37.3254, Reviews: Words that Wound: Tahmineh Tayebi; Vahid Parvaresh (2026)

[The LINGUIST List]

LINGUIST List: Vol-37-3254. Fri Oct 09 2026. ISSN: 1069 - 4875.

Subject: 37.3254, Reviews: Words that Wound: Tahmineh Tayebi; Vahid Parvaresh (2026)

Moderator: Steven Moran (linguist at linguistlist.org)
Managing Editor: Valeriia Vyshnevetska
Team: Helen Aristar-Dry, Daniel Swanson
Jobs: jobs at linguistlist.org | Conferences: callconf at linguistlist.org | Pubs: pubs at linguistlist.org

Homepage: http://linguistlist.org

Editor for this issue: Valeriia Vyshnevetska <valeriia at linguistlist.org>

================================================================


Date: 09-Oct-2026
From: Keighley Perkins [perkinskeighley at gmail.com]
Subject: Tahmineh Tayebi; Vahid Parvaresh (2026)


Book announced at https://linguistlist.org/issues/37-1840

Title: Words that Wound
Subtitle: Exploring Online Offensive Language
Publication Year: 2026

Publisher: Cambridge University Press
           http://www.cambridge.org/linguistics
Book URL:
https://www.cambridge.org/ch/universitypress/subjects/languages-linguistics/sociolinguistics/words-wound-exploring-online-offensive-language?format=HB&isbn=9781316517499

Author(s): Tahmineh Tayebi; Vahid Parvaresh

Reviewer: Keighley Perkins

SUMMARY
“Words that Wound: Exploring Online Offensive Language” by Tahmineh
Tayebi and Vahid Parvaresh investigates two fundamental questions in
digital discourse studies: what is the nature of online offence and
how is it linguistically constructed in computer-mediated interaction?
Situated primarily within pragmatics and (im)politeness theory, the
volume addresses an underexplored “middle ground” of digital
communication: language that harms, insults or marginalises
individuals while falling short of the legal threshold for hate
speech. The authors establish a pragma-linguistic rationale for
studying online hostility, arguing that offensive force cannot be
understood through isolated lexical items alone but emerges from
speaker intentions, hearer evaluations, contextual dynamics and
Austinian speech act layers of locution, illocution and perlocution.
Chapter 1 introduces this framework by mapping offensive language
across a continuum ranging from minor netiquette breaches
(“non-polite”) to everyday “offensive” speech and, finally, “grossly
offensive” or potentially illegal communication. It also details the
limitations of current automated content moderation tools which
struggle with overlapping definitions, annotator label bias and
implicit hostility.
Chapter 2 develops the book’s theoretical model of offence by tracing
the evolution of (im)politeness theory across three waves: from
classic rule- and face-based models (first wave) to discursive
approaches (second wave) and holistic, frame-based or interactional
frameworks (third wave). Building on third-wave pragmatics, Tayebi and
Parvaresh propose a dual model distinguishing between law-informed
offences (legal transgressions) and interaction-informed offences
(breaches of the moral order). Through an analysis of UK statutory
frameworks, such as Section 127 of the Communications Act 2003 and the
Online Safety Act 2023, and landmark judicial decisions, the authors
demonstrate the persistent evidentiary challenges in proving mens rea
(intention) in digital speech.
Chapter 3 outlines the study’s research design, data collection and
ethical considerations. To overcome interpreter and researcher bias,
the authors compile a unique dataset of naturally occurring comments
that have been explicitly flagged as offensive by social media content
creators who read their “mean comments” on camera. This chapter
details how this dataset was investigated using a multi-layered,
corpus-assisted methodology combining Sketch Engine and Wmatrix5 with
qualitative thematic and pragmatic analysis.
Chapters 4 through 6 apply the first analytical lenses to the corpus
data, examining structural, semantic and lexical patterns of verbal
aggression. Drawing on Foucault’s concept of discourse, Chapter 4
contrasts the target-flagged corpus against a reference sample to
identify recurring structural tendencies, noting that offensive
language online is characteristically intrusive, opinionated and
personal while the frequent use of symbolic masking (e.g., asterisks)
indicates both self-censorship and attempts at normalization. Chapter
5 utilises Wmatrix5 to analyse semantic domains, uncovering a
significant overrepresentation of key concepts related to dislike,
foolishness and negative evaluations of physical appearance. Chapter 6
investigates multi-word units and keyword patterns, examining direct
insults and drawing a critical distinction between “hard” insults
(which convey intense, explicit denigration) and “soft” insults (which
rely on milder or contextual disparagement).
The final two chapters turn to indirectness and digital modalities,
representing the volume’s most theoretically nuanced analytical
contributions. Chapter 7 investigates the continuum between
explicitness and implicitness, employing Grice’s Theory of Implicature
and Relevance Theory to explain how offensive meanings are conveyed
through hinting, insinuations, rhetorical questions and ostensibly
polite phrasing. The authors examine how weak implicatures create room
for plausible deniability on the part of the offender while placing a
heavy interpretive responsibility on the hearer. Finally, Chapter 8
explores digital linguistic creativity, examining how users refashion
language through tropes, figures of diction and platform-specific
multi-modal features, such as memes and emojis, to construct novel,
context-dependent forms of hostility that easily evade keyword-based
automated detection systems.
EVALUATION
As Tayebi and Parvaresh (2026) argue, to understand online offensive
language, it is important to examine it from the perspective of those
who have experienced it. However, it can be difficult to accurately
identify examples of online offensive language. Rather than compiling
examples that they themselves find offensive, the authors deliberately
work with language that has already been judged as offensive by its
targets by drawing on a trend featuring content creators or social
media influencers reading their mean, hateful or offensive comments.
This dataset, Tayebi and Parvaresh note, benefits from the occasional
accompanying commentary from the content creator in question,
explaining why they found the comment under discussion offensive. As
such, the participant-led definition of offence is probably the book’s
most important methodological contribution and should be commended as
it provides the audience with a thorough investigation of online
offensive language that can definitively be classed as such.
A note about this database is required, however. While content
creators make an innovative source of naturally occurring data, they
also represent a distinctive communicative population, which
complicates broader generalisations. For example, the sampling
decision to exclude activist creators as their discourse is too
politically polarising excludes a sub-category of influencers from
being explored and leads to lifestyle-related content creators being
over-represented in the dataset. These conditions are likely to
influence the language against them. Accusations concerning
authenticity (“Why am I the only one who thinks she is faking?”, p.
96), appearance (“You are ugly and have man shoulders”, p. 100) and
lifestyle (“This girl’s diet is shocking!!! Does she never eat fruit
and vegetable?”, p. 104), for example, may be particularly salient
because they are precisely the aspects of self that influencers
commodify. This does not remove the book’s ecological richness but
does raise questions as to the scope of the book’s representativeness.
The book itself begins from an explicitly pragmatic position:
offensive force depends not simply on words but on speaker intention,
hearer interpretation, contextual dynamics and patterns of use.
However, the original communicative event is not always completely
available. Instead, some comments survive as screenshots while others
are read aloud by creators. While the authors may know that the target
took offence and have access to the target’s subsequent metalinguistic
explanation, they may not necessarily have the whole interactional
history, such as preceding turns, the relationships between
participants, platform affordances and audience responses. This
limitation becomes increasingly apparent when the analysis moves into
an exploration of weak implicatures (e.g., “You are kind of pretty for
a chubby girl”, pp. 143-144) and intended meanings where the authors
argue that hinting (e.g., “Is your boyfriend still with you?”, p.
138), insinuating (i.e., “Your girlfriend is so hot!”, p. 146) and
implicature (e.g., “Some of your Instagram posts are good!”, p. 140)
cannot be understood from lexical content alone. The book, therefore,
provides productive tension between its insistence on the importance
of context and the amount of context that is recoverable from the
dataset.
Despite these interactional data limits, the multi-layered analysis
that takes place throughout the book is one of its strengths in that
it combines structural, lexical, semantic and discourse-pragmatic
analysis, rather than treating corpus statistics or pragmatics as
sufficient in isolation. In this way, one of the book’s central
arguments is foregrounded: no single linguistic level adequately
explains offence. This argument is demonstrated across chapters five
and six whereby a semantic-domain analysis (Chapter five) does not
identify kinship as significant while a subsequent keyword analysis
(Chapter six) reveals how family members can become resources for
attacking targets (e.g., “Imagine being a dad and seeing your son do
this”, p. 110). These examples demonstrate a limitation of semantic
tagging in that it is a process grounded in dictionary meanings and
cannot, therefore, reliably capture contextual or pragmatic meanings.
Through these examples, then, the authors demonstrate not only why
this approach is necessary but also why such a multi-layered approach
is valuable.
Beyond analysing how offence is caused, the book makes a major
theoretical contribution to third wave (im)politeness theory by
conceptualizing the act of taking offence as an active, goal-oriented
social action. Drawing on Garfinkel’s notion of the moral order and
Haidt’s moral foundations theory, Tayebi and Parvaresh argue that
taking offence is not merely an internal, passive emotional state
(i.e., “feeling offended”) but a public stance in which an individual
registers a breach of interactional norms and holds the speaker
morally accountable. In digital environments, this distinction becomes
especially salient. The authors demonstrate that public registers of
offence, such as call-out posts, metalinguistic commentary or hashtag
campaigns like #offended, function as strategic tools for norm
negotiation, identity construction and boundary enforcement. For
example, the book illustrates how social media users deploy #offended
not simply to express hurt feelings but to assign public blame,
sanction perceived injustice and urge remedial action. By framing
taking offence as an empowered social practice rather than mere
vulnerability, “Words that Wound” offers a refreshing
counter-narrative to popular media discourse surrounding the
“snowflake generation”, positioning digital interlocutors as active
agents in online norm negotiation.
The chapters on implicitness and creativity may provide the book’s
strongest theoretical contribution as they bring together many of the
earlier strands of the book to demonstrate the limits of automatic
detection. The book shows that offensive meanings can be constructed
through ostensibly polite language (“Just a friendly suggestion, if
you do not put your face on thumbnail, you will get more views”, p.
144), rhetorical questions (e.g., “What are you doing with your
life???”, p. 171), comparisons (“I am 3 years older than you and I
look like a baby beside you”, p. 134), irony (e.g., “I can tell by her
teeth she loves her tea”, p. 157) and other strategies in which
surface forms do not straightforwardly encode the harmful proposition.
The book openly acknowledges that weak implicatures become
increasingly indeterminate and place greater interpretive
responsibility on the hearer. The creativity chapter then extends the
problem further: creative offence may operate across lexical,
syntactic and pragmatic levels, meaning that systems based
predominantly on recognising offensive words can miss the underlying
proposition altogether. This approach gives the volume clear
significance beyond descriptive corpus linguistics. Its contribution
to computational work is perhaps less that it supplies a ready-made
solution for automatic moderation and more that it specifies,
linguistically, why the problem is so difficult.
As Tayebi and Parvaresh (2026) write, language that falls short of the
legal threshold of hate speech is not always examined in the
literature. Therefore, this book makes an important contribution by
turning attention to the more everyday forms of offensive
communication that can easily fall between existing analytical
categories. Its particular strength lies in demonstrating the
diversity of forms that such offence can take: from overt insults, to
meanings communicated through implication, ostensibly polite language
and linguistic creativity. By grounding its analysis in comments
identified as offensive by those targeted, the book also places the
experience and interpretation of offence at the centre of the
analysis. On the whole, ‘Words that Wound’ is essential reading for
anyone interested in online harms, offensive communication and digital
discourse, and will be valuable to scholars and students across
linguistics, communication studies and the wider social sciences.
ABOUT THE REVIEWER
Dr Keighley Perkins is a Research Associate at Cardiff University and
a researcher at Swansea University. Her work investigates the
intersection of language, politics, and digital media, with particular
interests in online communities, extremism, and online harms.
Utilising discourse analysis, corpus linguistics and forensic
linguistics, she examines how language constructs identity and harm
online. She has collaborated on projects analysing media impartiality,
online child sexual abuse and exploitation, and peer-to-peer sexual
harassment.
ORCID:  https://orcid.org/0000-0002-9667-9403



------------------------------------------------------------------------------

********************** LINGUIST List Support ***********************
Please consider donating to the Linguist List, a U.S. 501(c)(3) not for profit organization:

https://www.paypal.com/donate/?hosted_button_id=87C2AXTVC4PP8

LINGUIST List is supported by the following publishers:

Australian Linguistics Society https://als.asn.au/Home

Bloomsbury Publishing http://www.bloomsbury.com/uk/

Cambridge University Press http://www.cambridge.org/linguistics

Cascadilla Press http://www.cascadilla.com/

De Gruyter Brill https://www.degruyterbrill.com/?changeLang=en

Edinburgh University Press http://www.edinburghuniversitypress.com

European Language Resources Association (ELRA) http://www.elra.info

John Benjamins http://www.benjamins.com/

Language Science Press http://langsci-press.org

MDPI Languages https://www.mdpi.com/journal/languages

MIT Press http://mitpress.mit.edu/

Multilingual Matters http://www.multilingual-matters.com/

Narr Francke Attempto Verlag GmbH + Co. KG http://www.narr.de/

Netherlands Graduate School of Linguistics / Landelijke (LOT) http://www.lotpublications.nl/

Peter Lang AG http://www.peterlang.com

SIL International Publications http://www.sil.org/resources/publications

University of Toronto Press http://www.utpjournals.com/


----------------------------------------------------------
LINGUIST List: Vol-37-3254
----------------------------------------------------------