Al-ḥamdu li’llāhi fātiri’l-samāwāti wa’l-arḍi jā‘ili’l-malā’ikati rusulan
ūlā ajniḥatin Mathnā wa-thulātha wa-rubā‘a yazīdu fī’l-khalqi mā yashā’u
inna’llāha ‘alā kulli shay’in qadīr. Q. 35:1
Praise belongs to God, Originator of the heavens and the earth, who appointed the angels to be messengers,
having wings two, three and four, increasing creation as He wills.
Surely God is powerful.
The opening of Q. 35:1 (Surat al-Mala’ika or Surat al-Faṭir) attests to the creative power of God and describes the angels as winged messengers, the only aya (verse) where angels are portrayed in this way in the whole of the Qur’an. It is an important aya, as it is one of only a few which describe the relationships between God, humans and angels. However, two of the most significant words in this aya are often passed over without comment or consideration by modern translators and commentators alike: malak (mala’ika) and faṭir. These two words are usually given the translations ‘angel’ or ‘messenger’ and ‘Creator’ respectively. Western philologists believe that both of these words are of foreign origin and both are normally thought to be lexical borrowings from Ethiopic. The precise meanings of malak and faṭir are not the only words in this aya that appear to have posed problems. The three distributive adjectives found in the aya (mathna, thulath and ruba‘) tend in modern translations to be given interpretations not found in the classical exegeses of the aya. This article will therefore discuss whether these three adjectives should be interpreted following the modern translators or the classical exegetes, as well as discussing the implications of the different interpretations.
To this end, this paper will provide a detailed analysis of these three cruces, drawing on a number of medieval Islamic exegeses, focusing on those of al-Tabari (d. 310 AH / 923 CE), al-Zamakhshari (d. 538 AH / 1144 CE), al-Razi (d. 607 AH / 1210 CE), al-Qurtubi (d. 671 AH / 1273 CE ), al-Baydawi (d. 685 AH / 1286 CE) and al-Suyuti (d. 911 AH / 1505 CE). The main aim of this discussion is to show that a close reading of this aya is crucial for gaining an understanding of the relationships between God, His angels and His earthly creations. This aya has often been neglected, but as the only verse to give a description of both the physical appearance of the angels and their function, it remains a principal source for understanding their place in Qur’anic theology.
The Meaning of the Word malak
Q. 35:1 contains one of the numerous occurrences of malak (mala’ika); however, the origin and precise meaning of this word are widely disputed in both classical and modern scholarship. There are two main difficulties posed by the word malak: firstly, the root of the word; and secondly, the (Semitic) origin of the word. Modern scholarship has been concerned with both of these problems, whereas the classical Islamic lexicographers and exegetes were, on the whole, solely concerned with the question of the word’s root.
The classical Islamic lexicographical tradition places the word malak under the roots ’-l-k, l-’-k or m-l-k. However, the mim is not usually considered part of the trilateral root, but rather as one of the two other roots in a nominal form. Al-Tabari believes that the root of malak (also with the form mal’ak) comes from the root ’-l-k, explaining (i) the dropping of the glottal stop in the singular (i.e. malak) and (ii) the metathesis (i.e. ma’lak to mal’ak) with other examples where similar phonological and morphological changes occur. However, there is no strict consensus among the exegetes about whether the original root is ’-l-k or l-’-k: as al-Tabari notes, ‘some say that mal’ak is the maf‘al form
The general consensus in modern scholarship has followed that of the classical tradition, tending to take the root as ’-l-k, although some scholars have listed the noun under l-’-k, which again shows the ambiguity of the word’s root. The quotation from al-Tabari above shows clearly that the Arabic lexicographical tradition recognised two roots with subtly different meanings; although, as both have the same basic meaning (i.e. ‘to send a message’ or ‘to send a messenger’), there was no real need to go into further detail. Had there been two roots with slightly different meanings, there would have been a need to explore the theological implications of the use of each word. However, it should be noted that al-Tabari’s explanation of the roots may be an attempt to minimise the differences in the selection of one of the two roots – if the two roots have the same basic meaning, the choice between the two is largely irrelevant.
During the early-nineteenth and mid-twentieth centuries, there was particular interest in Europe in both comparative Semitic philology and Islam; it was during this period that much of the modern work on comparative philology was undertaken, and words of non-Arabic origin in the Qur’an were studied in great detail; these studies included work on the origins of malak. Languages as diverse as Canaanite, Hebrew, Aramaic, Ethiopic and Syriac have all been suggested as the potential Semitic origin of the word. The Ethiopic mal’ak is usually given as the most likely source of the Arabic malak, on account of the similarity between the Arabic broken plural (mala’ik [a] and the Ethiopic broken plural mala’ekt. Dillmann gives the root as l-’-k and mal’ak the meanings nuntius and legatus but notes that ‘in specie legatus Dei i.e. angelus, (Luc. 2,13; Job 1,6; Hebr. 1,4–7) …’ What is important here is that this shows that Ethiopic understands the root l-’-k and its noun mal’ak in terms of the action of sending a message. Secondly, the messenger is also not necessarily understood to be angelic, as the divine origin of the messenger can only be understood in context. This understanding of the semantic value of malak is common to most Semitic languages. The root l-’-k or ’-l-k is only found in the nominal form in the Hebrew Bible, and when it was translated into Greek in the Septuagint, the translation angelos was used for messengers of both divine and non-divine origin, which can also be found in New Testament Greek. In Latin, however, a distinction was made between human and divine messengers – this was firmly established in Jerome’s Vulgate which uses angelus for divine emissaries and nuntius for human messengers.
Surprisingly, the Qur’anic and classical Arabic usage of malak has far more in common with the Latin use of angelus than it does with the Hebrew, Ethiopic and Greek uses of mal’ak, mal’ak and angelos respectively. In Q. 35:1 the noun mala’ika is glossed clearly with rusul, the more common Arabic word for messenger. The aya reads: al-hamdu li’llah … ja‘ilu’l-mala’ikati rusulan. Here the Qur’an appears to be explaining the meaning of the term malak, suggesting that the notion of ‘sending a messenger or a message’ was not necessarily understood by the audience of the Qur’an as being a nominal form of the verbal root (be it l-’-k or ’-l-k). If the word were not understood to be a nominal form of a verb, the way in which malak is conceived is quite different. It means that in the Qur’an, the malak is no longer a messenger in the basic sense of the word (i.e. used of both human and divine messengers) as in Biblical Hebrew, Ethiopic and other Semitic languages, but it is a purely celestial being, occasionally entrusted with delivering messages from God to His creations, amongst many other things.
Although scholars such as Paul Bonsechi have shown that malak is a perfectly acceptable Arabic form, this does not necessarily mean that the word has the same semantic value as its cognates in other Semitic languages. The evidence suggests otherwise. In the Qur’an itself angels frequently appear without performing the function of a messenger: e.g. in the process of death; during the adoration of Adam and the fall of Iblis; and eschatological passages, to give just a few examples. A survey of the ahadith containing malak and mala’ika (which gives a useful guide to the semantic value of words in classical Arabic in general) supports the idea that malak was understood only in the sense of a divine messenger. In the majority of the ahadith malak is used: (i) with verbs such as nazala (which in context cannot refer to human messengers); (ii) in reference to Muhammad’s experience of revelation; or (iii) to events after death. All these indicate that in Arabic there is a break from the Semitic sense of mal’ak referring to both human and divine messages. There is a particularly strong break from the Ethiopic understanding of the word (which is important as Ethiopic is usually seen as the origin of the Arabic word), since Ethiopic makes frequent use of the verbal form, whereas the verbal use of ’-l-k or l-’-k is more limited in other Semitic languages such as Biblical Hebrew.
In the Qur’an there are three main references to angels in which the word malak is juxtaposed with rasul. If these two words are (as the classical lexicographers and some modern philologists say) synonymous, why is it necessary for the angel to be described as a rasul, rather than in terms of the base meaning of the root? The extreme rarity of the verbal form in classical Arabic would seem to suggest that it was, at the very least, not common at all. This is corroborated further by Q. 22:75, Allahu yastafi mina’l-mala’ikati rusulan wa-mina’l-nas (God chooses of the angels messengers and of mankind) in which it is said that God selects his rusul from mala’ika and nas: this can only suggest that malak is not understood in terms of messenger on its own merit, otherwise this aya would mean that God selects his messengers from messengers and people. The annunciation of the birth of Jesus to Mary is similar: when the Spirit in Q. 19:17 is sent to Mary from God (fa-arsalna ilayha ruhana [then We sent to her Our Spirit]), the root r-s-l is used and not ’-l-k. If the root ’-l-k were understood in Arabic to refer to the sending of messengers by God, the root ’-l-k might well be expected here; but it is not used, and the more standard arsala is chosen.
Whilst discussing Q. 81:19, innahu la-qawlu rasulin karim (truly, this is the word of a noble Messenger), in which a similar use of rasul is found, Montgomery Watt states that ‘in 81:19 it is used of an angel bearing a message to Muhammad (which is especially appropriate, since angel comes from the Greek word for “messenger”)’. This would seem to be a rather hasty remark, as Watt merely assumes that the word malak was understood to be a messenger. This assumption is also made by Gisele Webb, who simply says ‘malak (mala’ika) means “messenger”.’ Is this necessarily the case? The confusion about the triliteral root of malak – even amongst classical Islamic lexicographers and exegetes – would suggest that there is uncertainty over the exact meaning and use of the word. If malak were understood to mean ‘messenger’, why is malak glossed with rasul so often? And why do mala’ika not always perform the tasks that a malak should then be doing? There is a simple solution to the problem.
The Qur’an uses the word malak in much the same was as Latin speakers differentiated between a nuntius and an angelus. In the Qur’an the malak is understood to be a creature that is part of the divine world. Toshihiko Izutsu has argued the same, briefly, in his God and Man in the Koran. The noun malak does not mean ‘messenger’ and the verbal form seems to have largely fallen into obscurity – which is why the word is frequently glossed with rasul to explain the word in situations where the malak is delivering a message. Therefore, the correct translation in English is ‘angel’, but not ‘messenger’. English has conveniently inherited the term from Latin, and so makes a distinction between ‘messenger’ and ‘angel’. English has lost the primary sense of the original Greek angelos, in a similar way to that in which, according to this theory, Arabic has lost the primary sense of the original Semitic meaning of the root. Were the Arabic malak translated into Greek, Hebrew or any other similar Semitic language, much more caution would need to be taken in the selection of an appropriate translation.
The Meaning of faṭir
The root f-ṭ-r [Proto-Semitic p-ṭ-r] occurs ten times in the Qur’an: six of these usages are in the form of the active participle faṭir and the remaining four are: fiṭra (Q. 30:30), faṭara (Q. 30:30), fuṭur (Q. 67:3) and munfaṭir (Q. 73:18). Faṭir is normally translated as either ‘the Originator’ (e.g. Arberry, Yusuf Ali, Behbudi and Turner, Zidan) or ‘the Creator’ (e.g. Bell, Dawood, Khan, Létumy, Montet, Rodwell) . However, there is some debate about what the word means in both classical and modern scholarship. In Islamic exegesis, the word is normally glossed with khaliq (‘Creator’) or mubtadi’ (‘Originator’), but these are not the only glosses provided in the exegetical tradition. It is clear that the formula ‘faṭir al-samawat wa’l-ard’ is an attribute of God, and the English ‘Creator of the Heavens and the Earth’ would appear to be an obvious translation, so much so that were faṭir a hapax legemenon, ‘Creator’ would be the natural choice. However, the precise meaning of the root is important as it may affect the interpretation of the aya; and, furthermore, a better understanding of the semantic value of the word will aid translation, particularly, in view of the fact that faṭir is a relatively uncommon word in the Qur’an. Afnan H. Fatani argues that the relative obscurity of a word in a source language should be reflected in the target language. When considering the translation of falaq in Sura 113, he comments, ‘the translation process falaq-daybreak is more successful than falaq-dawn because it succeeds in reproducing some of the non-standardness of the SL [Source Language] lexeme’. The translation ‘Creator’ for faṭir does not reflect the relative rarity of its use in the Qur’an, but is ‘Originator’ the best option for an alternative to ‘Creator’?
The Proto-Semitic root p-ṭ-r has a wide range of semantic values. The root is attested in Akkadian, Ugaritic, Biblical Hebrew, Ethiopic, Aramaic and Syriac. Unfortunately, like malak, the root has not been found in any of the Arabian inscriptions which could have been useful. Akkadian, Ugaritic and Biblical Hebrew use p-ṭ-r to denote the physical action of ‘splitting’. In Biblical Hebrew the same root is also used in 1 Kings 16:18, 29 and 32 in reference to the construction of the Temple carvings. The use of p-ṭ-r in these circumstances indicates the physical action of carving. Another important meaning of p-ṭ-r is the notion of ‘letting go’: this occurs in Akkadian, Syriac, the Hebrew of the Dead Sea Scrolls, and in Mishnaic Hebrew. Mishnaic Hebrew includes a further extension to this meaning using the root as ‘to divorce’. Akkadian and Biblical Hebrew also use the root in expressions concerning ransoming, and both Biblical and Mishnaic Hebrew use it in the specific cultic expression peter rehem, which refers to the first-born of an animal (i.e. the animals for use in sacrifice). This is related to the notion of separation seen above: the animal is distinguished and set aside from the rest of the animal’s offspring.
Although the Proto-Semitic p-ṭ-r has a number of different meanings in these various languages, they all contain the basic notion of separation of one group from another, be it in the sense of leaving, carving or divorce. This would seem to suggest that the translation of faṭir as ‘Creator’ or ‘Originator’ does not convey the basic sense of the word. Many commentators have cited Ethiopic as the origin of the Arabic root, but Ethiopic does not, however, carry the Northwest Semitic association with separation, rather it is the word used for ‘creation’ when God is the subject. This means that there are two different possibilities for the meaning of faṭir: either meaning ‘to separate’ from the Northwest Semitic language group or ‘to create’ from Ethiopic. The question, then, is from which language group the Arabic use of faṭir is derived.
The two main Islamic lexicographical works, Murtadha al-Zabidi’s Taj al-‘arus and Ibn Manzur’s Lisan al-‘Arab, both cite the same hadith from Ibn ‘Abbas to gloss faṭir: ‘I did not know what faṭir al-samawati wa’l-ard meant, until I came across two Arabs working on a well. One of them said, “I have ‘faṭara’ed it”, meaning: I have started to dig it’ (‘ma kuntu adri ma faṭir al-samawati wa’l-ardi hatta atani A‘rabiyyan yakhtasimani fi bi’r fa-qala ahaduhuma ana fatartuha ayy ana ibtada’tu hafraha’). The use of this hadith shows that Ibn Manzur and al-Zabidi favour the gloss ibtada’a. This is confirmed by both of the lexicographers with a second hadith attributed to Abu ‘Abbas: ‘I heard Ibn al-‘Arabi saying, “I am the first to ‘faṭara’ this”, meaning: the one who began it’ (‘annahu sami‘a Ibn al-‘Arabi yaqulu ana awwal man faṭara hadha ayy ibtada’ahu’). These two glosses of faṭir are also the most frequent glosses in the exegetical tradition, where the most frequent synonyms are khaliq and mubtadi’. In the majority of the tafsir collections, the most detail regarding the semantic meaning of the word is given at its first occurrence at Q. 6:14. For example, al-Tabari gives khaliq, mubtadi’ and mubtadi‘ as glosses for faṭir, but in his tafsir of the other five uses only khaliq is given as the gloss. Similar glosses are used by al-Baydawi, the Jalalayn, al-Zamakhshari and al-Qurtubi. Al-Tabari, al-Razi, al-Zamakhshari and al-Baydawi also make use of the Ibn ‘Abbas hadith mentioned above.
It is upon these common glosses found in the traditional exegetical material that translators of the Qur’an have often relied. The reason for this lies in a number of different factors. Firstly, the most common gloss for faṭir is khaliq and other synonyms; secondly, the concept of the word faṭir in the context of the formula faṭir al-samawat wa’l-ardh implies the act of creation; and lastly, there is a comparative Semitic use (i.e. Ethiopic) meaning ‘to create’ with God as the subject. These three factors have led to the widely used translation ‘Creator’, or ‘Originator’ for those translators who have wished to show a difference between khalaqa and faṭara. The evidence for ‘Creator’ and ‘Originator’ seen above seems compelling at first; however, these glosses do encounter some problems of their own. Firstly, the creation glosses (khaliq, mubtadi’ and mubtadi‘) are not the only glosses found in the exegetical tradition. Despite the fact that the end result of khalaqa and faṭara may be the same, the Qur’anic text may be describing a more dynamic process by using a root with a more specific meaning than khalaqa – i.e. standardising faṭara to the generic term khalaqa may destroy the subtle nuance of the word and the aya. Secondly, the Ethiopic derivation may make the semantic meaning of faṭir easier to understand, but in Ethiopic the root is used only of God and Arabic already has a verb of creation that is only applicable to God: bara’a. The final problem posed by the adoption of the creation glosses is that the use of the active participle in the set phrase faṭir al-samawat wa’l-ardh differs greatly from two other forms in the Qur’an: futur (Q. 67:3) and munfaṭir (Q. 73:18). In these two instances it is impossible to use a creation gloss.
Whilst the creation glosses are the most prevalent glosses in the exegetical literature, some of the commentators use glosses with meanings far closer to the Northwest Semitic concept of separation. Al-Razi, al-Baydawi and al-Qurtubi use the verb shaqqa as a gloss. The use of shaqqa is a marked departure from the creation glosses as it describes God’s creative action very differently. If a creation gloss is used, the aya reads ‘Praise be to God, the Creator of heaven and earth’. If shaqqa is used as a gloss, the aya will now read ‘Praise be to God, the One who divided the heaven and the earth’. This gloss does not simply describe God as the creator of the world, but describes a part of the creative process and the relationship between the heavens and the earth, something also noted by Arnaldez: ‘En réalité, le mot fâtir qui a bien le sens de Créateur, vient de la racine faṭara qui signifie «fendre», De là résulte une interprétation plus satisfaisante pour le but que se propose al-Râzî. Dieu, dit-il, a fendu le ciel pour faire descendre les esprits angéliques, et il a fendu la terre pour en faire sortir les corps des ressuscités.’ When the uses of the root in Q. 67:3, fa’rji‘i’lbasara hal tara min fuṭur (return thy gaze; seest thou any fissure?), and Q. 73:18, al-sama’u munfaṭirun bihi kana wa‘duhu maf‘ulan (whereby the heaven shall be split, and its promise shall be performed), are considered, the dividing gloss gains more credibility. These ayas describe the tearing and rupturing of the sky, hailing the end of time and impending judgement. There is a very strong consensus in the Islamic exegetical tradition regarding the meaning of the root f-ṭ-r in these two ayas. Al-Tabari, al-Baydawi, al-Razi, the Jalalayn, al-Qurtubi and al-Zamakhshari all use shaqqa as a gloss, and some also use sada‘a. The two uses of the root in Q. 30:30, fa-aqim wajhaka li’l-dini hanifan fitrata’llahi allati faṭara’l-nasa ‘alayha (so set thy face to the religion, a man of pure faith – God’s original on which He originated mankind), are more complicated, as neither the creation glosses nor the dividing glosses provide a truly suitable translation of the root. Whilst the creation glosses are used more frequently, it is possible for a dividing gloss to be used, as Q. 30:30 describes the distinction between those who have true faith (hanif) and those that do not. The aya could equally read, so set thy face to the true religion, a man of pure faith – God’s means of division by which He divided mankind. That is, the fitra signifies the notion that true religion is the means by which God separates mankind into believers and non-believers. The Qur’an could be highlighting the distinction between the two groups, alluding to the separation of mankind into winners and losers on the Day of Judgement. There is also further allusion to separation (using faraqa) in Q. 30:32, which highlights the importance of separation and division in this particular passage of the Qur’an.
The second problem with the use of creation glosses is the fact that Arabic makes a clear distinction with the use of faṭir as opposed to khaliq. In Ethiopic the same word is used exclusively for the action of God. However, the Qur’an does not only use faṭara to describe the process of creation; but, rather, it has a number of different verbs: khalaqa, ja‘ala, bara’a, to name but a few. In fact the Qur’an uses the verb khalaqa in association with heaven and earth some 33 times. Why does the Qur’an decide to use faṭir in this sense a mere six times? This use of faṭir must indicate some distinction between the meanings of khalaqa and faṭara. Al-Zamakhshari includes an interesting comment about the meaning of the root in his exegesis of Q. 67:3. As seen above, Q. 67:3 is one of the uses of faṭara which clearly indicates the notion of splitting and dividing. Al-Zamakhshari writes ‘and from it the camel’s tusk “faṭara”s just as it is said that it splits [shaqqa] and tears [bazala], meaning that it splits the flesh so that it can be seen’ (‘wa-minhu faṭara nab al-ba‘ir kama shaqqa wa-bazala wama ‘anahu shaqqa al-lahm fa-tala‘a’). The concept of the tusk breaking through and splitting the skin is a significant distance away from the Ethiopic understanding of the word, and has a lot more in common with the Northwestern Semitic understanding of the root. This, along with the Ibn ‘Abbas and Abu ‘Abbas hadith above, shows that the subject of the verb is not exclusively God, but can also be animals or humans. This distances, in this particular case, the relationship between Arabic and Ethiopic. If the root were adopted from Ethiopic it would be expected to be used solely of the creative action of God, and not of humans and animals. Secondly, it could be possible that the Qur’an adopted the entire formula faṭir al-samawati wa’l-ard from Ethiopic. However, if this were the case, it would seem unlikely that the verb was assimilated into the Arabic language in the way in which these ahadith suggest. Whilst it is possible to gloss faṭara with a creation gloss found in the Ibn ‘Abbas hadith, it is impossible to apply one to the tusk hadith.
The use of comparative Semitic philology thus seems to indicate that the readings of ‘Creator’ and ‘Originator’ for faṭir may not be the most suitable translations. Only in Ethiopic does faṭara convey the notion of creation, but the way in which the word is used in both Qur’anic and classical Arabic differs greatly from the Ethiopic usage. The internal evidence in the Qur’an itself also seems to suggest that faṭir should not necessarily be understood in terms of creation, but could be understood in terms of dividing and splitting. The Qur’anic text has specifically selected faṭir over other, more common, words for creation.
How Many Wings?
The statement in Q. 35:1 that angels were created with the function of being messengers (i.e. ja‘ili’l-mala’ikati rusulan) is followed by the description of their physical bodies:…ula ajnihatin mathna wa-thulatha wa-ruba‘a yazidu fi’l-khalqi ma yasha’u (… who appointed the angels to be messengers, having wings two, three and four, increasing creation as He wills). This is the only finer detail that the Qur’an provides about the wings of the angels themselves. Again, there are some exegetical difficulties provided by the Arabic text as the precise meaning of mathna wa-thulatha wa-ruba‘ is, to some extent, disputed between classical exegetes and contemporary translators of the Qur’an. In Islamic tradition the interpretation of the phrase yazidu fi’l-khalqi ma yasha’u also took Islamic angelic imagery in an interesting and slightly unexpected direction.
The three adjectives mathna wa-thulatha wa-ruba‘ are distributive adjectives. In Arabic the relatively rare distributive adjectives take two forms: either fu‘al or maf‘al, and both are used in the Qur’anic text. These types of adjectives are used to designate the numbers of different groups of substantives; for example, in the phrase ‘the men entered in threes’, thulath would be used. Returning to Q. 35:1, a more literal translation of this phrase would be: who have wings: two at a time, and three at a time, and a four at a time. This would, at first sight, appear to be relatively straightforward; however, the disputes in the exegeses, and more particularly in contemporary translations of the Qur’an, centre around the question as to whether this clause refers to the angels themselves (i.e. ‘angels have two, three or four wings’) or the angels’ wings (i.e. ‘each individual wing comes two at time, three at a time, four at time; making four, six or eight wings’). Translators of the Qur’an have tended to either suggest pairs of wings (e.g. Yusuf Ali, Rodwell, Montet) or leave the text ambiguous (e.g. Arberry, Bell, Khan, Léturmy, Zidan). Although Bell’s translation keeps the text’s ambiguities, in his commentary he notes that the wings are said to be ‘double, treble and fourfold, i.e. two, three or four pairs’. Only Muhammad Baqir Behbudi’s exegetical translation gives the explicit interpretation: some of them fly with two wings, some with three, others with four. However, it should be stressed that this is not a strict translation of the Qur’an.
Turning to the classical exegetes, there is evidence of some debate about the precise meaning of mathna, thulatha and ruba‘. Al-Tabari begins his exegesis of this clause by stating his belief, that the phrase refers to individual wings: ‘the possesors of wings: meaning angels; thus, amongst them, one [group] has two wings, and amongst them one [group] has three wings, and amongst them one [group] has four’ (‘ashab ajniha ya‘ni mala’ika fa-minhum man lahu ithnayn min al-ajniha wa-minhum man lahu thulath ajniha wa-minhum man lahu arba‘a’). He supports his position with a tradition from Qatada, which uses the dual janahan, making the interpretation very explicit: ‘He said: some of them have two wings (janahan), some of them three wings, some of them four’ (‘qala ba‘dahum lahu janahan wa-ba‘dahum thalatha ajniha wa-ba‘dahum lahu arba‘a’). This hadith is also cited directly by al-Qurtubi and he also gives another, slightly confusing, gloss for the phrase mathna wa-thulatha wa-ruba‘: ‘Two, three, four – meaning two at a time, three at a time, four at a time’ (‘mathna wa-thulatha wa-ruba‘a ayy ithnayn ithnayn wa-thalatha thalatha wa-arba‘a arba‘a’). The repetition of the cardinal number is not used to refer to ‘two twos’, ‘three threes’ and ‘four fours’ etc., but it is a relatively rare, alternative way of expressing the distributive adjectives.
Al-Baydawi takes the same position as al-Tabari and al-Qurtubi, but adds further information, stating that the difference in the amount of wings reflects the angels’ differing status: ‘with wings: two, three, four; they have varying numbers of wings, different in respect of their rank’ (‘ula ajnihati mathna wa-thulath wa-ruba‘ dhuwa ajniha muta‘addida mutafawita bi-tafawut ma lahum min al-maratib’). Al-Razi also agrees that this refers to individual wings, stating that ‘[the angel] has two wings, and anything beyond that is an addition’ (‘yakuna lahu janahan wa-ma ba‘dahuma ziyadatan’). Al-Razi here shows particularly clearly that the angels only have two wings and not two pairs of wings, as does al-Zamakhshari who states: ‘their wings, two at a time, meaning that for every single one of them there are two’ (‘ajnihatihum ithnan ithnan ayy li-kulla wahid minhum janahan’). The post-nominal suffix – hum refers to the angels and not to the wings. A survey of the exegeses being considered here shows that the general consensus in classical Islamic exegesis is that this aya refers to individual wings, and not pairs of wings.
Q. 35:1 is not the only place where the phrase mathna wa-thulath wa-ruba‘ occurs, as it is also used in Q. 4:3, in the so-called ‘polygamy verse’: wa-in khiftum alla tuqsitu fi’l-yatama fa’nkihu ma taba lakum mina’l-nisa’i mathna wa-thulatha wa-ruba‘ (If you fear that you will not act justly towards the orphans, marry such women as seem good to you, two three, four). This use of mathna etc. provides a useful comparison: in this case, the meaning is clearly: ‘two at time’, ‘three at time’ and ‘four at a time’; it would be logical to assume that the same meaning is intended in Q. 35:1, indicating, as in the Qatada tradition, that the angels have two wings, three wings or four wings only. The fact that the use of these terms in Q. 4:3 raises little discussion on the meaning of the terms (cf. al-Tabari on this aya), instead focusing on the legal implications of the Qur’anic legislation, suggests that, despite the consensus amongst the classical exegetes, the use of mathna, thulath and ruba‘ in Q. 35:1 did pose problems.
The preference for the interpretation of ‘pairs of wings’ in the modern translations and commentaries appears to be based on a sense that angels would not be able to fly with an odd number of wings. For example, Yusuf Ali comments ‘we need not suppose that angelic “wings” have muscles and feathers, like the wings of birds. If they had, how could there be three, or any odd number? We may suppose “two, three, or four” to refer to pairs of wings.’ The question of whether angels could (or could not) fly with three wings is not found in the classical exegetical tradition, and it seems to reflect modern scientific sensibilities. While Yusuf Ali states that angelic wings should not be conceived as being similar to birds’ wings, he, amongst others, still applies basic avian (or more general) aeronautical principals to them. Such scientific approaches to the celestial, supernatural, supra-mundane world deprive the angels of their other-worldliness. There are numerous ahadith about angels that describe them in such terms that it would seem impossible for them to move. For example, Israfil is described by Aisha in this way: ‘[Israfil] is the Angel of God. There is nothing in his presence. He has a wing in the East and he has a wing in the West, and a wing is on the nape of his neck and the throne is on the nape of his neck’ (‘huwa malak Allah laysa dunihi shay’ janah lahu bi’l-mashriq wa-janah lahu bi’l-gharb wa-janah lahu ‘ala kahilihi wa’l-‘arsh ‘ala kahilihi’). Here it is said explicitly that the important angel Israfil, who will blow the trumpet announcing the Last Day, does have three wings. If read in a scientific, post-enlightenment frame of mind, it is hard to see how the wing on Israfil’s neck could aid flight, but the fact that the angel Israfil does have three wings, or even that Israfil has a wing attached to his neck, is not important. This hadith is trying to present the magnificence and power of the angel, and the angel’s relationship to God. Above all, it shows that the belief that angels could have an odd number of wings did not pose any difficulties for classical traditionists; a belief seen throughout the medieval Islamic exegeses of Q. 35:1.
In wider Islamic tradition, however, angels are not restricted to two, three or four wings, nor even to two, three or four pairs of wings. The most famous example of an angel with more than four wings is the Prophet Muhammad’s vision of Gabriel’s ‘real’ form during the mi‘raj (‘ascension’), in which the angel has 600 wings, e.g. al-Bukhari’s Sahih: Abd Allah, concerning [His Word] two bows’-length away, or nearer, then revealed to his servant that which he revealed [Q. 53:9] said, Ibn Mas‘ud related that [the Prophet] saw Gabriel, and he had 600 wings’ (‘[‘Abd Allah]: fa-kana qaba qawsayn aw adna fa-aw adna fa-awha ila ‘abdihi ma awha [Q. 53:9] qala haddathana Ibn Mas‘ud annahu ra’a Jibril lahu situmi’at janah’). This hadith, and others like it, are also found in most of the tafasir on Q. 35:1, including al-Qurtubi, and the increase in the number of wings is based exclusively on Q. 35:1 and the phrase yazidu fi’l-khalqi ma yasha’ (increasing creation as He wills). Al-Tabari includes a very clear explanation of this part of the aya: ‘and that is His – the Blessed and the Most High – addition to His creation, namely the angel’s wings on the extremities, as He wills; and He decreases [the angel’s] extremities as He desires’ (‘wa-dhalika ziyadatuhu tabarak wa-ta‘ala fi khalq hadha’l-malak min al-ajniha ‘ala’l-akhir ma yasha’u wa-nuqṣanihi ‘an al-akhir ma uhibb’). Gabriel can, therefore, have any number of wings. This creates what seems at first to be a rather paradoxical situation in which the Qur’an both limits the amount of angels’ wings to four wings (or possibly eight if the distributive adjectives are incorrectly read as pairs), and also allows angels to have any number of wings. Al-Tabari provides an answer to this problem: the forms of all creation are dependent on God, and God has the power and the authority to alter His creation as He wills: ‘And thus in all of His creation, He increases what He wills for His creation, whatever He wills for it; and He decreases as He wills, whatever He wills for a creation; His is the creation, and the power, and His is the ability and the authority’ (‘wa-kadhalika dhalika fi jami‘ khalqihi yazidu ma yasha’u fi khalq ma sha’a minhu wa-yanqasu ma sha’a min khalq ma sha’a lahu al-khalq wa’l-amr wa-lahu al-qadira wa’l-sultan’). Just as God created the angels, God can also change the form of the angels; a belief that is echoed in the final clause of the aya: inna’llaha ‘ala kulli shay’in qadir (Surely, God is powerful over everything).
Having established the number of wings that the angels have, the medieval commentators then proceed to explain the purpose of the wings. Al-Qurtubi comments simply: ‘they come down from heaven to earth with them, and they ascend from earth to heaven with them … and [God] created them messengers’ (‘yanzaluna biha min al-sama’ ila’l-ard wa-ya‘rajuna min al-ard ila’l-sama’ … wa-ja‘alahum rusulan’). These two comments establish a direct link between the creation of the angels with wings and their function as messengers. In this respect, Q. 35:1 becomes aetiological, explaining both the way in which God communicates with the human world, and why angels have wings.
This sense that the angels’ wings are linked to their role as messengers appears to be reflected in the grammar of the aya itself. The aya begins with a laudation of God (al-hamdu li’llah), which is followed by two titles (or descriptions) of God: (i) the one who divided the heavens and the earth (faṭiri’l-samawati wa’l-ard) and (ii) the one who made the angels as messengers (ja‘ili’l-mala’ikati rusulan). The references to the angels’ wings is followed by a description of them (ula ajnihatin mathna wa-thulatha wa-ruba‘). Finally, comes the main clause of the aya (yazidu fi’l-khalqi ma yasha’), followed by a concluding comment (inna’llaha ‘ala kulli shay’in qadir). The exegeses of al-Qurtubi and al-Baydawi both argue that the angels have wings so that they can descend to earth and return to heaven. Surely it is not simply coincidental that the statement that the angels were made the messengers of God is both preceded by the belief that God separated the divine and human worlds and followed by the description of the angels’ wings? This aya adopts a logical argument: heaven and earth are split and divided from each other, God needs intermediaries to communicate with the human world, the angels are made messengers for this purpose, and are given wings to perform this task. Above all, this aya stresses God’s ability to do whatever He wills to His creation, which is the general theme to the rest of Surat al-Mala’ika.
The classical exegetes extrapolate a number of important points about angels and their form from this short clause. Primarily, the normal physical form of the angels is described by the Qur’an; and this is especially significant as it is the only aya to provide evidence for angels having wings. Beyond this basic image of a winged angel, the exegetes debated their number. The various disputes about the meaning of the three distributive adjectives mathna, thulath and ruba‘ are, to a certain extent, not particularly crucial theologically speaking – does it matter whether angels have wings in pairs or not? The use of these adjectives clearly raised some concerns, and the exegetes all discuss their meaning in some detail except for the Tafsir al-Jalalayn. There are two main focus points for the exegetes: firstly, that God created angels with wings for a particular reason: to act as intermediaries between God and humans; and secondly, that God has power to change His creations as He wishes. These two points are stressed throughout the medieval commentaries on this aya.
Conclusions: God, the Angels, the Beginning and the End
Q. 35:1 may be short, but it contains much information about both Qur’anic and wider Islamic beliefs about angels. The aya contains a number of grammatical and lexicographical cruces, which need to be considered with some care: (i) the meaning of malak in the Qur’anic milieu; (ii) the meaning of faṭir and (iii) the precise meaning of the three distributive adjectives mathna, thulath and ruba‘. Islamic tradition, reflected in the tafasir on this aya, often takes the interpretation of this aya in unexpected directions, with the interpretation of yazidu fi’l-khalqi ma yasha’ (increasing creation as He wills) in wider Islamic tradition being particularly good example of this. The structure of the aya also seems to indicate a link between the fact that angels have wings and their role as intermediaries; a link which is highlighted in a number of the classical exegeses seen above.
So far, a study of the meanings of malak and faṭir has shown that their traditional interpretations may not be the most suitable. The general assumption that malak means ‘messenger’ has been seen to have very little evidence to support it. The malak is a creature of the divine world, in much the same way as jinn, humans and animals are part of the human world. There is also significant evidence to suggest that faṭir does not simply mean ‘Creator’. In Q. 35:1 the use of faṭir is significant, not necessarily in and of itself, but by the fact that it is neither ja‘il nor khaliq. This is supported in two ways: firstly, the phrase ‘Creator of the heavens and the earth’ is (as has been noted above) commonly phrased with ja‘ala and khalaqa; secondly, faṭir is juxtaposed with ja‘il in the same aya. But how do these new readings influence the understanding of the aya?
The use of intermediaries has long since been seen to support the idea that a particular cosmology portrays God as utterly transcendent: ‘A developed belief in angels is likely to be found in conjunction with a belief in a supreme God. If there be numerable deities, none is so great and unapproachable that a supernatural hierarchy between him and man is felt to be necessary.’ A transcendent view of God leads to a notion of spatial (or cosmic) dualism: the belief that the divine and human worlds are divided and distinct realms. This is not an unexpected theological position, especially in the context of Jewish cosmology, which also presents a worldview ‘where a contrast is drawn between heaven and earth, the mundane and the supra-mundane’. Furthermore, if malak is read in the English sense of angel rather than the general Semitic understanding of mal’ak, the separation and division between heaven and earth is more clearly defined. This language of division can also be seen in the Qur’anic vocabulary of revelation, as Stefan Wild notes: ‘the terms nuzul, tanzil and ’inzal only make sense in a space in which there is an above and a below. God sends a verse, a sura, the whole of the Qur’an down, because God is in heaven and the Prophet together with mankind is on earth.’
This aya also suggests that because of this division between heaven and earth, angels were both created messengers, and given wings so that they can deliver these messages. This theme is one that is expanded greatly in the exegetical literature. It should also be noted that the aya praises God on account of this: Praise belongs to God … who appointed the angels to be messengers. Although heaven and earth are separated, God creates a way for His will to be communicated to His earthly creations: the division may create space between the divine and human, but God ensures that separation does not cause abandonment and isolation. The dispute about the number of wings that angels have is one that is relatively academic. However, the trend to read mathna etc. as ‘pairs of wings’ appears to be a result of modern scientific sensibilities, and it is the classical exegetes who stress the wonders and mysteries of God’s creation. For the medieval exegetes, the variation in the number of wings that angels have is simply a testament to God’s creative power and authority over the created world. Hasan Hanafi comments that ‘everything on land is subject to Divine Will. Land is obedient to God and worships Him’. Angels, in the divine world, are similarly subject to God’s will, obedient to God and worship God, and the fact that God can increase (or decrease) the number of wings that an angel has, is evidence of this submission to God.
The Qur’an does not only have a concept of cosmic or spatial dualism, but there is an element of ethical dualism too. The relationship between opposites is an important part of Qur’anic language and idiom. Richard Martin comments that: ‘Quranic cosmology is a dynamic, structured complex. Binary oppositions such as divine/human, heaven/earth, acceptance/rejection, and the believer/unbeliever provide the contrasts within the text of Islamic cosmology, which we may suspect acts to symbolise … tensions felt within the concrete world of everyday life.’ Islamic eschatology does not escape this sense of duality, and in Creation and Termination Shinya Makino found that ‘the eschatological condition of the heaven may be correctly understood in relation to, or in contrast to the manner, in which the heaven was created and has been conserved’. The splitting or dividing of heaven and earth at the beginning of time may also reflect the splitting of the heavens at the end of time, creating a sense of cosmological unity: ‘in the disruption of the natural order as portrayed in the Qur’an one can see a reverse process of creation.’
Thus, in this short Qur’anic aya it is possible to gain a basic insight into Qur’anic cosmology. By understanding malak to be referring to a generic heavenly being, and not to a ‘messenger’, a clear distinction is made between the creatures of the heavenly and earthly realms. This is reinforced by a reading of faṭir as ‘divider’ or ‘splitter’: the division between ‘heavenly’ and ‘earthly’ is firmly established. As a result of this division between heaven and earth, God created messengers from the angels to act as intermediaries between the human and divine worlds out of His mercy and desire to communicate with His creatures on earth. The portrayal of the angels with wings simply explains, in a logical way, how the angels move from one realm to the other. In isolation, this aya suggests a very basic cosmogony, but when this aya is read in the context of the rest of the Qur’an, a great deal of symmetry can be seen between the Qur’anic descriptions of the beginning and the end of the world. Just as the division between this world and the next is destroyed at the Last Day, the cosmos is divided into two at the beginning of time, separating the creatures of the divine world from the creatures of the earthly world.
 Q. 35:1; Unless otherwise stated, this and all further translations of the Qur’an cited are from A.J. Arberry, The Koran Interpreted (Oxford: One World, 1998). I would like to thank Prof. Carole Hillenbrand for her helpful comments, and the AHRC for their financial support.
 The translations consulted have been selected to represent both older and more contemporary translations, and those by Muslims and non-Muslims. They include: Muhammad Baqir Behbudi and Colin Turner, The Qur’an: A New Interpretation (Richmond: Curzon, 1997); Richard Bell, The Qur’an: Translated with a Critical Arrangement of the Surahs (Edinburgh: T. & T. Clark, 1937); N.J. Dawood, The Koran: A Translation (Harmonsworth: Penguin, 1974); Muhammad Hamidullah and Michel Léturmy, al-Qur’an al-majid (Beirut: Salih Ozcan, 1973); M.Z. Khan, The Qur’an: The External Revelation Vouchsafed to Muhammad – Seal of the Prophets (London: Curzon, 1975); E. Montet, Le Coran (Paris: Payot, 1929); J.M. Rodwell, The Koran: Translated from the Arabic (London: J.M. Dent, 1909); A. Yusuf Ali, The Holy Qur’an (Lahore: Sheikh Muhammad Ashraf, 1938).
 Respectively: Abu Ja‘far ibn Jarir al-Tabari, Jami‘ al-bayan fi tafsir al-Qur’an (30 vols. Cairo: n.p., 1968); Abu’l-Qasim Jar Allah Mahmud ibn ‘Umar al-Zamakhshari, al-Kashshaf haqa’iq al-tanzil wa-‘uyun al-aqawil fi wujuh al-ta’wil (3 vols. Cairo: Muṣtafa al-Babi al-Halabi, 1900); Fakhr al-Din Muhammad ibn ‘Umar al-Razi, Mafatih al-ghayb tafsir al-Qur’an (8 vols. Cairo: al-Matba‘a al-Husayniyya al-Misriyya, 1909); Abu ‘Abd Allah Muhammad ibn Ahmad al-Qurtubi, al-Jami‘ li-ahkam al-Qur’an (20 vols. Cairo: Dar al-Qalam, 1966); ‘Abd Allah ibn ‘Umar al-Baydawi, Beidhawi Commentarius in Coranum ex Codd. Parisiensibus Dresdensibus et Lipsiensibus, ed. Winnand Fell (2 vols. Leipzig: Vogel, 1878); Jalal al-Din Muhammad ibn Ahmad al-Maḥalli and Jalal al-Din ‘Abd al-Rahman ibn Abi Bakr al-Suyuti (al-Jalalayn), al-Qur’an al-karim wa-bi-hamishihi tafsir al-imamayn al-Jalalayn (n.p.: Maktabat al-Muthanna, ca 1920).
 Malak/mala’ika occurs 81 times. See Fu’ad ‘Abd al-Baqi, al-Mu‘jam al-mufahras li-alfaz al-Qur’an al-karim (Cairo: Dar al-Hadith, 1996), pp. 771–2.
 For a brief introductory discussion of these issues, see D.B. MacDonald, art. ‘Mala’ika’ in Encyclopaedia of Islam, 2nd ed.
 Although some works, such as the Taj al-‘arus, do place the word under the entry for m–l-k; see Muhammad ibn Muhammad Murtada al-Zabidi, Taj al-‘arus min jawahir al-qamus (40 vols. Kuwait: Matba‘at Hukumat al-Kuwayt, 1965–2001), vol. 27, pp. 346–66.
 See al-Tabari, Jami‘ al-bayan, vol. 1, pp. 197–8. The first occurrence of malak is at Q. 2:32. The usual word selected as an example in the exegetical literature is sham’al/sha’mal, see e.g. al-Baydawi, Commentarius, vol. 1, p. 47; al- Qurtubi, al-Jami‘, vol. 1, p. 263 and al-Razi, Mafatih al-ghayb, vol. 1, p. 247.
 Al-Ṭabari, Commentary on the Qur’an, tr. J. Cooper (Oxford: Oxford University Press, 1987), p. 207. Most of the exegetes cite uses of these two roots in pre-Islamic poetry. See Albert Arazi, al-‘Iqd al-thamin fi dawawin al-shu‘ara’ al-sitta al-jahiliyyin (Jerusalem: Hebrew University, 1999), p. 1003.
 E.g. E.W. Lane, Arabic-English Lexicon, repr. (2 vols. Cambridge: Islamic Texts Society, 1984), pp. 81–2.
 E.g. Arne A. Ambros, with Stephan Procházka, A Concise Dictionary of Koranic Arabic (Wiesbaden: Reichert Verlag, 2004), p. 244; A. Fischer, Th. Nöldeke and H. Reckendorf, Wörterbuch der klassischen arabischen Sprache (Wiesbaden: Otto Harrassowitz, 1957–), vol. 2, part 1, pp. 53–5; Arazi, al-‘Iqd al-thamin, p. 1003; A. Fischer, Glossar in E. Rudolf and A. Fischer, Arabische Chrestomathie aus Proausschriftstellern (Wiesbaden: Otto Harraosswitz, 1984), pp. 1–157, p. 128, p. 118.
 As Jonathan Owens has noted, the lexicographical and exegetical sciences arose out of the need to explain difficult passages in the Qur’an; see Jonathan Owens, Early Arabic Grammatical Theory (Amsterdam: Benjamins, 1990), p. 28.
 No other Semitic language has a similar position whereby the two roots l-’-k and ’-l-k have two different, but largely synonymous, meanings.
 That is not to say that the origins of foreign words was not of interest to some Muslim scholars, as Lothar Kopf notes: ‘yet most of the earliest exegetes had admitted that foreign words are found in the Kor’an, and considering the high esteem in which they were held, their The Angels in Surat al-Mala’ika opinion could not simply be rejected’ (Lothar Kopf, ‘The Treatment of Foreign Words in Mediaeval Arabic Lexicology’ in Uriel Heyd (ed.), Studies in Islamic History and Civilisation (Jerusalem: Magnes Press, 1961), pp. 191–205, pp. 202–3).
 E.g. MacDonald suggests Canaanite (MacDonald, art. ‘Mala’ika’); Hirschfeld proposes Hebrew or Aramaic (Hartwig Hirschfeld, Beiträge zur Erklärung des Korân (Leipzig: Otto Schulze, 1886) p. 45); Bell and Pautz suggest Ethiopic origin (Richard Bell, Introduction to the Qur’an (Edinburgh: Edinburgh University Press, 1953), p. 144; Otto Pautz, Muhammads Lehre von der Offenbarung (Leipzig: J.C. Heinrichs’sche Buchhandlung, 1898), p. 69, n. 1); and Mingana proposes Syriac (A. Mingana, ‘Syriac influence on the Style of the Kur’an’ Bulletin of the John Rylands Library 11 (1927), pp. 77–98, p. 85). There are no occurrences of the word in the Arabian inscriptions: see A.F.L. Beeston et al., Dictionnaire Sabéen (Anglais-Français-Arabe) (Louvain-le-Neuve: Editions Peeters, and Beirut: Librarie du Liban, 1982), p. 5, p. 81, pp. 85–6; Karlous Conti Rossini, Chrestomathia Arabica Meridionalis Epigraphica (Rome: Instituto per l’Oriente, 1931), p. 105, p. 171, p. 178; Joan Copeland Biella, Dictionary of Old South Arabic – Sabean Dialect (Chico, CA: Scholars’ Press, 1982), p. 18, pp. 277–8.Biella does find an occcurence of the root l-’-k (p. 256), but it is a hapax legomenon and its meaning is unclear.
 See Thomas O. Lambdin, Introduction to Classical Ethiopic (Ge‘ez) (Missoula: Scholars’ Press, 1978), p. 18; August Dillmann, Lexicon Linguae Aethiopicae (Leipzig: T.O. Weigel, 1865), col. 47–9. Jeffrey comments that ‘there can be little doubt, however, that the source of the word is the Eth. malʾak with its characteristic plu. mala’ekt which is the common Eth. Word for angelos, whether in the sense of angelus or nuntius, and thus corresponds exactly with the Hebrew mal’ak, Phon. mal’ak; Syr. malak’ (Arthur Jeffrey, The Foreign Vocabulary of the Qur’an (Baroda: The Oriental Institute, 1938), p. 269). For more on broken plurals in Ethiopic see Robert R. Ratcliffe, The ‘Broken’ Plural Problem in Arabic and Comparative Semitic (Amsterdam: John Benjamins Publishing Company, 1998), pp. 166–72.
 Dillmann, Lexicon, col. 48.
 E.g. (i) Akkadian: Bruno Meissner and Wilfram von Soden, Akkadisches Handwörterbuch (3 vols. Wiesbaden: Otto Harrassowitz, 1966–81), vol. 2, p. 595; (ii) NW and W Semitic Inscriptions: J. Hoftijzer and K. Jongeling, Dictionary of North-West Semitic Inscriptions (Leiden: E.J. Brill, 1965), p. 151; Charles-F Jean and J. Hoftijzer, Dictionnaire des Inscriptions Sémitiques de l’Ouest (Leiden: E.J. Brill, 1965), p. 151; (iii) Biblical Hebrew: Francis Brown, S.R. Driver and C.A. Briggs, Hebrew and English Lexicon of the Old Testament (Oxford: Oxford University Press, 1952), p. 521; (iv) Biblical Aramaic: Brown, Driver and Briggs, Lexicon, p. 1098; (v) Mishnaic Hebrew and Aramaic: Marcus Jastrow, A Dictionary of the Targumim, the Talmud Balbi and Yerushalmi, and the Midrashic Literature (London: E. Shapiro, 1926), p. 786; (vi) Samaritan Aramaic: Abraham Tal, A Dictionary of Samaritan Aramaic (2 vols. Leiden: E.J. Brill, 2000), vol. 2, p. 468; and (vii) Syriac: J. Payne Smith, A Compendious Syriac Dictionary found upon the Theasaurus Syriacus of R. Payne Smith (Oxford: Clarendon Press, 1903), p. 275.
 Both Dillmann and Lambdin give examples of verbal usage in Ethiopic, see Dillmann, Lexicon, col. 47; Lambdin, Introduction, p. 411; also it should be noted that the Gospel writer’s name, Loukan, which bears no philological relationship to the root l-’-k is associated with the root in Ethiopic (i.e. meaning ‘apostle’ or ‘evangelist’, that is ‘one sent with a message’) – see Lambdin, Introduction, p. 411.
 See Freedman-Willoughby and Fabry, art. ‘mal’ak’ in The Theological Dictionary of the Old Testament (12 vols. Cambridge: William B. Eerdmanns Publishing Company, 1997), vol. 8, pp. 308–25.
 A good example of this is Acts 12:16: ‘kai idou angelos kuriou espestē kai phōs elampsen en to oikēmati pataxas de tēn pleuran tou Petrou ēgeiren auton legōn, Anasta ex taxei’ which is normally translated: ‘suddenly an Angel of the Lord appeared and a light shone in the cell. He tapped Peter on the side and woke him saying, “Get up quickly”’ (Acts 12:16 (New Revised Standard Version; Oxford: Oxford University Press, 1995), p. 127). The divine origin of the angel is not understood immediately.
 Cf. Gen. 16:7 and 32:4; R. Kittel, A. Alt, O. Eissfeldt and P. Kahle (eds), Biblia Hebraica Stuttgartensia (Stuttgart: Deutsche Bibelgesellschaft, 1997), p. 22, p. 51; Alfred Rahlfs (ed.), Septuaginta (Stuttgart: Privilegierte Württembergische Bibelanstalt, 1935), p. 20, p. 51; Bonifatio Fischer, Iohann Gribonont, H.F.D. Sparks and W. Thiele (eds), Biblia Sacra Vulgata (Stuttgart: Württembergische Bibelanstalt, 1969), p. 22, p. 47. For early uses of angelus and nuntius see Philip Burton, The Old Latin Gospels (New York: Oxford University Press, 2000), p. 93.
 Praise be to God … who appointed angels to be messengers.
 He concludes: ‘I hope to have clearly proved that, from the morphological point of view, the noun ma’lak > mal’ak > malak is a native Arabic word’ (Paul Boneschi, ‘Is malak an Arabic word?’, Journal of the American Oriental Society 65 (1945), pp. 107–11, p. 111).
 E.g. Q. 4:97; Q. 6:93; Q. 8:50; Q. 16:28, 32; Q. 32:11; and Q. 47:27.
 E.g. Q. 2:34; Q. 7:11; Q. 15:30; Q. 17:61; Q. 18:50; Q. 20:116; and Q. 38:73.
 E.g, Q. 39:75; and Q. 69:17.
 See A.J. Wensinck, Concordance et indices de la tradition musulmane (8 vols. Leiden: E.J. Brill, 1936–88), vol. 6, pp. 260–72.
 Q. 2:98; Q. 4:136; and Q. 35:1. This trend is also found in other texts, such as Acts 12:1 in the Mount Sinai Codex 151 (see Harvey Staal (ed. and tr.), Mt. Sinai Arabic Codex 151 (Louvain: E. Peeters, 1983–4)). Cf. 1 Peter 1:1; 2 Peter 1:1; and Philippians 2:25. The Apocalypse of Baruch 6:6; 55:2; and IV Ezra 4:1, 52; 5:31; 7:1; and 14:4 have similar juxtapositions (see The Arabic Text of the Apocalypse of Baruch, ed. and tr. F. Leemhuis, A.F.J. Klijn and G.J.H. van Gelder (Leiden: E.J. Brill, 1986)); Adriana Drint (ed. and tr.), The Mount Sinai Arabic Version of IV Ezra (Louvain: E. Peeters, 1997)). See also Meira Pollack, ‘Alternate Renderings and Additions in Yeshu‘ah ben Yehydah’s Arabic Translation of the Pentateuch’, Jewish Quarterly Review 84 (1993), pp. 209–25, pp. 220–2.
 Lane, Lexicon, pp. 81–2; there are also no occurrences of the verbal root in a large majority of ahadith – see Wensinck, Concordance.
 W. Montgomery Watt, Bell’s Introduction to the Qur’an (Edinburgh: Edinburgh University Press, 1970), p. 25.
 Gisele Webb, art. ‘Angel’ in The Encyclopaedia of the Qur’an.
 See Toshihiko Izutsu, God and Man in the Koran (Tokyo: The Keio Institute of Cultural and Linguistic Studies, 1964), pp. 16–17.
 Q. 6:14; Q. 12:101; Q. 14:10; Q. 35:1; Q. 39:46; and Q. 42:11.
 Arberry, The Koran, p. 138; Yusuf Ali, The Holy Qur’an, p. 846; Behbudi and Turner, The Qur’an, p. 260; Ahmad Zidan and Dina Zidan, The Glorious Qur’an: Text and Translation (Cairo: Islamic House for Publishing and Distribution, 1996), p. 434.
 Bell, The Qur’an, p. 429; Dawood, The Koran, p. 304; Khan (‘… who is about to fashion …’), The Qur’an, p. 427; Hamidullah and Léturmy, al-Qur’an al-majid, p. 571; Montet, Le Coran, p. 581; and Rodwell (‘ …. maker …’), The Koran, p. 289. Daniel Gimaret also comments ‘il n’est pas dobteux que, dans tous ses cas, faṭir et faṭara sont pours equivalent de haliq et halaqa’ (Daniel Gimaret, Les Noms Divins en Islam (Paris: Les Editions du Cerf, 1988), p. 290).
 Afnan H. Fatani, ‘The Lexical Transfer of Arabic Non-Core Lexicon: Sura 113 of the Qur’an – al-Falaq (The Splitting)’, Journal of Qur’anic Studies 4:2 (2002), pp. 61–81, p. 71; The Angels in Surat al-Mala’ika W.F. Albright also highlights the fact that the base meaning of creation words is not necessarily ‘to create’: ‘many Semitic words for “create, fashion, form” mean primarily to “cut into shape”’ (W.F. Albright, ‘Notes on Egypto-Semitic Etymology III’, Journal of the American Oriental Society 47 (1927), pp. 81–98 and pp. 215–55, pp. 217–18.
 See Beeston, Dictionnaire, p. 47; Conti Rossini, Chrestomathie, p. 218; Biella, Dictionary, p. 413.
 Meissner and von Soden, Handwörterbuch, vol. 2, p. 849; for an example in Ugaritic see Michael C. Astour, ‘Two Ugaritic Serpent Charms’, Journal of Near East Studies 27 (1968), pp. 13–36, pp. 33–4; for Biblical Hebrew see Brown, Driver and Briggs, Lexicon, p. 809.
 See H. Niehr, art. ‘paṭar’ in The Theological Dictionary of the Old Testament (12 vols. Cambridge: William B. Eerdmanns Publishing Company, 1997), vol. 11, p. 529.
 (i) Akkadian: Meissner and von Soden, Handwörterbuch, vol. 2, pp. 849–51; (ii) the root also appears in Old Babylonian (la-ap-ṭu-ru-tum), but the meaning is uncertain; Albrecht Goetze suggests faṭara in the sense of ‘to split’ ‘invent’ or ‘create’ – see Albrecht Goetze, ‘An Old Babylonian Itinerary’, Journal of Cuneiform Studies 7 (1953), pp. 51–72, p. 57; (iii) Syriac: Payne Smith, Dictionary, pp. 442–3; (iv) the Dead Sea Scrolls: Niehr, art. ‘paṭar’; (iv) Mishnaic Hebrew: Niehr, art. ‘paṭar’.
 Jastrow, Dictionary, p. 1157.
 (i) Akkadian: Meissner and von Soden, Handwörterbuch, vol. 2, pp. 849–51; (ii) Biblical Hebrew: Brown, Driver and Briggs, Lexicon, p. 809.
 E.g. Ex. 34:19–20; see Niehr, art. ‘paṭar’.
 The lexical influence of Ethiopic over Qur’anic Arabic is important; see C.H.M. Versteegh, The Arabic Language (Edinburgh: Edinburgh University Press, 1997), pp. 59–60. For the lexical borrowing of faṭir from Ethiopic, see al-Tabari, Jami‘ al-bayan, vol. 29, p. 138; C.C. Torrey, The Jewish Foundations of Islam (New York: Jewish Institute of Religion Press, 1933), p. 53; Lane, Lexicon, pp. 2415–16.
 Dillmann, Lexicon, col. 1383–5; Lamdin, Introduction, p. 400.
 Murtaza al-Zabidi, Taj al-‘arus, vol. 13, p. 326. This hadith is also cited by al-Qurtubi, al-Jami‘, vol. 14, pp. 318–19; and by al-Razi, Mafatih al-ghayb, vol. 2, p. 29.
 Al-Zabidi, Taj al-‘arus, vol. 13, pp. 326–7.
 ‘Bi-qawlihi «faṭir al-samawat wa’l-arḍ» mubtadi’ wa-mubtadi’uhuma wa-khaliquhuma’ (al-Tabari, Jami‘ al-bayan, vol. 6, p. 158).
 Cf. al-Tabari, Jami‘ al-bayan, vol. 13, p. 73 (Q. 12:101), p. 190 (Q. 14:10); vol. 24, p. 11 (Q. 39:46); vol. 25, p. 11 (Q. 42:11); vol. 29, p. 2 (Q. 67:3).
 Al-Baydawi, Commentarius, vol. 1, p. 275, p. 472; vol. 2, p. 148, p. 227; al-Jalalayn, Tafsir al-Jalalayn, p. 106, p. 203, p. 211, p. 364, p. 389, p. 406; al-Zamakhshari, al-Kashshaf, vol. 1, p. 446; and vol. 2, p. 454; al-Qurtubi, al-Jami‘, vol. 9, p. 269, p. 346; vol. 14, pp. 318–19; and vol. 16, p. 7.
 Al-Tabari, Jami‘ al-bayan, vol. 7, pp. 158–9; al-Razi, Mafatih al-ghayb, vol. 4, p. 15; al-Zamakhshari, al-Kashshaf, vol. 1, p. 445; al-Baydawi, Commentarius, vol. 1, p. 275.
 Al-Razi, Mafatih al-ghayb, vol. 4, p. 16; al-Baydawi, Commentarius, vol. 2, p. 142; al-Qurṭubi, al-Jami‘, vol. 14, pp. 318–19; Gimaret citing Halimi, talks of the idea that fatara could signify splitting comparing it to the division of light and dark in Genesis 1; however, he concludes ‘mais cette hypothèse est tout de même peu credible, car il y a quantité d’autre racines parallèles signifiant paraillement «fender, disjoindre, séparer» … alors que, de toute cette série, seul faṭara a aussi le sens de «créer»’ (Gimaret, Noms Divins, p. 291).
 R. Analdez, ‘La Louange à Dieu dans le Coran d’après le Commentaire de Fakhr al-Dîn al-Râzî’, Annales Islamologiques 4 (1963), pp. 79–92, p. 81.
 Al-Tabari, Jami‘ al-bayan, vol. 29, pp. 137–9; al-Baydawi, Commentarius, vol. 2, p. 344, pp. 365–6; al-Jalalayn, Tafsir al-imamayn, p. 478, p. 490; al-Qurtubi, al-Jami‘, vol. 18, p. 208; vol. 19, p. 50; al-Zamakhshari, al-Kashshaf, vol. 3, p. 201.
 Al-Jalalayn, Tafsir al-imamayn, p. 478; al-Zamakhshari, al-Kashshaf, vol. 3, p. 201.
 Toshihiko Izutsu’s comments on this aya in God and Man in the Koran highlights the difficulty that these two uses of the root pose. Izutsu translates them without a creation or a dividing gloss: ‘this pure monotheistic belief symbolized by the name of Abraham is the “true religion”, the “natural disposition” (fiṭrah) to which God predisposed (faṭara) mankind’ (Izutsu, God and Man, p. 112).
 khalaqa: Q. 6:1, 73; Q. 7:54, 185; Q. 9:36; Q. 10:3, 6; Q. 11:7; Q. 14:19, 32; Q. 16:3; Q. 17:99; Q. 20:4; Q. 25:59; Q. 27:60; Q. 29:44, 61; Q. 30:8; Q. 31:10 (heaven only); Q. 31:25; Q. 32:4; Q. 36:81; Q. 39:5, 38; Q. 41:9 (earth only); Q. 43:9; Q. 45:22; Q. 57:4; Q. 64:3; Q. 65:12; Q. 67:3; Q. 71:15 (seven heavens only); khaliq: Q. 35:3. From ‘Abd al-Baqi, al-Mu‘jam, pp. 296–7.
 al-Zamakhshari, al-Kashshaf, vol. 3, p. 203.
 See W. Wright, A Grammar of the Arabic Language (Third Edition), repr. (Cambridge: Cambridge University Press, 2004), §333, vol. 1, pp. 262–3.
 Yusuf Ali, The Holy Qur’an, p. 1152; Montet, Coran, p. 581; Rodwell, The Koran, p. 288.
 Arberry, The Koran, p. 444; Bell, The Qur’an, p. 429; Khan, The Qur’an, p. 427; Hamidullah and Léturmy, al-Qur’an al-majid, p. 571; Zidan, Qur’an, p. 434.
 (My emphasis) Richard Bell, A Commentary on the Qur’an, C.E. Bosworth and M.E.J. Richardson (eds), Journal of Semitic Studies Monograph, 14 (2 vols. Manchester: University of Manchester Press, 1991), vol. 2, p. 125.
 Behbudi and Turner, The Qur’an, p. 260.
 Al-Tabari, Jami‘ al-bayan, vol. 22, p. 114.
 Al-Tabari, Jami‘ al-bayan, vol. 22, p. 114.
 Al-Qurtubi, al-Jami‘, vol. 14, p. 319.
 Al-Qurtubi, al-Jami‘, vol. 14, p. 319.
 As William Wright notes: ‘the distributive adjectives are expressed by the repeating the cardinal numbers once; or by the forms (fu‘al) and (maf‘al), either singly or repeated’ (Wright, Grammar, vol. 1, p. 262).
 Al-Baydawi, Commentarius, vol. 2, p. 148.
 Al-Razi, Mafatih al-ghayb, vol. 7, p. 29.
 Al-Zamakhshari, al-Kashshaf, vol. 2, p. 455.
 Yusuf Ali, The Holy Qur’an, p. 1152, n. 3871.
 Jalal al-Din al-Suyuti, al-Haba’ik fi akhbar al-mala’ik (Beirut: Dar al-Kutub al-‘Ilmiyya, 1988), §94, p. 35.
 Abu ‘Abd Allah Muhammad ibn Isma‘il al-Bukhari, Sahih al-Bukhari (Riyadh: Bayt al-Afkar al-Dawliyya li’l-Nashr wa’l-Tawzi‘, 1419/1998), §4856, p. 954.
 Al-Qurtubi, al-Jami‘, vol. 14, pp. 319–20; al-Baydawi, Commentarius, vol. 2, p. 148.
 Al-Tabari, Jami‘ al-bayan, vol. 22, p. 114.
 Al-Tabari, Jami‘ al-bayan, vol. 22, p. 114.
 Al-Qurtubi al-Jami‘, vol. 14, p. 319; cf. al-Baydawi, Commentarius, vol. 2, p. 148.
 See especially al-Razi, Mafatih al-ghayb, vol. 7, p. 29.
 T.G. Soares, ‘The Religious Ideas of Judaism from Ezra to the Macabees’, Biblical World 13 (1899), pp. 380–8, p. 385; see also Gerald Foot Moore, ‘Intermediaries in Jewish Theology’, Harvard Theological Review 15 (1922), pp. 41–85.
 For more on different types of dualism see John G. Gammie, ‘Spatial and Ethical Dualism in Jewish Wisdom and Apocalyptic Literature’, Journal of Biblical Literature 93 (1974), pp. 356–85.
 Gammie, ‘Spatial and Ethical Dualism’, p. 358.
 Stefan Wild, ‘“We have sent down to thee the Book with the Truth”: Spatial and Temporal Implications of the Qur’anic Concepts of nuzul, tanzil, and ’inzal’, in Stefan Wild (ed.), The Qur’an as Text (Leiden: E.J. Brill, 1996), pp. 137–53, p. 141; see also Josef van Ess, ‘Vision and Ascension: Surat al-Najm and its Relationship with Muhammad’s mi‘raj’, Journal of Qur’anic Studies 1 (1999), pp. 47–62.
 Hassan Hanafi, ‘World-Views of Arab Geographers’, GeoJournal 26 (1992), pp. 153–6, p. 153.
 Richard C. Martin, ‘Understanding of the Qur’an in Text and Context’, History of Religions 21 (1981), pp. 361–84, p. 372.
 Shinya Makino, Creation and Termination (Tokyo: The Keio Institute of Cultural and Linguistic Studies, 1970), pp. 32–3.
 Jane Idleman Smith and Yvonne Yazbeck Haddad, The Islamic Understanding of Death and Resurrection (Albany: State University of New York Press, 1981), p. 66; Annemarie Schimmel also notes that ‘creation and judgment are the two poles of the world which human beings can experience. Before creation lies azal, eternity without beginning; after Judgment, abad, eternity with end’ (Annemarie Schimmel, ‘Creation and Judgment in the Koran and Mystico-Poetical Interpretation’ in A. Schimmel and A. Falaturi (eds), We Believe in One God (London: Burns and Oates, 1979), pp. 149–77, p. 150.