Information Packaging in Arabic Private and Business Letters (8^th to 13^th c. CE): Templates, Slots and a Cascade of Reduction and Rearrangement

1. Documentary private and business letters[1]

Only a few hundred documentary private and business letters in Arabic script have been published to date.[2] The major editions[3] comprise a total of about 600 items. Of these, there are 500 with a full translation and scan, 100 without scan.[4] About two third of them are business letters, the other third private letters, although this differentiation is in many instances problematical. Nearly all documents stem from Egypt, although some were written outside Egypt and sent to partners within Egypt. For most items, the discovery location is unknown. However, the available information from the collections or the documents themselves, indicates a dominance of Upper Egypt and the Fayum. About 60% of these letters are written on papyrus (7^th to beginning of 10^th c.) and 40% on paper (after the beginning of 10^th c.). Less than half a percent is covered by other writing materials.[5] The dominance of papyrus documents does not reflect the actual holdings of most collections but may be explained by the general easier readability of papyrus documents and that the early documents are especially popular among editors, for only a few primary sources of this time have survived to this day.

Only a handful of letters are dated or can be assigned to a short time period with any great certainty. The overwhelming majority were assigned a span of one or more centuries. The paper documents pose a greater challenge to editors in this respect. Although the dating of the documents is often difficult, if we do not insist on palaeography alone, there are numerous developments that can be traced down over the centuries and allow in most instances a relative dating. In this paper I shall concentrate on the topic of structural developments in the opening part of documentary letters from Egypt up to the 13^th c. CE.[6]

2. Approach and methodology: charting and crucial concepts

If we compare medieval Arabic letters with each other, even letters from the same period or the same person, it is striking that, although some very widespread formulae exist, the variance in expression is large. There are innumerable ways of saying basically the same thing. How can one thus describe the letter structures without simply constructing huge lists containing all these expressions and their innumerable combinations? One solution is to consider the formulaic associations and seek the underlying concepts viewing the epistolary formulary in Arabic letter writing as a kind of algorithm, i.e. an order of operation that resorts not only to the high level of organisation of the text but includes also templates for smaller sections within the document. Within these templates, there are obligatory and optional slots with demands and constraints. In describing the features of these slots some very general concepts proved to be important. These are specified further as we move deeper into the structure. Concepts of an overall importance are for example blessings, religious formulae in a broader sense, reference to written communication ("you mentioned in your letter"), and linguistic reference to sender and addressee ("Your letter, my brother, has reached me"). A slot within a template may thus be constrained to expressions exhibiting a given concept; the actual expression however is not determined.

In order to describe and compare structural phenomena I annotated documents by means of charting.[7] The text of a document is hereby represented in a column in short phrases and then annotated on its left and right hand side. The annotations include marking for formal, grammatical, semantic and structural features in columns or with colors on the Arabic text. The categories that turned out to be crucial emerged iteratively during my work with the documents. By annotating letters with the help of these charts and thus assigning underlying concepts to expressions and by breaking a text into smaller or larger portions, I created an instrument to compare texts and to make statistically relevant conclusions.

3. Building blocks, templates and slots: The blessing section of a 9^th c. letter

Medieval Arabic private and business letters exhibit to a greater or lesser degree, structures that are evocative of "building-blocks." These units are self-contained and arranged into larger entities following particular templates in a conventional order: The letter itself is a unity consisting of different obligatory and optional parts with clearly marked boundaries, even though the different parts are related and interact. The building blocks themselves also consist of different parts, both obligatory and optional, independent and dependent ones. As an illustration, we will take a closer look at the blessing section within a 9^th century private or business letter.

A typical letter of this period starts with the invocation (expression is fixed). It is followed by a part where the sender bestows blessings on behalf of the addressee (expression is not fixed but with some common formulae), e.g.:

The whole blessing consists of several single blessings in one grammatical sentence with the subject allāh "God."[9] To have at least one blessing with the subject allāh "God" is obligatory, the wording itself is not fixed. As illustrated above, a single blessing may just consist of the verb plus the subject allah "God" or also include an object like in wa-ʾatamma niʿamahu ʿalayka "fulfil His favors upon you." There may also be further extensions. Compare (1) to the next letter:

ʾakramaka llāhu "may God honor you" is already self-contained, but in the second letter further specified with an optional bi-ṭāʿatihi "by (your) obedience towards Him."

After blessings of the obligatory type with God as subject like ḥafiẓaka llāhu "May God protect you" or ʾakramaka llāhu "May God honor you" there is always the possibility to close the passage by an optional bi-raḥmatihi "by His mercy" or a similar expression, which can again – but does not have to – be further extended by fī d-dunyā wa-l-ʾāḫirati "in this world and the hereafter" with slight variations in front of bi-raḥmatihi " by His mercy." Compare the examples above with the following two documents:

Although the expressions bi-raḥmatihi "by His mercy" and fī d-dunyā wa-l-ʾāḫirati "in this world and the hereafter" are both optional, fī d-dunyā wa-l-ʾāḫirati " in this world and the hereafter" is in the majority of cases combined with bi-raḥmatihi "by His mercy."

The structure of the blessing part may thus be visualized as follows:

4. Slide-in-blessings[10] and the second-position slot

Slide-in-blessings – short blessings inserted parenthetically into the text – are very characteristic of Arabic letter writing. They appear typically in the second position of a thematic unit after a grammatical reference to the addressee and on behalf of the addressee. In the following example the reference to the addressee is established by an imperative – a typical beginning of an instruction unit:

As a consequence of using slide-in-blessings parenthetically in the second position of a unit after reference to the addressee, it seems that their function of marking a unit gains the upper hand in many instances and we find therefore slide-in-blessings without explicit reference to an addressee:[11]

However, slide-in blessing are not constrained to the second position of a unit but may also occur within a unit and also on behalf of a third party. In these cases they are considered as a mere device of courtesy without further function of structuring the text:[13]

This example also shows that not only is a blessing sometimes inserted parenthetically into the text, but also other phrases like ʾin šāʾa llāhu "God willing!" wa-llāhi "by God" or direct addressing of the addressee by means of the vocative particle yā occur often parenthetically, and – not surprisingly – mostly in the second position of a unit or sentence.[14]

The following passage comprises many of the aspects just referred to:

At the end of line 9 a new sentence opens, and a slide-in-blessing on behalf of the addressee is inserted (l. 10). This is done after grammatical reference to the addressee by means of the personal suffix -ka "you" between verb and its direct object – a very close grammatical relationship. The sentence goes on and after mentioning a certain Abū Yaḥyā a slide-in-blessing on his behalf is inserted. The new sentence on line 11, beginning with fa-raʾyaka "do me the favor," exhibits again a slide-in-blessing on behalf of the addressee in the second position. Note that the next reference to the addressee wa-raddihi "and send it back" does not trigger a slide-in-blessing.

The sentence that then follows is especially intriguing. The slide-in-blessing ʾaṭāla llāhu baqāʾaka "may God prolong your life!" shares the parenthetical second-position slot with wa-llāhi "by God!" and direct addressing yā bā l-ḫayri "oh Abū l-Ḫayr!" These three elements separate the verbal particle qad from the verb it belongs to. At this point, the writer complains about a man named Stenuris and instead of bestowing upon him a blessing, he curses him with a parenthetical lā raḥimahu llāhu "may God not have mercy on him!" The writer of the letter must have been very angry – for such cursing is extremely rare! This short passage illustrates, that slide-in-blessings are an important device in Arabic letter writing: They express attitudes towards persons mentioned in the text and have in the same time a structural function.[16]

5. Diachronic developments in the opening of a letter

Over time, the functional load, demands, and constraints of a particular slot vary. This will be illustrated in looking at the openings of letters throughout the centuries. The oldest private letters we have originate from the 8^th c. CE. Surprisingly, they show a very elaborate prescript, which does not show a great linguistic variation in its realisation. It consists of invocation, prescript, and blessings in favourite of the addressee. The unit to follow is typically well being where the sender expresses that he is fine and gives his thanks to God. See below a scheme with hierarchical templates[17] and as example P.Jahn 3.[18] Obligatory sections are marked grey.[19]

The block well being starts with a formula referring to written communication that expresses both sender and addressee morphologically. After an inserted, optional slide-in-blessing, the main text of the unit follows, which can be formulaic or free. Well being is closed by one or several religious formulae.[26] Also sections in the body of a personal letter tend to follow this template.

The style of 8^th c. letters is subjective and the structure of most units rather clear-cut (mostly short, coordinated sentences). Within the internal address of the letter, both parties are named with proper names. However, no titles or further specifications are added. Might this be a reflection of egalitarian Islamic views? Even though some elements, e.g. internal address and greeting, are evocative of Greek or Coptic epistolary formulary, the Arabic type of prescript represented here seems to be a genuine Arabic one, brought to Egypt by the new rulers and not simply modelled after a Greek archetype.[27]

From the 8^th c. to the 9^th c. letters change drastically: Section Internal address disappears as a whole, there are no intermediate stages.[28] Section blessings however gains prominence and is now obligatory. See below a 9^th c. letter until the section confirmation, that exhibits also the template just discussed:

Thus, letters from the 9^th century do not mention any sender or addressee within their text. They remain unknown unless an external address is preserved. Possible direct addressing within the text is effected by yā (vocative particle) + relational noun, e.g. yā ʾabī "oh my father" in second positions of a unit. In the 10^th c. exactly these slots were strengthened and the direct addressing appears frequently, at least in the first two sections of a letter.[29] It seems then that this double application of addressing plus blessings yielded to a reanalysis of the now unwieldy opening section. See below a typical 11^th c. letter:

The constraints on the former slide-in-blessing position are relaxed. The slide-in-blessing becomes a parenthetical blessing and "takes over" the blessings that were previously in a separate blessing unit right after the invocation.

Whereas early letters overtly display the distance between sender and addressee (e.g. internal address), we observe a shift in this conception from the 9^th c. onwards. In leaving out the address and starting with prayers, a sender deals with his addressee as if he would stand personally in front of him.[30] Although the epistolary perfect was in use throughout all these centuries, the introductory formulae for the body of the letter were widely replaced by verbs in the imperfect like yunhī "he reports" or like in P.Berl.Arab. II 83.2 (Fig. 3) ʾuʿrifuka – I make known to you or literally "I make you know."[31] Letters are viewed as to replace their senders.

The typical 13^th century-letter is difficult to describe, as there is a rather big variance in structure (see below two examples). Further developments draw on the structure of 11^th and 12^th c. letters and are heavily influenced by official letters and petitions (e.g. formula of prostration,[32] see below second document). When it came to the 13^th c., a distinct unit for the sender had been established, graphically set off but grammatically mostly integrated. The addressee’s section is further developed by a bulk of titles and honorifics.[33] Direct reference to the addressee by means of the vocative particle yā and grammatical 2^nd person has however vanished in the formulaic parts. Objective style throughout these sections intensifies the construction of (social) distance: The conception of being received in audience at the addressee’s clearly emerges.[34]

A new optional link (siwā ḏālika lit. "other than that") may remind us of 8^th century letters (link ʾammā baʿd lit. "for what is after"), but is rather different structurally and grammatically: it follows the blessings parenthetically and is not preceding them as in the 8^th century. In further contrast, sender and addressee are not only named (as in the 8^th c.) but the proper names are furnished with titles, or humble expressions (e.g. mamlūkuhu "his slave"). By the Mamluk period (1250–1517) the world of addressing and related blessings developed into a whole system – probably being an indicator of the social and political changes in that time. Regarding style, letters fluctuate between mixed and objective. Purely subjective style is no longer found.

6. Summary

What we have seen is a cascade of reduction and rearrangement. Though the structure of the 13^th century may somehow be evocative to the one of the 8^th century, it is grammatically much more complex and exhibits a different underlying structure.

If the structuring principles of templates and slots are taken into account then the developments in the opening section of letters are traceable. The concept of parenthetical elements in the second position of units was especially important in this process: Loosened restrictions regarding kind and length of possible contents contributed to an opening section becoming structurally more and more complex and diverse over time.