Made-to-measure Qur’anic quotations: the incomplete verses of al-Qaeda in the Islamic Maghreb

Made-to-measure Qur’anic quotations: the incomplete verses of al-Qaeda in the Islamic Maghreb

Original version in Spanish: Literalidad coránica a medida: las aleyas incompletas de al-Qaeda en el Magreb Islámico


More than half the Qur’anic quotations used by al-Qaeda in the Islamic Maghreb (AQIM) in its propaganda are faithful reproductions of the text but incomplete or truncated.


A digitalised corpus comprising more than 200 official documents released by al-Qaeda in the Islamic Maghreb (AQIM) between 2004 –when it still styled itself the Salafist Group for Preaching and Combat (GSPC)– and 2017 enables an evidence-based answer to be given to the question of how religious its rhetoric really is. Moreover, by systematically collating all the Qur’anic quotes that have been recorded –more than 1,200– it emerges that, although the organisation never alters or modifies the Qur’an, enabling all accusations of the sacred texts being manipulated to be definitively rejected, there is a notable use of tailor-made quotation. Despite being unrestricted in terms of both time and space, on more than half of the occasions when AQIM has quoted the Qur’an in its propaganda over the course of the last 14 years – 52.42% of the total – it has done so using incomplete or shortened quotations.


The literal and decontextualised interpretation made of Islam’s sacred texts by the various terrorist organisations adhering to jihadist Salafism is well known, but to what extent is their rhetoric purely religious? Moreover, with the goal of discrediting and undermining the legitimacy of such organisations, it is still argued with a certain frequency that they manipulate and distort religious doctrine. Can such an assertion be verified? If so, bearing in mind the consensual nature regarding the sacredness of the main sources of revelation, it would be relatively simple to devise strategies aimed at exploiting such distortions.

By systematically mapping the more than 1,200 Qur’anic quotations used by al-Qaeda in the Islamic Maghreb (AQIM) between 2004 –when it still styled itself the Salafist Group for Preaching and Combat (GSPC in its French acronym)–1 and 2017, this analysis seeks to throw light on some aspects linked to the use of the sacred texts by the terrorist organisation. To this end an extensive monolingual, diachronic and digitalised corpus has been created covering a total of 203 official records released by the group.2 Among these there are 179 audiovisual recordings (more than 65 hours of audio and video), 20 text documents including magazines, books, articles and essays (approximately 800 pages of text) and the complete transcripts of four interviews given by prominent members of the organisation (see Figure 1). The dataset that has been compiled includes all the official audiovisual recordings3 published by the terrorist organisation in the aforementioned timeframe, as well as other official written documents whose content has been judged to be relevant by the author.4 Brief communications, used mainly to claim authorship of successful terrorist operations, panegyrics, opinion articles and other secondary content published on platforms such as Ifriqīyā al-Muslima were not included in the dataset.Figure 1. Annual distribution of the material used to create the corpus of Qur’anic quotations found in AQIM propaganda, 2004-17

YearAudios and videosMagazinesBooks and essaysInterviews
201521 (37)2
201710 (26)3

Notes: (1) the difference in the number of documents for 2015 arises from the inclusion in brackets of all the documents associated with the ‘غزوة أحد في القرآن ’ series, a collection of lessons delivered by Abu Hassan Rashid al-Bulaydi; (2) the difference in the number of documents for 2017 arises from the inclusion in brackets of all the documents associated with the ‘شرح كتاب الصيام من بلوغ المرام ’ series, a collection of lessons delivered by Abu Hassan Rashid al-Bulaydi.

Specialised corpuses, such as the one used here, focus on a particular variety or register of language, in this case the jihadist Salafist rhetoric of AQIM. Up until now, the few attempts to analyse jihadist Salafist rhetoric using a quantitative approach have run up against an obstacle that is difficult to overcome, namely the technical challenges inherent in Arabic. Such attempts are outnumbered by various qualitative studies and analyses, mainly since languages other than Arabic started to appear in Islamic State propaganda. However, without wanting to delve too deeply into the debate, Arabic continues to be not only the language used in the bulk of jihadist propaganda but also, of course, the cornerstone on which all ideological and doctrinal debate in this current of thought is founded and erected.

The majority of studies aiming to analyse jihadist rhetoric or narrative, in one way or another, do so on the basis of publications such as DābiqRumiyah and Inspire. These publications, although they have greater impact in the West, have a specific target audience, and this, while they adhere the basic principles of jihadist Salafist ideology, determines their nature and conditions the language used. Moreover, the translations of documents composed in Arabic, including those translated in-houseby the terrorist organisations themselves, show a greater degree of bias given that the process of translating jihadist propaganda is not systematic.

Meanwhile, there are a significant number of studies that use different perspectives to focus on the operations and relations of the militants –mainly munāṣirūn, ‘fanboys’ and official propaganda platforms– on such social media as Twitter and Facebook. This type of research –such as that conducted by J.M. Berger and Charlie Winter, among others–5 has cast light on the way the propaganda dissemination apparatus and the broadcasting of its message on social media works, as well as the number of militants active on the Internet and the central subject matter of each individualised audiovisual product. The present analysis, on the other hand, uses a quantitative approach to focus mainly on the way AQIM uses the Qur’an.

When AQIM quotes the Qur’an

Offering an evidence-based answer to the question of how religious jihadist rhetoric is represents a genuine challenge. The corpus assembled for this analysis is made up of more than 810,000 words; it has been codified in such a way as to enable all the Qur’anic quotations to be extracted from the text for their subsequent classification, focusing on any criteria that are specified. Thus, exactly 1,219 Qur’anic quotations are found in the 203 documents that have been included, but 932 if only those quotations extracted from documents published before the organisation started issuing its publications under the current AQIM name are considered. Although it may not be the most exact way of performing the calculation, these 1,219 quotations account for a total of 27,367 words, or 3.37% of the total corpus. Although they do not form part of the present analysis, if quotations of Hadiths –accounting for 7,014 words of the total corpus– are added to the calculation, a total of 34,381 words would be obtained. Thus, 4.24% of the total AQIM discourse consists entirely of quotations from the Qur’an and the Sunnah, the two main sources of revelation and the fundamental core of Islam.

That said, the purely religious discourse cannot be reduced solely to the Qur’anic quotations. Although Qur’anic quotations are sometimes introduced to convey a certain patina of religiosity to the discourse in order to justify a particular act or stance, depending on the document, it is normal for the Qur’anic quotations to be accompanied by a contextualised religious explanation. This exercise, which sometimes takes the form of exegesis and is sometimes simply didactic or clarificatory, constitutes a fundamental element of AQIM communication. Using the coding system described above, heat maps such as the one shown below were created to represent the relative weight of the group’s strictly religious discourse in the most faithful way possible.

The outcome of mapping all the documents in the corpus offers disparate results depending on the type of document in question, but they confirm that AQIM’s strictly religious discourse does indeed account for a significant fraction of the total. The results fluctuate between 2% and 4% on average for those documents aimed at publicising military capabilities –for example the bulk of the ‘ظلال السيوف ’ series– and more than 50% for documents of a doctrinal or ideological nature. Reducing it all purely to figures and percentages, AQIM quotes the Qur’an once every 665 words, which is to say once in slightly more than every page of running text, and although the distribution is quite some way from being uniform, each mapped document has an average of six Qur’anic quotations.

AQIM’s incomplete āyāt

One fact emerges in a particularly striking way after conducting an exhaustive comparison of all the Qur’anic quotations employed by AQIM in its media output.6 More than half the Qur’anic quotations to which AQIM refers in the mapped documents are partial or incomplete quotations, accounting for 52.42% of the total. Using systematic comparison of the sample of the 1,219 Qur’anic quotations assembled in the corpus with the Qur’an it emerges that 639 quotations are incomplete. This does not mean that the content of the quotations was manipulated (taḥrīf), just that the quotation, whether it was from a single āyah or a set, is incomplete.

As seen in Figure 4, which shows the distribution of the mapped Qur’anic quotations by year, the tendency to cite incomplete āyātconstitutes the norm, with 64.07% being the highest percentage of incomplete quotations used, in 2004, and 35.96% the lowest, in 2010. It should also be noted that the percentage of incomplete quotations fell below 50% of the total in only three years: 2007, 2010 and 2016.

Reciting incomplete āyāt or verses of the Qur’an is a habitual practice in Islam. Given that the prohibition of this practice does not find support either in the Qur’an or in the Sunnah, the basic principle accepted by the various schools of jurisprudence suggests that it is permitted to quote incomplete āyāt on condition that the meaning is complete. Many āyāt, especially the longest ones such as Qur’an 2:282 –the longest in the Qur’an– encompass various complete and independent ideas and meanings; it is not uncommon therefore to come across incomplete quotations. Indeed, this practice receives a certain amount of support, for example in the following Hadith:

‘Abdullah ibn Masa’ud recalled: the Messenger of Allah, ﷺ , said: “Whoever recites a letter from the Book of Allah will receive one good deed as ten good deeds like it. I do not say that Alif-Lām-Mīm7 is one letter, but rather Alif is a letter, Lām is a letter and Mīm is a letter”.’ Hadith classified as ṣaḥīḥ, or authentic, Sunan al-Tirmidhī 2910.8

Moreover, the debate about how and to what extent the Qur’an should be quoted, whether in the context of prayer or simply as an authoritative argument in rhetoric, is covered by the hermeneutic Islamic literature, as the following extract demonstrates:

‘As far as reading the Qur’an is concerned, apart from the obligatory prayers, according to Abu Hanifa, it should start with at least one āyah, even if it is short. Ibn Abbas is of the same opinion. He said: “recite what you can of the Qur’an, because nothing in the Qur’an is insignificant”. On the other hand, Abu Yussuf said: “recitation should be no less than a long āyah, such as the Throne Verse [Qur’an 02:255], or at least three short verses, because less than this would go against custom and would not demonstrate how miraculous the Qur’an is”.’9

The Qur’an, which can literally be translated as ‘recitation’, is a complex book and it is of primordial importance to understand its nature. It is not a chronological compilation designed to tell a story, as might be said of the book of Genesis, and therefore it should not be viewed as a sequential narrative. Its verses, or āyāt, are not standard either in length or in meter and according to Muslim tradition the start and end of each one do not relate to the arbitrary decision of men but the dictation of God. Thus, as a general rule, each verse –and sometimes a set of verses– deals with a particular subject; dividing them therefore may lead to the modification of the global meaning and limit or distort the context of the literal revelation.

Responding to whether it is permitted to split a verse into various parts, the erudite Saudi Salafist Muhammad Salah Al-Munajjid (born in Aleppo) stated that it was possible unless it should lead to an inappropriate meaning, but stipulated that it would be preferable to complete the quotation.10 The ulema,11 currently in prison in Saudi Arabia, based his response on the fact that the earliest generations of believers viewed reciting a complete Surah as mustaḥab,12 without stopping, so that it would be appropriate to extend the practice to the āyāt too. He based his conclusion on the following hadith:

‘Jabir ibn Abdullah recalled: we proceeded in the company of the Messenger of Allah, ﷺ , for the battle of Dhat ar-Riqa. One of the Muslims killed the wife of one of the unbelievers. He (the husband of the woman killed) took an oath saying: “I shall not rest in peace until I kill one of the companions of Muhammad”. He went out following the footsteps of the Prophet, ﷺ . The Prophet, ﷺ , encamped at a certain place. He said: “Who will keep watch on us?”. A person from the Muhājirūn and another from the Anṣār responded. He said: “Go to the mouth of the mountain-pass”. When they went to the mouth of the mountain-pass, the man from the Muhājirūn lay down while the man from the Anṣār stood praying. The man (enemy) came to them. When he saw the person he realised that he was the watchman of the Muslims. He shot him with an arrow and hit the target. But he(took the arrow out and) threw it away. He (the enemy) then shot three arrows. Then he (the Muslim) bowed and prostrated and awoke his companion. When he (the enemy) perceived that they (the Muslims) had become aware of his presence, he ran away. When the man from the Muhājirūn saw the (man from the Anṣār) bleeding, he asked him: “Glory be to Allah! Why did you not wake me up the first time when he shot at you?”. He replied: “I was busy reciting a chapter of the Qur’an. I did not like to leave it”.’ Hadith classified as ḥasan, Sunan Abu Dawud.13

Moreover, in his work Al-Itqān fī ʻUlūm al-Qurʼān, considered a fundamental linguistic and stylistic tool for understanding the meanings of the Qur’an, Jalal ad-Din al-Suyuti, delving into the correct pronunciation when it came to reciting the Qur’an, referred to the work of Uthman al-Dani, a linguist and exegete from Al-Andalus. The latter quoted a Hadith of al-Hakim,14 who put into the mouth of Zayd ibn Thabit, scribe of the Prophet according to tradition and one of the Anṣār, the following words: ‘The Qur’an was revealed in order to be recited in full’.15

The biased and tendentious use that is made of religion by the jihadist Salafist organisations is well known and, while it is not the aim of this paper to delve into the exegesis of the sacred texts, it does seek to open the way to the construction of counternarratives. Unlike other groups, such as imams, TV preachers, etc., AQIM –like other terrorist organisations– is totally unrestricted in terms of time and space when it comes to creating its message; instead it has the tools needed to design its communication strategy and the content of its message as it sees fit. It does not broadcast live and therefore is not subjected to scrutiny or debate with third parties in real time. Yet even without restrictions when it comes to conveying its message, it chooses to use partial quotations on more than half the occasions when it invokes the Qur’an.

The two verses most frequently used by AQIM, Qur’an 02:217 and Qur’an 08:36, provide an example of the above phenomenon. There is no particular controversy surrounding these āyāt and there is a broad consensus (ijmāʿ) about their meaning at the core of the Ummah. Verse 217 of the Surah of the Cow is the most used by AQIM: 22 times in all according to the data extracted from the corpus, all of them in an incomplete way. The underlined section of the translation shown below does not appear in any of the 22 quotations used by AQIM in the documents from which the corpus is compiled. Moreover, without going into the matter in detail, it is worth mentioning that āyah 36 of the Surah of Repentance [Qur’an 09:36], a verse that in the opinion of several mufassirūn abrogates Qur’an 02:217 and whose content restricts the possibilities of fighting the infidels during the sacred months only to occasions when the Muslims are attacked first, appears on just three occasions in the AQIM discourse, none of them unedited.

[Qur’an 02:217]: ‘They will question thee concerning the holy month, and fighting in it. Say: “Fighting in it is a heinous thing, but to bar from God’s way, and disbelief in Him, and the Holy Mosque, and to expel its people from it – that is more heinous in God’s sight; and persecution is more heinous than slaying”. They will not cease to fight with you, till they turn you from your religion, if they are able; and whosoever of you turns from his religion, and dies disbelieving – their works have failed in this world and the next; those are the inhabitants of the Fire; therein they shall dwell forever.’16

Verse 39 of the Al-Anfāl Surah (The Spoils of War) is the second most-used verse by AQIM, quoted 21 times in all according to the corpus data, 19 of them in an incomplete way. The underlined section of the translation shown below forms part of the only two complete quotations of this verse that are to be found in the corpus.

[Qur’an 08:39]: ‘Fight them, till there is no persecution and the religion is God’s entirely; then if they give over, surely God sees the things they do.’17

By means of a systematic collation of the Qur’anic quotations used by AQIM over a 14-year period it may be stated that they do not resort to manipulating the Qur’an or altering its content (taḥrīf) at any time. What have been noted however, albeit only on a handful of occasions, are unintentional errors, failures to observe conventions –mainly in the pronunciation of certain terms, in breach of the rules of the tajwid18 and minor ellipses, all them errors that are common in oral discourse.

However, as became clear at the start of this section, such a systematic comparison does enable a figure to be placed on the number of incomplete quotations; and this, for an organisation that finds the raison d’être for its activities, according to its own interpretation, in the word of God, is undeniably high. It adheres strictly to the text in a literal sense, but in a made-to-measure way. This finding, combined with the manifold possibilities inherent in a contextualised study of each Qur’anic quotation and the way their use develops over time, represents a significant step enabling more in-depth study to be conducted into the religious discourse emanating from jihadist Salafism.


Despite the growing institutional interest in the development of counternarrative tools and the consequent academic steps that have been made in the same direction, the rewards that have been reaped, in light of the militant mobilisation witnessed in recent years, cannot be described as anything other than insufficient. Part of the problem consists in the fact that, in order to create effective counternarratives, it is necessary to have in-depth knowledge of the jihadist Salafist narrative, and this is a field in which, although there are numerous qualitative studies, quantitative or mixed-approach research has barely been conducted, particularly if we confine ourselves to Arabic as the main language for conveying the jihadist Salafist message.

Systematising the analysis of religious discourse by employing automated tools not only enables irrefutable data to be obtained, it should also facilitate the creation of counternarrative strategies and tools of greater efficacy as well as making it possible to finetune those that already exist. The use of analytical techniques based on empirical data makes it possible to produce a reliable picture of the reality of the discourse, contextualise the findings and, consequently, provide a much more precise study of the evolution of various organisations’ narrative over a particular period of time.

This analysis, by means of data extracted from a monolingual, diachronic corpus that is representative of the AQIM discourse –thereby ensuring the consistency of the results obtained– enables the notion that the organisation manipulates or alters the content of the Qur’an to be categorically dismissed. This is simply and plainly false; we may be able to agree that the extremely strict interpretations that are embraced are biased, unjustified, decontextualised and lack the approval of the majority of Muslims, but on no occasion do they breach the limits of exegesis historically accepted by the bulk of Sunni Islam. They flirt with the boundaries of the rules, as it were, but remain within the rules. It goes without saying that the conclusions drawn here cannot necessarily be extrapolated to other jihadist organisations, although they prepare the ground for future comparative studies.

Secondly, and this undoubtedly constitutes the most important conclusion of this study, it has been found that more than half the Qur’anic quotations used by AQIM in its propaganda output during the period under review are incomplete. As mentioned, the biased reading and use that jihadist organisations make of religion is well known, favouring certain passages over the rest, thereby distorting the overall significance of the revealed message. Apart from the overrepresentation of certain verses compared to others and the reliance on the most belligerent form of exegesis, however, one aspect stands out after quantitatively analysing how AQIM quotes the Qur’an and comparing the results with the original source. The terrorist organisation quotes the Qur’an rigorously, but in an incomplete way, shortening more than half the Qur’anic quotations it uses, obviously in a way that suits its agenda: the quotations are faithful, but made-to-measure.

Future contextualised analysis of the way these incomplete quotations are used, as well as research into how their use changes over time, will undoubtedly serve to continue not only in the further acquisition of in-depth knowledge of the religious discourse of jihadist Salafism, but also in making progress towards the creation of new counternarrative strategies and the refinement of those that already exist.

Sergio Altuna Galán
Associate Analyst, Violent Radicalisation and Global Terrorism Programme, Elcano Royal Institute | @wellesbien

1 The year 2004 was chosen as the starting point of the dataset precisely to rule out the possibility of the absorption of the Salafist Group for Preaching and Combat involving substantial changes in the use of Qur’anic quotations. The same tear also represents the start of Abdelmalek Droukdel’s time as the head of the organisation, which, in January 2007, having merged with al-Qaeda, came to be known as al-Qaeda in the Islamic Maghreb, but still under his leadership.

2 A corpus can be defined as a database, an extensive collection of authentic texts that have been compiled and digitalised following a specific set of criteria enabling it to be used as a representative sample of a linguistic reality. See T. McEnery, R. Xiao & Y. Tono (2006), Corpus-Based Language Studies. An Advanced Resource Book, Routledge, London & New York.

3 All the audiovisual documents published by the Media Committee of the Salafist Group for Preaching and Combat (اللجنة الاعلامية للجماعة السلفية للدعوة والقتال ), the Media Committee of the al-Qaeda in the Islamic Maghreb (اللجنة الاعلامية لتنظيم القاعدة ببلاد النغرب الإسلامي ) and the al-Andalus Foundation.

4 The written documents whose inclusion in the corpus has been judged as relevant are all those official documents that include a minimal ideological and doctrinal component.

5 See J.M. Berger & J. Morgan (2015), The ISIS Twitter Census Defining and Describing the Population of ISIS Supporters on Twitter, The Brookings Project on US Relations with the Islamic World, The Brookings Institution; and C. Winter (2015), The Virtual ‘Caliphate’: Understanding Islamic State’s Propaganda Strategy, Quilliam International.

6 For the purposes of comparing data, the author used the materials made available through the Qur’an Digitalisation project run by the King Saud University in Riyadh, Saudi Arabia.

7 [Qur’an 02:01].

8 2910 سنن الترمذي، كتاب فضائل القرآن، حديث رقم

9 المحيط البرهاني في الفقه النعماني، كتاب الطهارات – الصلاة، ص. 298. دار الكتب العلمية 2004

10 Islam Q&A, رقم السؤال 206946 .

11 It would be more correct to use ʿAlīm, but in English it is acceptable to use ulema in the singular.

12 Recommended, favoured.

13 سنن أبي داود، كتاب الطهارة، حديث 198

14 Abu Abd-Allah Muhammad ibn Abd-Allah al-Hakim al-Nishapuri, the Persian ulema and respected compiler of Hadiths.

15 جلال الدين السيوطي، الإتقان في علوم القرآن، ص. 129.

16 [Qur’an 02:217]. Translation from Arthur John Arberry.

17 [Qur’an 08:39]. Translation from Arthur John Arberry.

18 Rules governing the recitation of the Qur’an, or tilāwa.