This paper explores the integration of Corpus Linguistics and Discourse Analysis methodologies, specifically the use of keyword analysis to pinpoint salient discourse representations of given social identities. The particular identity examined here is that of 'kuffar', which is a derogatory term used by jihadist-ideology groups to describe people who do not share their faith. Two measures (Log Likelihood and %DIFF) are used to determine the keyness value of the term kuffar in the propaganda magazines produced by two such groups: Al Qaeda (Inspire) and Islamic State (Dabiq). Although they yield different keyness values, both confirm the salience of this social identity within jihadist propaganda. The results of a Key Word in Context concordance analysis of the term kuffar show how negative representations thereof are mainly legitimized on impersonal authority grounds (Van Leeuwen, 2007). The results also reveal a trend towards reminding Muslims of their “duty” to fight and kill kuffar individuals. This is supported by positive expressions to describe Muslims who perform this duty, and negative expressions and traits associated with the non-believers. This case study supports the view that CL and Discourse Analysis can offer a 'useful synergy' (Baker et al. 2008, 2015) to research into, amongst other areas, representation and ideology in language.
The research in this paper was carried out as part
of a research project funded by Swansea
University and directed by professors Stuart
Macdonald and Nuria Lorenzo-Dus. The project
examined jihadist propaganda from different
disciplinary perspectives, including Terrorism
Studies and Linguistics. All available online
publications of five jihadist magazines released
between January 2009 and the end of June 2015
were collected. These publications are Jihad
Recollections, Gaidi Mtaani, Inspire, Dabiq and
Azan. To be included in the data set, publications
had to be written in English and meet the project’s
definition of a magazine (Macdonald et al., 2015).
The reason this paper focuses on Dabiq and
Inspire, produced by IS and Al-Qaeda,
respectively, is that these two groups currently
pose the biggest terrorist threat to the Western
world. Terrorist groups are also often regarded as
acting the same, especially if they follow a jihadist
ideology. However, they might target different
social groups and compete for recruits in fighting
a common enemy – usually an out-group who
does not share their ideology and faith.
Comparative research within these groups is
therefore most useful to understand their
similarities and differences. The aim of this paper
is twofold: (1) to contribute to the current
academic debate about one of the main analytic
tools in Corpus Assisted Discourse Studies
(CADS): keywords; and (2) to show how IS and
Al-Qaeda discursively construct “kuffar” in their
propaganda magazines and legitimize this as a
negative social identity. “Kuffar” is a derogatory
term, which describes individuals on religious
grounds, or lack thereof. Analyzing this term
provides insights into the role of religion in
jihadist ideologies, which is still heavily debated.
1.1 CADS
Corpus-Assisted Discourse Studies (CADS) is the
application of Corpus Linguistics (CL) tools such
as keywords, wordlist, and concordance line
analysis in Discourse Studies. The field has been
evolving since the 1990s, when new Corpus
Linguistics tools were developed. Until 2004 not
many studies used a CADS approach
Keyword analysis is a much used technique of CL
Log-likelihood is a test that shows statistical
significance (Baker et al., 2006b). It assigns a
pvalue to every word and measures how significant
the word is in comparison to a reference corpus.
According to Biber et al. the p-value ‘represents
the probability that this keyness is accidental’
Effect size measures ‘the difference of the
frequency of a word in a study corpus when
compared to that in the reference corpus’
This paper will first compare the results of a keyword analysis of the corpus (see 3.1), focusing on the term ‘kuffar’1, by applying log-likelihood and %DIFF measures. It will then report the results of a KWIC analysis for “kuffar” that shows how its negative discourse representation is legitimized by the magazine authors.
3.1 Data The corpus for this study consist of 22 issues of Dabiq (185,951 words) and Inspire (304, 347 words). 13 of these issues are from Inspire and were published between the summer of 2010 and the spring of 2015. The remaining nine issues were published by IS between the summer of 2014 and the spring of 2015. The digital versions of the issues were collected in the summer of 2015, coinciding with the onset of the project.
Aim (1) entailed calculating keywords using
Loglikelihood and %DIFF (see section 4 and
Appendix A and B for details). In order to further
our understanding of how “kuffar” is used and
constructed as a social identity for aim (2), it was
necessary to carry out a Key Word in Context
(KWIC) analysis, which requires manual
examination of concordance lines generated by
AntConc
The corpus was not lemmatized because one of the project’s aims was to examine stylistic variation, including at the spelling level, between the different magazines, especially regarding use of Arabic terms.
In CADS, the analysis usually starts with a keyword analysis. As can be seen in Tables 1 and 2, the log-likelihood for “kuffār” is the same in both tables, even though in the first Dabiq is the study corpus, and in the second Inspire is the study corpus.
word kuffār kuffar frequencies SC 28 47
RC
LL 125 %DIFF
-86.31 1335.81 The highest %DIFF with Inspire as the study corpus was the word inspire with a %DIFF of 28738.42, however since this is the title of the magazine as well as a verb, this is not surprising and does not reflect what the text is about. It appeared once in Dabiq probably as a verb. The word “kuffar” is 30th in the list with a %DIFF of 1335.81. The word is only used twice in Dabiq, which suggests that this is the preferred spelling in Inspire. This is an interesting finding, considering kuffār is only used 28 times in Inspire and 125 times in Dabiq. Its %DIFF is also not very high, compared to the 100th highest, which still has a %DIFF of 342.96.
Note that if two or more words have the same %DIFF, they are all the top Xth word, for example the words “road”, “verily”, “al-bukhari”, “depend”, “dry”, “perspective”, “proceed”, “race” and “reveal” all have the same %DIFF of 358.24 and are therefore all the top 95th word. These words also happen to have a Log-likelihood of 5.87. According to this measure, they are not very relevant to the corpus. Similarly, “kuffar” would 2 For further details see a forthcoming paper by the author not have been in the top 100 keywords, if the keywords had been measured by log-likelihood; it has a LL of 31.99 as can be seen in Table 1.
In table 2 the %DIFF and Log-Likelihood of the words “kuffār” and “kāfir”3 are shown. “kuffār” is not in the Top 100 keywords, it has a %DIFF of 630.67, but it is not significant in Dabiq, since it is also used in Inspire. The top keyword in Dabiq is “shām”, which is a region in Syria near the city of Damascus, with a %DIFF of 28542.34. This is logical, because IS are in Syria fighting a war for territory among other things. This word is only used once in Inspire, which could be because it is not that important to AlQaeda, or because there are other spellings they use.
frequencies word kuffār kāfir %DIFF “kāfir” is the 30th keyword, with a %DIFF of 2791.51 and a log-likelihood of 82.23. It probably would not have been in the top 30 keywords if they had been compiled using log-likelihood.
Two words with a very high loglikelihood are “wa“ with 707.35 and “abū” with 641.25. Abū means “father of”, and is often used as part of a name. “Wa” is part of the phrase “salla llāhu ʿalayhi wa-alehe wa-sallam”, which translates to “peace be upon him” and is used immediately after mentioning the prophet Mohammad. Both terms are therefore not what the two texts are about, they are frequently used as part of or following names. They cannot be considered keywords, although they would be likely at the top of the list according to LogLikelihood.
Keywords indicate what the text is about.
Despite the different keyness values, “kuffar”
emerges as a salient keyword using both methods.
Other keywords mentioned in this section did not
occur as keywords using both methods. It is worth
pursuing other case studies to determine
advantages and disadvantages of both methods.
This is especially relevant to CADS Researchers,
who usually only analyze the top keywords. It is
important to ensure that these words are indeed
the most important ones in the given corpus
aforementioned KWIC analysis. In the context of terrorist magazines, impersonal authority is defined as follows: Legitimation through laws, rules and regulations (such as laws of Islam, the Koran, scriptures) or unwritten guidelines made by the terrorist groups. The following are two typical examples of impersonal authority as legitimation in Inspire:
Every Muslim who carries the creed of pure Tawheed, alliance to Muslims and disavowal from kuffār and lives in the West has a duty to find a suitable method to severely kill the kuffār and waylay4 them at every place of ambush until they cease killing Muslim women and their children. (Inspire, issue 13) We are proud that every muslim [sic] believes in the Islamic unity and waging jihad against kuffar who occupy our lands. (Inspire, issue 10) This example is a general appeal to Muslims living in the West, who identify with a jihadist ideology such as the one Al-Qaeda puts forward, to carry out their duty to kill non-believers (kuffar). This example shows that Inspire uses the words “duty” to remind Muslims what they have to do. The example uses the oneness of Allah (Tawheed) as a reminder and it might want to inflict guilt on Muslims who do not follow this “duty”. In fact, it attempts to do this by explicitly stating that in order to belong to the terrorist group, one must not associate with non-believers. The impersonal authority in this case comes from Al-Qaeda, who have ruled that Muslims living in the West have to kill kuffar.
The second example shows the unity and sense of community one can expect when joining Al-Qaeda. It expresses pride in waging war (jihad) against the non-believers) and makes joining the fight sound quite appealing to Muslims. This is a softer approach to reminding Muslims of their duty, playing on the fact that Muslims who share Al-Qaeda’s ideology might feel isolated and discriminated against and long to belong to a group of like-minded individuals. This example portrays the terrorist very positively and the kuffar as negatively, especially by describing them as “occupy[ing]” their lands. The impersonal authority is again expressed by Al-Qaeda based on their ideology.
The analysis will now focus on Dabiq. These are typical examples of legitimation through impersonal authority in Dabiq:
The Muslims in the West will quickly find themselves between one of two choices, they either apostatize and adopt the kufrī religion propagated by Bush, Obama, Blair, Cameron, Sarkozy, and Hollande in the name of Islam so as to live amongst the kuffār without hardship, or they perform hijrah to the Islamic State and thereby escape persecution from the crusader governments and citizens. (Dabiq, issue 7) A halt of war between the Muslims and the kuffār can never be permanent, as war against the kuffār is the default obligation upon the Muslims only to be temporarily halted by truce for a greater shar’ī interest, as in the offer of truce from the Prophet (sallallāhu ‘alayhi wa sallam) to the mushrikīn of Makkah in Hudaybiyah. (Dabiq, issue 8) But due to the deviance and arrogance of the enemies of Allah, they plot against His religion and His allies. Their plots almost cause the mountains to collapse out of shock that the kuffār dare to oppose the Lord of the heavens and the earth. (Dabiq, issue 9) The first example describes the two choices Muslims living in the West have. They can adopt Islam the way non-believers spread it and not face any problems from the government. The other option is to join the Islamic State and flee pursuit from Western governments. The words “apostatize” and “kufrī religion” imply that this is not the option Muslims should choose. The impersonal authority is expressed by IS, which is explicitly mentioned in the text.
The second example describes the fact that according to Dabiq, Muslims are always at war with the non-believers (kuffar). It emphasizes their obligation to fight the kuffar. A cease-fire can only be temporary, which happened in a historic situation that is outlined in the example, involving the Prophet, which indicates a second layer of legitimation. This example supports the call to join IS in their fight, which is outlined in the previous example.
In the third example, kuffar are associated with being deviant and arrogant and it is explicitly stated that they oppose Allah, which might inflict fear in Muslims who are living in the West or have not joined IS’s jihad. All three examples show that IS portray their jihad as the only true way to obey Allah and fight the non-believers, who oppose him.
This paper has compared %DIFF with the
traditional technique Log-Likelihood and pointed
out that kuffar emerges as a salient term using
both methods. In this case, the method did not
change the salience of the keyword in question.
This paper did not seek to prescribe to use one
method over the other, it simply presented %DIFF
as an appropriate technique to select keywords for
further analysis, especially to compare corpora of
different sizes, since it takes into account the size
of both corpora, unlike Log-Likelihood.
LogLikelihood will likely continue to be the
predominantly used method. However, the newest
version of WordSmith (WordSmith 7.0) includes
%DIFF as an option of “statistical tests” to
calculate keywords
The paper then analysed the term “kuffar” in the terrorist propaganda magazines Inspire and Dabiq. The findings include a trend towards reminding Muslims of their “duty” or “obligation” to fight and kill non-believers (kuffar). This is supported by positive expressions and notions to describe Muslims who perform this duty, and negative expressions and traits, such as being deviant, arrogant and occupying lands, associated with the non-believers. Al-Qaeda portrayed making hijra (travelling to the territory of war) and joining them as something to take pride in. They also imply a sense of belonging in their propaganda, which may speak to individuals who feel isolated. Their propaganda seems welcoming. Contrastingly, IS focus on the negative traits associated with non-believers and emphasise that kuffar and Muslims can never be at peace. They draw on historical examples of the Messenger and Allah to legitimise their discourse and evoke fear and guilt in Muslims who have not joined their jihad.
To conclude, Corpus linguistics techniques are useful tools to apply in discourse analysis to narrow the text down to a few areas of interest. The CADS approach combines quantitative and qualitative analyses as outlined above, which ensures that the findings are reliable.
Many thanks go to Nuria Lorenzo-Dus for the support, expertise and comments that greatly improved this paper. I’d also like to thank the Cyberterrorism Project and Swansea University. Thanks also go to Matteo Di Christofaro for insights into effect size (%DIFF).
Michael Stubbs. (2010). Three concepts of keywords.
In Marina Bondi and Mike Scott (Eds.),
Mike Scott and Christopher Tribble. (2006). Textual Patterns - Key words and corpus analysis in language education. Amsterdam: John Benjamin Publishing Company. Mike
Scott. (2016b). Get started guides for WordSmith. Retrieved from WordSmith Tools: http://www.lexically.net/downloads/version7 /HTML/index.html?keywords_calculate_info .htm Paul Baker. (2006a). Usind Corpora in Discourse
Analysis. London: Continuum.
Paul Baker, Andrew Hardie and Tony McEnery. (2006b). A Glossary of Corpus Linguistics.
Edinburgh: Edinburgh University Press.
Paul Baker, Costas Gabrielatos, Majid Khosravinik, Michal Krzyzanowsk, Tony McEnery and Ruth Wodak. (2008). A useful methodological synergy? Combining critical discourse analysis and corpus linguistics to examine discourses of refugees and asylum seekers in the UK press. Discourse & Society, 273-307.
Paul Baker, Costas Gabrielatos and Tony McEnery. (2013). Sketching Muslims: A Corpus Driven Analysis of Representations Around the Word ‘Muslim’ in the British Press 1998–2009.
Applied Linguistics, 34(3), 255–278.
Paul Baker. (2015). Introduction to the Special Issue.
Discourse & Communication, 9(2), 143-147. Stuart Macdonald, Nuria Lorenzo-Dus, Matteo di Cristofaro, David Mair, Anina Kinzel, Saffron Lee, Nyasha Maravanyika, David Nezri, Jodie Parker, Elliot Parry, Kate Thomas and Luke Walker. (2015). Online Terrorist Magazines. Cyberterrorism Project Research Report (No. 5). Available via: www.cyberterrorism-project.org Theo Van Leeuwen. (2007). Legitimation in discourse and communication. Discourse & Communication, 1(1), 91-112.
Tony McEnery, Mark McGlashan and Robbie Love. (2015). Press and social media reaction to ideologically inspired murder: The case of Lee Rigby. Discourse & Communication, 9 (2), 237–259.
1 66.74997 109.2500316 1 43.9943 72.00570265 1 32.61646 53.38353817 1 32.61646 53.38353817 1 23.13493 37.86506778 1 20.85937 34.14063488 1 19.34232 31.65767962 4 76.99002 126.0099796 1 17.06675 27.93324672 2 31.09942 50.90058291 1 13.27414 21.72585856 2 25.03124 40.96876186 1 12.13636 19.86364211 1 11.7571 19.2429033 2 21.23863 34.7613737 2 20.4801 33.51989606 1 10.24005 16.75994803 1 9.48153 15.5184704 1 9.48153 15.5184704 1 9.48153 15.5184704 1 9.48153 15.5184704 1 8.723007 14.27699277 4 34.13351 55.86649344 2 16.68749 27.3125079 1 8.343746 13.65625395 8 64.85366 106.1463375 20 160.8067 263.193258 1 7.964485 13.03551514 1 7.585224 12.41477632 24 180.1491 294.8509376 2 14.41192 23.58807501 2 14.41192 23.58807501 2 14.41192 23.58807501 1 7.205962 11.7940375 1 7.205962 11.7940375 1 7.205962 11.7940375 3 21.23863 34.7613737 2 14.03266 22.96733619 2 13.6534 22.34659738 1 6.826701 11.17329869 1 6.826701 11.17329869 1 6.826701 11.17329869 1 6.826701 11.17329869 4 26.92754 44.07245593 3 20.10084 32.89915725