Detecting Hate Speech Acts in Social Media

Rahmadsyah Rangkuti

, Andi Pratama Lubis

and Zulfan

Department of English, Faculty of Cultural Sciences, Universitas Sumatera Utara, Medan, Indonesia

Department of Arabic, Faculty of Cultural Sciences, Universitas Sumatera Utara, Medan, Indonesia

Department of Linguistic, Faculty of Cultural Sciences, Universitas Sumatera Utara, Medan, Indonesia

Keywords: Hate Speech, Social Media, Speech Act.

Abstract: Hate speech acts that occur in the online realm expressed with words of prejudice and negative feelings are far

more dangerous than in the offline realm. Hate speech is a new area in the study of illocutionary speech acts.

This new area of speech acts becomes more interesting because every hate speech has various meanings or

illocutionary forces based on speaker’s intention. This study aims to analyze the classifications and aims of

illocutionary acts and illocutionary forces of hate speech contained in the facebook group account related to

Batu Bara district’s local election. The research method is descriptive qualitative. The data of this research

are utterances/speeches of facebook users in Batu Bara district’s local election group account. Data were

collected using the refer method by technique of record. This method is used to observe the expression of

speech of the facebook users’ hatred on issues related to social, cultural and political background on each

candidate. Record techniques are used to record the utterances and expression of hatred uttered by the net

citizens.

1 INTRODUCTION

This research is motivated by the rise of hate speech

phenomena in the life of Indonesian multicultural

society both in the offline domain and in the online

domain that can trigger social conflict and cause the

vulnerability of the integrity of the State (Ahnaf and

Suhadi, 2015).

Hate speech can be understood as any form of

expression that is propagated to incite, promote or

justify racial hatred, xenophobia, anti-Semitism, or

other forms of hatred rooted in intolerance, including

intolerance expressed by nationalism and aggressive

ethnocentrism, discrimination and hostility towards

minorities and migrants (Weber, 2009). A somewhat

similar definition is also conveyed by national human

rights commission of the Republic of Indonesia that

what is meant by hate speech is any action and effort

either directly or indirectly based on hatred on tribe,

religion, religious stream, belief/faith, race,

interclass, color skin, ethnicity, gender, disability, and

sexual orientation which is incitement towards

individuals and groups to discriminate, violence,

disappearance of life and /or social conflict through

various means (Komisi Nasional Hak Asasi Manusia

Republik Indonesia, 2016).

The Police of Republic of Indonesia cyber crime

unit stated in 2015 there were 671 reports related to

hate speech (Nadia, 2016). That number continues to

rise in 2016 which amounted to 1829 cases, and until

the end of 2017 reached 3325 cases (Medistiara,

2017). The number shows that there is an increase in

hate speech cases on the online sphere.

To anticipate the above phenomena, the

Indonesian government has actually created two laws

and regulations that substantially regulate hate

speech, the Criminal Code (KUHP) and the

Information and Electronic Transaction Act of 2008

(Rongiyati, 2015). However, the more massive

phenomena of hate speech contribute to the

emergence of a Circular Letter of the Chief Police of

the Republic of Indonesia No. SE / 6 / X / 2015 on the

handling of hate speech. In the circular letter, as a law

enforcement officer, the Indonesian National Police

describes that hate speech may takes various forms

such as criminal offenses, defamation, blasphemy,

unpleasant acts, provocation, incitement and

spreading of hoax.

The description of various forms of hate speech is

a new subject for linguistic studies, especially in

discourse analysis (Gagliardone, Patel & Pohjonen,

2014) and pragmatics (Assimakopoulos, Baidar &

1242

Rangkuti, R., Lubis, A. and Zulfan, .

Detecting Hate Speech Acts in Social Media.

DOI: 10.5220/0010070912421247

In Proceedings of the International Conference of Science, Technology, Engineering, Environmental and Ramiﬁcation Researches (ICOSTEERR 2018) - Research in Industry 4.0, pages

1242-1247

ISBN: 978-989-758-449-7

Millar, 2017, or speech acts related studies. However,

Özarslan (2014) stated that speech act theory can be

applied in analyzing hate speech. According to him,

in the concept of speech acts, as proposed by Austin

(1962) that when someone says something, he also

does something, and may also do hate speech acts

related to social and cultural context.

As the core study of speech acts, Searle develops

and classifies speech acts into five types namely, 1.

Representatives (Assertives) are those kinds of

speech acts that state what a speaker believes to be the

case or not 2. Commissives are those kinds of speech

acts that speakers use to commit themselves to some

future action 3. Directives are those kinds of speech

acts that speakers use to get someone else to do

something 4. Declarations are those kinds of speech

acts that change the world via their utterance and 5.

Expressives are those kinds of speech acts that state

what the speaker feels (Searle, 1979). The proposed

classification is based on three main principles

namely, illocutionary point or the purpose of the

speaker, words adjusted to the reality of the world

(direction of fit), and psychological states/ sincerity

condition.

Table 1. Searle's Classification of Speech Acts (adapted

from Yule, 1996)

Speech act

type

Direction of fit S = Speaker

X = Situation

Assertives make words fit

the worl

S believes X

Commissives make the world

fit words

S intends X

Directives make the world

fit words

S wants X

Declarations word changes the

worl

S causes X

Expressives make words fit

the worl

S feels X

Searle has also proposed the idea that the

realization of speech acts is influenced by four

conditions, namely (1) the propositional content

condition, (2) the preparatory condition, (3) the

sincerity condition, and (4) the essential condition

(Searle, 1969). According to him, each of these

conditions can distinguish the intention of each form

of speech acts.

In relation to this, there are several studies that

have been done based on the perspective of linguists

in assessing the phenomena of hate speech. For

example, research conducted by Mintonawati (2016)

associated with defamation cases on Facebook's

social network using linguistic forensic studies

involving lexical semantic analysis, grammatical

semantics and analysis of illocutionary speech acts.

To sum up, this research has not found specific type

of illocutionary acts and forces from the data.

Towsend (2014) in his research, also did not mention

clearly the types of illocution of hate speech cases he

discussed.

Slightly different from the two earlier studies,

Virginia and Olanrewaju (2017) in their research

based on speech analysis of the politician's language

in Nigerian magazines and newspapers in 2012-2015

show that violence, extortion, provocation and

intimidation. However, their research has not

specified the type of illocutionary acts of the

politicians' speeches.

Another study was Linawati's research (2017),

which was specifically done in analyzing the

netcitizen's utterances in the Online Tribunnews.com

newspaper comment column. Using the theory of

speech acts, she has successfully found various forms

of hate speech in the Online Tribunnews.com

newspaper comment column such as, humiliation,

defamation, unpleasant acts, provocation and

incitement and based on speech act types she found

assertive, directive and commissive. Linawati's

research has not found the expressive type of hate

speech.

The last is a study done by Octaviani (2017).

Based on her research conducted using ethnography

of communication, it was found that there were forms

of hate speech in the instagram account of former

Governor of the Special District of Jakarta, Basuki

Tjahaja Purnama. The forms of hate speech include

humiliation, defamation, blasphemy, provocation,

spreading hoax, and incitement. According to her, all

forms of hate speech have impacts such as

excommunication, discrimination, violence, hatred

towards groups, and group extermination. However,

this research has also not found the expressive type of

hate speech.

Based on the above explanation of hate speech

from linguistic perspective, this research tries to find

the classification and aims of hate speech acts

contained in Facebook group account related to Batu

Bara district local election. This is important because

the propensity for spreading hate speech is more

massive when entering the period of direct election of

regional heads and closely related to the discourse to

seize or maintain the power (George, 2017). More is

known at this time that one of the four pairs of

candidates for regional head of Batu Bara distrct is

Chinese descendant. This fact is increasingly

encouraging the occurrence of hate speech in the

social media such as Facebook done by each

supporters.

Detecting Hate Speech Acts in Social Media

1243

2 METHOD

The research method used is descriptive qualitative.

This research method aims to describe, summarize

various conditions, various situations, or various

phenomena of social reality. The study also attempts

to draw the reality to the surface as a characteristic,

character, trait, model, sign, or description of a

particular condition, situation, or phenomenon

(Bungin, 2017).

The data are utterances or speech in the form of

words, phrases and sentences. The data are from the

utterances or speech of the members and net citizens

in the Facebook group account KOMBUR-

KOMBUR PILKADA BATU BARA 2018

MENUJU MASYARAKAT EKONOMI BB

BERSINAR (Towards the prosperity of Batu Bara)

and Facebook group account KOMBUR-KOMBUR

PILKADA BATUBARA 2018. Data collected using

the refer method (Sudaryanto, 2015). This method is

used to observe the speech of net citizen's hatred on

issues related to the social, cultural and political

background of each candidate. The technique of

record is used to record the expression of hate speech

by the net citizens (Sudaryanto, 2015).

The data collected then analyzed using contextual

analysis method (Rahardi, 2009). A contextual

analysis is simply an analysis of a text (in whatever

medium, including multi-media) that helps us to

assess that text within the context of its historical and

cultural setting, but also in terms of its textuality – or

the qualities that characterize the text as a text.The

context referred to in this method is the language

environment. Linguistic environment may be

physical environment or nonphysical environment. In

other words the context can be understood as the

background knowledge shared by speakers and

hearers, so the hearers understand what the speakers

mean (Leech, 1983).

3 RESULTS AND DISCUSSION

In accordance with the objectives of the study, then in

this section results and discussion of research that

includes two things will be presented, namely: 1),

classification and, 2) aims of hate speech found in

Facebook group account KOMBUR- KOMBUR

PILKADA BATU BARA 2018 MENUJU

MASYARAKAT EKONOMI BB BERSINAR

(Towards the prosperity of Batu Bara) and Facebook

group account KOMBUR-KOMBUR PILKADA

BATUBARA 2018.

3.1 Results

Based on the speech acts theory described earlier,

three types of illocutionary acts are found. The three

types of illocutionary act have their own meanings

and parameters. They were found from thirteen

utterances of net citizens in the Facebook account. Of

the thirteen utterances three illocutionary forces are

found. All the findings can be seen in the following

table.

Table 2. List of hate speech found in Facebook group

accounts KOMBUR-KOMBUR PILKADA

BATUBARA 2018 MENUJU MASYARAKAT

EKONOMI BB BERSINAR (Towards the prosperity

of Batu Bara) and KOMBUR-KOMBUR PILKADA

BATUBARA 2018

No Hate speech Illocutionary

Acts

1 Cino dah makin merajolela

di Indonesia ini di seluruh

enjurunyo […]

Assertive

2 Kojap lagi Indonesia pun

men

adi ne

aro cino[…]

Assertive

3 Kalau bukan punyo asing

mako

a asen

[…]

Assertive

4 Kalau cino menghalalkan

segala cara demi harta

karna mereka mati bawak

harta tidak kenal riba[…]

Assertive

5 Saya pribadi tak ondak dan

tak rela apobilo btu bra ni

di kuasai oleh cino.

Assertive

6 Selain dia yg di dukung itu

thionghoa kturunan.

Assertive

7 Semua nya dri cina

kemunis

Assertive

8 Dai dulu kami ondak milih

pak zahir tapi chino di

belakangnya jadi malas

odan.

Assertive

9 Ini salah satu cabup

batubara yg di tunggangi

anak cina sbg cawabup

nya.

Assertive

10 Ganyang CINA Directive

11 Pukimak kau china Ex

ressive

12 Kau bukan ada hak dalam

negeri kami

Assertive

13 USIRRRR CHINA DARI

INDONESIA

Directive

3.2 Discussion

Based on the analysis that has been done on both

Facebook group accounts, three classifications of

illocutionary acts are found, assertive, directive and

expressive.

ICOSTEERR 2018 - International Conference of Science, Technology, Engineering, Environmental and Ramiﬁcation Researches

1244

3.2.1 Assertive

Assertive is a speech act involving the speaker on the

truth of the expressed proposition. For example,

states, notices, prides, complains, demands, and

reports. Therefore, every act of assertive can be

judged to be true. In other words, that in every

assertive, speaker speaks the words to the world

(belief). Then every speech from every speaker

within the two Facebook group accounts expresses

the speakers' true belief on the information,

experience and evidence they have related to Chinese

ethnic in Indonesia. In this case it is the truth that the

Chinese descendants are trying to dominate Indonesia

in any way. Based on the truth as believed by each

speaker, the assertives found in this study can be

described as follows.

Table 3. Illocutionary assertive acts found in

KOMBUR-KOMBUR PILKADA BATUBARA 2018

MENUJU MASYARAKAT EKONOMI BB

BERSINAR (Towards the prosperity of Batu Bara) and

KOMBUR-KOMBUR PILKADA BATUBARA 2018

Illocutionar

Assertive Act

Propositional

content

truth of the information, experience

and evidence that Chinese

descendants try to dominate

Indonesia

(

)

Preparatory 1. Speaker (S) has facts, reasons,

evidences on Pc

2. Hearer has no clear reasons on

Sincerit

Speaker believes Pc

Essential Pc shows that the information is true

There are three sub types of assertive found in this

case, namely;

a. Statement. This sub type indicates the speaker’s

belief in stating something. There are two aims of this

sub type, to state and to inform. The parameter of this

category is when th e speaker speaks, the speaker

states that the Proposition is: i). stating that there is

information, ii). experience and iii). evidence that he

believes that Chinese is an ethnic who is trying to

dominate Indonesia in any way. The explanation of

this sub-type can be seen as follows,

(1) Cino dah makin merajolela di Indonesia ini

diseluruh penjurunyo [...] (HS.01) ‘Chinese are

rampant all over Indonesia’. The utterance (1) is

statement sub type of assertive.

Through the speech (1) speaker tries to declare

that Chinese in Indonesia are trying to dominate

Indonesia. It seems that this prejudice arose after the

presidential election of Indonesia in 2014. It

continued after the act to protect Islam against

defamation I, II, and III which asked the government

to punish Basuki Tjahaja Purnama who had defamed

Islam. Meanwhile, the news of a large wave of

foreign workers from the People's Republic of China

(PRC) is increasingly leads to situations and

sentiments between ethnics and religions in

Indonesia. One way to think about the speech acts

being performed via utterances is to assume that

underlying every utterance (U) there is a clause,

containing a performative verb (Vp) which makes the

illocutionary force explicit (Austin, 1962). This is

known as performative hypothesis and the basic

format of the underlying clause is ‘I (hereby) Vp you

(that) U’ (Yule, 1996). Applying paraphrase

technique as proposed by Sudaryanto (2015) on hate

speech (1) the true aim of statement of assertive sub

type can be exercised, “I hereby {state/ inform} you

that Chinese are rampant all over Indonesia”.

b.Predicting. This sub-type predicts what will happen

in the future based on the speaker's belief. The

parameter of this sub type is when speaker utters the

speech, the speaker predicts the Proposition (Pc): i).

Predicting that something will happen as evidence

that the information, experience and evidence of

ethnic Chinese ambition that he has is true.

Explanation of this sub-type can be seen as follows,

(2) Kojap lagi Indonesia pun menjadi Negara Cina

[...] (HS.02) In the near future Indonesia will become

China [...]. The sub-type truth can be tested by the

performative hypothesis and paraphrase technique as

described earlier. Consider the following illustration

: (2) I hereby {predict} that in the near future

Indonesia will become China [...]

Based on the above paraphrase technique, it can

be said that the speech (2) is an assertive of predicting

sub-type, the speaker intended to predict what would

happen if the Chinese and the foreign workers from

China had succeeded in carrying out Indonesia.

3.2.2 Directive

Directive is is a speech act intended by the speaker to

make the hearer does something. For example,

ordering, begging, requesting, suggesting, and

advising. Therefore, the proposition content of the

directive speech act is a future act of the hearer. Based

on this, it is understood that in each directive speech,

the speaker seeks to make the world fit words.

Related to that, then every directive speech within the

two Facebook accounts express the actions of hearers

in the future that is a solicitation to murder Chinese

descendants and an order to expel them from

Indonesia.

Detecting Hate Speech Acts in Social Media

1245

Table 4. Illocutionary Directive acts found in Facebook

group accounts of KOMBUR-KOMBUR PILKADA

BATUBARA 2018 MENUJU MASYARAKAT

EKONOMI BB BERSINAR (Towards the prosperity

of Batu Bara) and KOMBUR-KOMBUR PILKADA

BATUBARA 2018

Illocutionar

Directive Acts

Propositional

content

An action (A) that will be carried

out by the hearer in the future

Preparatory 1. The hearer (H) can do an action

(A), and the speaker believes

that the hearer is able to do an

action (A)

2. It is not clear to the speaker and

the hearer that the hearer is able

to do an action as directed by

the s

eake

Sincerity The speaker wanted the hearer to

do an action (A)

Essential To persuade the hearer to do an

action (A)

Based on the description above, one sub type of

directive is found namely solicitation.This sub-type

is based on the speaker's desire for the hearer to do

something in the future. The parameter is when telling

the speech, the speaker persuades the hearer to

perform an action, if the speaker expresses: i). the

desire that the hearer does an action (A), ii). that

action (A) is the speaker's desire. An explanation of

this sub-type can be seen as follows, (3) Ganyang

CHINA. (HS.03) 'Kill Chinese descendants'. Speech

(3) is a solicitation sub-type of illocutionary directive

acts. Through speech (3) speaker tries to persuade

hearer to kill Chinese descendants. To examine the

truth of this sub-type the formula "S verb (O) that X"

or "S verb O to Y" is used (Yule, 1996). Consider the

following: (3) I persuade you to kill Chinese

descendants. After using paraphrase technique, it can

be said that speech (3) is a solicitation sub-type of

illocutionary directive acts, in which the speaker

persuades the hearer to kill Chinese descendants.

3.2.3 Expressive

Expressive is a speech act that expresses the speaker's

psychological attitude to a situation. It could be an

expression of gratitude, joy, hatred, anger,

forgiveness, blame, praise, condolence, and so on.

Yule (1996) explains that when speakers use

illocutionary expressive acts they attempt to fit words

to the world. Based on this, it can be understood that

every illocutionary expressive acts found in this study

show resentment of speakers to Chinese descendants

based on a state of experience, information, and

evidence that already possessed by speakers.

Table 5. Illocutionary Expressive acts found in

Facebook group accounts of KOMBUR-KOMBUR

PILKADA BATUBARA 2018 MENUJU

MASYARAKAT EKONOMI BB BERSINAR

(Towards the prosperity of Batu Bara) and KOMBUR-

KOMBUR PILKADA BATUBARA 2018

Illocutionary Expressive acts

Propositional

coentent

An event has been experienced and

influenced the speaker's attitude

towards Chinese descendants.

Preparatory 1. Speaker has fact, reason and

evidence on Pc.

Sincerit

Essential Speaker denotes hatred

Based on the description above, one sub type of

expressive is found namely offensive. This sub-type

is based on the expression of the speaker's hatred for

Chinese descendants based on a state of experience,

information, and evidence that the speaker has. The

parameter is when the speech is spoken, speaker

insults hearer by expressing: i. Speakers' dislike

toward Chinese descendants, ii. and the speaker's

hatred on all Chinese activities. An explanation is as

follows: (4) Pukimak kau China (HS.04) ‘Fuck

you....’. Speech (4) is an offensive sub-type of

illocutionary expressive act. Through speech (4)

speaker shows his dislike and hatred towards Chinese

descendants.

4 CONCLUSIONS

Based on the results of research and discussion, it can

be concluded that hate speeches found in Facebook

group accounts KOMBUR- KOMBUR PILKADA

BATU BARA 2018 MENUJU MASYARAKAT

EKONOMI BB BERSINAR (Towards the prosperity

of Batu Bara) and KOMBUR-KOMBUR PILKADA

BATUBARA 2018 are classified into assertive,

directive and expressive. In addition, the thirteen

speeches analyzed have illocutionary forces that are

insulting, inciting and discriminating, especially on

Chinese descendants. This happened because one

candidate for Batu Bara local election is Chinese

descendant.

This study concludes that legal cases on hate

speech can be interpreted and solved linguistically

provided that fairness and social justice are the ends

to be achieved.

ICOSTEERR 2018 - International Conference of Science, Technology, Engineering, Environmental and Ramiﬁcation Researches

1246

ACKNOWLEDGEMENTS

The authors gratefully acknowledge that the present

research is supported by Ministry of Research and

Technology and Higher Education Republic of

Indonesia. The support is under research grant DRPM

of year 2018 contract number:

89/UN5.2.3.1/PPM/KP-DRPM/2018.

REFERENCES

Ahnaf, M. Iqbal & Suhadi. 2015. Isu-Isu Kunci Ujaran

Kebencian (Hate Speech): Implikasinya Terhadap

Gerakan Sosial Membangun Toleransi. Harmoni:

Jurnal Multikultur Multireligius. 3 (13), pp.153-154.

Assimakopoulos, Stavros, Baider, Fabienne H, & Millar,

Sharon. 2017. Online Hate Speech in The European

Union: A Discourse Analytic Perspective. Springer.

Switzerland.

Austin, J.L. 1962. How to Do Things With Words. London:

Cambridge University Press.

Bungin, Burhan. 2007. Penelitian Kualitatif: Komunikasi,

Ekonomi, Kebijakan Publik, dan Ilmu Sosial Lainnya.

Jakarta: Kencana.

Gagliardone, Ignio, Patel, Alisha & Pohjonen, Matti. 2014.

Mapping And Analysing Hate Speech Online:

Opportunities and Challenges for Ethiopia. Oxford.

George, Cherian. 2017. Pelintiran kebencian: Rekayasa

Ketersinggungan Agama Dan Ancamannya Bagi

Demokrasi. PUSAD Yayasan Paramadina dan IIS

UGM. Jakarta.

Kepolisian Negara Republik Indonesia. 2015. Surat Edaran

Kepala Kepolisian Negara Republik Indonesia Nomor:

SE/6/X/2015 tentang Penanganan Ujaran Kebencian

(Hate Speech). Jakarta.

Komisi Nasional Hak Asasi Manusia Republik Indonesia.

2016. Buku Saku Penanganan Ujaran Kebencian (Hate

Speech). Jakarta.

Leech, Geofrey. 1983. Priciples of Pragmatics. New York:

Longman.

Medistiara, Yulid 2017, ‘Selama-2017-polri-tangani-3325-

kasus-ujaran-kebencian’, Detik News, 29 Desember,

viewed 23 Juli 2018, < https://news.detik.com/berita/d-

3790973/selama-2017-polri-tangani-3325-kasus-

ujaran-kebencian>.

Mintonawati, 2016, ‘Pencemaran Nama Baik: Kajian

Linguistik Forensik’, PARAMASASTRA, 3 (2), pp.

27-37.

Nadia, Ambaraie 2017, ‘2016 konten berisi ujaran

kebencian paling banyak diadukan’, Kompas.Com, 26

March, viewed 23 Juli 2018, <

https://nasional.kompas.com/read/2017/03/26/084656

11/2016.konten.berisi.ujaran.kebencian.paling.banyak.

diadukan.ke.polisi.>.

Octaviani, Aulia, 2017, ‘Ungkapan Kebencian Pada

Tuturan Heaters Di Akun Instagram BasukiBTP Dan

Relevansinya Terhadap Pembelajaran Bahasa

Indonesia Di SMA’, eprints.ums.ac.id, 03 August,

viewed 23 Juli 2018,

http://eprints.ums.ac.id/id/eprints/54308.

Özarslan, Zeynep. 2014. Introducing Two New Terms Into

The Literature of Hate Speech: “Hate Discourse” And

“Hate Speech Act” Application of “Speech Act

Theory” Into Hate Speech Studies in The Era of Web

2.0. Haziran. (20), pp.53-75.

Rahardi, Kunjana. 2009. Sosiopragmatik. Erlangga.Jakarta.

Rongiyati, Sulasi. 2015. Surat Edaran Kapolri Tentang

Ujaran Kebencian: Menjaga Kebebasan Berpendapat

Dan Harmonisasi Kemajemukan. Info Singkat Hukum,

Vol.II, No.21.pp.1-5

Searle, John R. 1979. Expression and Meaning: Studies in

The Theory of Speech Act. Cambridge: Cambridge

University Press.

Searle, John R. 1969.

Speech Acts an Essay in The

Philosophy of Language. Cambridge. Cambridge

University Press.

Sudaryanto. 2015. Metode dan Aneka Teknik Analisis

Bahasa Pengantar Penelitian Wahana Kebudayaan

secara Linguistis. Yokyakarta: Sanata Dharma

University Press.

Townsend, Emma. 2014. Hate Speech or Genocidal

Discourse? An Examination of Anti-Roma Sentiment

in Contemporary Europe. Portal. 11 (1). 1-23

Virginia, Okafor C & Olarenwaju, Alabit Taofeek, 2017,

‘A Speech Act Analysis of Hate Speeches in the 2015

General Election Campaign in Nigeria’, IMPACT:

Internasional Journal of Research In Humanities, Arts

and Literature. 5 (6). 61-72.

Weber, Anne. 2009. Manual On Hate Speech. France.

Council Of Europe.

Yule, George. 1996. Pragmatic. Oxford. Oxford University

Press.

Detecting Hate Speech Acts in Social Media

1247