The Dotting Of A Script And The Dating Of An Era: The Strange Neglect Of PERF 558

Alan Jones

Islamic Culture, 1998, Volume LXXII, No. 4, pp. 95-103.

© Islamic Culture, All Rights Reserved.

It is now over forty years since I first came across information about the papyrus known as PERF 558, a bilingual Greek and Arabic document from Ihnas (known in Greek as Herakleopolis) in Upper Egypt. I wondered then at the general failure to recognise the central importance of the document, and over the years my wonder has grown.

PERF 558 is just one of a crucial group of twenty-two papyri from Egypt, written mainly in Greek but with the odd one in Greek and Arabic, and dating from the period A.H. 22-57, that found their way in the nineteenth century into the Erzherzog Rainer Papyrus Collection in Vienna.* Their authenticity has never been challenged, nor can I see any valid basis for such a challenge. The wealth of circumstantial detail that they contain is such that they can only come from the period indicated. The minutiae that we find in them would simply have not been available to anyone writing at any later stage of history.

The whole group has had the attention of some papyrologists, notably Adolf Grohmann, and of some of those interested in the early Muslim administration of Egypt,[1] but real attention seems to have stopped there. I suspect that some of the problem lies with the fact that Caetani does not appear to have been aware of PERF 558 when he compiled volume 4 of the Annali, though he did know of others in the group.[2] It is probably as a result of this that PERF 558 is only occasionally mentioned in more general historical writings by Western scholars, and then only in passing, for example in footnotes by Crone and Cook in Hagarism[3] and by Hoyland in Seeing Islam As Others Saw It.[4] References in Muslim sources are even rarer.[5] Yet, as I hope this brief paper will show, it should be well known to all those interested in early Islam.

The first serious examination of PERF 558 is to be found in the Corpus Papyrorum Archiducis Rainer, III I, pt. 2,[6] published in Vienna in 1923-24, edited by Adolf Grohmann. However, it was another work by Grohmann published in 1932 that made the document more accessible. This was his Aperçu de papyrologie arabe, published in Cairo.[7] It was through the Aperçu that I first made my acquaintance with PERF 558. Grohmann 's transcriptions and translations of both the Greek and the Arabic versions are a model of their kind, though one might question a couple of his readings. The facsimile plate, however, is hardly legible. Later plates were better,[8] but the most legible depiction of the original Arabic text is a tracing by Beatrice Gruendler in The Development of the Arabic Scripts,[9] to which I would refer those who wish to get an idea of how the original looks. The subject of the texts of the papyrus is mundane. It is simply the acknowledgement of the requisition of sixty-five sheep from Herakleopolis by the forces led by `Abd-Allah b. Jabir,** to be set off against the year's taxes. It has been suggested that the Greek version was written first, but this is by no means certain. The Greek occupies lines 1-3 and half of line 5 (5a) on the recto, and there is also a line on the verso. The Arabic occupies line 4, half of line 5 (5b) and lines 6-8 on the recto. The Arabic is not a direct translation of the Greek, but the core of the contents is the same in both versions. The Greek version refers to the Arab invaders as magaritai (i.e., muhajirun).[10] Despite its age PERF 558 is relatively easy to read. There are two reasons for this: it is written in a clear cursive hand; and it contains a fair sprinkling of dots.*** As will be seen, there are dotted forms of six letters [jim, kha', dhal, za', shin and nun, all of which are also to be found without dots]; there are some long vowels [a, i, and u are all to be found, though a is more frequently omitted]; and there are some examples of alif maqsurah.

In the following transliteration a line underneath a letter has been used to indicate that the letter has no dot in the original. The transliteration also uses the following non-standard signs, to allow all letters to be represented by a single character and to avoid using diacritics that might fuse with the underlining:

Q = tha'$ = shinG = ghayn

H = ha' S = sad h = ha'

C = kha't = ta' O = ta' marbutah.

d = dhalE = `aynA is used to transliterate alif.

[ ] surround an addition.








In the name of the Merciful, Compassionate God. These are the animals for slaughtering that `Abd-Allah ibn Jabir and his companions took from Ihnas:§ we took from the representatives of Theodor[akios], the elder son of Abu Qir, and from the representative of [Chr ]istofor[ os ], the younger son of Abu Qir, fifty sheep from the animals for slaughter and fifteen other sheep, which were butchered for the men on his ships, his cavalry and his infantry in the month Jumada I of the year Twenty-Two. The scribe was Ibn Hadid.§§

Though the main contents of the document cannot be described as vital, except to the sheep and the expeditionary force, the Arabic version contains two features that are of outstanding significance. The first is the script that is used; the second is that the date is given as the year A.H. 22.

The year 22 is the first Islamic year for which any dated documents written in Arabic survive, and there are only two of those: P Berol 15002,[11] which is unfortunately fragmentary, and PERF 558, the first complete Arabic document of the Muslim era.

The orthographic features¥ of PERF 558 set out above would be of great interest at any time during the first century, but the earlier the date the more important they become. Further, the way that they are used in PERF 558 indicates that they are hardly likely to be a new development. Gruendler sums up the position succinctly: "the first cursive influence must therefore be expected several decades earlier."[12] The fact that they can be traced back to A.H. 22 and earlier means that the traditional view that these orthographic features date from the period when al-Hajjaj ibn Yusuf was governor of Iraq (694-714)¤ cannot be correct. This traditional view is neatly set out for us in Hitti, History of the Arabs, p. 219:

He [al-Hajjaj] contributed to the development of diacritical marks in Arabic orthography to distinguish such similarly written letters as ba', ta', and tha', dal and dhal, and to the adaptation from Syriac of vowel signs, damma (u), fatha (a) and kasra (i), inserted above and below the consonants. In this orthographic reform he was prompted by the desire to prevent errors in the recitation of the sacred text, of which he evidently prepared a critical revision.

On the evidence of PERF 558 and other papyri, much of the above paragraph is misleading and must be discarded. The use of signs for short vowels appears to have been new, but this cannot have been the case with the other features, which must have been available to the earliest scribes of the Qur'an (whether they were used or not).

A possible explanation may be that less cursive styles than that exhibited in PERF 558 were used for the writing of the Qur'an, and that dotting did not feature in these; but the most that al-Hajjaj could have insisted on was the revival and regular use of earlier features.

This combination of date and script would be remarkable enough in itself, but the papyrus has yet another nugget of priceless information. The Greek version has its own Byzantine date[13] in line 5a of the recto of the papyrus. This has the form: "30 Pharmouthi of the indiction year 1". Such dates are commonly found in Greek papyri, and indeed there are half a dozen other documents in the Erzherzog Rainer Papyrus group that have indiction dates for the period A.D. 642-43.[14] The only problem is that the indiction cycle is a relatively short one of fifteen years, and thus great care is needed in working out the appropriate date in the Christian era. Grohmann worked out "30 Pharmouthi of the indiction year 1" to be 25th April, 643 A.D.

The date fits with two of the most commonly available conversion tables: those of Caetani[15] and of Freeman-Grenville.¥¥ They make the last day of Jumada I equate with 26th April, 643 A.D. Given the virtual inevitability of imprecisions of dating at the period, the fit is remarkably good.¤¤

* The reference codes for these papyri come from the Catalogue of the Erzherzog Rainer Papyrus Collection put together by J. von Karabacek and others inl Vienna in 1894. The papyri, numbered 552-573 in the catalogue, now form part of the collections of the Österreichische Nationalbibliothek.

I know of no one who has suggested that the papyri might have some later origin or suggested that carbon dating or other scientific tests be applied. For my part I am confident that scientific tests would confirm the age of the papyri.

** Abd-Allah ibn Jabir is known only from the papyri. He is referred to in PERF 555, 556, 558, 559 and 561. See Caetani, Annali, 4, A.H. 21, para. 99; Caetani and Gabrieli, Onomasticon Arabicum, 2, entry 11436.

*** There are sixteen ordinary dotted letters. In addition, in line 7 the alif of ASH[A]B is preceded by a dot. There is a similar instance in an inscription dated A.H. 80. (See Gruendler, op. cit., pp. 32 and 18.) Both examples may indicate hamzah. Hamzah may also lurk in the cluster WKT[A]'Bh in line 7, which has a spike that might be considered to be either a ya' or the bearer of a hamzah. The odds are slightly in favour of a bearer, as one might expect a similar spike in the next cluster WQQLA[ ' ]h if a ya' were intended. This possible indication of hamzah must remain uncertain.

Of the ta' marbutah of later times there is of course no sign; final ha' is employed where necessary.

§ Literally "This is what `Abd-Allah ibn Jabir and his companions took from lhnas from the animals for slaughtering."

The translation gives the name forms recoverable from the Greek version, which may be translated as:


1. In the name of God. [From the] Amir `Abd-Allah to you Christophor[os and] Theodor[akios], pagarchs of Herakle[opolis],

2. I have taken from you for the purpose of feeding the Saracens who are with me in Herakle[opolis] 65 sheep (sixty-

3. five and no more); and I have had the present document written to make this clear .

5a Written by me, Ionnes, no[tary] and off[icial], 30th day of the month of Pharmouthi of the ind[iction year] 1.


1. Document about sheep given to the muhajirun (Greek magaritai) and other [new] arrivals, towards the payment of the taxes of the Ind[iction year] 1.

Literally "which the men on his ships. . . had butchered."

§§ The name lbn Hadid is followed by a further waw, the import of which is not clear, though there are parallels in the Safaitic inscriptions.

¥ The orthographic features can be seen in other early papyri, including P Berol 15002, but they are at their clearest in P558.

¤ Al-Hajjaj ibn Yusuf (c. A.D. 661-714) is usually credited with the improvement of the script used for writing the Qur'an (See, for example, Nöldeke, Geschichte des Qorans, 3, pp. 260 ff). It may be that the basis of this was just one facet of his general quest for stability in all areas of life, but he seems to have pursued the matter with his usual vigour. He wanted the new text to supersede any other, and he seems to have succeeded. In a handful of readings it appears to be his preference rather than that of the `Uthmanic recension that has prevailed (see lbn Abi Da'ud, Kitab al-Masahif, ed. A.Jeffery, pp. 117-118)).

¥¥ Freeman-Grenville, Tables, leaves itto the reader to work this out, and so I may be wrong.

¤¤ Whilst such tables as those produced by Caetani and Freeman-Grenville and others may have a spurious certainty (and this is perhaps even more so with recent versions on CD-ROM), it would appear that they are not likely to be far out.


[1] See, for example, Kosei Morimoto, "Taxation in Egypt under the Arab Conquest," Orient (Tokyo) 15 (1979), pp. 71-99, and The Fiscal Administration of Egypt in the Early Islamic Period (Kyoto, 1981), though the emphasis is on other documents. One might have expected something useful in Frazer's revision of Butler's Arab Conquest of Egypt (A.J. Butler, The Arab Conquest of Egypt, 2nd edition revised by P.M. Frazer, Oxford, 1978); but though Frazer's introductory notes have a reasonably good section on the papyrological evidence (pp. lxxvi-lxxxiii), PERF 558 is not dealt with.

[2] Published in 1911. Caetani was aware of PERF 555, 556, 559 and 561.

[3] P. Crone and M. Cook, Hagarism (Cambridge, 1977), p.157.

[4] R. Hoyland, Seeing Islam As Others Saw It, Studies in Late Antiquity and Early Islam, 13 (Princeton, 1997), p. 688.

[5] The two that I have seen are in works still under consideration for publication.

[6] Corpus Papyrorum Archiducis Raineri III, Series Arabica I, Pt. 2, XXI-XXVI (Vienna, 1923-24).

[7] A. Grohmann, "Aperçu de papyrologie arabe", Études de papyrologie, I, Société Royale Égyptienne de papyrologie (Cairo, 1932), pp.41-43 and plate IX.

[8] See N. Abbott, The Rise of the North Arabic Script (Chicago, 1939), plate IV; A. Grohmann, From the World of Arabic Papyri (Cairo, 1952), pp. 113-115 and plate 11. Idem, Arabische Chronologie. Arabische Papyruskunde, 91 no.1 and plate 11.1. Handbuch der Orientalistik, Abt. l, Erg. Bd. II, 1 (Leiden/ Köln, 1966).

[9] B. Gruendler, The Development of the Arabic Scripts, Harvard Semitic series, no.43 (Harvard, 1993).

[10] The fullest discussions of the Greek term, of which there are variant spellings, are to be found in P. Crone and M. Cook, Hagarism, passim.

[11] This papyrus is now housed in the Staatsbibliothek in Berlin.

[12] The Development of the Arabic Scripts, p. 135.

[13] For detailed information on Byzantine chronology and in particular on the indiction cycles, see V. Grumel, La chronologie byzantine (Paris, 1958). I am grateful to Dr James Howard-Johnston of Corpus Christi College, Oxford, for drawing my attention to this invaluable work.

[14] PERF 553 (January-February, 642), 564 (19th July, 642), 555 (26th December, 642), 554 (25th February, 643), 559 (1st June, 643), 561 (29th November, 643).

[15] Caetani, Annali, 4, 674.

Arabic Papyri | Qur'anic Manuscripts | Qur'anic Studies