Coloured Dots and the Question of Regional Origins in Early Qur’ans (Part II) Alain George UNIVERSITY OF EDINBURGH Part I of the present article (Journal of Qur’anic Studies, 17:1) began with a study of remarks about regional patterns of vocalisation offered by the Andalusi scholar of the Qur’an al-Dānī (371–444/982–1053), primarily in al-Muḥkam fī naqṭ al-maṣāḥif. These assertions were then confronted with extant manuscripts of the third to fourth/ ninth to tenth centuries that can be ascribed to the broad region between Syria, Iraq, and Iran. This analysis largely confirmed al-Dānī’s observations, thereby suggesting the existence of a dominant norm for vocalisation in this area, which he identifies as the Mashriq. An increase in the complexity of notation systems over time also emerged. The same approach will now be applied to Qur’ans from the Maghrib. As shown in Part I, al-Dānī asserted that his region of origin possessed a distinctive vocalisation system with Medinan roots; on a few points, he also drew contrasts between earlier Maghribī conventions and those prevalent in his own day.1 Abbreviations: $ BNF: Bibliothèque Nationale de France (Paris) $ CBL: Chester Beatty Library (Dublin) $ Khalili: Nasser D. Khalili Collection of Islamic Art (London and Geneva) $ Met: Metropolitan Museum of Art (New York) From Theory to Praxis: The Manuscripts (continued from Part I) 2. Medina and the Maghrib The Qur’an of Faḍl (in or before 295/907) This small Qur’an (10.5 × 16 cm for the largest pages) has only been published in scattered and incomplete fragments;2 but it is of potential importance because of its Journal of Qur’anic Studies 17.2 (2015): 75–102 Edinburgh University Press DOI: 10.3366/jqs.2015.0196 # Centre of Islamic Studies, SOAS www.euppublishing.com/journal/jqs 76 Journal of Qur’anic Studies link to al-Qayrawān. It is written in style D.III with 6 lines to the page, and was initially bound in 30 volumes corresponding to as many ajzāʾ. Several hundred folios, most of them still sewn in quires, are preserved at the Musée National d’Art Islamique in Raqqada, near al-Qayrawān (the largest section is preserved under shelfmark R64a); four additional folios are at the Bibliothèque Nationale de France (Arabe 5178m, ff. 18–21).3 Each volume originally opened with a waqfiyya written on a double-page spread, followed by a double spread of illumination, then by the Qur’anic text. The text of the waqfiyyāt reads as follows, with minor variations between different volumes:4 Bi’smi’llāh al-raḥmān al-raḥīm, hādhā mā ḥabasat Faḍl mawlāt Abī Ayyūb Aḥmad b. Muḥammad raḥimahu’llāh ṭ[ā]liban li-thawāb Allāh wa’l-dār al-ʾākhira. Raḥima’llāh man qaraʾa fīhā wa-daʿā li-ṣāḥibatihā wa-katabat Faḍl bi-khaṭṭihā fī’l-Muḥarram sanat khams wa-tisʿīn wa-māʾatayn. In the name of God, the clement, the merciful. This was endowed by Faḍl, mawlāt of Abū Ayyūb Aḥmad b. Muḥammad, God have mercy on him, seeking the reward of God and the abode of the afterlife. God, have mercy on whoever reads from it and prays for its owner. Faḍl wrote in her hand in Muḥarram of the year 295 [October-November 907]. Most writers have thus far assumed that Faḍl was the scribe of this Qur’an, which would probably entail a provenance in al-Qayrawān, even though the city is not named in the deed. This hypothesis stems from the statement that ‘Faḍl wrote in her hand’; yet the script of the waqfiyyāt is stylistically unrelated to the calligraphy of the Qur’anic text, and rather appears a precursor of the New Style, but with relatively uneasy execution. The strokes lack the clarity and assurance of Qur’anic calligraphy, and tend to be hazy. Within the few sentences of the waqfiyya, Faḍl also omitted several words or letters, which she then added between the lines.5 It seems unlikely that she was a professional scribe or that she wrote the main text: she probably acquired this manuscript from an unknown source and donated it to the Great Mosque of al-Qayrawān. Indeed, the waqfiyyāt refer to her as the ‘owner’ on whom blessings should be brought (the use of the feminine makes this association unambiguous). Thus the local origin of this Qur’an cannot be ascertained, and its chronology could potentially extend to the decades before the waqfiyyāt were drawn up. A radiocarbon analysis has been performed on a parchment sample from the manuscript in Lyon at the request of François Déroche: the test indicates a date range between 716 and 891 A.D., with a 95% probability.6 The upper limit of this range thus predates the waqfiyyāt by 16 years: this points to the possibility of an earlier production, without being sufficient in itself to establish it. At any rate, the manuscript Coloured Dots and the Question of Regional Origins (Part II) 77 cannot serve as firm evidence of regional trends in early Qur’ans. The main features of its vocalisation will nevertheless be outlined for reference: $ Red dots for vowels, hamza, and tanwīn; $ A red dot to the right of alif for hamza, and to its left for hamza preceded or followed by madd; $ Green for variants: green dots for variants involving vowels, the pronunciation or elision of a hamza, and for letters pronounced with imāla by some readers; a vertical green line for variants involving alif maḥdhūfa, e.g. al-mushāriq wa’l-mughārib instead of al-mashriq wa’l-maghrib in Q. 2:115 (fig. 1, lines 1–2; variant not listed in the literature). This vocalisation system is close to that of a Qur’an completed at Isfahan in 383/993,7 with the difference that several cases of variants are noted through the same green dot: as in the earlier Qur’ans studied previously,8 more reliance was thereby placed on the training of readers and their ability to identify the implications of each sign. The manuscript, on the other hand, does not match al-Dānī’s description of Maghribī vocalisation. Several explanations suggest themselves. Such vocalisation could conceivably have been used at al-Qayrawān, given its relative proximity to Egypt and the Mashriq. This possibility is brought to mind by the Qur’an of the Nurse, copied in that city in 410/1020, which notably uses green hamza in the modern form along with a range of other signs, including green horizontal strokes for ṣila.9 The existence of numerous fragments that do carry ‘Maghribī’ vocalisation in the al-Qayrawān collection might imply that different conventions coexisted in that centre. But just as plausibly, the Qur’an of Faḍl may have been produced in Egypt or the Mashriq and later brought to al-Qayrawān. Indeed, some fragments brought into European collections from the Great Mosque of Fusṭāṭ in the nineteenth century appear to match others in the al-Qayrawān collection, which suggests some degree of circulation between these two cities.10 Another witness of this circulation may be BNF Arabe 376b, which was made a waqf at the Great Mosque of Fusṭāṭ in 366/977: its script is an idiosyncratic derivative of the D styles, moving away from classical forms, and its vocalisation closely matches al-Dānī’s characterisation of Maghribī conventions, notably the use of a yellow dot for hamza, a green dot paired with a horizontal stroke for alif al-waṣl, and a vertical red stroke for alif maḥdhūfa.11 In the final analysis, because it lacks a definite geographical attribution, the Qur’an of Faḍl raises more questions than it answers for our present purposes. The Palermo Qur’an (Palermo, 372/983) This manuscript was written in a round version of the New Style in 372/983 in the capital of Muslim Sicily, then under the nominal rule of the Fāṭimid dynasty. A statement about the uncreated nature of the Qur’an in its opening illuminations suggests that it was 78 Journal of Qur’anic Studies Fig. 1. Page from the Qur’an of Faḍl (Q. 2:115–16). Paris, Bibliothèque Nationale de France, Arabe 5178m, f. 18r. Page dimensions 10.2 × 14.7 cm. produced by a Sunnī community.12 The verse markers are, in all likelihood, original, since an empty space has been left for them while writing the manuscript. An impressive amount of information has been fitted into a text area of only 9.5 × 16 cm: seventeen lines of text, extensive vocalisation, and in some pages illuminations. Their successful integration suggests a concerted design, an idea confirmed by the unusual care taken by the illuminator not to cover any part of the calligraphy, however small, even if this meant adapting his work to the form of existing letters (see the golden hāʾ in fig. 4, line 4). I was unable to observe any clear overlaps between the vocalisation and illumination that would provide information about their chronological sequence. Verse and fifth verse markers are decorated with green and red dots: from reproductions, the latter appear to be the same as in the vocalisation, to the extent that in cases of contiguity, it can be difficult to tell them apart (golden hāʾ, fig. 6, line 2). There is, in sum, every reason to think that the vocalisation and illumination are original. In the vocalisation: $ Red dots indicate vowels and tanwīn; $ Yellow dots indicate hamza; as a general rule, the dot is placed to the upper right of initial alif for hamza with fatḥa; to its upper left for hamza followed by madd; to its lower left for hamza with ḍamma; and directly below it for kasra; Coloured Dots and the Question of Regional Origins (Part II) 79 Fig. 2. Folio from the Palermo Qur’an (Q. 22:21–30). London and Geneva, Nasser D. Khalili Collection, QUR261, f. 4a. Page dimensions 17.6 × 25 cm. Reproduced courtesy of the Khalili Family Trust. Fig. 3. Detail of folio from the Palermo Qur’an (Q. 22:13). London and Geneva, Nasser D. Khalili Collection, QUR261, f. 3b. Reproduced courtesy of the Khalili Family Trust. Fig. 4. Detail of folio from the Palermo Qur’an (Q. 25:32–6). London and Geneva, Nasser D. Khalili Collection, QUR261, f. 6b. Reproduced courtesy of the Khalili Family Trust. 80 Journal of Qur’anic Studies Fig. 5. Detail of folio from the Palermo Qur’an (Q. 26:101–13). London and Geneva, Nasser D. Khalili Collection, QUR368, f. 10b. Reproduced courtesy of the Khalili Family Trust. Fig. 6. Detail of folio from the Palermo Qur’an (part of Q. 23:44 and Q. 23:45). London and Geneva, Nasser D. Khalili Collection, QUR261, f. 8b. Reproduced courtesy of the Khalili Family Trust. $ A red shīn indicates shadda; $ A vertical red line indicates alif maḥdhūfa (e.g. fig. 4, line 1, rattalnāhu); $ An oblique red line inclined to the right indicates cases of alif maḥdhūfa preceded by lām (e.g. fig. 3, al-mawlā); the sign is placed before the lām, even though it is pronounced after it;13 $ A red horizontal stroke in the position of the last short vowel of the previous word indicates ṣila, e.g. fig. 4, line 4, fa-qulnā ʾidhhabā, recited fa-qulnā’dhhabā; $ Blue dots indicate the position of the hamza on alif al-waṣl, for cases where the reader will have marked a pause (waqf) at the previous word (e.g. fig. 4, line 4, fa-qulnā [pause] ʾidhhabā);14 because they are linked to ṣila, these dots normally appear next to a red horizontal stroke; $ In la-biʾsa’l-mawlā wa-la-biʾsa’l-ʿashīr (Q. 22:13; fig. 3), the same two signs appear below the hamza of la-biʾsa; this seems to indicate two readings: one with hamza (blue dot), i.e. the standard convention, and the other with yāʾ (la-bīsa, red stroke), as in the reading of Warsh, Abū Jaʿfar, al-Iṣfahānī, al-Azraq, and Abū ʿAmr;15 $ A red semi-circle, sometimes with one red dot (or two red dots for tanwīn), indicates idghām; when pointing upwards, it is pronounced with fatḥa; when placed above the line and pointing downwards, with ḍamma; and when Coloured Dots and the Question of Regional Origins (Part II) 81 placed below the line and pointing downwards, with kasra; e.g. fig. 4, line 3, ʾulāʾika sharrun makānan wa-ʾaḍallu sabīlan (Q. 25:34), recited ʾulāʾika sharrum-makānaw-wa-ʾaḍallu sabīlan (letters carrying the sign are underlined; note that the word ʾaḍallu features an unusual instance of the same sign being used to simply mark shadda); $ Small red circles denote sukūn, e.g., rattalnāhu (fig. 4, line 1); ʾulāʾika (line 3); jaʿalnā (line 4); fa-qulnā (line 4); faʿalnā (line 4); $ Small yellow and blue circles denote different cases of sukūn followed or preceded by hamza (in which cases the latter is noted either through a yellow or a blue dot), e.g. fig. 2, line 12, yaʾtūka; fig. 4, line 1, wa-lā yaʾtūnaka; $ An oblique red or blue stroke, often associated with a blue circle and blue dot, marks cases in which the hamza is dropped in some readings, e.g. fig. 5, line 2, muʾminīn/mūminīn; line 7, nuʾmin/nūmin; fig. 6, line 1, yuʾminīn/yūminīn;16 $ Fāʾ/qāf could not be studied, since the manuscript only has very rare diacritical marks; these are noted in red, typically to distinguish yāʾ from tāʾ at the beginning of a verb. The manuscript corroborates al-Dānī’s assertions about the Maghrib in its fundamental aspects: the red, yellow, and blue dots; the small red circles for sukūn; and the vertical red lines for alif maḥdhūfa. Small divergences can also be noted: the ‘Iraqi’ convention of placing the dot to the left of initial alif for madd has been followed (although here this dot remains yellow, rather than red); a small shīn rather than a small dāl denotes shadda; the use of a blue dot and red stroke do mark ṣila, as al-Dānī asserts, but also the occasional variant. Some signs also fall beyond the scope of his observations: the red semi-circles for idghām; the blue and yellow circles for different cases of sukūn; and the cases of alif maḥdhūfa marked by an oblique red stroke, rather than a vertical one. Their role remains limited to amplifications on the same set of notational features. 4. A Kufic-Maghribī Corpus Scattered leaves from a small corpus of Kufic manuscripts display a distinctive array of orthographic features associated with the Maghrib (figs 8–12). I have identified over a dozen of these manuscripts, two of them dated to the late fourth to early fifth/ early eleventh century. Some were only published in black-and-white, but a comparison with colour images suggests that we are dealing with a consistent pattern. Throughout this sample one can observe: $ Red dots for vowels and tanwīn; $ A yellow dot for hamza; $ A short red horizontal stroke for sukūn; 82 Journal of Qur’anic Studies $ A short red stroke for ṣila, in the position of the last short vowel of the previous word; $ A green or blue dot for the hamza of alif al-waṣl, for cases in which the reader has stopped at the previous word; $ A vertical red line for alif maḥdhūfa; in some cases, this is a fully formed miniature alif with lower return (fig. 11; fig. 12, line 12); $ Semi-circles drawn in a thin red (or sometimes green) line for shadda: like the idghām signs of the Palermo Qur’an, they are open upwards for fatḥa; downwards and above the line for ḍamma; downwards and below the line for kasra; in some cases, the semi-circle encloses a red dot; I was unable to determine the reason for adding or omitting this dot; $ A long red horizontal stroke for madd, in some manuscripts only. These features match al-Dānī’s observations about Maghribī Qur’ans in almost every detail. The only discrepancy lies in the notation of shadda with a semi-circle, whereas he describes it as a dāl. Even this minor difference might be explained by the angle often seen at the inflection of the sign: this makes it resemble a dāl, which may have been its original form.17 Indeed the conventions that govern its use, with a sign opening upwards for fatḥa and placed below the letter for kasra, echo al-Dānī’s assertion that ‘mushaddad letters carry a dāl, and this dāl opens upwards’ and that ‘for kasr, [it was placed] below the letter’.18 This small corpus also features two conventions which specifically distinguished the Maghrib from Medina, according to al-Dānī: the use of a red stroke, rather than a red circle, for sukūn; and of a red stroke and green dot, rather than a red circle, for alif al-waṣl. Finally, the manuscripts rarely carry diacritical signs, but those that do have the dash or dot below fāʾ and above qāf.19 The vocalisation of these manuscripts thus points to a Maghribī origin, an idea supported by idiosyncratic features in their calligraphy and decoration that set them apart from the mainstream of the Kufic tradition. In several of the manuscripts, fifth verse markers are noted as a golden hāʾ, a common formula in Kufic, but distinguished here by an upper part in the shape of a semi-palmette (fig. 12, line 6); similar forms also appear in the Palermo Qur’an and in later Maghribī manuscripts.20 At the level of calligraphy, these Kufic Qur’ans display a relatively constant tendency to thicken the body of compact letters, such as wāw, qāf, hāʾ, rāʾ, and the curving horizontal part of final jīm and final yāʾ. This gives a basic visual unity to an otherwise heterogeneous group. The full description of these writing styles falls beyond the scope of the present study, and would require the analysis of larger manuscript samples. For the time being, a few specimens will be highlighted in order to give the reader a sense of the aesthetic tendencies involved. Coloured Dots and the Question of Regional Origins (Part II) 83 A one-volume Qur’an dated 393/1003 (fig. 8): This small manuscript with fourteen lines to the page is held at the Tareq Rajab Museum (Kuwait).21 In its calligraphy (fig. 7): $ Independent alif has a medium-size, rounded lower return, as in D.I; $ Nūn has a curved head and horizontal return of medium length; $ The two eyes of medial hāʾ are aligned vertically and grouped in the right half of the letter, whereas the left half is filled with ink; the letter is in the shape of a slightly inclined drop, and bulges below the baseline; $ Initial ʿayn has a thick lower stroke and thin upper stroke, joined at an approximate right angle. The style is thus a relatively close derivation of D.I, with slightly divergent forms for hāʾ and nūn. The emblematic monumentality of the original style has given way to miniature calligraphy executed in a thin pen, with a tendency towards sharper angles at junctions between strokes. Fragments from one or several other Qur’ans bear the same vocalisation and are written in a closely related script.22 The Rajab manuscript is dated through a colophon executed in yellow or gold Kufic script outlined in a thin black line, the common chrysography technique of the period (fig. 8). Its writing style is again close to D.I, but with rounded hooks that reach below the horizontal baseline, as in style D.V and in the Palermo Qur’an.23 It reads (with erased parts in curly brackets): Yāʾ wāw nūn. Fī’l-nabratayn al-maksūratayn qulta al-ʾūlā yāʾ. Wa fī’l-maḍmūmatayn qulta al-ʾūlā wāwan, wa fī’l-maftūḥatayn yusqaṭ24 al-ʾūlā, wa’l-maftūḥatayn min kalima wāḥida yudkhal {ʿalayhā} alifan. Wa-tamma fī shahr Rajab sanat thal[ā]th wa-tisʿīn wa-thal[ā]th māʾa fa-raḥima’llāh kātibahu wa-kāsibahu {…} wa-ṣalla (sic) Allāh ʿalā Muḥammad wa-ʾālihi wa-sallama taslīman. Yāʾ wāw nūn. In cases of two hamzāt with kasra, you say the first one as a yāʾ; for two [hamzāt] with ḍamma, you say the first one as a wāw; for two [hamzāt] with fatḥa, the first one is dropped; and in cases of two [hamzāt] with fatḥa within the same word, an alif should be introduced. This was completed in the month of Rajab of the year 393 [May–June 1003]. May God have mercy on its scribe and its owner; … may the peace and blessings of God be on Muḥammad and his family. The word nabra, at the beginning of the text, is a relatively rare synonym of hamza; it occurs, among other places, in the Muḥkam.25 The different cases of two consecutive hamzāt evoked here are also mentioned in al-Dānī’s treatise.26 84 Journal of Qur’anic Studies Fig. 7. Double page from a Kufic-Maghribi Qur’an dated 393/1003 (Q. 15:57–16:1). Kuwait, Tareq Rajab Museum, no shelfmark. Page dimensions 12.8 × 19 cm. Coloured Dots and the Question of Regional Origins (Part II) 85 Fig. 8. Colophon of a Kufic-Maghribi Qur’an dated 393/1003. Kuwait, Tareq Rajab Museum, no shelfmark. Page dimensions 12.8 × 19 cm. He notably records the convention, in some schools in the Maghrib, of adding a red alif or red stroke (maṭṭa) between two hamzāt that carry a fatḥa if these are in the same word; and of dropping the first hamza in cases of two hamzāt with fatḥa in consecutive words: these precisely correspond to the injunctions of the colophon. Al-Dānī goes on to discuss cases in which two consecutive hamzāt both carry a kasra (or ḍamma). However he only refers to the elision or softening of either hamza, as opposed to the instruction in the colophon to replace them with a full letter of the same sound (i.e. yāʾ for hamza maksūra, wāw for hamza maḍmūma). On the other hand, the latter readings do appear in the qirāʾāt literature, where they are attributed to Qālūn, an early Medinan authority with a wide following in the Maghrib repeatedly cited by al-Dānī. Indeed the word nabra appears in the Muḥkam next to one of these citations, to which it probably belongs.27 Thus these indications are in broad agreement with the idea of an origin in the Maghrib, whilst providing a first element of chronology for this Kufic corpus. The closing invocation in favour of the scribe and owner imply that they were two different persons, hence that the manuscript was either commissioned by a 86 Journal of Qur’anic Studies patron or produced for the market. The formula used to designate them (kātibahu wa kāsibahu) is the same as in a Maghribī Qur’an of 432/1040 (fig. 10), although too few other colophons of the period are known to determine whether this formulation was a specificity of the Islamic West. A Fragment at the Bibliothèque Nationale de Tunisie (fig. 9): In this fragment, independent alif has a long, relatively flattened lower return, as in D.III; this feature is shared by other specimens in the Kufic-Maghribī corpus.28 The round letters, notably wāw and final hāʾ, have particularly thickened bodies; they are often of nearly the same height as the tall letters, such as alif and lām. Final nūn is thickened at the top and base, and has a thin vertical shaft. This writing style is closely related to that of a colophon written on 6 Ṣafar 432/16 October 1040 (fig. 10).29 It records the completion of a Qur’an written in an early type of ‘classical’ Maghribī calligraphy, of which the last few lines can be seen on the same page. This document thus confirms a link between this style and the Maghrib, in addition to providing a second chronological pointer. The colophon is (expectedly) unvocalised, but the elements of vocalisation that can be seen in the adjoining Maghribī calligraphy (namely shadda, ṣila, and sukūn) do follow the Kufic-Maghribī conventions outlined above. Fig. 9. Qur’an folio in Kufic–Maghribi calligraphy (Q. 6:83–6). Tunis, Bibliothèque Nationale de Tunisie, unknown shelfmark. Page dimensions 18.2 × 25.5 cm. Coloured Dots and the Question of Regional Origins (Part II) 87 Fig. 10. Final page of a Maghribi Qur’an dated 432/1040 (Q. 113:4–Q. 114:6 and colophon). Istanbul, Türk ve İslam Eserleri Müzesi, ŞE13644/1. Page dimensions 14.4 × 17.5 cm. BNF Arabe 6982 (fig. 11): This manuscript has a long, rounded and tapering lower return for independent alif, which recalls D.I; a medial hāʾ that bulges below the baseline, as in B.II and D.IV; a widely open initial ʿayn reminiscent of in D.III. This eclectic style is further distinguished by the uniformly accentuated thickness of its strokes and its prominent triangular endings, as in final nūn, in independent bāʾ/tāʾ and in the upper stroke of dāl and initial kāf. Closely related writing styles occur in a manuscript that belonged to the Madrasa of Ibn Yūsuf (Marrakesh) in the early twentieth century and a quire offered for sale at Drouot (Paris) in 1975.30 Khalili KFQ70 (fig. 12): In this manuscript, alif has a variable lower return, from short to long. Final mīm has a round body and a very thin tail, either straight or curved. Final nūn has a thickened head and nearly no lower return. Small rounded hooks occasionally reach underneath the baseline for letter connections, as in the Palermo Qur’an (fig. 12, middle of line 9, 88 Journal of Qur’anic Studies Figure 11. Folio from a Kufic-Maghribi Qur’an (Q. 13:43–Q. 14:5). Paris, Bibliothèque Nationale de France, Arabe 6982, f. 21v. Page dimensions 20 × 28 cm. ghaniyyan fa’l-yastaʿfif). Closely related fragments have been sold at London auction houses.31 Additional manuscripts bearing the same vocalisation were written in scripts representing further variations on classical Kufic. For example, BNF Arabe 337c has some formal characteristics that suggest a parallel with D.IV, but its thickened strokes and accentuated angles result in a distinct style.32 In a leaf from a private collection, the writing style is loosely reminiscent of E.I, but with slanting vertical strokes, an amplified return of final nūn that evokes the beginning of a loop, and a lower return for independent alif, whereas E.I is characterised by the absence of this return.33 In all of these scripts, elements rooted in classical Kufic calligraphy are evident, but they have been reshuffled and more or less radically transformed to give rise to a different aesthetic: these derivative styles must have evolved out of the D group in Spain or the broader Maghrib. More examples probably remain to be found in the collections of al-Qayrawān, Marrakesh, and Fes. The available elements of date converge to place these manuscripts around the lifetime of al-Dānī, in the fourth to fifth/tenth to eleventh century, although some may be earlier or later. These findings bring to mind a remark by the historian Ibn al-Fayyāḍ, who wrote in the fifth/eleventh century: ‘In the eastern suburb of Cordoba, there were 170 women Coloured Dots and the Question of Regional Origins (Part II) 89 Fig. 12. Qur’an folio in Kufic-Maghribi calligraphy (Q. 4:3–7). London and Geneva, Nasser D. Khalili Collection, KFQ70, verso. Page dimensions 18.3 × 25.5 cm. Reproduced courtesy of the Khalili Family Trust. who copied the Qur’an in Kufic calligraphy.’34 Although he does not elaborate on the meaning of ‘Kufic’, Ibn al-Fayyāḍ might have been referring to the above script types, which belong to the same period and have clear stylistic connections to the central Islamic lands. This passage also highlights the existence of women copyists of the Qur’an in the Maghrib; those mentioned here appear to have worked for the market, and were presumably not religious scholars. 3. Classical Kufic Fragments Vocalised in the Medinan-Maghribī Tradition Some manuscripts belonging squarely in the Kufic tradition do also display distinctive Medinan-Maghribī traits in their vocalisation. Thus the catalogue of early Qur’ans at BNF lists a dozen manuscripts with red dots for their vowels and yellow dots for hamza.35 Déroche classified these under various styles: B.II (Arabe 329d, 6982); D.I (Arabe 348a, 354b, 366b, 5178c); D.II (Arabe 349e, 352b, 356h, 378); D.IV (Arabe 337c), and common D (Arabe 343, 376b). But some of these specimens are in fact derivative in their calligraphy, as shown above for Arabe 6982 and Arabe 337c; the same is true of Arabe 343 and Arabe 376b, written in distinctive styles related to D. Amongst classical Kufic fragments proper, three configurations can be distinguished: manuscripts with extensive Maghribī vocalisation, comparable with that of the above corpus; manuscripts with more limited Maghribī vocalisation; manuscripts with vocalisation signs essentially reduced to red and yellow dots. 90 Journal of Qur’anic Studies Kufic Manuscripts with Full Maghribī Vocalisation A few classical Kufic fragments display the full range of conventions observed in the above Kufic-Maghribī corpus. This is the case with CBL Is. 1411, written in a style close to D.III (fig. 13); its Maghribī vocalisation has already been noted by Sheila Blair.36 Here one encounters yellow hamza; shadda marked with a semi-circle, including cases in which the tashdīd is caused by idghām; a blue dot and horizontal red stroke for ṣila; and a horizontal red stroke for sukūn. Alif maḥdhūfa is written in red. Interestingly, takhfīf and ḥuruf zawāʾid are noted with a small red circle, the convention described by al-Dānī for Medina and the early Maghrib. For example, the phrase ʾa-raʾayta (Q. 96:9 and 11; fig. 13, f. 6b, lines 5 and 8) has a red alif after the rāʾ and a small red circle before the yāʾ: this reflects the reading ʾā-rāyta, with the addition of one alif and the elision of a hamza (al-Azraq ʿan Warsh).37 The manuscript has fifth verse markers consisting of a golden hāʾ ending in a semipalmette, which again fit the decorative pattern set out above. The sura illuminations are largely conventional, their marginal ornament being made of symmetrical palmettes articulated around a central stem, although the colour palette is richer than normally in Kufic. In a slight variation on this pattern, a fragment in D.II (BNF Arabe 378) has red dots for vowels and tanwīn, yellow dots for hamza, a red shīn for shadda, a vertical red stroke for alif maḥdhūfa, a horizontal red stroke for madd, and horizontal green strokes for ṣila.38 The fifth verse markers are, again, comparable to the above. Two explanations may be suggested: either these manuscripts are earlier than the above corpus, and belong to a phase when the differentiation of Maghribī styles had not begun in earnest; or they are contemporaneous with these styles but reflect the production of a centre closer to the central Islamic lands, such as al-Qayrawān. It is, at this stage, not possible to bring an answer to this question. Kufic Manuscripts with Limited Maghribī Vocalisation Several other Qur’ans written in conventional Kufic styles have a more limited range of vocalisation that still reflects Maghribī conventions. Their most essential features are the yellow dots for hamza and red dots for vowels and tanwīn; but the notation goes a step further than this in the distinction from Medina and the Mashriq. Thus Khalili QUR372 is written in style D.II, but with thickened letter bodies that prefigure the aesthetic tendencies of the above corpus. In the vocalisation, in addition to the red vowels and yellow hamzāt, a blue dot combined with a thin horizontal red dash mark ṣila together with the position of hamza in cases of waqf, as in the Palermo Qur’an. An oblique red stroke followed by a blue dot appears on the wāw of the words yuʾmin, muʾmina, yuʾminū (fig. 14, right-hand page, line 6; left-hand page, lines 1 and 6): it appears to indicate the elision of hamza Coloured Dots and the Question of Regional Origins (Part II) Fig. 13. Double page from a Kufic Qur’an in style D.III with Maghribi vocalisation (right, Q. 96:2–16; left, Q. 102:1–Q. 103:3). Dublin, Chester Beatty Library, Is. 1411, ff. 6b–7a. Page dimensions 12.8 × 20.2 cm. # The Trustees of the Chester Beatty Library, Dublin. 91 92 Journal of Qur’anic Studies and its replacement by a wāw sākina, as evoked above for the Palermo Qur’an.39 As already noted by Déroche, the full-page illuminations also closely resemble those of the Palermo Qur’an, which reinforces the presumption of a Maghribī origin.40 This range of vocalisation occurs again in a very similar Qur’an in style D.II (BNF Arabe 349e). Shadda is marked by a small red shīn. Once again the blue dot and red stroke sometimes indicate, besides cases of waṣl, readings in which the hamza may be softened or elided.41 In at least one place (f. 113r), a lām omitted in standard orthography has been added to the word alladhīna, even though this would have been obvious to any fluent Arabic speaker. Nearly the same pattern can be observed in BNF Arabe 352b and Arabe 356h, two fragments each consisting of a single folio, both also in D.II: here one encounters yellow dots for hamza, a thin horizontal stroke for ṣila and a thin vertical stroke for alif maḥdhūfa. Blue dots are absent from both folios, although the fragments are too small to conclude that they did not exist in their original manuscripts. Both fragments again have shaddāt in the form of a red shīn: this sign, which also occurs in the Palermo Qur’an, was probably more widespread in the Maghrib than al-Dānī’s treatise would imply. The type of semi-palmette identified in the Kufic-Maghribī corpus can again be seen in Arabe 349e (e.g. f. 106r) and Arabe 352b (f. 2r). Five of the six Kufic manuscripts considered so far were written in style D.II, the exception being CBL Is. 1411. Thus the vocalisation seems to corroborate the hypothesis raised by Déroche several years ago on the basis of illumination that this small stylistic group originated in the Maghrib.42 However a minority of D.II specimens are vocalised in red or red and green, either because some schools in the same region used different conventions, or because D.II was also written further east.43 Kufic Manuscripts with Medinan-Maghribī Vocalisation Several fragments in style D.I, all of them with seven lines to the page, but with different dimensions, display vocalisation based solely on red dots for the vowels and tanwīn and yellow dots for hamza.44 Their sober, masterly D.I script laid down on large parchment leaves represents the apex of the classical Kufic tradition, much as in the Qur’ans of Amājūr and ʿAbd al-Munʿim.45 In these fragments (as also in Khalili QUR372 and BNF Arabe 349e), the yellow dot is placed before alif for hamza and after it for hamza followed or preceded by madd – the same convention applied with red dots in Mashriqī Qur’ans.46 Diacritics tend to be sparse, and I was unable to observe a fāʾ or qāf amongst the available pages.47 Alif maḥdhūfa, waṣl, waqf, and shadda are simply noted as red dots, as in the Iraqi-Mashriqī tradition. Coloured Dots and the Question of Regional Origins (Part II) Fig. 14. Double page from a Kufic Qur’an in style D.II (Q. 2:220–1). London and Geneva, Nasser D. Khalili Collection, QUR372, ff. 67b–68a. Page dimensions 28.2 × 20.6 cm. Reproduced courtesy of the Khalili Family Trust. 93 94 Journal of Qur’anic Studies A leaf at the Metropolitan Museum of Art (fig. 15) has additional green dots and dashes to mark variants, notably: $ Line 4 (Q. 47:37), yukhrij aṣghānakum (red, most readers) / takhruj aṣghānukum (green, Ibn ʿAbbās, Mujāhid, and others); $ Line 5 (Q. 47:38), hā ʾantum with a green and yellow dot on the alif; several possible readings exist for this phrase; the ones noted here might be ha ʾantum (yellow hamza, no madd; Ibn Mujāhid, Ibn Shannabūdh, and others) and hā ʾantum (green hamza with madd; Ibn Kathīr, Ibn ʿĀmir, ʿĀṣim, Ḥamza, al-Kisāʾī, and others).48 Such uses of green have otherwise been observed in Mashriqī rather than Maghribī vocalisation.49 Finally, in the same page, a blue dot sits on the line at the end of the word ʾantum, possibly to mark ḍamm mīm al-jamʿ. These D.I fragments are thus close not only in script type, but also in vocalisation, to the Qur’ans of Amājūr and ʿAbd al-Munʿim, the main difference being their yellow hamzāt. This suggests that they reflect a relatively early stage in the process of differentiation between vocalisation systems: more than any other manuscripts reviewed so far, they lend themselves to the hypothesis of a Medinan origin, and echo al-Dānī’s assertion that red and yellow were used at the exclusion of any other colour Fig. 15. Page from a Qur’an in style D.I (Q. 47:37–8). New York, Metropolitan Museum of Art, 30.45. Page dimensions 23.8 × 33.3 cm. Coloured Dots and the Question of Regional Origins (Part II) 95 in the vocalisation of Medina. But one cannot completely rule out, at this stage, an early derivation in the Maghrib, or indeed elsewhere. 5. Beyond al-Dānī’s Horizon Al-Dānī’s descriptions of different vocalisation systems have, in sum, proven remarkably accurate when confronted with extant manuscripts. His primary aim was to highlight the main conventions of his own region, to establish their Medinan roots, and to distinguish them from the customs of Iraq, the beating heart of the religious and grammatical sciences. As a consequence, there are many gaps left in our grasp of this phenomenon. A significant number of early Qur’ans do not carry any vocalisation at all. Expectedly, many others do not fall within al-Dānī’s categories, either because they are from other regions or because their notation was only used in some circles in the regions that he does mention. Numerous specimens, for example, have red dots for vowels, green dots for hamza, and in some cases blue or yellow dots for shadda.50 This echoes the statement in Ibn Mujāhid’s Kitāb al-naqṭ, as cited by al-Dānī, that some vocalisers marked vowels with red dots, hamza with green dots, and shadda with yellow dots for personal copies of the Qur’an (hence, one might infer, not for public copies).51 While this convention therefore seems to have been known in Iraq, we do not know whether it was widespread in this region or specific to it. Some Qur’ans also fall completely outside the framework provided by al-Dānī: thus in one manuscript written in ten lines of gold D.IV to the page, fatḥa is noted in red above the line, ḍamma in green in front of the line, and kasra in blue below the line.52 This brings to mind another assertion by Ibn Mujāhid, again cited in the Muḥkam, that ‘readers comprehend this [dots in different colours] more quickly than dots in one colour’.53 But again it cannot be linked to a specific region. Conclusion Al-Dānī’s treatise, while not exhaustive, makes it possible to raise hypotheses about certain types of vocalisation. These can be briefly summarised as follows: $ A Qur’an vocalised with red dots alone for vowels, tanwīn, and hamza follows Iraqi conventions, and may have been produced in Iraq, Iran, or Greater Syria; $ A Qur’an vocalised with red dots for vowels and tanwīn but with yellow dots for hamza follows Medinan conventions, and may have been produced in Medina, or possibly the early Maghrib; additional signs in a thin red pen might have sometimes been used in Medina and the early Maghrib; $ A Qur’an vocalised with red dots for vowels and tanwīn, yellow dots for hamza, a green or blue dot and/or a red stroke for ṣila, and a red stroke or circle for sukūn follows Maghribī conventions, and may have been produced in the Maghrib; 96 Journal of Qur’anic Studies $ A Qur’an vocalised with red dots for vowels, tanwīn, and hamza, and with the addition of green dots and signs for variants, may have been produced in Iraq, Iran or Greater Syria; $ The presumption of a Maghribī origin may be reinforced by other features associated with this region by al-Dānī; by the notation of fifth verse markers as a stylised hāʾ resembling a vertical semi-palmette; and by the notation of fāʾ and qāf with a single diacritical sign placed respectively below or above the letter; $ The presumption of a Mashriqī origin may be reinforced by the notation of hamza through a red dot to the right of alif; by the notation of hamza followed or preceded by madd though a red dot to the left of alif; by the absence of a red dot for alif alwaṣl; and by the notation of fāʾ and qāf respectively with one or two dots above the letter. Some limitations ought to be recognised. First, the number of dated and localised manuscripts available for study remains small; one can only hope that more of these will eventually come under scrutiny. It is the strong convergence of their notation with the assertions of al-Dānī’s treatise that makes it possible to draw some inferences. Second, unless overlaps with the illumination can show that the vocalisation is original, any attribution made on this basis should, in theory, be applicable to the vocalisation rather than the whole manuscript. In practice, most manuscripts were probably vocalised in their region of origin, whether at the time of production or afterwards. Some may have travelled and eventually have received this layer of notation in a different region of the Islamic world; but the proportion of early Qur’ans produced without any vocalisation in one region, then carried into a different region and vocalised there must have remained relatively small. What can be reached, in cases where no overlaps have been observed, is therefore a reasonable hypothesis rather than an absolute certainty. In the same line of thought, both masters and pupils also travelled, which may have led certain vocalisers to adopt conventions different from the dominant usage of their region, as suggested by al-Dānī’s constant references to the usage of a ‘majority’ (ʿāmmat ahl baladinā, ʿāmmat ahl al-mashriq), and his allusions to different schools, especially in Iraq. For example, the Baghdadis and Syrians whom he met in Cairo may notionally have vocalised manuscripts there, and have gained a local following, even if limited; conversely, they could (just as theoretically) have learned Egyptian conventions (whatever these may have been) and brought them back to a given circle in their respective homelands. The nuances observed in the application of the same basic formulae, for example in the Kufic-Maghribī corpus and in related classical Kufic manuscripts, do suggest localised usages, some of which must have been associated with a given teacher and his circle. At the image of al-Dānī, religious scholars were particularly prone to travel Coloured Dots and the Question of Regional Origins (Part II) 97 to Mecca for the ḥajj, and to seek knowledge of the Qur’anic sciences from respected authorities as they journeyed. An individual could thereby emulate a convention learned at the other end of the Islamic world in his region of origin, and eventually bring his own modifications to it, in a movement sustained in different directions, generation after generation. Books carried by these people could also serve as a record of different conventions. An awareness of such complexities remains necessary when analysing these patterns. Interestingly, the above evidence has only suggested few clear correlations between vocalisation and script types. It seems probable that D.I was common in the central provinces of the ʿAbbāsid empire, in a range of regions comprising Syria, the Hijaz, and presumably Iraq; but it remains difficult to assess how widespread this style was elsewhere. The closely related D.II, as we have seen, bears a distinct affinity with the Maghrib, though it is again too early to say whether it was exclusively written in this region. A manuscript with extensive Maghribī vocalisation has been noted in D.III, but many others in this style also use the Mashriqī system.54 Thus certain styles may have been more prominent in some regions than in others, but their geographical range might still have been wide for classical Kufic and indeed for NS. The main exception to the rule, the Kufic-Maghribī corpus, with its distinctive blend of calligraphy and vocalisation, is broadly datable to the fourth to fifth/tenth to eleventh century; its derivation from classical Kufic styles suggests that the latter might initially have been practised in the same area. In NS, some manuscripts written in a similar style to that of the Palermo Qur’an are vocalised solely in red;55 and the famous Qur’an of the Nurse was copied in 410/1020 at al-Qayrawān using a monumentalised script comparable to that of the Isfahan Qur’an. Finally, the aural dimension of early Qur’ans is brought into relief by the study of their vocalisation. Modern scholarship has tended to approach these manuscripts as the record of a text and as a flat canvas crafted for the eyes. But these were full objects meant to be displayed, studied, and used for recitation. They conveyed not only an image, but also a sound that carried into the public sphere, at the heart of daily life in cities and towns. The contrast between these two expressions of the Qur’an – the aural and the visual – was expressed through this aphorism cited by Ibn Abī Dāwūd (d. 316/929) on the authority of early religious scholars: ‘Its most beautiful ornament is its correct recitation.’56 These manuscripts await to be more fully studied as witnesses of the history of the Qur’anic text and its recitation. NOTES 1 For al-Dānī’s statements about the Maghrib, see George, ‘Coloured Dots: Part I’, pp. 7–11 and Table 1. 2 Ross, ‘Some Rare MSS Seen in Tunis’, pp. 614–15 and Pl.; Shabbūḥ, ‘Sijil qadīm li-maktabat jāmiʿ al-Qayrawān’, p. 339 and Pl. 1; Roy, Poinssot, and Poinssot, Inscriptions 98 Journal of Qur’anic Studies arabes de Kairouan, p. 34 (No. 9 bis); ʿAbd al-Wahhāb, Shahīrāt al-Tūnisiyyāt, pp. 30–1; al-Wartatānī, al-Burnus fī Bārīs, p. 202. One page attributed to this manuscript has also been published by al-Rammāḥ, ‘Tasāfīr maktabat al-Qayrawān al-ʿatīqa’, p. 142 and Pl. 2. But its calligraphy and vocalisation seem to differ from those of the Qur’an of Faḍl: since the quality of the reproduction does not allow a detailed analysis, it has been left aside for our present purposes. 3 I thank François Déroche for bringing the Paris fragment to my attention, and for generously sharing his knowledge of this manuscript, including most of the above bibliographical references. For a brief description of Arabe 5178m, see Déroche, Les manuscrits du Coran (I, 1), p. 124 (No. 197). The full fragment can be consulted on Gallica, gallica.bnf.fr (accessed 29/09/2012). 4 Reading by the present writer from unpublished images by François Déroche. See also Ross, ‘Some Rare MSS Seen in Tunis’, Pl. (where the date was erroneously read as 275); Roy, Poinssot, and Poinssot, Inscriptions arabes de Kairouan, p. 34; ʿAbd al-Wahhāb, Shahīrāt al-Tūnisiyyāt, pp. 30–31. 5 For example amongst the waqfiyyāt of R64a, three of the four letters forming the word Allāh were initially forgotten in two different pages, as was the name of the month in another page; all of them were then added above the line, in the same hand. 6 Personal communication from François Déroche; unpublished analysis performed at the Centre de Datation pour le Radiocarbone, CNRS UMR 5138, Lyon (France). 7 See George, ‘Coloured Dots: Part I’, pp. 27–33. 8 See George, ‘Coloured Dots: Part I’, pp. 17–21. 9 See for colour images from different collections Roxburgh, Writing the Word of God, figs 12–13 and cover images; http://www.metmuseum.org/Collections/search-the-collections/ 140014390 (accessed 20/12/2012); http://www.davidmus.dk/en/collections/islamic/materials/ calligraphy/art/252003 (accessed 20/12/2012). About the manuscript, cf. also Déroche, Le livre manuscrit arabe, p. 54. 10 Personal communication from François Déroche. 11 For a description, see Déroche, Les manuscrits du Coran (I, 1), p. 129 (No. 216). 12 Déroche, The Abbasid Tradition, p. 146 (No. 81); Déroche, ‘Cercles et entrelacs’, pp. 596–604; George, The Rise of Islamic Calligraphy, pp. 142–3. 13 The rationale for distinguishing this case remains unclear. Some but not all of these cases may have been pronounced with imāla – for a definition, see Levin, art. ‘ʾImāla’. 14 As already noted by Dutton, ‘Red Dots, Green Dots: Part 2’, p. 13. 15 al-Khaṭīb, Muʿjam, vol. 6, p. 89. 16 al-Khaṭīb, Muʿjam, vol. 1, p. 30, p. 310. 17 This is particularly clear in BNF Arabe 337c, where this sign appears in green (images available on Gallica). 18 See George, ‘Coloured Dots: Part I’, p. 8. 19 E.g. Christie’s (London), 14 October 1997, Lots 39, 40; Bonhams (London), 21 April 1999, Lot 513. 20 See also Christie’s, 20 April 1999, Lots 303, 304; Bonhams, 21 April 1999, Lot 513. For later Maghribī examples, see Déroche, Les manuscrits du Coran (I,2), Pl. III–B (c. 807/1405), XIV–B (703/1304), XV–B (ninth/fifteenth century); Black and Saidi, Islamic Manuscripts, p. 17 (Cat. 5, sixth/twelfth century); Roxburgh, Writing the Word of God, figs 14–15 (both seventh–eighth/thirteenth–fourteenth century). Coloured Dots and the Question of Regional Origins (Part II) 99 21 Safwat et al., The Harmony of Letters, pp. 36–7. 22 Christie’s (London), 14 October 1997, Lots 38, 41; Christie’s (London), 20 April 1999, Lot 304 (now in Kuwait, Dar al-Athar al-Islamiyyah, LNS 318 MS); Christie’s (London), 10 October 2000, Lot 4. 23 I thank Mr. Rajab for his permission to publish this image and for sharing the initial transcription which he had carried out. 24 The two diacritical dashes below the initial yāʾ are a grammatical error: this word should presumably read tusqaṭ. 25 George, ‘Coloured Dots: Part I’, p. 36 (n. 38). 26 al-Dānī, al-Muḥkam, p. 97, p. 110, p. 112. 27 al-Khaṭīb, Muʿjam, vol. 1, p. 76; vol. 8, p. 513; al-Dānī, al-Muḥkam, p. 8. It is difficult, in this case, to determine the exact end of the citation: the editor placed it at the end of the previous sentence, but it could equally include this word. 28 The Tunis fragment was published in Paris, De Carthage à Kairouan, Cat. 354. For related fragments, see Bonhams (London), 11 October 2000, Lots 3a, 3b. 29 Déroche, ‘Deux fragments’, pp. 231–2. 30 Bergsträsser and Pretzl, Geschichte, Abb. 3; Boisgirard, Hôtel Drouot (Paris), 23 June 1975, Lot 104. 31 For KFQ70, cf. also Déroche, The Abbasid Tradition, p. 123, p. 125 (No. 68). Related fragments: Bonhams (London), 17 October 2001, Lot 1; Christie’s (London), 28 April 1998, Lot 19; Quaritch (London), Catalogue 1366, Islamic manuscripts (2008), item L. 2. 32 This singular aspect was noted by Déroche, Les manuscrits du Coran (I, 1), p. 109 (No. 146). Images consulted on Gallica. 33 Christie’s (London), 14 October 1997, Lot 39. For dated waqfiyyāt in style E.I, see Déroche, ‘Collections de manuscrits anciens’, Pl. Ia, IVa. 34 ‘Kāna bi’l-rabaḍ al-sharqī min Qurṭuba māʾa wa-sabʿūn imraʾa kulluhunna yaktubna al-maṣāḥif bi’l-khaṭṭ al-Kūfī’. Cited from the History of Ibn al-Fayyāḍ by al-Marrākushi, Kitāb al-muʿjib fī talkhīṣ akhbār al-Maghrib, p. 270. Cf. also Imamuddin, Muslim Spain 711–1492 A.D., p. 143; Déroche, Le livre manuscrit arabe, p. 49. 35 Déroche, Les manuscrits du Coran (I, 1). Déroche’s script classification is laid out in Déroche, Les manuscrits du Coran (I, 1); Déroche, The Abbasid Tradition; George, The Rise of Islamic Calligraphy, Appendix. 36 Blair, Islamic Calligraphy, pp. 122–4. Lām-alif, in this manuscript, has two arms that cross in the shape of an X in a way that could recall B.II; the horizontal tapering of intial jīm and ʿayn is also less marked than in classical D.III; nevertheless in its other aspects, the script of Is. 1411 remains close to D.III. For colour images, see Wright, Islam, pp. 102–3 (figs 65–6). Another dispersed specimen written in style D.Vc was noted by Blair as a possible Maghribī Qur’an (Blair, Islamic Calligraphy, pp. 121–2). However in this manuscript, hamzat al-qaṭʿ is noted in green rather than yellow, whilst alif al-waṣl is left unvocalised; in the page published by Blair, the word rabwatin(Q. 23:50) has three dots, blue, red, and yellow on the letter rāʾ (which carries neither a hamza nor a shadda): this reflects a convention different from what has been observed in the Maghribī tradition. 37 al-Khaṭīb, Muʿjam, vol. 10, p. 506, p. 507. See also for instance the red circles on f. 5a, lines 4, 6, ʾulāʾika (unpronounced wāw); f. 6a, line 3, radadnāhu (Q. 95:5, no shadda). 38 Arabe 378 can be seen on Gallica and in Guesdon and Vernay-Nouri, L’art du livre arabe, p. 90 (Cat. 59). Cf. also Déroche, Les manuscrits du Coran (I, 1), p. 99 (No. 118). 39 The rule is set out in al-Khaṭīb, Muʿjam, vol. 1, p. 30. 100 Journal of Qur’anic Studies 40 Déroche, The Abbasid Tradition, pp. 72–5 (no. 24). See also the images in Piotrovsky, Earthly Beauty, Heavenly Art, p. 110 (Cat. 54). 41 E.g. f. 109v, line 6: man shāʾa ʾan (Q. 25:57) can also be read man shā ʾan or man shāʾ ān, with the elision or softening of one hamza; cf. al-Khaṭīb, Muʿjam, vol. 6, p. 368. The word yuʾmin and its derivatives do not appear in Arabe 349e, and thus could not be checked. 42 Déroche, The Abbasid Tradition, p. 37; Déroche, ‘Cercles et entrelacs’, pp. 605–6. 43 See Déroche, Les manuscrits du Coran (I, 1), p. 98 (Arabe 325e, with green hamza and orange shadda; Arabe 352d, red alone); Paris, De Carthage à Kairouan, Cat. 339 (red alone). 44 BNF Arabe 5178c, esp. f. 5v (accessed on Gallica, 12/10/2012); double folio from anonymous private collection, Roxburgh, Writing the Word of God, fig. 6; folio from the Mohammad Afkhami collection, Sotheby’s (London), 24 October 2007, Lot 8; and George, The Rise of Islamic Calligraphy, cover image; Cambridge, MA, Harvard University Art Museums, Sackler Museum, 1956.197; Roxburgh, Writing the Word of God, fig. 7; New York, Metropolitan Museum of Art, 30.45, http://www.metmuseum.org/Collections/search-thecollections/140005627 (accessed 10/07/2012). The Harvard and New York fragments were previously in the collection of Rudolph M. Riestahl but do not appear to be from the same manuscript, as the latter is markedly larger than the former. 45 George, ‘Coloured Dots: Part I,’ pp. 17–21. 46 E.g. Roxburgh, Writing the Word of God, fig. 6, right–hand page, ʾābāʾakum ʾaw ʾashadda dhikran(Q. 2:200): the first two hamzāt, which correspond to cases of madd, are after the alif; the last two, which do not, are before the alif. In the Harvard fragment, one green dot appears in the phrase wa’llāhu yaʿlamu mā fī al-samāwāt (Q. 49:16). Its function could not be ascertained from a single occurence. 47 Initial letters occasionally have their diacritics marked in red, e.g. George, The Rise of Islamic Calligraphy, cover image, line 5, tabtaghū (Q. 2:198). 48 Khaṭīb, Muʿjam, vol. 9, pp. 34–7. 49 George, ‘Coloured Dots: Part I,’ pp. 17–20, pp. 26–7, pp. 32–3. 50 For example, Kuwait, Maṣāḥif Ṣanʿāʾ, p. 39 (No. 59); Piotrovsky, Earthly Beauty, Heavenly Art, p. 108 (Cat. 49). See also Dutton, ‘Red Dots, Green Dots: Part 1’, pp. 132–3 (Nos 8, 11); Dutton, ‘Red Dots, Green Dots: Part 2’, p. 1 (No. 15), pp. 3, 7–8 (No. 17), pp. 10– 11 (No. 21). In the latter two cases, yellow dots are also used for variants and other functions. 51 al-Dānī, al-Muḥkam, p. 23; Dutton, ‘Red Dots, Green Dots: Part 1’, p. 119. 52 Christie’s (London), 9 October 1990, Lot 44; Sotheby’s (London), 12 October 2000, Lot 3; Paris, De Carthage à Kairouan, Cat. 337. Cf. also Déroche, Les manuscrits du Coran (I, 1), p. 87 (Arabe 348e, No. 71), for an example of the same system with a different colour pattern. 53 al-Dānī, al-Muḥkam, pp. 23–4. 54 Cf. Déroche, Les manuscrits du Coran (I, 1), p. 103 (Arabe 349c), p. 105 (Arabe 357), p. 106 (Arabe 364b, 375d, 381e), p. 107 (Arabe 5178k). 55 For example, Déroche, The Abbasid Tradition, pp. 138–9 (no. 75). 56 George, ‘Calligraphy, Colour and Light’, p. 103. Bibliography ʿAbd al-Wahhāb, Ḥasan Ḥusnī, Shahīrāt al-Tūnisīyyāt: baḥth tārīkhī adabī fī ḥayāt al-nisāʾ al-nawābigh bi’l-quṭr al-Tūnisī (Tunis: al-Maṭbaʿa al-Tūnisiyya, 1966). Coloured Dots and the Question of Regional Origins (Part II) 101 Bergsträsser, Gotthelf, and Otto Pretzl, Geschichte des Qorâns von Theodor Nöldeke. III: Die Geschichte des Korantexts, 2nd edn (Leipzig: Dieterichische Buchhandlung, 1938). Black, Crofton, and Nabil Saidi, Islamic Manuscripts, Catalogue 22, Sam Fogg Rare Books & Manuscripts (London: Sam Fogg Rare Books and Manuscripts, 2003). Blair, Sheila, Islamic Calligraphy (Edinburgh: Edinburgh University Press, 2006). al-Dānī, Abū ʿAmr ʿUthmān b. Saʿīd, al-Muḥkam fī naqṭ al-maṣāḥif, ed. ʿIzzat Ḥasan, 1st edn (Damascus: Wizārat al-Thaqāfa wa’l-Irshād al-Qawmī, 1960). ——, al-Muqniʿ fī rasm maṣāḥif al-amṣār, maʿa Kitāb al-naqṭ, ed. Otto Pretzl (Beirut: Orient-Institut, 2009). Déroche, François, Le livre manuscrit arabe: Préludes à une histoire (Paris: Bibliothèque Nationale de France, 2004). ——, ‘Cercles et entrelacs: format et décor des corans maghrébins médiévaux’, Comptes Rendus des Séances de l’Académie des Inscriptions et Belles Lettres (2001), pp. 593–620. ——, The Abbasid Tradition: Qurʾāns of the 8th to 10th Centuries A.D. (London: The Nour Foundation, 1992). ——, ‘Deux fragments coraniques maghrébins anciens au Musée des Arts Turc et Islamique d’Istanbul’, Revue des Etudes Islamiques 59 (1991), pp. 229–35. ——, Les manuscrits du Coran: Du Maghreb à l’Insulinde, Catalogue des manuscrits arabes Deuxième partie: manuscrits musulmans. Tome I, 2 (Paris: Bibliothèque Nationale de France, 1985). ——, Les manuscrits du Coran: Aux origines de la calligraphie coranique, Catalogue des manuscrits arabes. Deuxième partie: manuscrits musulmans. Tome I, 1 (Paris: Bibliothèque Nationale de France, 1983). ——, ‘Collections de manuscrits anciens du Coran à Istanbul: Rapport préliminaire’ in Janine Sourdel-Thomine (ed.), Etudes médiévales et patrimoine turc (Paris, 1983), pp. 145–65. Dutton, Yasin, ‘Red Dots, Green Dots, Yellow Dots and Blue: Some Reflections on the Vocalisation of Early Qur’anic Manuscripts. Part II’, Journal of Qur’anic Studies 2:1 (2000), pp. 1–24. ——, ‘Red Dots, Green Dots, Yellow Dots and Blue: Some Reflections on the Vocalisation of Early Qur’anic Manuscripts. Part I’, Journal of Qur’anic Studies 1:1 (1999), pp. 115–40. George, Alain, ‘Coloured Dots and the Question of Regional Origins in Early Qur’ans (Part I)’, Journal of Qur’anic Studies 17:1 (2015), pp. 1–44. ——, The Rise of Islamic Calligraphy (London: Saqi, 2010). ——, ‘Calligraphy, Colour and Light in the Blue Qur’an’, Journal of Qur’anic Studies 11:1 (2009), pp. 75–125. Guesdon, Marie-Geneviève and Annie Vernay-Nouri, L’art du livre arabe: Du manuscrit au livre d’artiste (Paris: Bibliothèque Nationale de France, 2001). 102 Journal of Qur’anic Studies Imamuddin, S.M., Muslim Spain 711–1492 A.D.: A Sociological Study (Leiden: Brill, 1981). al-Khaṭīb, ʿAbd al-Laṭīf Muḥammad, Muʿjam al-qirāʾāt (11 vols. Damascus: Dār Saʿd al-Dīn, 2002). Kuwait, Maṣāḥif Ṣanʿāʾ: Catalogue of an Exhibition at the Kuwait National Museum (Kuwait: Dār al-Athar al-Islamiyyah, 1985). Levin, Aryeh, art. ‘ʾImāla’ in Kees Versteegh, Mushira Eid, Alaa Elgibali, Manfred Woidich, and Andrzej Zaborski (eds), Encyclopedia of Arabic Language and Linguistics (Leiden: Brill, 2006–7). al-Marrākushi, ʿAbd al-Wāḥid b. ʿAlī al-Tamīmī, Kitāb al-muʿjib fī talkhīṣ akhbār al-Maghrib (The History of the Almohades), ed. Reinhart Pieter Anne Dozy (Leiden: London Society for the Publication of Oriental Texts, 1847). Paris, De Carthage à Kairouan: 2000 ans d’art et d’histoire en Tunisie (Musée du Petit Palais de la Ville de Paris, 20 octobre 1982–27 février 1983) (Paris: Association Française d’Action Artistique, 1982). Piotrovsky, Mikhail (ed.), Earthly Beauty, Heavenly Art: Art of Islam (Amsterdam: Lund Humphries Publishers, 1999). al-Rammāḥ, Murād, ‘Tasāfīr maktabat al-Qayrawān al-ʿatīqa’ in R. ʿInānī (ed.), Dirāsāt al-makhṭūṭāt al-Islāmiyya bayn iʿtibārāt al-mādda wa’l-bashar (London: al-Furqān Islamic Heritage Foundation, 1997), pp. 135–50. Ross, Edward Denison, ‘Some Rare MSS Seen in Tunis’, Bulletin of the School of Oriental Studies 3:3 (1924), pp. 610–15. Roxburgh, David, Writing the Word of God: Calligraphy and the Qur’an (Houston: The Museum of Fine Arts, 2007). Roy, Bernard, Paule Poinssot, and Louis Poinssot, Inscriptions arabes de Kairouan (Paris: Librairie C. Klincksieck, 1950). Safwat, Nabil, Géza Fehérvári, and Mohamed Zakariya, The Harmony of Letters: Islamic Calligraphy from the Tareq Rajab Museum (Kuwait) (Singapore: National Heritage Board, 1997). Shabbūḥ, Ibrāhīm, ‘Sijil qadīm li-maktabat jāmiʿ al-Qayrawān’, Majallat maʿhad al-makhṭūṭāt al-ʿArabīyya (Revue de l’Institut des Manuscrits Arabes) 2:2 (1956), pp. 339–72. al-Wartatānī, Muḥammad al-Miqdād, al-Burnus fī Bārīs: Riḥla ilā Faransā wa-Suwīsrā, 1913, ed. Saʿīd al-Fāḍilī (Abu Dhabi, Beirut: Dār al-Suwaydī, al-Muʾassasa al-ʿArabiyya li’l-Dirāsāt wa’l-Nashr, 2004).