miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
17
EVALUATIVE LANGUAGE IN THE (RE)PRODUCTION
AND RESISTANCE OF DISCOURSES
OF SEXUAL VIOLENCE ON X (TWITTER)
EL LENGUAJE EVALUATIVO EN LA (RE)
PRODUCCIÓN Y RESISTENCIA DE DISCURSOS
SOBRE VIOLENCIA SEXUAL EN X (TWITTER)
<https://doi.org/10.26754/ojs_misc/mj.202510725>
PATRICIA PALOMINO-MANJÓN
Centro Universitario de la Defensa de Zaragoza
ppalomino@unizar.es
<https://orcid.org/0000-0002-6548-8022>
Abstract
This paper investigates the role of evaluative language in the (re)production
and resistance of discourses concerning sexual violence on X (formerly Twitter).
Drawing on Appraisal Theory (Martin and White 2005) as the analytical
framework, the present paper identifies linguistic patterns that either reinforce
or challenge patriarchal ideologies, practices and gendered power dynamics in
society. Using allegations of sexual assault made against Brett Kavanaugh during
his confirmation proceedings to the United States Supreme Court in 2018 as a
case study, the analysis illustrates that evaluative language was used to (re)enact
opposing discourses and (re)negotiate traditional rape scripts and experiences of
sexual violence. The findings also reveal the interplay between conflicting narratives
—perpetrator vs. victim-survivor— on X and how communication on this platform
shapes and reflects societal attitudes toward sexual violence and aggression against
women in both North American society and institutions.
Keywords: X (Twitter), evaluative language, Appraisal Theory, sexual violence,
online feminism.
Resumen
Este artículo investiga el papel del lenguaje evaluativo en la (re)producción y
resistencia de discursos sobre violencia sexual en X —antes Twitter—. Basándose
en la Teoría de la Valoración (Martin and White 2005) como marco analítico, este
trabajo identifica patrones lingüísticos que refuerzan o desafían las ideologías y
Patricia Palomino-Manjón
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
18
prácticas patriarcales, así como las dinámicas de poder de género en la sociedad.
Utilizando las acusaciones de agresión sexual de Brett Kavanaugh durante su
proceso de confirmación ante el Tribunal Supremo de los Estados Unidos en 2018
como estudio de caso, los resultados ilustran que el lenguaje evaluativo se usa para
(re)crear discursos opuestos y (re)negociar narrativas tradicionales sobre violación
y experiencias de violencia sexual. Los resultados también revelan la interacción
entre narrativas opuestas —perpetrador vs. víctima/superviviente— en X y cómo
la comunicación en esta plataforma da forma y refleja las actitudes sociales hacia la
violencia sexual y la agresión contra las mujeres tanto en la sociedad como en las
instituciones estadounidenses.
Palabras clave: X (Twitter), lenguaje evaluativo, Teoría de la Valoración, violencia
sexual, feminismo digital.
1. Introduction
Sexual violence has become a simultaneously public and private issue after the
emergence of digital platforms, even though it was traditionally viewed as a
private matter (Bou-Franch 2013). Research on online aggression against women
has examined the use of various digital platforms to spread discourses derived
from rape culture, which are used to deny the existence of such violence (Bou-
Franch 2013; Bou-Franch and Garcés-Conejos Blitvich 2014). More precisely, the
microblogging platform X —formerly Twitter— has been singled out as the most
sexist and (sexually) aggressive social media platform (Jane 2017; Mendes et al.
2018). Several studies have shown that Twitter is employed to (sexually) threaten
women (e.g. Hardaker and McGlashan 2016; Frenda et al. 2019), as well as to
victimise victim-survivors1 (e.g. Stubbs-Richardson et al. 2018) while portraying
perpetrators as the real victims.
Despite this negative view of X, the platform has also given victim-survivors a
relatively safe space to contribute to digital feminism. X’s most popular function,
the hashtag (#), is used as a tool for socio-political resistance and to form online
communities, even if users never interact directly or know each other (Zappavigna
2012, 2018). X users challenge traditional rape myths and scripts by offering
support and validating personal narratives of sexual violence (Loney-Howes
2018), which has been key to the establishment of the fourth wave of feminism
(Blevins 2018). Therefore, it is unsurprising that the study of online networked
feminism on X is gaining momentum in linguistics and discourse analysis (e.g.
Morikawa 2019; Bouvier 2020; Jones et al. 2022; Palomino-Manjón 2022,
2024).
Evaluative Language in the (Re)production and Resistance of Discourses
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
19
Even though research has focused on both digital (sexual) violence against women
and feminist resistance on X separately, there is much less information of a linguistic
nature on the different discourses concerning misogyny and (anti)feminism when
these coexist within the same digital platform. Therefore, this paper examines the
discourses and ideologies of gendered violence on X and the evaluative resources
used to enact such discourses and ideologies.
The present paper is divided as follows. Section 2, which follows this introduction,
begins by outlining the literature on gender and digital communication, especially
from a critical perspective on gendered violence. The paper then introduces the
case study employed to carry out the research. Section 3 presents the analytical
framework used to examine evaluative language and continues by explaining the
data and procedures carried out during the analysis. Section 4 goes on to discuss
the findings of the study. Lastly, Section 5 summarises the main findings of the
analysis, discusses the paper’s implications for the study of violence against women
on digital platforms, and provides some concluding remarks.
2 Literature Review
2.1. Gender and Digital Communication
The rise of new technologies anticipated more democratic communication, as
social factors such as gender, race and class would be invisible to Internet users.
However, research from a variety of fields, including linguistics, media, sociology
and psychology (e.g. Herring 1999; Jane 2017; Stubbs-Richardson et al. 2018)
has suggested that digital platforms are used to harass and intimidate women, thus
bringing attention to pre-existing gender differences in the offline world.
Since the early stages of research on digitally mediated communication, scholars
focusing on gender inequality have pointed out that women and other socially
marginalised groups, such as people of colour and LGBTQIA+ communities,
are often the targets of sexism and hate speech, which, in turn, reflects a (white,
heterosexual, able-bodied) male-dominated Internet culture (Jane 2017). As
previously mentioned, X is considered an aggressive platform towards women,
since it hosts communities —sometimes formed in the manosphere (see Jaki et al.
2019)— that encourage hostile and misogynistic attitudes, thus enforcing toxic
masculinities (Jane 2017; Mendes et al. 2018). Linguistic research has shown
how X is used to send death threats (e.g. Hardaker and McGlashan 2016) and to
perpetuate victim-blaming and sexist attitudes (e.g. Stubbs-Richardson et al. 2018;
Frenda et al. 2019), hindering women’s participation in digital communities and
degrading and dehumanising female users.
Patricia Palomino-Manjón
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
20
Despite this, research has also shown that language can help victim-survivors of
(cyber)abuse and their allies to resist and challenge digital practices and ideologies
derived from rape and patriarchal cultures. For instance, Dynel and Poppi (2020)
examine the rhetorical strategies used by Stormy Daniels when she was mentioned
in hateful posts relating to her job (i.e. pornographic film actress). Additionally,
the platform also allows individuals like her to connect with other users who share
similar experiences. Victim-survivors take advantage of X’s widespread popularity
to engage in online feminist activism to bring attention to (verbal) sexual violence.
Linguists have analysed how the use of specific linguistic patterns, especially
evaluative language, helps victim-survivors to share self-narratives of sexual
violence (e.g. Jones et al. 2022; Palomino-Manjón 2022, 2024).
X has become a tool to engage in online feminist activism and has contributed to
the development of a ‘call-out culture’ (Lawrence and Ringrose 2018) in which
sexism, misogyny and rape culture are challenged. However, it has also brought
attention to victim-survivors and left them “vulnerable to criticisms, threats,
and harassment from trolls who are often participating for the sole purpose of
antagonising feminists” (Blevins 2018: 94). Therefore, it is evident that X has
evolved into a platform where diverse and opposing discourses and ideologies
coexist. Bearing in mind the aim of this paper, the following subsection introduces
the case study, which will guide the research questions.
2.2. AsJ Kavanaugh’s Sexual Assault Allegations
Following Associate Justice (AsJ) Anthony Kennedy’s retirement announcement
on 27 June 2018, the President of the United States at the time, Donald
Trump, nominated former Judge Brett Kavanaugh to fill the vacancy on the
Supreme Court of the United States (SCOTUS). On 30 July 2018, Senator
Dianne Feinstein received an anonymous letter in which a woman explained
that she had been sexually assaulted by Judge Kavanaugh in 1982. The writer
of the letter also reached out to The Washington Post’s lawyer, Debra Katz, who
recommended she take a polygraph test so that she could not be accused of
lying.
Since the press tried to reveal her identity, the accuser, college professor Dr
Christine Blasey Ford, went public in an interview with The Washington Post
on 16 September 2018. In the interview, she described the encounter with
Kavanaugh and a friend of his, Mark Judge. She recounted that, when she was
15, the two men attempted to rape her, pinning her to a bed while trying to
remove her clothes. As both men were heavily intoxicated, she managed to
escape and lock herself in the bathroom. On 27 September 2018, Dr Ford and
Judge Kavanaugh testified in a televised hearing. Republican senators attracted
Evaluative Language in the (Re)production and Resistance of Discourses
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
21
attention when they accused the Democratic Party of tarnishing the nominee’s
reputation with false allegations. In the end, charges were not pressed against
the accused, and the Senate ultimately confirmed Kavanaugh as Associate Justice
on 6 October 2018.
Dr Ford faced harassment and (death) threats from Internet users for speaking out
against AsJ Kavanaugh, especially on the microblogging platform X. She was also
accused of being part of a political strategy to bring down his nomination as well
as Trump’s administration. However, she stated she had no regrets and hoped her
testimony would empower other women to upend traditional rape narratives and
patriarchal societal structures.
It is against this backdrop that the present paper takes AsJ Kavanaugh’s
confirmation process and sexual allegations as a case study to examine the different
discourses surrounding gendered violence on X, including the linguistic and
evaluative patterns that contribute to the negotiation of both Dr Ford and AsJ
Kavanaugh’s identities. Therefore, the research questions that guided this study
are the following:
RQ1: What evaluative resources did X users employ to signal different gender
ideologies and patriarchal discourses during AsJ Kavanaugh’s confirmation
process?
RQ2: To what extent can these resources be employed to perpetuate or
challenge gender (in)equality, patriarchal practices and sexual violence?
3. Analytical Framework and Methodology
3.1. Appraisal Theory
Appraisal Theory (Martin and White 2005) was used as the analytical framework
to identify linguistic evaluative patterns and discourses around sexual violence,
perpetrators and victim-survivors. This framework was developed to examine the
social function of language within Systemic-Functional Linguistics and is employed
to examine how authors of a text use evaluative language to express their position
and stance towards “both the material they present and those with whom they
communicate” (Martin and White 2005: 1). The theory has also been proven to be
an effective tool for analysing and categorising discourses and language (Bednarek
2008) and to understand how online communities are built (Zappavigna 2012,
2018). Appraisal Theory is divided into three meaning domains: Attitude,
GrAduAtion and enGAGement. This paper focuses on Attitude and GrAduAtion
(see Table 1).
Patricia Palomino-Manjón
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
22
Affect
Happiness cheer Unhappiness misery
affection antipathy
Security quiet Insecurity disquiet
trust distrust
Satisfaction interest Dissatisfaction ennui
pleasure displeasure
Inclination desire Disinclination non-desire
Surprise positive Surprise negative
Judgement
Social esteem normality Social sanction veracity
capacity propriety
tenacity
Appreciation
Reaction impact Composition balance
quality complexity
Valuation
Graduation
Focus sharpen Force intensification
soften quantification
Table 1. Appraisal Theory categorisation (adapted from Martin and White 2005; Bednarek 2008)
Attitude concerns the use of evaluations to express emotional reactions,
judgements of behaviour and the worth or aesthetics of things. It is further
subdivided into three domains: Affect, JudGement and AppreciAtion.
Affect is used to express “positive and negative feelings: do we feel happy or
sad, confident or anxious, interested or bored?” (Martin and White 2005: 42).
Affect can convey the emotions of the producer of a text or a third party, and
it can be implied or directly conveyed depending on the expression of explicit
emotional states (e.g. happy, sad, worried) or the physical behaviours of the person
experiencing the emotion (e.g. ‘rushed breath’ or ‘to shake uncontrollably’)
Evaluative Language in the (Re)production and Resistance of Discourses
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
23
(Martin and Rose 2007). In addition, emotions can also be conveyed through
metaphors (e.g. cold as ice or dull like the dead). Affect is further subdivided into
four subcategories, namely (Un)Happiness (misery, antipathy, cheer and affection),
(Dis)Satisfaction (ennui, displeasure, interest and pleasure), (In)Security (disquiet,
surprise, confidence and trust) and (Dis)Inclination (fear and desire).
However, since Appraisal Theory is still considered hypothetical (Martin and
White 2005: 46), this paper adopts Bednarek’s (2008, 2009) refinements to the
Affect subsystem. This change not only introduces a new type of emotion (i.e.
Surprise), but also modifies the subsystems proposed by Martin and White (2005).
Consequently, Bednarek’s framework reorganises the original categorisation and
expands it into five subsystems: 1) (Un)Happiness; 2) (In)Security; 3) (Dis)
Satisfaction; 4) (Dis)Inclination and 5) Surprise (see Table 1).
JudGement involves the evaluation of human actions and behaviour not only from
individuals but also from organisations and institutions such as governments,
courts and legislative bodies. The framework distinguishes between evaluations
of social sanction, which are based on a set of rules or regulations (i.e. how legal
or moral someone’s actions are), and social esteem (i.e. admiring and criticising
someone’s actions without legal or moral implications) (Martin and White 2005;
White 2011). It is relevant for this study to note that JudGement can be negotiated
in context. However, according to Martin and Rose (2007), legal lexis such as
victim, crime, perpetrator, guilty and innocent cannot be separated from their
evaluative role in specific contexts.
Lastly, AppreciAtion concerns the evaluation of aesthetics, human creations,
situations and natural phenomena. People can also be appraised when describing
their physical appearance.
Additionally, ApprAisAl can be conveyed explicitly or implicitly. Explicit
evaluation, or attitudinal inscription, refers to the use of words or fixed phrases
that carry positive or negative meaning even when removed from their context
(White 2004, 2011). This contrasts with attitudinal tokens or invocation, in
which no single item carries positive or negative value independently of its
context. This distinction between attitudinal inscription and invocation becomes
relevant when the boundaries between the Attitude subsystems are not clear-
cut (Bednarek 2009). For example, the terms disgust or revolt may convey
JudGement and Affect simultaneously. Martin and White (2005) name these
instances ‘hybrid realisations’, whereas Thompson (2014) suggests the concept
‘Russian doll’ to illustrate that one ApprAisAl resource can function as a token or
indirect expression of another subsystem. Nevertheless, the authors agree that
Affect should be considered the primary attitudinal value as it constitutes the
basis of Attitude.
Patricia Palomino-Manjón
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
24
On the other hand, GrAduAtion is related to the intensification or weakening of
someone’s evaluations. It is subdivided into Focus (i.e. the graduation of non-
scalable lexis) and Force (i.e. the varying degrees of intensity and quantity). Whereas
Focus is used to make “something that is inherently non-gradable gradable”
(Martin and Rose 2007: 46), Force refers to resources that intensify meaning,
such as qualifiers (e.g. very, extremely) and attitudinal lexis, which includes degrees
of intensity (e.g. happy/ecstatic).
Research that has applied Appraisal Theory to the analysis of digital communication
has shown that the framework is an effective tool for examining how Internet users
bond and build online communities based on shared values (see Zappavigna 2012,
2018). The theory has also been applied to the study of gender and sexuality and
its potential to facilitate the creation of online feminist networks (e.g. Palomino-
Manjón 2022) as well as to deride and abuse verbally (del Saz-Rubio 2024a, 2024b)
and to spread anti-feminist and misogynist ideologies and discourses (e.g. Heritage
and Koller 2020; Krendel 2020, 2023). This illustrates the versatility of the theory
as a qualitative framework for understanding the different ways in which online
interactions and communities form around topics relating to gender-based violence.
3.2. Data
A corpus of X posts published during Kavanaugh’s confirmation process was
compiled, including specific hashtags concerning the confirmation process, namely
#KavanaughConfirmation and #NoKavanaughConfirmation. These hashtags
were selected for their structure, which suggests two contrasting views regarding
his nomination.
Posts under the #KavanaughConfirmation hashtag were manually obtained using
X’s TweetDeck application, which allows for manual, advanced search functionalities
using Boolean terms and filtering options, such as language, date, number of
reposts, etc. On the other hand, posts containing the #NoKavanaughConfirmation
hashtag were collected through Google Sheets’s add-on Twitter Archiver (Agarwal
n.d.), which automatically retrieves metadata about the tweets.
Posts and reposts in languages other than English were filtered out and excluded
from the dataset. The complete dataset covers tweets published over 23 days:
from Dr Ford’s public statement (16/09/2018) to the day after Kavanaugh was
publicly confirmed as an Associate Justice of SCOTUS (8/10/2018). In total,
there were 112,428 tweets (N = 2,924,498 words). In addition, bearing in mind
the methodological approach taken in this paper (i.e. corpus-assisted discourse
analysis), X conventions such as hashtags (#), (manual) reposts and mentions (@)
were removed using the software R (R Core Team 2020). This resulted in two
corpora made up of 1,474,172 words (#KC) and 417,639 words (#NoKC).
Evaluative Language in the (Re)production and Resistance of Discourses
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
25
3.3. Procedure
The present study adopts a mixed methodology that combines corpus linguistics
tools and Appraisal Theory (Martin and White 2005). The corpora were analysed
separately and then compared to identify the linguistic patterns and ApprAisAl
resources used to (re)produce discourses concerning sexual violence.
The first step involved the analysis of the 100 most frequent words in each corpus
(RQ1). The software R (R Core Team 2020) was used to obtain wordlists and
frequencies. After carefully analysing the terms obtained in the analysis through
the reading of their concordance lines, the terms were grouped according to
discussion topics. For the sake of brevity, the analysis scrutinised only the most
commonly occurring words related to social actors, gender and violence in both
corpora.
The second analysis identified and computed the different evaluative resources
employed by X users (RQ2). Based on the frequency analysis results, ten
subcorpora were created around the search terms Kavanaugh, Ford, women, men
and sexual. To avoid cherry-picking, a technologically randomised selection of
100 concordance lines of each search word was obtained using SketchEngine’s
(Kilgarriff et al. 2014) concordance tool (Baker and Levon 2015). This tool was
also used to expand the concordance lines and retrieve complete tweets (Baker and
Levon 2015; Jones et al. 2022). The resulting subcorpora comprised 100 tweets
each (i.e. 1,000 X posts in total; see Table 2).
Subcorpus Number of X posts Number of words
#NoKC-Kavanaugh 100 3,096
#NoKC-Ford 100 3,135
#NoKC-Men 100 3,309
#NoKC-Women 100 3,208
#NoKC-Sexual 100 3,383
#KC-Kavanaugh 100 3,102
#KC-Ford 100 3,078
#KC-Men 100 3,254
#KC-Women 100 3,423
#KC-Sexual 100 3,114
Total 1,000 32,102
Table 2. Information about the subcorpora examined in the ApprAisAl analysis
Patricia Palomino-Manjón
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
26
The tweets were analysed qualitatively to examine ApprAisAl resources. The
identified resources were manually coded on a document and classified. After the
coding of resources, the ApprAisAl values were quantified to identify the most
frequent ApprAisAl (sub)system and the polarity of the evaluative resources. Then,
these resources were grouped according to their potential to convey discourses
relating to (sexual) violence against women in the different subcorpora.
Following Page (2022), an intra-analysis was conducted and measured using the
test-retest reliability correlation coefficient (Pearson correlation). The results show
that both analyses were highly correlated, with a correlation indicator of ≥ 0.9,
which indicates excellent reliability (see Table 3).
Subcorpus Total occurrences
(1st analysis, March 2022)
Total occurrences
(2nd analysis, May 2022)
#NoKC-Kavanaugh 273 273
#NoKC-Ford 235 268
#NoKC-Men 333 334
#NoKC-Women 212 215
#NoKC-Sexual 265 267
#KC-Kavanaugh 255 255
#KC-Ford 241 241
#KC-Men 221 221
#KC-Women 210 210
#KC-Sexual 143 143
Corr. coef. 0.979964863
Table 3. Test-retest reliability correlation coefficient
As stated by Page (2022), reanalysing ApprAisAl resources helped identify coding
errors and re-code ambiguity in the resources. The subcorpus with the most
significant changes in the coding was the #NoKC-Ford subcorpus, and all the
changes were associated with ambiguous cases of implicit positive Veracity.
4. Analysis and Results
4.1. Frequency Analysis
An examination of the 100 most frequent words in each corpus was carried out to
identify prevalent linguistic patterns. As illustrated in Appendix A and B, closed-
class words accounted for over three-quarters of the total. The qualitative reading
Evaluative Language in the (Re)production and Resistance of Discourses
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
27
of these words in context revealed that both the #NoKC and #KC corpora shared
similar thematic categories (see Table 4).
Thematic category #NoKC #KC
Pronouns
you, he, your, we, his, they,
I, my, him, her, she, our, me,
them, us, he’s, I’m
you, I, he, they, your, his, her,
she, their, my, him, me, our,
I’m, the, us
Gender pronouns he, his, she, him, her, he’s he, his, her, she, him
Social actors Kavanaugh, women, Ford, FBI,
Dr, Trump, men, GOP, people
Kavanaugh, women, he,
Ford, people, judge, Brett,
democrats, man, Dr, FBI,
senate
Legal field vote, court, investigation,
supreme, SCOTUS, assault
vote, court, judge, supreme,
senate, investigation, justice
(Political) authorities FBI, Trump, SCOTUS, GOP,
supreme, court
FBI, court, supreme, judge,
senate, democrats, Brett,
supreme, court, senate
Gender and violence sexual, assault sexual
Miscellaneous thank, want, need time, today, good, know, want
Table 4. Classification of the 100 most frequent grammatically open-class words and gendered
pronouns in the dataset
Thematic categories encompassed lexis concerning (gender) pronouns, social actors,
politicians and authorities, terms related to the legal field, vocabulary related to
gender and violence and miscellaneous words. Some terms overlapped in different
categories, for instance, Trump as a social actor and authority, or judge as a title and
an authority (e.g. Judge Kavanaugh), a verb relating to the legal field or as a social
actor (e.g. Mark Judge).
In addition, the presence of gendered pronouns reveals that male social actors
were more frequently discussed than female individuals. The male pronouns he,
his and him and the pronoun plus verb he’s appeared in both frequency wordlists.
Male pronouns were more frequently used to refer to Kavanaugh, but X users also
employed them to refer to the male senators who participated in the hearings and
to then President Donald Trump. However, the qualitative analysis unveiled that
the possessive his also referred to Dr Ford as part of the n-gram his accuser, thus
rendering her identity as related to Kavanaugh (van Leeuwen 2008). This specific
word sequence is particularly noteworthy, as explicit references to Dr Ford were
scant in the list of the 100 most frequent words.
Patricia Palomino-Manjón
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
28
Social actors included not only individuals but also assimilated and collectivised
actors (van Leeuwen 2008) such as political groups and organisations (e.g. FBI,
Senate, Supreme, Court, SCOTUS, GOP, democrats). The most interesting result
is that GOP (i.e. Grand Old Party, the Republican Party) was more prevalent in
#NoKC, whereas democrats only appeared in #KC, which suggests the ideological
and political leanings of X users in each hashtag-specific dataset.
The frequency wordlists encompassed a range of male social actors, including
general terms like man and men, as well as specific and individualised male figures
such as Kavanaugh, Brett, Judge (Dr Ford’s second perpetrator’s surname) and
Trump. In contrast, the female social actors included women and the surname
Ford, which highlights the scarce presence of explicit references to Dr Ford. The
presence of the singular noun man in the frequency list, while the noun woman
was absent, suggests that men were often individualised, as opposed to women
(Pearce 2008). Another shared category between both wordlists was lexis related
to gender and violence. However, whereas the #NoKC corpus featured the words
assault and sexual, the #KC corpus only included the adjective sexual.
Bearing in mind the aim of this paper, the five terms related to victims and
perpetrators and sexual violence which were obtained from both corpora during
the frequency analysis were selected for further scrutiny. The surnames Kavanaugh
and Ford, along with the general gendered identities women and man/men, were
chosen to investigate the negotiation of victim-perpetrator identities. Furthermore,
the adjective sexual, which ranked among the top 100 most frequent words in each
corpus, was also examined.
The selected lexis showed a similar normalised frequency per thousand words (ptw)
in both corpora (see Table 5), which shows that they were not only patterns of
potentially frequent topics of debate in each corpus, but were also common when
discussing the events of the confirmation process on X. Therefore, a qualitative
analysis of these words in context is of special relevance to the objective of this paper.
Word #NoKC #KC
Kavanaugh 5.19 5.89
Ford 1.44 2.37
men 1.07 0
man 0 1.47
women 3.38 2.94
sexual 1.77 1.75
Table 5. Normalised frequencies per thousand words (ptw)
Evaluative Language in the (Re)production and Resistance of Discourses
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
29
The following subsection (4.2) features the analysis of these terms in context,
drawing on Appraisal Theory (Martin and White 2005; see Section 3). While it is
true that the #KC corpus featured the singular male form man and the #NoKC
corpus included the plural form men, both corpora contained the plural female
noun women. Consequently, the plural form men was selected to compare the
results to those of its female counterpart.
4.2. Evaluative Resources
4.2.1. Quantitative Analysis
As can be seen in Appendix C, more than half of the total occurrences of ApprAisAl in
each subcorpus were negative JudGement, making up over three-quarters of the total
resources in the #KC-Kavanaugh (80.39% 205 instances) and #NoKC-Kavanaugh
(73.26% 200 instances) subcorpora. In contrast, negative JudGement accounted for
fewer than half of the identified resources in the #NoKC-Ford subcorpus (48.51%,
130), despite being the most prevalent ApprAisAl resource in that subcorpus. This
might be attributed to higher positive Affect and JudGement frequencies in the
#NoKC-Ford subcorpus, with 34 (12.49%) and 72 (26.87%) instances, respectively.
These results suggest a prevalence of negative evaluative prosodies to discuss the
actions of the social actors involved in the confirmation process.
JudGement was the only ApprAisAl value implicitly conveyed aside from
AppreciAtion in the #NoKC-Kavanaugh subcorpus, which yielded one instance
of a hybrid realisation between Affect and AppreciAtion. Implicit JudGement
resources were frequently found to express judgements of Veracity and Propriety
through hybrid realisations and factual statements. However, explicit JudGement
values were more common in all subcorpora than implicit realisations, as illustrated
in Table 6.
Subcorpus Explicit Implicit Total
#KC-Kavanaugh 161 (70.61%) 67 (29.39%) 228 (100%)
#NoKC-Kavanaugh 141 (58.51%) 100 (41.49%) 241 (100%)
#KC-Ford 161 (84.74%) 29 (15.26%) 190 (100%)
#NoKC-Ford 108 (53.47%) 94 (46.53%) 202 (100%)
#KC-Men 95 (54.29%) 80 (45.71%) 175 (100%)
#NoKC-Men 136 (55.74%) 108 (44.26%) 244 (100%)
#KC-Women 99 (57.89%) 72 (42.11%) 171 (100%)
#NoKC-Women 119 (63.98%) 67 (36.02%) 186 (100%)
Patricia Palomino-Manjón
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
30
Subcorpus Explicit Implicit Total
#KC-Sexual 44 (37.29%) 74 (62.71%) 118 (100%)
#NoKC-Sexual 133 (64.88%) 72 (35.12%) 205 (100%)
Table 6. Explicit and implicit Judgement occurrences in the examined subcorpora
On the other hand, GrAduAtion was frequently found to enhance Attitude
resources rather than to downscale them (see Table 7). Force was the most frequent
resource across all subcorpora and was mainly used for booster evaluations. Force-
Intensification, which took up more than half of the GrAduAtion resources in all
subcorpora, mostly involved repetitions of attitudinal lexis, superlatives and capital
letters.
Subcorpus Force Focus Total
#KC-Kavanaugh 39 (88.64%) 5 (11.36%) 44 (100%)
#NoKC-Kavanaugh 45 (97.83%) 1 (2.17%) 46 (100%)
#KC-Ford 36 (92.31%) 3 (7.69%) 39 (100%)
#NoKC-Ford 31 (100%) 0 31 (100%)
#KC-Men 36 (97.3%) 1 (2.7%) 37 (100%)
#NoKC-Men 35 (94.59%) 2 (5.41%) 37 (100%)
#KC-Women 52 (100%) 0 52 (100%)
#NoKC-Women 69 (97.18%) 2 (2.28%) 71 (100%)
#KC-Sexual 27 (79.41%) 7 (20.59%) 34 (100%)
#NoKC-Sexual 47 (97.92%) 1 (2.08%) 48 (100%)
Table 7. Use of grAduAtion resources in the examined subcorpora
Following this qualitative examination of these ApprAisAl resources, the next
subsection analyses and delves into the different discourses identified in the corpora.
4.2.2. Overview of Discourses of Sexual Violence
The qualitative reading of ApprAisAl values revealed various discourses concerning
sexual violence. However, they are not always clear-cut and occasionally intersect
within the same X post, thus conveying more than one discourse at times.
Due to the nature of the event, Dr Ford’s and AsJ Kavanaugh’s identities were
intertwined, since his portrayal as a perpetrator contributed to her depiction as
a victim-survivor. It is worth highlighting that all subcorpora included more
instances of posts that focused on the construal of AsJ Kavanaugh as a perpetrator,
as shown in Example 1 below:
Evaluative Language in the (Re)production and Resistance of Discourses
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
31
(1) Post65/NoKC-Sexual: A nominee for the Supreme Court committed
sexual assault [-JudGement;propriety] and the President is a pussy-grabber
[-JudGement;propriety] (in his own words). And Trump sits in the Oval
Office [-token of JudGement;propriety]. Gotta love [-Affect;unhappiness-
antipathy] republicans! #NoKavanaughConfirmation2,3
As can be seen in Example 1, the representation of AsJ Kavanaugh as a perpetrator
was commonly constructed by drawing parallels between Donald Trump and him.
Both were depicted as perpetrators of sexual violence. In the example, the X user
conveys Propriety to denounce the presence of authoritative figures as well as a
lack of action to prevent perpetrators from occupying seats in U.S. institutions
and politics.
(2) Post82/KC-Kavanaugh: The Brave [+JudGement;tenacity] woman who
lying [-JudGement;veracity] #SCOTUS nominee #BrettKavanaugh tried
to rape [-JudGement;propriety] just came out with her story [+token of
JudGement;tenacity] she has also [+GrAduAtion;force-intensification]
taken a lie detector test which shows she was being truthful
[+JudGement;propriety] Will #kavanaugh volunteer a lie detector test too
[-token of JudGement;veracity] #KavanaughLied [-JudGement;veracity]
#KavanaughConfirmation
Example 2 describes Dr Ford as a determined woman for coming forward with her
story through a discourse of feminism. The use of the adjective brave is linked to
the evoked Tenacity resource came out with her story. This is due to the fact many
feminists and allies of feminism consider the telling of stories of sexual violence
as an act of boldness (Clark-Parsons 2019; Palomino-Manjón 2022, 2024). This
contrasts with the portrayal of AsJ Kavanaugh, which emphasised falsehood
and negative discourses by using Veracity and Propriety values. In this example,
Kavanaugh is constructed as a perpetrator as he is positioned as the agent of the
negative Propriety resource (tried to) rape. Hence, Dr Ford is presented as the
object of the crime and, therefore, portrayed as a victim.
Discourses of falsehood were frequent in all corpora as they intertwined with
discourses of political violence. Dr Ford and AsJ Kavanaugh were both portrayed
as victims of a political process since many X users considered they were being
manipulated as pawns by both the Republican and the Democratic Parties for their
political gain. However, the discourses differed when referring to the accused and
the accuser, as can be observed in Examples 3 and 4 below:
(3) Post7/NoKC-Ford: So it’s all a lie [-JudGement;veracity] and a sham
[-JudGement;veracity] [+GrAduAtion;force-intensification]. There
was NEVER [+GrAduAtion;force-intensification] any intention
Patricia Palomino-Manjón
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
32
of a fair [-JudGement;propriety] hearing for Dr. Ford. You disgust
[-Affect;dissatisfaction-displeasure] me, @SenateMajLdr, @lisamurkowski
@SenatorCollins @JeffMerkley -Fuck you all [-Affect;dissatisfaction-
displeasure]. #TakeBackTheSenate #IStandWithChristineBlaseyFord [+token
of JudGement;veracity] #NoKavanaughConfirmation #BlueTsunami
(4) Post100/KC-Kavanaugh: Such a FAKE ATTEMPT [-JudGement;veracity]
[+GrAduAtion;force-intensification] to dishonor [-JudGement;propriety]
a very [+GrAduAtion;force-intensification] smart [+JudGement;capacity]
and very [+GrAduAtion;force-intensification] fine [+JudGement;normality]
man Judge Kavanaugh [-token of JudGement;propriety], Diane
Feinstein YOU [+GrAduAtion;force-intensification] employed a
CHINESE SPY [+GrAduAtion;force-intensification] 4 20 yrs [-token of
JudGement;propriety] YOU [+GrAduAtion;force-intensification] are the
threat [-JudGement;propriety] to America not Judge Kavanaugh [+token
of JUDGEMENT;propriety] #KavanaughConfirmation
Dr Ford was often portrayed as a victim of politicians. For instance, Example 3
illustrates how some users expressed disapproval of the outcome of the hearing.
This user employs Veracity resources (lie and sham) to evaluate the confirmation
process and then proceeds to provide an ethical condemnation of the GOP and the
Republican senators by questioning the course and the credibility of the hearing. As
opposed to this, the user in Example 4 evaluates the allegations of sexual assault as
deceitful and a political strategy to discredit and ruin AsJ Kavanaugh’s reputation
through an intensified use of Veracity resources as well as several Propriety values.
Then, they provide a positive appraisal of AsJ Kavanaugh through the use of Social
Esteem (i.e. Capacity and Normality) to create a positive discourse prosody. This
helps the user to depict Senator Feinstein as an unethical politician and a political
perpetrator, while they present AsJ Kavanaugh as a political victim.
Discourses of political violence also constructed North American women as victims
of the patriarchal system that prevails in the country’s institutions. These posts
intertwined with feminist discourses frequently, as illustrated in Example 5 below:
(5) Post39/NoKC-Men: @peterdaou THEY. HAVE. NO. CONSCIENCE
[-JudGement;propriety] [+GrAduAtion;force-intensification]. They
gaslighted [-JudGement;propriety] Ford so bad [+GrAduAtion;force-
intensification], that as a victim I worry [-Affect;insecurity-disquiet] that
she is questioning [-Affect;insecurity-disquiet] all of her memories of the
event [-token of JudGement;capacity]. That is what we do was survivors
[-token of JudGement;capacity]. BECAUSE. MEN. AND. WOMEN IN.
POWER. REFUSE [-Affect;disinclination-non-desire]. TO. BELIEVE.
Evaluative Language in the (Re)production and Resistance of Discourses
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
33
US. [-token of JudGement;propriety] [+GrAduAtion;force-intensification]
#NoKavanaughConfirmation #VoteThemOut #MeToo
Example 5 above contains a disclosure of sexual assault which builds Dr Ford’s
identity as that of a political victim. The user employs judgements of Propriety,
strengthened by different GrAduAtion resources, to denounce the behaviour
of male senators during the hearing and portray them as political perpetrators.
The user also denounces rape myths and victim-blaming attitudes used against
Dr Ford that make victim-survivors question their memories, which deepens
their traumatic wound (Palomino-Manjón 2022). This is conveyed with implicit
negative Capacity to refer to victim-survivors, as well as negative Affect and
implicit negative Propriety to blame the use of such patriarchal discourses by
North American authorities and institutions.
Lastly, all subcorpora were found to include two opposing discourses: feminism and
male victimhood. Moreover, discourses of feminism were divided into discourses
of empowerment and discourses of emotional pain.
(6) Post99/NoKC-Men: Women are strong [+JudGement;capacity]
and truly [+GrAduAtion;force-intensification] unpredictable
[+JudGement;normality]. Mechanistic [-JudGement;propriety] old men
[-JudGement;propriety] [+GrAduAtion;force-intensification], not all men,
are weak [-JudGement;capacity], insecure [-Affect;insecurity-disquiet]
[-token of JudGement;capacity] and scared [-Affect;insecurity-disquiet]
[+GrAduAtion;force-intensification] of strong [+JudGement;capacity]
women. November we show them what we are made of, we will not give-up
[+token of JudGement;tenacity]. #NovemberIsComing #ProtectOurCare
#NoKavanaughConfirmation
Example 6 shows the discourse of empowerment surrounding women in
the subcorpora. The X user begins by expressing positive Capacity (strong)
and Normality (unpredictable)4 to depict women as powerful social actors.
Furthermore, the user defines an out-group of men surrounded by negative
prosody associated with conservatism, sexism and misogyny as the cluster old men
is used to convey negative Propriety values (Palomino-Manjón 2024). In addition,
the adjective mechanistic, which defines a patriarchal ideology that considers men
as the foundation of society and human nature (Hultman and Pulé 2018), also
amplifies the negative depiction of this out-group of men. Finally, they evoke
positive Tenacity to portray women as determined to bring an end to Trump’s
Republican administration.
A second feminist discourse was concerned with emotional pain. AsJ Kavanaugh’s
confirmation to SCOTUS caused concern and fear among some female users
Patricia Palomino-Manjón
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
34
because of his conservative views. Their distress and sadness were evident in Affect
resources, as shown in Example 7 below:
(7) Post71/KC-Women: This is beyond [+GrAduAtion;force-quantification]
maddening [+GrAduAtion;force-intensification] [-AFFECT;dissatisfaction-
displeasure], sad [-Affect;unhappiness-misery], unbelievably [+GrAduAtion;
force-intensification] disappointing [-Affect;unhappiness-misery] and a
wake up call [-Affect;surprise] for women all around the globe [+GrAduAtion;
force-quantification]. #wematter #KavanaughConfirmation
The X user in Example 7 employs different Affect resources to convey unease,
anger and emotional distress. This is conveyed through feelings resources of Misery
(sad, disappointing) and Displeasure (maddening). Additionally, these emotions
are intensified with GrAduAtion values (beyond, unbelievably and maddening),
emphasising the emotional discomfort experienced by the user.
As expected from networked feminism in social media, the subcorpora #NoKC-
Sexual and #KC-Sexual also featured personal narratives of sexual violence. The
employment of Affect values allowed victim-survivors to express their traumatic
experiences to other users.
(8) Post17/NoKC-Sexual: Not really going to explain all the sexual assault
[-JUDGEMENT;propriety] that I’ve gone through but I blamed myself
[-AFFECT;dissatisfaction-displeasure] for years [+GrAduAtion;force-
quantification]. I was ashamed [-Affect;insecurity-disquiet]. I told no one
for so long because I didn’t know who to trust [-Affect;insecurity-distrust].
It never goes away nor will it ever [+GrAduAtion;force-quantification]. Its
a life long [+GrAduAtion;force-quantification] pain [-Affect;unhappiness-
misery] I will live with. #NoKavanaughConfirmation
The user in Example 8 employs different Affect resources to express their trauma.
Among these, Dissatisfaction-Displeasure is found to convey a self-blaming
attitude. Affect is also used to communicate feelings of Insecurity to express
shame (Disquiet) and fear of trusting others (Distrust) regarding the crime. All
these values are intensified with Quantification resources (for years, it never goes
away nor will it ever and life long), which shows the long-term damage that sexual
violence causes to victim-survivors.
In contrast, the subcorpora #KC-Men was heavily shaped by a discourse of male
victimhood, depicting men as victims of feminist movements and, particularly, the
#MeToo Movement. The user in Example 9 promotes a discourse of fear through
an Insecurity-Disquiet value (safe), which, in turn, results in a negative JudGement
of Propriety of the #MeToo Movement and elites. Additional implicit negative
Evaluative Language in the (Re)production and Resistance of Discourses
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
35
Propriety is presented to blame the movement for victimising men, as well as to
accuse the elites (i.e. politicians, the press) for exploiting it to further deprecate men.
(9) Tweet78/KC-Women: Men are not safe [+Affect;insecurity-disquiet]
anywhere [-token of JudGement;propriety]. The #MeToo movement has
successfully launched the “Men are Evil [-JudGement;propriety]” narrative
[-token of JudGement;propriety] and there are enough dishonest
[-JudGement;veracity] politicians and media on the left to pump up this
narrative [-token of JudGement;propriety]. A out of control [+GrAduAtion;
force-intensification] [-JudGement;propriety] MeToo movement is bad
[-JudGement;propriety] for both men and women [+GrAduAtion;force-
intensification]. #KavanaughConfirmation
Overall, the analysis of evaluative language revealed highly frequent use of negative
prosodies to convey discourses of sexual violence and to depict the different social
actors involved in AsJ Kavanaugh’s confirmation process. A dominant pattern
of negative evaluative prosody which prevailed in all subcorpora was frequently
employed to construct discourses of violence, truth and falsehood in the context
of the confirmation process. Negative evaluations were primarily conveyed
through JudGement resources to condemn social actors’ morality and ethics
and denounce their dishonesty during the hearings. Emotional expressions also
contributed significantly to creating negative prosodies, since a range of Affect
lexis was employed to show anger, fear and sadness as the confirmation process
unfolded. Notably, the recurrent use of GrAduAtion resources helped intensify
and strengthen these evaluations and display widespread discomfort and emotional
distress among (female) X users and victim-survivors of sexual violence.
5. Conclusion
The present research aimed to examine the use of evaluative language in the (re)
production and (counter) resistance of discourses of sexual violence and patriarchal
practices and ideologies on X. The use of such resources was expected to construct
discourses that condemned rape culture and gendered power structures in North
American society. To do so, the sexual assault allegations against AsJ Kavanaugh
during his confirmation process to SCOTUS were taken as a case study.
Overall, the results indicated the coexistence of opposing discourses on X during
the confirmation process. On the one hand, the subcorpora obtained from the
#KavanaughConfirmation corpus featured hegemonic discourses that denied and
invalidated Dr Ford’s testimony, which in turn allowed the spread of anti-feminist
and victim-blaming discourses. These reflected the power imbalance that persists
in the offline world (Herring 1999) and resulted in the dismissal of Dr Ford’s
allegations by both Republican authorities and X users.
Patricia Palomino-Manjón
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
36
On the other hand, X users employed counter-hegemonic discourses to resist
(online) patriarchal discourses and practices by demystifying the rape scripts and
myths (Loney-Howes 2018) that invalidated Dr Ford’s experience of sexual
assault. This also resulted in the sharing of personal narratives by victim-survivors
to support her testimony (Jones et al. 2022; Palomino-Manjón 2022). These
results suggest that the hashtag #NoKavanaughConfirmation also served as a form
of networked feminism, even if it was not exclusively designed for this end.
In addition, the study of these online interactions not only contributed to
examining gendered discourses surrounding X but also provided insights into its
users’ online identities and ideologies. These identities were constructed based
on the discourses they employed to express support or opposition to the social
actors and political groups involved in the process through the (re)production
and defiance of feminist and patriarchal discourses (van Dijk 2006; Bou-Franch
and Garcés-Conejos Blitvich 2014). While each group of supporters used negative
values of JudGement, those advocating for Dr Ford’s testimony were characterised
bytheir use of a greater number of positive JudGement resources as well as different
values of Affect to foster a more supportive and empathetic stance.
Nevertheless, the generalisability of these results is limited. Whereas this study
provides relevant perspectives on the use of specific discourses to negotiate, (re)
produce, challenge and sustain patriarchal discourses and gendered asymmetry in
society, the results cannot be representative of all discourses concerning (sexual)
violence on X, as the corpus was limited to a relatively small sample of North
American society.
Notwithstanding these limitations, this paper highlights the complex role of X in
shaping and spreading discourses of sexual violence and rape culture. The findings
reflect social dynamics in which the combination of anonymity and easy access to
the Internet (Herring 1999) enables conflicting discourses to coexist on the same
platform. While some users find support and a platform to share their experiences,
others display dismissive and hostile behaviours that contribute to perpetuating a
pervasive rape culture in society. This illustrates the potential and limitations of X
as a space for feminist and social activism.
Notes
1. The terms ‘victim’ and ‘survivor’ are considered as two ends of a continuum
that carry negative and positive connotations, respectively. Consequently, this paper takes the
merged term ‘victim-survivor’ to refer to the people who have been the object of any type of
gender-based violence.
Evaluative Language in the (Re)production and Resistance of Discourses
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
37
Works cited
AgArwAl, Amit. n.d. Twitter Archiver. Google Addons. <https://www.labnol.org/>.
BAker, Paul and Erez levon. 2015. “Picking the Right Cherries? A Comparison of Corpus-based
and Qualitative Analyses of News Articles about Masculinity”. Discourse & Communication 9
(2): 221-236. <https://doi.org/10.1177/1750481314568542>.
BednArek, Monika. 2008. Emotion Talk Across Corpora. Palgrave Macmillan.
BednArek, Monika. 2009. “Language Patterns and ATTITUDE”. Functions of Language 16 (2): 165-
192. <https://doi.org/10.1075/fol.16.2.01bed>.
Blevins, Katie. 2018. “bell hooks and Consciousness-Raising: Argument for a Fourth Wave of
Feminism”. In Vickery, Jacqueline and Tracy Everbach (eds.) Mediating Misogyny: Gender,
Technology, and Harassment. Palgrave Macmillan: 91-108. <https://doi.org/10.1007/978-3-319-
72917-6_5>.
Bou-FrAnch, Patricia. 2013. “Domestic Violence and Public Participation in the Media: The Case of
Citizen Journalism”. Gender and Language 7 (3): 275-302. <https://doi.org/10.1558/genl.
v7i3.275>.
Bou-FrAnch, Patricia and Pilar gArcés-coneJos Blitvich. 2014. “Gender Ideology and Social Identity
Processes in Online Language Aggression against Women”. Journal of Language Aggression
and Conflict 2 (2): 226-248. <https://doi.org/10.1075/bct.86.03bou>.
Bouvier, Gwen. 2020. “From ‘Echo Chambers’ to ‘Chaos Chambers’: Discursive Coherence and
Contradiction in the #MeToo Twitter Feed”. Critical Discourse Studies 19 (6): 1-17. <https://doi.or
g/10.1080/17405904.2020.1822898>.
clArk-pArsons, Rosemary. 2019. “‘I SEE YOU, I BELIEVE YOU, I STAND WITH YOU’: #MeToo and
the Performance of Networked Feminist Visibility”. Feminist Media Studies 21 (3): 362-380.
<https://doi.org/10.1080/14680777.2019.1628797>.
del sAz-ruBio, Mª Milagros. 2024a. “An Approach to the Rhetoric of Boebert and Díaz-Ayuso’s
Political Tweets on X: A Corpus-Assisted Analysis”. Alicante Journal of English Studies / Revista
Alicantina de Estudios Ingleses 41: 75-103. <https://doi.org/10.14198/raei.2024.41.04>.
del sAz-ruBio, Mª Milagros. 2024b. “From ‘Low-class’ and ‘Talentless’ to ‘Narcissist and
Pathological Liar’: a Functional-pragmatic Approach to Meghan Markle’s Negative Evaluation
on X”. Poznan Studies in Contemporary Linguistics 60 (3): 311-348. <https://doi.org/10.1515/
psicl-2023-0068>.
dynel, Marta and Fabio I. M. poppi. 2020. “Arcana Imperii: The Power of Humorous Retorts to
Insults on Twitter”. Journal of Language Aggression and Conflict 8 (1): 57-87. <https://doi.
org/10.1075/jlac.00031.dyn>.
2. Explicit Appraisal resources are underlined, whereas implicit resources are
highlighted in italics.
3. Examples do not include usernames to ensure authors’ anonymity and avoid the
traceability of the original message. Grammar, spelling and punctuation have not been altered.
4. Although Martin and White (2005) categorise unpredictable as a value of Non-
normality, the context in which this adjective appears helps convey a positive evaluative prosody.
Patricia Palomino-Manjón
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
38
FrendA, Simona, Bilial ghAnem, Manuel montes-y-gómez and Paolo rosso. 2019. “Online Hate
Speech against Women: Automatic Identification of Misogyny and Sexism on Twitter”. Journal of
Intelligent and Fuzzy Systems 36 (5): 4743-4752. <https://doi.org/10.3233/JIFS-179023>.
hArdAker, Claire and Mark mcglAshAn. 2016. “‘Real Men don’t Hate Women’: Twitter Rape Threats
and Group Identity”. Journal of Pragmatics 91: 80-93. <https://doi.org/10.1016/j.
pragma.2015.11.005>.
heritAge, Frazer and Veronika koller. 2020. “Incels, In-groups, and Ideologies: The Representation
of Gendered Social Actors in a Sexuality-based Online Community”. Journal of Language and
Sexuality 9 (2): 151-177. <https://doi.org/10.1075/jls.19014.her>.
herring, Susan C. 1999. “The Rhetorical Dynamics of Gender Harassment On-Line”. The
Information Society 15: 151-167. <https://doi.org/10.1080/019722499128466>.
hultmAn, Martin and Paul M. pulé. 2018. Ecological Masculinities: Theoretical Foundations and
Practical Guidance. Routledge.
JAki, Sylvia, Tom de smedt, Maja gwóźdź, Rudresh pAnchAl, Alexander rossA and Guy de pAuw.
2019. “Online Hatred of Women in the Incels.me Forum: Linguistic Analysis and Automatic
Detection”. Journal of Language Aggression and Conflict 7 (2): 240-268. <https://doi.org/10.1075/
jlac.00026.jak>.
JAne, Emma A. 2017. Misogyny Online: A Short (and Brutish) History. SAGE.
Jones, Lucy, Małgorzata chAłupnik, Jai mAckenzie and Louise mullAny. 2022. “‘STFU and Start
Listening to How Scared We Are’: Resisting Misogyny on Twitter via #NotAllMen”. Discourse,
Context & Media 47: 1-10. <https://doi.org/10.1016/j.dcm.2022.100596>.
kilgArriFF, Adam, Vít BAisA, Jan BuštA, Miloš JAkuBíček, Vojtěch kovář, Jan michelFeit, Pavel rychlý
and Vít suchomel. 2014. “The Sketch Engine: Ten Years On”. Lexicography 1: 7-36. <http://doi.
org/10.1007/s40607-014-0009-9>.
krendel, Alexandra. 2020. “The Men and Women, Guys and Girls of the ‘Manosphere’: A Corpus-
Assisted Discourse Approach”. Discourse and Society 31 (6): 607-630. <https://doi.
org/10.1177/0957926520939690>.
krendel, Alexandra. 2023. “The Reddit Manosphere as a Text and Place: A Three-part Analysis”. In
Esposito, Eleonora and Majid KhosraviNik (eds.) Discourses in the Digital Age: Social Media,
Power, and Society. Routledge: 69-88.
lAwrence, Emilie and Jessica ringrose. 2018. “@Notofeminism, #Feministsareugly, and Misandry
Memes: How Social Media Feminist Humor is Calling out Antifeminism”. In Keller, Jessalynn
and Maureen E. Ryan (eds.) Emergent Feminisms Complicating a Postfeminist Media Culture.
Routledge: 211-232.
loney-howes, Rachel. 2018. “Shifting the Rape Script: ‘Coming Out’ Online as a Rape Victim”.
Frontiers: A Journal of Women Studies 39 (2): 26-57. <https://doi.org/10.1353/fro.2018.a698452>.
mArtin, James R. and David rose. 2007. Working with Discourse: Meaning behind the Clause.
Continuum.
mArtin, James R. and Peter R. R. white. 2005. The Language of Evaluation: Appraisal in English.
Palgrave Macmillan.
mendes, Kaitlynn, Jessica ringrose and Jessalynn keller. 2018. “#MeToo and the Promise and
Pitfalls of Challenging Rape Culture through Digital Feminist Activism”. European Journal of
Women’s Studies 25 (2): 236-246. <https://doi.org/10.1177/1350506818765318>.
Evaluative Language in the (Re)production and Resistance of Discourses
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
39
morikAwA, Nora. 2019. “#YesAllWomen’s Language: Women’s Style Shifting in Feminist
Discourse on Twitter”. Discourse, Context & Media 28: 112-120. <https://doi.org/10.1016/j.
dcm.2018.11.001>.
pAge, Ruth. 2022. “Analyzing Multimodal Interactions in Social Media Contexts”. In Vásquez,
Camilla (Ed.) Research Methods for Digital Discourse Analysis. Bloomsbury Publishing: 159-
176.
pAlomino-mAnJón, Patricia. 2022. “Feminist Activism on Twitter: The Discursive Construction of
Sexual Violence and Victim-survivors in #WhyIDidntReport”. Journal of Language Aggression
and Conflict 10 (1): 140-168. <https://doi.org/10.1075/jlac.00049.pal>.
pAlomino-mAnJón, Patricia. 2024. “Challenging Discourses of Sexual Violence on X: The Linguistic
Representation of Victims and Perpetrators in Kavanaugh’s Confirmation Process”. Alicante
Journal of English Studies / Revista Alicantina de Estudios Ingleses 41: 105-127. <https://doi.
org/10.14198/raei.2024.41.05>.
peArce, Michael. 2008. “Investigating the Collocational Behaviour of MAN and WOMAN in the
BNC Using Sketch Engine”. Corpora 3 (1): 1-29.
r core teAm. 2020. R: A Language and Environment for Statistical Computing. Version 4.0.1.
RStudio, PBC.
stuBBs-richArdson, Megan, Nicole E. rAder and Arthur G. cosBy. 2018. “Tweeting Rape Culture:
Examining Portrayals of Victim Blaming in Discussions of Sexual Assault Cases on Twitter”.
Feminism & Psychology 28 (1): 90-108. <https://doi.org/10.1177/0959353517715874>.
thompson, Geoff. 2014. “AFFECT and Emotion, Target-value Mismatches, and Russian Dolls:
Refining the Appraisal Model”. In Alba-Juez, Laura and Geoff Thompson (eds.) Evaluation in
Context. John Benjamins: 47-66.
vAn diJk, Teun A. 2006. “Ideology and Discourse Analysis”. Journal of Political Ideologies 11 (2):
115-140. <https://doi.org/10.1080/13569310600687908>.
vAn leeuwen, Teun. 2008. Discourse and Practice: New Tools for Critical Discourse Analysis.
Oxford U.P.
white, Peter R. R. 2004. “Subjectivity, Evaluation and Point of View in Media Discourse”. In
Coffin, Caroline, Ann Hewings and Kieran O’Halloran (eds.) Applying English Grammar:
Functional and Corpus Approaches. Routledge: 229-246.
white, Peter R. R. 2011. “Appraisal”. In Zienkowski, Jan, Jan-Ola Östman and Jef Verschueren
(eds.) Discursive Pragmatics. John Benjamins: 14-36.
zAppAvignA, Michelle. 2012. Discourse of Twitter and Social Media: How We Use Language to
Create Affiliation on the Web. Bloomsbury.
zAppAvignA, Michelle. 2018. Searchable Talk: Hashtags and Social Media Metadiscourse.
Bloomsbury.
Accepted: 11/07/2025
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Received: 19/06/2024
Accepted: 10/01/2025
Patricia Palomino-Manjón
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
40
Appendix A: Frequency list for the #NoKC corpus
#NoKC corpus
Rank Word Frequency Rank Word Frequency
1the 15364 51 right 1096
2to 11482 52 out 1094
3and 7775 53 him 1091
4you 7717 54 sexual 1083
5is 6868 55 from 1051
6of 6580 56 why 1047
7this 4901 57 one 1038
8for 4559 58 her 1029
9he 4286 59 like 1025
10 in 3971 60 now 1007
11 not 3609 61 investigation 1000
12 that 3606 62 it’s 997
13 on 3440 63 know 995
14 are 3353 64 can 993
15 it 3274 65 please 969
16 kavanaugh 3178 66 their 963
17 be 2989 67 there 958
18 your 2559 68 would 955
19 no 2551 69 people 937
20 we 2495 70 she 913
21 his 2450 71 our 909
22 have 2091 72 ford 879
23 women 2069 73 when 861
24 with 2052 74 more 853
25 will 1961 75 fbi 853
26 all 1878 76 assault 833
27 they 1875 77 me 793
28 do 1870 78 dr 757
29 if 1840 79 time 741
30 vote 1808 80 get 738
31 was 1770 81 never 728
32 who 1727 82 trump 726
33 what 1690 83 them 717
34 so 1613 84 us 690
35 about 1612 85 want 667
Evaluative Language in the (Re)production and Resistance of Discourses
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
41
#NoKC corpus
Rank Word Frequency Rank Word Frequency
36 an 1421 86 supreme 663
37 at 1320 87 gop 661
38 i1298 88 men 655
39 has 1291 89 did 648
40 my 1285 90 because 627
41 just 1258 91 need 613
42 as 1215 92 any 612
43 how 12 11 93 he’s 608
44 should 119 4 94 even 603
45 up 119 3 95 thing 596
46 or 1174 96 i’m 595
47 by 113 8 97 say 594
48 court 1138 98 too 582
49 don’t 113 2 99 scotus 582
50 but 112 8 100 thank 574
Patricia Palomino-Manjón
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
42
APPENDIX B: Frequency list for the #KC corpus
#KC corpus
Rank Word Frequency Rank Word Frequency
1the 79881 51 she 5010
2to 50076 52 out 4990
3a39392 53 how 4920
4and 32611 54 their 4803
5is 32384 55 it’s 4766
6of 31273 56 my 4580
7you 22833 57 don’t 4461
8in 21032 58 when 4430
9this 20624 59 can 4409
10 for 20317 60 court 4353
11 i18892 61 should 4178
12 that 16366 62 ford 4159
13 on 16087 63 get 4127
14 be 13439 64 one 4102
15 it 13096 65 would 4038
16 are 13076 66 people 3982
17 not 12156 67 why 3744
18 have 10465 68 more 3663
19 kavanaugh 10326 69 him 3631
20 he 9915 70 know 3568
21 will 9571 71 right 3528
22 with 9193 72 there 3521
23 we 9037 73 time 3469
24 they 8759 74 judge 3458
25 if 8442 75 me 3429
26 all 8253 76 supreme 3393
27 what 8101 77 our 3385
28 was 7735 78 i’m 3249
29 your 74 11 79 senate 3138
30 vote 7403 80 fbi 311 2
31 as 7326 81 today 3066
32 no 7315 82 sexual 3065
33 about 7196 83 been 3008
34 his 6703 84 investigation 2990
35 so 6701 85 brett 2938
Evaluative Language in the (Re)production and Resistance of Discourses
miscelánea 72 (2025): pp. 17-44 ISSN: 1137-6368 e-ISSN: 2386-4834
43
#KC corpus
Rank Word Frequency Rank Word Frequency
36 who 6510 86 democrats 2861
37 just 6422 87 them 2838
38 has 6150 88 think 2828
39 at 6050 89 us 2810
40 but 6044 90 did 2789
41 now 5737 91 over 2779
42 do 5691 92 these 2757
43 an 5661 93 going 2697
44 like 5562 94 want 2626
45 her 5551 95 justice 2600
46 by 5523 96 because 2580
47 or 5348 97 man 2579
48 from 5220 98 good 2522
49 women 5152 99 dr 2501
50 up 5106 100 after 2489
APPENDIX C: Percentages of instances of ApprAisAl resources in each subcorpus.