a right-to-left Semitic script, used by adherents of Manicheism to write texts in Middle Persian, Parthian, Sogdian, Early New Persian, Bactrian, and Uighur (Old Turkish). It is closely related to the Palmyrene script of Aramaic and the Estrangelo script of Syriac; some of its orthographical conventions are also to be found in the Mandaean script.


MANICHEAN SCRIPT, a right-to-left Semitic script, used mainly to write Middle Iranian languages and Uighur (Old Turkish). It is closely related to the Palmyrene script of Aramaic and the Estrangelo script of Syriac; some of its orthographical conventions are also to be found in the Mandaean script (see Naveh, 1982, pp. 151-52; see also MANDAIC). The Manichean script was used by adherents of Manicheism to write texts in (Manichean) Middle Persian, Parthian, Sogdian, Early New Persian, Bactrian, and Uighur; these texts were found in Central Asia and are now for the most part in the Turfan Collection in Berlin (see the catalogues by Boyce, 1960; Wilkens, 2000; and the short survey in Sundermann, 1993). Some Aramaic texts on ‘magic bowls’ found in Mesopotamia are written in a script very close to the Manichean script (Montgomery, 1912; Segal, 2000, Plate 137, pp. 239 and 149). The term “Manichean” was introduced by F. W. K. Müller, the first scholar in modern times to read the script in 1903-04, because of the use of the script in Manichean texts.

Description. The Manichean script (see TABLE 1) contains letters with final and non-final forms only, though not all letters have distinct final forms. Final forms only occur of those letters that have a final stroke pointing to the left (b, g, l, δ, s, p, and letters derived from these) or curling upwards (m, c, q). These are the letters that also have non-final joining forms, i.e., the stroke (minus the final globule or curve) joins these letters to the following one in a word. The only exception to this is n, where the non-final and final forms are somewhat more distinct. The variant final forms of hò and m show the final stroke pointing vertically downwards. The variant form of has no significance. The letters d and r occur in two forms: the more complex form coincides with the Mandaean letter dy. The transliterations w: and u are different conventional transliterations used freely for both of the letter forms which are composed of the letter w with diacritics. The joiner stroke (often terminating with a globule) has a decorative use in ordinary texts; it is also used in cantillations (the schematic presentation of hymns indicating syllables elongated in singing). No ligatures occur except for θ (a ligature of δδ). The final stroke of the letter c tends to enclose a following y or n, but without distorting either letter; the final stroke of k and related letters can extend to the end of the word in which the letter occurs. Various script styles occur: normal (Table 1), cursive, larger initial, and headline letters and decorative vine-like outline letters in some headlines.

The basic letters are: ʾ, b, g, d, ẖ w, z, ǰ, h, ṯ, y, k, l, δ, m, n, s, ʿ, p, c, q, r, š, t. These are the 22 Aramaic letters supplemented by ǰ (the only ‘new’ letter) and δ (clearly related to l). The order is confirmed by alphabetical hymns (abecedaria) in Parthian and Middle Persian, though there is a tendency for h and to change positions and for x and f to be used for k and p. The letter ǰ, which only occurs in Parthian hymns, is therefore placed in Table 1 after z rather than after c as in Henning’s list (Andreas and Henning, p. 911). Most Parthian and Middle Persian texts are written with these letters, supplemented at most by x and f. Letter names are attested only for the first letter (Parthian ʾʾlyf, Middle Persian ʾrb and ʾ), the last basic letter (Parthian), and possibly for m and s (mʾm, syn).

Slight differences exist between the Iranist and Turcologist conventions of transliteration, e.g., the letter (or) in Iranian texts is transliterated t in Uighur texts (for further Uighur conventions, see von Gabain, 1974, table on p. 17; illustration on p. 30).

These basic letters are extended: (1) three letters differentiated by a single point above: x, f, qµ; (2) seven letters differentiated by two points above (here represented by “:” after the letter): β, ҳ, k: (or q:), ʿ: (in Early New Persian texts, for Arabic), ǰ: (possibly ź, see ś), q: (in Turkish texts), ś (N. Sims-Williams suggests this for the Bactrian fragment in Manichean script [personal communication]); and (3) and one letter differentiated by a loop at the end of the stroke extending below the line: γ. Many of these letters are of very restricted use. The gutterals q and k with one or two points are characteristic of Uighur texts.

Additionally, in Middle Persian and Parthian texts much use is made of two abbreviations indicated by double dots above or, in the case of w, also above and below the letter; they are: w (also transliterated as u) for ʾwd “and,” š: and plural š:ʾn for ʾwš and ʾwšʾn “and he/she,” “and they.” Also “elision” or “plural” double dots are used, placed above or below or between letters (these forms are not shown in Table 1); and a single dot represents a miniature y. Other signs occurring in Manichean manuscripts are various punctuation marks and flowers (three of which are given in Table 1) and the numerical signs for 1, 5 (in the shape of ʿ), 10, 20 (in the shape of p, but in fact Aramaic k), and 100 (in the shape of m, but in fact Aramaic q).Manichean manuscripts are generally very well written (see PLATE 1). A careful distribution of the letters on the page (often using columns) is achieved by stretching letters or the joining line; compressing the letters; and using abbreviations, usually marked by elision dots, which in effect are shortened forms of the written word. These dots are particularly common for plural forms squeezed in at the end of a line, but they are also used for other words. Word boundaries are consistently indicated by a small gap; punctuation marks (sometimes in red ink) divide up the text; blank lines, captions and headlines (often decorated and in a different color) are also used.

One side of a double page with the signature M 172 (= MIK III 196; shown on Plate 1) contains, on the right, part of a confession text in Uighur (entry no. 408 in Wilkens, 2000, p. 357) and, on the left, part of a bilingual text in Middle Persian and its Sogdian translation, written in alternating sentences. The Middle Persian text is one of Mani’s own texts, from the beginning of his Evangel. This page was edited by D. N. MacKenzie (1994).

Orthographic conventions. As can be expected, the representation of vowels is limited in range: w, besides being a consonant indicates /u/ and /o/ but without designating length, though a long vowel is more likely to be written than a short one. The letter ʾ is used for /a/ but also to indicate any initial vowel: ʾwd = /ud/. The letter ʿ is almost wholly restricted to the initial position to designate an initial front vowel (a trait the Manichean and the Mandaean scripts share): Middle Persian and Parthian ʿspyd = /ispēd/ “white.” Through the use of points (e.g., f) and some digraphs (e.g., wx for a consonant peculiar to Parthian, xw for a similar one in Middle Persian) an adequate representation of the consonants is achieved. Besides the graphical clarity of the Manichean script (i.e., sufficient differentiation and legibility of the letters), the language material written in this script is all the more valuable for being written without conventions that hide the actual linguistic values. In this respect, for example, Manichean Middle Persian contrasts with Zoroastrian Middle Persian that uses ‘historical’ writing and heterograms.

Origin. The attribution “Manichean,” being a modern one (Müller, 1904 first called the script “Estrangelo”), gives no indication of the origin of the script. It seems likely that Mani or his disciples used a script already available to them and similar to, if not identical with, the Palmyrene script (see Montgomery, Lidzbarksi, Naveh, and Skjærvø) and that they, at most, simply took the significant step of writing non-Semitic languages with this script. (Henning, 1958, p. 73, thought Mani first used it to write down his Middle Persian works.) The initial steps in this direction may have been taken by scribes before Mani and his disciples, though there is (apart from the magic bowls) no direct examples of the use of the Manichean script in Mesopotamia for likely purposes (e.g., merchants’ accounts and correspondance, archival entries, legal documents and, possibly, inscriptions). The Manichean script as attested in the texts in Central Asia shares conventions with the script of Aramaic inscriptions such as those in Hatra, a border area between Aramaic and Iranian linguistic spheres (see Durkin-Meisterernst, 2000). This is not unexpected, because the Manichean Middle Persian and Parthian texts found in Central Asia (at Turfan) were for the most part brought there from areas of the Sasanian empire where Middle Persian and Parthian were spoken.

In common with the Mandaean script, the Manichean script shows a remarkably pragmatic approach to the languages written in it. This is all the more impressive because the Manichean script was used for Iranian languages that already had written traditions, though apparently these were of quite restricted use. This applies in particular to Parthian, for which the texts in Manichean script are the major source besides some inscriptions in Parthian native script. Middle Persian is well attested in its own particular form developed from an Aramaic script that led to a highly complex writing system using heterograms (see IDEOGRAPHIC WRITING) and historical spellings, two things that the Manichean script disposed of except for some conventional spellings that, at least in later material, became historical. Similarly, Sogdian had its own script, with a few heterograms and conventional (historical) spellings; here it seems that Manichean Sogdian follows most of the general historical conventions of the native script. The single extant page of the Bactrian language in Manichean script is written with conventions that differ from those in the ‘native’ Bactrian script (e.g., the regular word-final vowel of the latter is missing in the former). Overall, the contrast between the comparative lack of convention in the Manichean script and the complex older scripts is greatest in the case of Middle Persian (and the less well attested Parthian). Why Mani or his followers chose to use the Manichean script for their religious texts in a state that already had various script traditions is hard to establish. The Semitic background is certainly significant and is confirmed by the few orthographic conventions of the Manichean script which can be explained only by reference to this (see Durkin-Meisterernst, 2000). If deliberate diversification and differentiation from neighboring scripts was an important motive (Lidzbarski, 1916 and others talk of a “sectarian script”), this would apply in the first instance to use of the Manichean script to write Aramaic, but, for lack of evidence about use of the script apart from the Iranian languages, this question cannot be decided. It may be no less important that religious literature was not written down at all in Sasanian Iran of the 3rd century CE. The desire to fix a religious tradition in writing was one not shared by Zoroastrian priests of that time—one that may even have arisen, among other reasons, under the pressure of rapidly spreading Manichean literature; there was no reason for the Manicheans to employ the scribal traditions of 3rd century Iran for this purpose. Therefore it would have been only natural for them to fill this vacuum with their own script, which, in this new context, was a script restricted to them alone.



Facsimiles of manuscript pages written in Manichean script can be found in most of the more recent editions, but especially in W. Sundermann, Iranian Manichaean Turfan Texts in Early Publications (1904-1934), Photo Edition, Corp. Inscr. Iran., Supplementary Series III, London, 1996, and D. Weber, Iranian Manichaean Turfan Texts in Publications since 1934, Photo Edition II, Corp. Inscr. Iran., Supplementary Series IV, London, 2000.

As part of the ongoing digitization of the texts in the Turfan Collection in Berlin, color images of the texts in Manichean script are available online at http://www.

F. C. Andreas and W. Henning, Mitteliranische Manichaica aus Chinesisch-Turkestan. III, SbPAW 1934, pp. 848-912 (= W. B. Henning, Selected Papers I, pp. 275-340).

M. Boyce, A Catalogue of the Iranian Manuscripts in Manichean Script in the German Turfan Collection, Deutsche Akademie der Wissenschaften zu Berlin, Institut für Orientforschung, Veröffentlichung Nr. 45, Berlin, 1960.

D. Durkin-Meisterernst, “Erfand Mani die manichäische Schrift?” in R. E. Emmerick, W. Sundermann, and P. Zieme, eds., Studia Manichaica IV. Internationaler Kongreß zum Manichäismus, Berlin, 14.-18.

Juli 1997, Berlin 2000, pp. 161-178.

A. von Gabain, Alttürkische Grammatik, 3. Aufl., Wiesbaden, 1974.

W. B. Henning, “Mitteliranisch,” in Handbuch der Orientalistik, Iranistik I, Leiden, 1958, pp. 20-130.

M. Lidzbarksi, “Die Herkunft der manichäischen Schrift,” Sitzungsberichte der Preußischen Akademie der Wissenschaften, Berlin, 1916, pp. 1213-22.

D. N. MacKenzie, “I, Mani ..,” in H. Preißler and H. Seiwert, eds., Gnosisforschung und Religionsgeschichte. Festschrift Rudolph, Marburg, 1994, pp. 183-198.

J. A. Montgomery, “A Magical Bowl-Text and the Original Script of the Manichaeans,” JAOS 32 1912, pp. 434-38.

F. W. K. Müller, “Handschriften-Reste in Estrangelo-Schrift aus Turfan, Chinesisch-Turkistan [I],” Sitzungsberichte der Preußischen Akademie der Wissenschaften, Berlin, 1904, pp. 348-52.

Idem, Handschriften-Reste in Estrangelo-Schrift aus Turfan, Chinesisch-Turkistan. II Teil, Anhang zu Abhandlungen der Preußsichen Akademie der Wissenschaften, Berlin, 1904.

J. Naveh, Early History of the Alphabet. An Introduction to West Semitic Epigraphy and Palaeography, Jerusalem-Leiden 1982.

J. B. Segal, Catalogue of the Aramaic and Mandaic Incantation Bowls in the British Museum. With a contribution by E. C. D. Hunter, London, 2000.

P. O. Skjærvø, “Aramaic Scripts for Iranian Languages,” in P. T. Daniels and W. Bright, The World’s Writing Systems, New York and Oxford, 1996, pp. 515-35 (esp. pp. 519 and 530-33).

W. Sundermann, “The Manichaean Texts in Languages and Scripts of Central Asia,” in S. Akiner and N. Sims-Williams, (eds., Languages and Scripts of Central Asia, London, 1993, 39-45.

J. Wilkens, Alttürkische Handschriften Teil 8. Manichäisch-türkishe Texte der Berliner Turfansammlung, Verzeichnis der Orientalischen Handschriften in Deutschland XIII, 16, Stuttgart, 2000.


October 14, 2005

(Desmond Durkin-Meisterernst)

Originally Published: July 20, 2005

Last Updated: July 20, 2005