ‘intra-group’ (‘Kike’ is considered more offensive than ‘Yid’, even though both refer to the same group), and ‘use-variation’, where the level of the offence varies across different uses of the same slur, contingent on contextual factors such as the interlocutor’s identity, situation, tone of voice, etc.

• Appropriation: In so-called ‘appropriated contexts’ (Hom 2008), namely, among members of the derogated group, slurring utterances cease to elicit offence. Instead, they are seen as fostering camaraderie, group identity, and solidarity (Jeshion 2013). In some cases, slurs permanently shift their standard negative effect to a positive one, enabling members of external groups to employ the term harmlessly (as exemplified by ‘queer’). Nevertheless, as Popa-Wyatt (2017) points out, the fact that a slur is used in appropriated contexts does not automatically make it inoffensive. A gay speaker can use a homophobic term to derogate another gay individual, for instance.

3.2. Pejorative nicknames, in contrast

How similar are nicknames and slurs? Sceptics may argue that pejorative nicknames lack some of the attributes associated with slurs, including projection. At first sight, in a conditional utterance such as ‘If F studied, then he will pass the exam’ (where F is a pejorative nickname), ‘F’ elicits offence towards the individual it references, thus displaying a projective behaviour. However, it could be argued that the nickname F, like any other name, presupposes the existence of an F. Therefore, by hitchhiking this presupposition, F would achieve conversational effects that scope out from the conditional's antecedent. However, it is worth noting that presupposing the existence of a reference does not preclude the projection of derogation, should the derogation exist. Hence, the sceptic’s argument is inconclusive.

The sceptic may also argue that nicknames and slurs fall into distinct grammatical categories. Nicknames are grammatically similar to the original names from which they derive, so they cannot be pluralised. For example, for any nickname F, it is infelicitous to say, ‘There are many F’s in the party’, unless one means that there are different individuals in the party whose names are similar. Moreover, nicknames are non-predicative. Namely, for any nickname F, saying ‘John is F’ signals that ‘John’ and ‘F’ are co-referential. In contrast, slurs are typically nouns and thus can be felicitously pluralised, predicated, etc. However, it should be noted that the class of slurs does not constitute a unified grammatical class as well (Sennet and Copp 2020), as it also encompasses adjectives (e.g. ‘slutty’) and verbs (e.g. ‘to jew’), thus making the grammatical distinction between slurs and nicknames somewhat superfluous.

Lastly, one of the most prominent distinctions between nicknames and slurs is that the former does not provoke the same ‘moral’ offence associated with the latter. In the hierarchy of offensive language, slurs are placed much higher than other forms of profanity. Thus, even though nicknames may have lasting adverse effects on their targets, they may not fundamentally differ from other forms of aggression or disrespect, such as consistently attributing negative traits to individuals (e.g. being lazy or incompetent). Unlike the schoolyard bullies’ use of nicknames to taunt the vulnerable, slurs tap into condemnable racial or gender-based social injustices that transcend generations and societies. I will address this point in the following sub-section.

3.3. Slurs and pejorative nicknames, intermingled

Several scholars have conceived derogatory epithets for marginalised groups as nicknames. For instance, Carter (1944) considers that minority groups in America ‘have acquired nicknames which have given them a fixed status in their relation to the majority or dominant group’ (241). Busse (1983), in turn, observes that ethnic heritage plays a role in the ‘nicknaming’ process of groups, which is employed in attempts to facilitate the control of social and cultural ‘outsiders’, i.e. marginalised groups. Furthermore, Allen (1983, 309) posits that the creation and use of ‘generic nicknames for ethnic persons’ within the ‘macrocosm’ of society is analogous to the formation and use of nicknames in the ‘microcosm’ of small social spheres, such as among schoolchildren. In a nutshell, both inter and intra-group nicknames are observed to contribute to the enforcement of social norms and hierarchies.

These authors were heading in the right direction. Slurs and nicknames’ offence-generation profiles exhibit remarkably similar mechanisms. First, the mere pronouncement of a pejorative nickname can elicit strong emotional responses independently of whether the speaker is unaware of the term’s connotations or does not intend to cause harm (Autonomy). Second, nicknames can be offensive when mentioned. During interviews, even ethnographers were forbidden from quoting certain nicknames (Projection). Third, the offensiveness of nicknames is contingent on factors such as the speaker’s identity, the utterance context, the pronunciation, etc. Inter-individual word variation can be inferred from Crozier and Skliopidou’s (2002) study, where individuals rated the names they received as moderately, very, extremely, or quite hurtful. In contrast, intra-individual variation is not explicitly documented. Still, it may arise when the same individual is assigned different nicknames, with one being more offensive due, for example, to its descriptive connotations (Variation). Lastly, although this process warrants further investigation, evidence suggests that targets sometimes embrace their nicknames to enhance pride and build a stronger identity (Reclamation).⁴

It may still be (rightly) pointed out that slurs have historically served as tools of discrimination and atrocities. Therefore, while the harm inflicted by schoolyard nicknaming upon its recipients is undeniable, categorising such nicknames as ‘slurs’ risks trivialising the efforts of those who fight against deeply ingrained forms of oppression. However, as seen earlier, nicknames can inflict severe and enduring consequences on their targets. Hence, it is possible to simultaneously maintain that, at their core, both slurs and nicknames enact the same type of offence while also recognising that their perlocutionary consequences can significantly differ. On the one hand, both are grounded in social hierarchies, where targets are placed as subordinates. On the other hand, both operate at different social scales or networks, so their perlocutionary effects are expected to differ exponentially.

To illustrate this idea, consider the fact that both slurs and pejorative nicknames are, at their core, ‘action-engendering’, that is, incite other forms of violent actions. Tirrell (2012) argues that group-based slurs in Rwanda were inextricably linked to genocide, not merely precursors to it. Labelling the Tutsi as ‘inyenzi’ (cockroach) stripped them of their humanity, legitimising hatred and authorising killings. Similarly, the numerous studies on verbal bullying reviewed earlier corroborate that some pejorative nicknames often develop into other forms of aggression, particularly physical violence. In small social spheres, labels can dehumanise their targets, too, rendering it acceptable to inflict harm upon them. Therefore, pejorative nicknames can effectively operate as slurs within the groups in which they circulate. This revision of the empirical domain sets the basis for a unified account of the structurally similar offensive characteristics of slurs and pejorative nicknames (hereafter referred to as ‘slurring tags’ for lack of a better name). This unified perspective identifies a single phenomenon that operates at different social scales: the imposition of labels attempting to assign a subordinate role within a hierarchy.

4. A unified proposal

Speakers employ slurring tags to express a broad spectrum of emotions. First, slurring tags express the speaker’s contempt towards their targets in hostile situations. Second, when used privately, in a jocular or camaraderie-inducing vein, slurring tags can foster a sense of unity among individuals sharing prejudiced views. Third, despite extreme contempt or loathing, specific slurring tags are often perceived as unoffensive, mainly when targeted at individuals or groups holding a dominant position. Lastly, slurring tags exchanged within the targeted group or among close acquaintances can still display derogatory attitudes without necessarily offending.

These observations seem to undermine theories that associate the offensiveness of slurring tags with any affective states that the speaker may be expressing through their use. For instance, Jeshion’s (2013) influential theory of slurs for groups, which posits that all of them are conventionally linked to the expression of contempt, has faced criticism because the expression of contempt alone is neither sufficient nor necessary to account for the distinct type of denigration that slurs inflict upon their targets. An expression can be profoundly contemptuous without qualifying as a slurring tag. Conversely, a slurring tag can convey other emotions (e.g. disgust, amusement, or even admiration) while qualifying as highly offensive. Therefore, should we conclude that the offensiveness of slurs is independent of the emotions they convey? In the ensuing discussion, I argue that this is not the case.

The proposal developed in this section is based on three independently motivated assumptions: (i) that emotions can be delineated in terms of basic dimensions such as valence, arousal and dominance rather than discrete categories such as contempt; (ii) that affective signs can be linked to these dimensions by employing an indexical, rather than conventional, perspective of affective meaning; and (iii) that lay people infer the underlying states of agents by integrating multiple affective cues, including linguistic signs, employing Bayes theorem. In a nutshell, I argue that, through Bayesian updating, slurring tags index values derived from affective dimensions but are ultimately interpreted based on what is assumed about the speaker’s affective states in the utterance context.

4.1. Background

4.1.1. Affective dimensions

Affective states can be conceived not only as discrete categories such as ‘contempt’ or ‘joy’ but also in terms of more basic dimensions. According to Mehrabian and Russell (1974), affective experiences can be measured using the following three orthogonal, continuous, and bi-polar dimensions:

• Pleasure (also called ‘Valence’): This dimension pertains to the emotional experience’s hedonic aspect. It ranges from negatively valenced states (e.g. sadness) to positively valenced ones (e.g. joy).

• Arousal: This dimension relates to the physiological component and ranges from low mental alertness (e.g. boredom) to high mental alertness (e.g. excitement).

• Dominance: This dimension reflects how much control subjects perceive over a stimulus and corresponds to a scale ranging from the sensation of being submissive (e.g., frustration) to being dominant (e.g., anger).

Following Mehrabian (1996), all possible combinations of low and high values in each dimension result in eight prototypical temperaments, that is, predispositions to feel in particular ways when the appropriate conditions are met. For instance, as Figure 1 illustrates, Hostility corresponds to the [− P, + A, + D] state and Anxiety to the [− P, + A,− D] state.

Figure 1. PAD model’s space of temperaments adapted from Mehrabian (1996), where all possible combinations of low and high values in each dimension result in eight prototypical temperaments

Affective dimensions can be interpreted in various ways. As Colombetti (2005) observes, for any emotion, we can distinguish between its valence, the valence of its experience, and the valence of its outcomes (e.g. its behavioural, physiological, or teleological aspects). Arguably, dominance can also assume different theoretical roles. When directed towards a situation, dominance pertains to an individual’s perceived capability to surmount a challenge. In contrast, within social interactions, dominance correlates to measurable interpersonal hierarchies (Oosterhof and Todorov 2008). Hence, dominance may denote an individual or comparative trait of social organisms.⁵

4.1.2. Indexical fields

How does the link between signs and affective dimensions is established? As previously noted, we can assume that the link between slurring tags and affective states is indexical rather than conventional. Indexical associations are typically instantiated by variables, that is, contrast sets which include alternative ways of ‘saying the same thing’ (Labov 1972). These associations are grounded in the perceived co-occurrence of a sign and some attribute of the speaker, emerging from factors such as co-presence, causality, or another form of spatial or temporal contiguity (Silverstein 1976).

For example, even though the following utterances have the same truth-conditions, the different ways of pronouncing (ING) have been observed to index different social traits of speakers (Campbell-Kibler 2005). Specifically, the use of ‘-ing’ tends to be linked with competence (e.g. being educated or articulate) but aloofness (e.g. being formal or unfriendly). Conversely, as (1) illustrates, the use of ‘-in’ tends to be associated with being incompetent but friendly.

Table

(1)	a.	John is fishing
	b.	John is fishin’.

In sociolinguistics, the set of qualities associated with a variant is referred to as its ‘indexical field’, defined as a ‘constellation of ideologically related meanings, any one of which can be activated in the situated use of the form’ (Eckert 2008, 453). In other words, how variables are interpreted in a context heavily depends on what other properties are believed to hold of the speaker. While some people who employ the ‘-in’ variant can be seen as more easy-going or friendly, others as insincere or condescending, contingent on the listener’s background assumptions.

4.1.3. Affective cognition

How do background assumptions influence affective reasoning? We assume that emotions arise as reactions to specific events. If someone wins the lottery, we think they are likely experiencing happiness. However, we also consider other people’s beliefs and desires. If we know that the lottery winner views money as useless, our initial inference is defeated. Moreover, we anticipate that emotions trigger specific actions (e.g. fleeing) or expressions (e.g. frowning) and use them during the inference process. Consequently, even though the causal flow from events and mental states to emotions and, in turn, from emotions to expressions and actions is unidirectional (see the arrows in Figure 2), information can flow in different directions (Hess and Hareli 2015; 2017; Ong, Zaki, and Goodman 2015):

Figure 2. Lay theories of emotions, indicating the causal flow from emotions’ causes to their outcomes (based on Ong, Zaki, and Goodman 2019). Depending on the situation, the agent’s utterances can be analysed as ‘actions’, akin to fleeing, or ‘expressions’, akin to facial configurations.

One way to represent this flow of reasoning employs Bayes’ theorem (Ong, Zaki, and Goodman 2015; 2019; Saxe and Houlihan 2017; Zaki 2013). For example, to infer an individual’s emotion ‘e’ after they perform an action ‘a’ (i.e. ‘P(e|a)’, read as ‘the probability that someone is feeling e based on action a’), we combine the likelihood that the action a is performed given that the agent is experiencing an emotion e (i.e. ‘P(a|e)’) and the prior probability that emotion e occurs (i.e. ‘P(e)’), and then divide the result by the probability of action a occurring in the first place (i.e. ‘P(a)’), as shown in the formula below. In more complex scenarios, agents infer an individual’s emotion e from multiple cues, such as their actions a and expressions ‘x’ (i.e. ‘P(e|a, x)’).

Table

(2)	$\mathrm{P}(\mathrm{e}|\mathrm{a})=\frac{\mathrm{P}(\mathrm{a}|\mathrm{e})\times \mathrm{P}(\mathrm{e})}{\mathrm{P}(\mathrm{a})}$

These multiple cues can be either in harmony or in conflict with each other. In cases where cues are in harmony (e.g. after winning the lottery, the subject smiles), both cues reinforce and fine-tune each other. However, when they contradict each other (e.g. when an individual appears sad after winning a lottery), interpreters need to rely more on some cues than others. It has been standardly assumed that, across contexts, some cues ‘dominate’ others, e.g. that facial expressions are more reliable than past events. However, evidence suggests that the weight interpreters assign to each cue varies depending on the situation. Sometimes, an individual’s body posture or the background scenery can be perceived as more reliable than their facial expressions (Hess and Hareli 2015; Kayyal, Widen, and Russell 2015).

4.2. Assembling the pieces⁶

How might we characterise the conversational effects of slurring tags (namely, their offensive character) through the lens of the PAD dimensions? First, when a speaker utters a slurring tag, listeners are likely to infer the speaker’s negative evaluation of its referent. Hence, slurring tags index negative pleasure/valence. Second, slurring tags can be uttered in both heated and festive contexts. That is, they do not come across as infelicitous when the speaker does not experience strong feelings towards the target. Hence, slurring tags index neutral arousal. Third, by uttering a slurring tag, the speaker displays that it perceives its referent as lesser, thereby attempting to present itself as superior in relation to it. Hence, slurring tags index a high degree of dominance.⁷

As mentioned earlier, slurs can express emotions of opposite valence and still be regarded as offensive. For instance, they can express contempt towards the target group or amusement at their expense. Nevertheless, in both cases, slurs invariably express that targets are beneath others within a dominance hierarchy. Hence, our multidimensional analysis provides a straightforward account of why slurs are offensive. As slurs are linked to high-dominance affective states, their use warrants offence to those who perceive any form of oppression as detrimental to society. In other words, slurring utterances provide moral justification for taking offence to those who reject unjust forms of hierarchy, whether it is among groups or within the same group.

It might be contended that the expression of high dominance, which involves deeming targets as low in worth, necessarily presupposes a negative evaluation, thus implying that both aspects cannot be genuinely disentangled. However, despite the frequent co-occurrence of high dominance and low pleasure in various contexts, they appear dissociated in many others. Indeed, it is entirely possible to evaluate an individual or group negatively without conveying that they are inferior as persons (e.g. when characterising people as unpunctual). Conversely, it is possible to express that someone is lesser than others without negatively appraising them (e.g. as exemplified by utterances like ‘Chinks are much smarter than us’). Being evaluated as good in some aspect does not preclude being simultaneously judged as inferior and vice-versa.

Drawing inspiration from Burnett (2017, 2019), I operationalise affective dimensions in a theory of meaning by positing a structure ⟨Q,>⟩, where ‘Q’ is the set of relevant affective qualities and ‘>’ denotes relations of incompatibility between them (e.g. that an individual cannot be in a [P-] and [P + ] state simultaneously, etc.). As noted earlier, slurs do not correlate with a specific degree of arousal, so this dimension is omitted:

Table

(3)	$\begin{array}{rl}& \mathrm{Q}=\{[+\mathrm{P}],[\text{–}\mathrm{P}],[\text{–}\mathrm{D}],[+\mathrm{D}]\}\\ & \mathrm{a}.[+\mathrm{P}]>[\text{–}\mathrm{P}]\\ & \mathrm{b}.[+\mathrm{D}]>[\text{–}\mathrm{D}]\end{array}$

Based on this structure Q, four types of affective states α are derived: the [− P, + D] affective state, labelled contempt; the [+ P, + D] state, labelled amusement; and so forth. Importantly, these labels assemble various discrete emotional categories. For example, contempt encompasses [− P, + D] states in general (e.g. rage, hostility, etc.) and not merely contempt (see Table 1).

Then, I posit that for a given slurring tag F, there is an alternative tag F* that references the same target without carrying its pejorative connotations. For example, ‘Spic’ and ‘Hispanic’ or ‘Furry’ and ‘Jennifer’ are such alternative terms, with the former of each pair emerging as a hypocoristic variant of the latter. Note that it is unnecessary to assume that F and F* are fully co-referential or etymologically related. It suffices that they represent salient lexical choices for the speaker within the conversational interaction. Moreover, F* does not need to be construed as a ‘neutral’ counterpart of F, just like ‘-ing’ does not need to be the neutral counterpart of ‘-in’. Both alternatives index qualities that go beyond their truth-conditional encoded content (see Section 4.1.2).

Table 1. Affective states α ∈ aff.

aff	affiliation	amusement	anxiety	contempt
α	[+P, − D]	[+P, + D]	[− P, − D]	[− P, + D]

How can we characterise the link between the alternatives F/F* and the affective states α ∈ aff that they express? Given the instability of slurring tags across utterance contexts, I posit that the link between F/F* and affective states is indexical. That is, it is grounded on the statistical correlation between the use of F/F* and various PAD qualities, any of which may be activated within a particular context. Specifically, I propose that slurring tags exhibit a stronger correlation with [+ D] states, such as contempt, as opposed to [− D] states, such as affiliation. To capture these regularities, I assign a probability distribution Pr(F|α) to F, representing the likelihood of uttering F given an affective state α. As Table 2 illustrates, the non-slurring alternative F* is associated with the distribution Pr(F*|α) = 1 – Pr(F|α).⁸

Table 2. Affective-indexical meaning of F and F*.

aff	affiliation	amusement	anxiety	contempt
Pr(F|α)	0.3	0.6	0.4	0.7
Pr(F*|α)	0.7	0.4	0.6	0.3

Then, I postulate that slurring tags are interpreted based on the listener’s prior beliefs regarding the speaker’s affective disposition towards the target of the insult. Following Burnett's (2017, 2019) framework, I represent the listener’s prior beliefs as a probability distribution Pr(α), read as ‘the probability distribution that the speaker feels an affective state α with respect to the target’. When the listener has no expectations about the speaker’s emotional stance towards the target (e.g. due to lack of familiarity), we represent Pr(α) as a uniform distribution over affective states (see Table 3).

Finally, once the speaker utters a slurring tag F aimed at an individual or group, the listener updates their prior beliefs by conditioning Pr(α) on F’s affective meaning, Pr(F|α). In other terms, the interpretation process involves (i) combining the likelihood of F’s signalling an affective state α with the listener’s prior beliefs about the speaker’s affective stance towards the target and then (ii) readjusting the resulting measure with a normalising constant, i.e. the sum of these terms computed for all affective states α ∈ aff (see 4).

Table

(4)	$Pr(\alpha |\mathrm{F})={\displaystyle \frac{Pr(\alpha )\times \mathrm{Pr}(\mathrm{F}|\alpha )}{{\sum }_{\alpha \in \mathrm{AFF}}^{}Pr(\alpha )\times \mathrm{Pr}(\mathrm{F}|\alpha )}}$

The critical conjecture of the model presented in this section is thus the following: the affective information expressed by using a slurring tag, as inferred by an audience member, is constrained by the perceived affective relationship – according to that particular audience member – between the speaker and the target of the slurring tag. Put differently, reasoning about the speaker’s potential emotions towards the target can influence the weighting assigned to the affective states α ∈ aff within a particular context. In what follows, I shall put this model into work by examining how it accounts for the analogous offence-generation profiles of slurs for groups and individuals.

5. Applying the model

5.1. Autonomy

There is a disconnection between the offensiveness of slurring tags and the speaker’s lack of ill will in uttering them. To illustrate this phenomenon, let us use a term that can assume the simultaneous role of slur and nickname. For instance, a person of African descent might be colloquially referred to as ‘Blackie’, a hypocoristic term which, due to its racial connotations, can develop into a slurring tag for that individual. In a scenario where the speaker S utters (5), we expect the slurring tag to be interpreted as offensive by the listener L.

Table

(5)	Blackie was promoted.

To derive this interpretation, we assume L lacks prior expectations about S’s affective disposition towards the target. Hence, we plug the uniform distribution in (Table 3), along with the indexical field associated with ‘Blackie’ (as outlined in Table 2), into the formula presented in (4). The upshot is that L is more likely to interpret (5) as a manifestation of S’s contempt towards the promoted person (cf. the fourth raw in Table 4). Then, because the display of high dominance within interpersonal interactions warrants offence, the model will predict this utterance to be offensive for L.

A noteworthy feature of this explanation is that regardless of whether L assumes that S holds positive or negative sentiments towards the target, the update will consistently favour [+ D] states. If S seems to be feeling positive, they will be interpreted as finding amusement at the target's expense and people of African descent in general. If S seems to feel negative, they will be interpreted as harbouring hostility towards the target. The speaker may attempt to counteract the latter interpretation by adding affective cues signalling low dominance (such as explicitly following (5) with ‘He deserves it!’ or declaring that they were ignorant of the term’s connotations). However, the ultimate decision on how to weigh these cues rests with L. Hence, L’s inference remains justified independently of the speaker’s beliefs, intentions, or another psychological state.

Table 3. L’s prior beliefs about S’s affective stance α.

aff	affiliation	amusement	anxiety	contempt
Pr(α)	0.25	0.25	0.25	0.25

Table 4. Out-group uses.

aff	affiliation	amusement	anxiety	contempt
Pr(α)	0.25	0.25	0.25	0.25
Pr(F|α)	0.30	0.60	0.40	0.70
Pr(α)·Pr(blackie|α)	0.08	0.15	0.10	0.18
Pr(α|blackie)	0.15	0.30	0.20	0.35

5.2. Projection

Indexicality is not restricted to lexical items but applies to the phenomenon of variation between alternatives more generally. Virtually any facet of human behaviour, like clothing styles, habits, or activities, has the potential to index social (or, as we are currently discussing, affective) qualities as long as they evoke a noticeable contrast between relevant alternatives. When someone opts for one style of clothing over another or engages in one type of sport rather than another, they often unconsciously present themselves in a particular light. Hence, indexicality can account for the ‘projective’ character of slurring tags, meaning their capacity to offend even when they occur under the scope of entailment-cancelling operators. To wit, repeatedly employing ‘-in’ rather than ‘-ing’ (e.g. saying ‘cookin’’ rather than ‘cooking’) within conditionals or speech reports is not less indicative than its unembedded counterpart that the speaker is probably friendly but incompetent. Likewise, the mere utterance of a slurring tag, whether inside or outside of embeddings, does not block the expression of dominance.

5.3. Variation

The degree of offensiveness triggered by slurring tags fluctuates. A term like ‘Kike’ is more offensive than ‘Yid’, even though they refer to the same group. A hypothetical nickname like ‘Blackie’ may be deemed more offensive than ‘Chinaman’ within some discourse networks too. How can we understand this type of variation? According to the indexical viewn presented here, expressions gradually develop different indexical fields that reflect the pleasure, arousal, and dominance values they regularly co-occur with. Among these, expressions that consistently co-occur with [+ D] feelings, such as contempt, will acquire the status of slurring tags. Over time, some of these tags will eventually index [+ D] states with more strength than others due to external factors such as the additional presence of other forms of aggression (e.g. physical violence) or internal factors such as the target’s explicit prohibition of the term’s use. Consequently, some slurring tags will be regarded as more reliable indicators of high dominance states, triggering a greater sense of offence.

The offensiveness of slurring tags can also dissipate depending on various contextual factors. In some cases, targets may grant permission to specific users to create and use tags for them. In others, the proper intonation may counteract the expression of dominance, making the utterance harmless. The agent’s identity plays a crucial role in this regard. When the speaker S is recognised as embodying a particular identity, such identification is interpreted as giving S reasons to feel in specific ways and act upon those feelings. For instance, should S be recognised as Catholic, a listener L may surmise that S is disposed to feel positive about the Catholic church and favour its teachings. When L considers S a good friend, they assume S does not harbour contempt or consider L morally inferior. Although these assumptions may be proven incorrect, S’s social identities initially guide how L thinks about them.

For instance, in a scenario where L is aware that S is of African descent, L will expect S to feel [+ P] rather than [− P] states towards the target. L will also expect S not to experience [+ D] feelings towards the target, as they belong to the same social group. Therefore, when S utters (5), we plug in the formula in (4) a probability distribution which favours [+ P] and [− D] states (e.g. affiliation) and the probabilistic indexical field associated with ‘Blackie’. As a result, we deduce that, upon hearing (5), L will interpret S as more likely expressing affiliation (e.g. affection, friendship, etc.) towards the target, thereby explaining its non-offensive character (cf. fourth row in Table 5). This example illustrates how slurring tags can be uttered without harm when the speaker also belongs to the social group being insulted.

Table 5. In-group uses.

aff	affiliation	amusement	anxiety	contempt
Pr(α)	0.60	0.15	0.15	0.10
Pr(F|α)	0.30	0.60	0.40	0.70
Pr(α)·Pr(blackie|α)	0.18	0.09	0.06	0.07
Pr(α|blackie)	0.45	0.23	0.15	0.17

Numerous other interpretations, including the most unorthodox, can arise from this inferential process. For example, McCready and Davis (2017) report an instance where the use of a group-based slur among African Americans in the presence of a white person was deemed offensive. In such situations, it is possible that the speaker, an African American, is trying to accommodate the idea that he is, in truth, not of African descent, as such an idea seems required to interpret his utterance as expressing that African American people are inferior. Thus, if the listener agrees that the speaker does not see themselves as African American, they will likely interpret the use of the slur as an expression of contempt. Conversely, if the listener does not accommodate this idea, they will perceive the speaker as expressing anxiety instead.

5.4. Reclamation

We can distinguish two ways of dismantling the act of subordination enacted by slurring tags. This stands as the inverse process outlined in the case of word-variation (Section 5.3). Under our framework, reclamation constitutes the process by which a term that acquired a relatively strong indexical association with [+ D] states gradually loses it over time. The development of this process occurs in two, often interconnected, ways:

• Diachronic: When slurring tags are frequently used to express non-dominant affective states, the indexical field linked to them may be gradually reinterpreted by the members of the speech network in which they circulate. As a result, the likelihood that a term user holds derogatory views falls, leading to the attenuation of the signal’s capacity to dehumanise. At this point, employing the term is equally likely to express an affiliative or denigrating attitude with similar probability (Bolinger 2017).

• Synchronic: As observed in Section 2, individuals who exhibit more significant distress in response to hearing their pejorative nicknames are more susceptible to being ridiculed by them. Similarly, as Rappaport (2019) points out, campaigns to prohibit group-based slurs, although instrumental in generating an environment characterised by inclusivity and respect, may increase the strength of a slur. Then, the targets’ embracement of the imposed label inaugurates the process of attenuating its harmful effects. In some cases, targets may actively adopt their slurring tags, thereby neutralising their status as a symbolic imposition.

6. Forecasting alternative analysis of slurring tags

In this section, I analyse how standard semantic and pragmatic theories of group-based slurs can explain how nicknames generate offence. If the analogy between slurs and pejorative nicknames holds, providing an accurate account of nicknames’ offence-generation profile becomes a crucial requirement for an adequate explanation of the offensiveness of slurring tags in general.

6.1. Conventional accounts

A Hom (2008)-style explanation would posit that nicknames elicit offensiveness because they ascribe derogatory properties to their targets. In this framework, a nickname would be analysed as denoting a socially constructed property such as ‘ought to be subject to p₁ + … + p_n because of being d₁ + … + d_n all because of being X’, where ‘d₁ + … + d_n’ are the negative stereotypes attributed to the target X by an ideology. Since it is false that anyone ought to be subject to any negative evaluation on account of their identity (e.g. their race, gender, quirks, etc.), the truth-conditional account predicts that pejorative nicknames have a null extension.

However, as we have seen in Section 2, there is extensive evidence that nicknames are seldom descriptive, and when they are, their descriptive connotations are considered irrelevant. Moreover, the last component of the property putatively ascribed (‘ … all because of being X’) may fit nicely to explain the offensive profile of group-based slurs (e.g. ‘ … all because of being Chinese/Black/Jew’, etc.) but not of pejorative nicknames. To wit, a component such as ‘ … all because of being John/Mary/Alex’ falls short of explaining why nicknames are often particularly pungent. Lastly, holding that pejorative nicknames are semantically vacuous contradicts the intuition that nicknames do indeed refer to their targets.

Alternatively, a Jeshion (2013)-style explanation would identify three components of nicknames’ semantics: (i) a truth-conditional component, (ii) an expressivist component, and (iii) an identifying component. The truth-conditional component of a pejorative nickname would capture the fact that it refers to the same individual referenced by its non-pejorative counterpart. The expressivist component would capture nicknames’ ability to express contempt towards individuals due to their behaviours, attitudes, or quirks. Finally, the identifying component would ascribe a property to the individual perceived as central to its identity.

A central issue with this view pertains to the nature of the identifying component. On standard accounts of names (including nicknames), these do not semantically encode any extension-determining property. Alternatively, one may adopt a predicativist view of nicknames. Under this framework, a nickname F would be conventionally associated with a property determining its individual referent I. One may conceive of such encoded property as either (i) a description or cluster of properties of the individual target I (e.g. ‘being black’, ‘being effeminate’, etc.) or (ii) simply as the property ‘being the individual I’. However, the former option lacks empirical support, as nicknames do not satisfactorily determine their referents by whoever falls into the linked description (Jeshion 2021). In turn, the latter option is explanatorily innocuous, as no one would object to the property ‘being I’ as central in thinking of all the other attributes that I instantiates.

6.2. Conversational accounts

A Bolinger’s (2017)-like pragmatic account would propose that the offensiveness of pejorative nicknames is a function of the contrastive choice made by the speaker between the nickname and the target’s official name. When speakers have the free choice between such referentially equivalent expressions, one being neutral and the other tainted with discriminatory attitudes, and choose the latter option, they signal their affiliation to the bigoted attitudes of the people who typically use it. Under this framework, the content of the nickname would be determined by co-occurrence expectations, as some nicknames regularly co-occur with contexts in which speakers express derogatory attitudes.

This mechanism bears various similitudes to the indexical one proposed in this paper. Both proposals are based on expressions signalling attitudes to varying strengths based on the regular co-occurrence of a nickname and contexts where derogatory attitudes prevail. Thus, both proposals explain the categorisation of a term as a slurring nickname (or its ceasing to be so) based on the statistical robustness of the signal. Moreover, in both views, the degree of offensiveness of the nickname entirely depends on the interpreter. That is, it varies with the speaker’s confidence that using the nickname signals the speaker’s endorsement of offensive attitudes.

However, in the account proposed in this paper, the utterance of nicknames signals speakers’ affective states towards the target rather than their endorsement of an attitude. This difference stems from our consideration of the speaker’s expression of high dominance as the explanatory factor for the particular kind of offence elicited by slurring tags (in contrast to that elicited by other forms of verbal aggression, such as calling someone ‘stupid’). Moreover, the indexical approach I propose explicitly outlines how various contextual cues are integrated in determining the listener’s degree of ‘confidence’ that the utterance signals the speaker’s expression of dominance.

Lastly, an Anderson and Lepore’s (2013)-like sociocultural explanation would posit that nicknames are offensive because they are taboo. In a nutshell, societal norms would prohibit pejorative nicknames, rendering their use offensive for those who adhere to such prohibitions. Hence, this view would correctly accommodate the arbitrariness of nicknaming practices. As observed in Section 2, nicknames can indeed offend when they are laudatory and, conversely, can be endearing even when they evoke the hardest insults. However, a drawback to this prohibitionist view is that nicknames can have long-lasting psychological effects on their victims, even in the absence of taboos or social prohibitions ruling their use. Within close-knit communities, such as high schools, educators, peers, and even the target himself may fail to acknowledge the oppressive effects of pejorative nicknaming, dismissing it as playful banter. Hence, prohibitions are not always relevant within pejorative nicknaming practices.

7. Conclusion

Analogies come in various forms, one of which consists of ‘proportional metaphors’. These are analogies that imply a ratio and where phenomena are mutually interchangeable. In this context, phenomenon A is compared to B and vice versa. In this paper, I used this type of analogy. I mapped slurring onto nicknaming, aiming to systematise aspects of pejorative nicknaming practices that we typically find scattered in ethnological observations. Conversely, I mapped nicknaming onto slurring practices aiming to dissect the subtleties of a phenomenon where manifold societal and psychological issues intersect. This paper thus underscores the advantages of simultaneously considering the broader social and political landscape alongside the intricate web of interpersonal relationships instead of exclusively focusing on one of these aspects.

Using this proportional metaphor, I also proposed that slurs and nicknames’ offensiveness’ originate from a common source and, therefore, can be given a structurally similar explanation. To summarise the view proposed, I argued that slurs are indexically associated with affective qualities derived from the pleasure, arousal, and dominance dimensions. According to this perspective, slurring tags’ affective qualities are expressed during the interpretation process based on (i) how the tag is typically used within a speech network and (ii) how the listener perceives the speaker’s affective predispositions towards the target. The affective states expressed warrant (moral) offence insofar as they manifest a high degree of dominance, which makes them tools of oppression within interpersonal or intergroup dynamics. Importantly, this view explains why slurring tags can warrant offence while expressing a broad (and often contradictory) spectrum of valenced states.

This work aimed to bring attention to nicknaming practices, suggesting new avenues to understand instances where they may inflict harm on single individuals. In the future, further research could refine the predictions offered by the indexical proposal by utilising advancements in Bayesian computing and the emerging field of affective cognition. Additionally, examining slurs and nicknames in parallel can help us identify situations where nicknames display a slurring function, such as in cases of bullying. Although governments have formulated legislation to address hate speech and discrimination against social groups, it is also necessary to establish measures regarding nicknaming practices in contexts where there is a high risk of long-term adverse effects on individuals’ psychological well-being.

Acknowledgements

I am grateful to Márta Abrusán, Elisa Paganini, Heather Burnett, Elin McCready, Isidora Stojanovic, and Pyro Suarez for their insights and feedback on previous versions of this paper. Any remaining errors are my own.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was funded by the German Research Foundation (DFG) as part of the CRC TRR 212 (NC³) – Project number 316099922.

Notes

1 It is noteworthy to observe that while nicknames may be descriptive in nature, their referent is not necessarily determined by whoever 'fits' the description (Jeshion 2021). This phenomenon becomes evident in situations where nicknames are extended from one individual to their offspring. Antoun (1968), for example, reports of various individuals within an Arab village identified by the nickname 'Sevener', a term derived from a common ancestor who was born prematurely in the seventh month.

2 Almost every ethnographer working in Mediterranean-Hispanic areas reports incidents of violent responses to nicknames’ utterances. Foster (1964), for example, tells of a case in which a Mexican villager knifed a neighbor simply for using his meaningless nickname face-to-face.

3 Gilmore also observes that the mere mention of pejorative nicknames during interviews elicited emotional reactions: 'I was warned against using the names openly because most people take offence.' (1982, p. 693). Participants were generally reluctant to talk about their own pejorative nicknames and of absent third parties as well, thus indicating that the pejorative character of a nickname persists when these are not used as weapons.

4 It is also worth noting that nicknames’ typology can readily apply to slurs. Slurs can be either descriptive or non-sensical (e.g., “Beaner” vs “Kike”), originate from official names or not (e.g., “Jap” vs “Guido”), and be based on norms or not (e.g., “slut” vs “Spic”) (see Jeshion 2021, for an in-depth analysis).

5 Dominance has often been characterised as 'peripheral' to emotions or 'core affect' (Russell and Barrett 1999). Nonetheless, whether dominance or the other dimensions constitute primary or secondary aspects of emotional experiences is irrelevant to our current purposes.

6 The content of this sub-section draws from and enhances (Carranza-Pinedo 2023).

7 The reason why slurring tags come to express high dominance is as multifaceted as why they signal negative valence. The indexing of high dominance may primarily arise in situations where there is an arbitrary power imbalance that unfairly determines some or all aspects of the life of a group or individual. However, it can also emerge when such relations are not yet consolidated. An expression can attain the status of a slurring tag when the target perceives it as threatening their capacity to autonomously build their own identity (Anderson and Lepore 2013).

8 This distribution fluctuates contingent upon the relation between the specific slurring tag and its alternative. For example, gender-based tags have alternatives which explicitly carry normative connotations (Ashwell 2016). Hence, in this case, the differences between the indexical fields associated with F and F* are likely to be less pronounced.