The Format of the Tablets

In describing the discovery and excavation at Vindolanda of wooden leaf tablets with ink writing in the 1970s we suggested that such finds might be more frequent in the future.1 It is Vindolanda itself which has fulfilled this prediction most spectacularly, as the contents of the present volume show, but tablets of this type from the Flavian period have been discovered at two other important military sites in Britain - Caerleon and Carlisle.2

The very large number of wooden leaf tablets found in the excavations of the 1980s and their continued preponderance compared to the number of stilus tablets confirms the view that their use must have been very widespread and that this medium was the counterpart of papyrus at least in the north-western provinces of the empire.3 The practice of scoring and folding the leaves and the frequent presence of notches and tie-holes, which writers often avoid, shows that the tablets were manufactured in this way before the texts were written. It is interesting to note that there are two tablets with writing by the same hand which have distinctive semi-circular notches in the left margin; if these belong to the same letter it must have been written on more than one leaf, but it is perhaps more likely that they are from two different letters written on leaves from the same batch (214, 216).4 Discussion of the new evidence concerning the format of the leaf tablets can be divided as before, treating the letters and the documents and accounts separately.5

Letters

From the point of view of format the letters perhaps offer the more interesting evidence. In the layout which occurs most commonly the leaf is used with the broad dimension running horizontally and the writing along the grain. The broadest examples we have found are of the order of 250mm.6 The text is normally written in two columns, the first on the left-hand side of the leaf, the second on the right, the first often having longer lines than the second and extending beyond the centre of the leaf. The leaf is scored vertically down the centre and folded, forming a diptych. This is what we previously described as the "letter format" and it bears emphasising that it is used in a great number of letters which evidently came from a variety of different places.7 The address is normally written on the back of the right-hand half of the diptych (see below). The two-column format has parallels in Greek letters, but it is extremely rare; indeed, it is very uncommon to find a Greek letter on papyrus written in more than one column. In papyrus letters, too, it is conventional to write the address on the back but normally along the fibres at a right-angle to the writing on the front. The two-column format may well be commoner in Latin papyrus letters but so few examples survive (and those biassed towards the military context) that it is impossible to be sure.8

This arrangement is, however, not universal and the variations are interesting. A letter with the opening on a right-hand portion may be either very short or may be continued on another leaf (243). A letter may be written on two or three leaves, folded together; the first part of a letter from Claudia Severa is written in one column across the whole diptych, the remainder in two columns on two further leaves and the closing greeting is actually on the back of the middle leaf (292). A letter with a closure on what appears to be a left-hand half may have been written on more than one diptych (247). Octavius writes to Candidus on two leaves, two columns on each leaf, but the sequence runs from right to left and the easiest explanation of this is that Octavius was a left-hander (343); placing the first column on the right would enable him to see what he had written as he moved on to the second column at the left. Some writers conclude their text by writing at a right-angle down the left-hand margin, usually between the columns (302,TVII 311), but in one case in the margin beside the first column (316); others continue on the back of the leaf (305, 307, 340).

As to other features of the layout and appearance of the text, the heading is usually written on the first two lines of the left-hand column with salutem or suo salutem on line 2 and justified at the right.9 There are, again, a few interesting exceptions. On one occasion the whole address is written in the first line (212); on another salutem seems to be more or less in the centre of line 2 (275); a third appears to have suo justified at the left and salutem justified at the right (250). Two examples have more effusive openings occupying (probably) five and three lines respectively (214, 310).10 The first line of the message proper is more often than not indented, with an enlarged initial letter,11 and the subsequent lines are often further indented. Occasionally the first line of the second column begins further to the left than the subsequent lines (e.g. 215).

Some of the letters found at Vindolanda did not come from other places. The correspondence of Flavius Cerialis includes examples written by him which we have described as drafts or file copies (225-241). Our conception of a draft is based on 225, in which the name of the sender, certainly Flavius Cerialis, is omitted from the heading. Both sides of the leaf are used and there are several erasures; we think that this hand, which occurs in 8 texts (225-232), is that of Cerialis himself; it is just conceivable that 226 contained drafts of at least two separate letters. There is another draft with erasures which we are unable to assign to any group of letters (317) and another which was found in the Period 4 barracks (344). The latter is a draft of a petition or letter and is written on three sections of a leaf or leaves which have an account on the other side (180); since its beginning is missing, it must have commenced on another leaf. Our notion of a "file copy" is best exemplified by 234 which has a clear instance of the correction of a dictation error (see note to ii.2). This is a letter from Flavius Cerialis with the opening address in its full form but lacking a closure and an address on the back. Another example has a brief account written across the grain from top to bottom on the right-hand portion; this is likely to be an earlier text which was re-used for the letter (233). We regard the lack of an address on the back or outside of the leaf as an essential feature of both drafts and file copies and for this reason we prefer to regard one example of a letter from Cerialis which includes this element (as well as a closure) as a letter which was sent to someone elsewhere and brought back to Vindolanda by the addressee (242). As the dictation error in 234 suggests, the distinction between "drafts" and "file copies" is not completely clear-cut and involves an element of arbitrary judgement on our part. Furthermore, one could easily envisage a "draft" being used as a file copy after the final version had been despatched.

One of the most important features of the evidence in the letters concerns the form and content of the addresses written on the back of the leaves. As has been stated, letters were nearly always written in two columns and addresses were written on the back of the right-hand side of the leaf which normally contained the second column. We have noticed only one exception to this: 319 appears to have the address on the back of the left-hand side of the letter, though as only a fragment remains, there may be some special explanation for this.12 In 221 the address on the back is upside down in relation to the writing on the inner faces. 343 is also anomalous: it has on the back of one of its two leaves the word Vindol (on which see below), but there is no sign of the name of the addressee on the remainder of the leaf.

As a minimum an address consisted of the name of the addressee in the dative.13 Usually the name consists of both a gentilicium and a cognomen. The addition of a praenomen is most unusual and occurs only in 352. It is almost as rare to find the gentilicium omitted. Except in the case of slaves or probable slaves, this omission is only found in 299, 310 and (probably) 250. As a further aid to identification a title or other description was often added; thus in the case of Cerialis it is usual to find him described as praef(ecto) or praef(ecto) coh(ortis). Much more rarely is the name of the unit given as well (242, 248?, 263, 271, 281, 284, 311). Verecundus is described as praef(ecto) coh(ortis) in 210 and 211. Other ranks or positions are mentioned in 299, 310 and 312. Cerialis' wife is addressed by her full name in 291 and 292, with the name of her husband added. Similarly slaves are addressed by their own name plus the name of their owner (e.g. 301). The name of the addressee is usually, but by no means always, written in elongated letters, to which we have given the name "address script" (Vol. II, Ch. 4). This script is not used for any further description of the addressee nor is it used for either of the other two parts of the address, which we are about to discuss.

The first of these is the name of the sender. This is found in nearly half of the tablets and may of course have occurred much more frequently and simply not have survived. It is written below the address proper, usually in the bottom left-hand part of this half of the tablet. It is frequently written diagonally upwards, starting in the corner. It begins with the preposition a/ab followed, usually, by the sender's gentilicium and cognomen.14 The gentilicium can be omitted: for examples other than those involving slaves see 260, 263, 310, 312; Severa also omits her gentilicium, see 291-293. To the name a title may be added (215, 255, 258, 263, 284) or the word collega (210, 260, 341, 345);15 fratre is found only in 310. Slaves give their owner's name plus seruo (301, 347). In 292 Severa may add her husband's name, and other tablets probably record the unit in which the sender was serving (281, 299, 319).

Very occasionally there is one further element in addresses, the occurrence of a place-name. This happens only four or five times in the tablets published in this volume: 310 has Londini; 312, Coris with some further description (see note ad loc.); 338 and 343, Vindol (i.e. an abbreviated form); 271 may also have a place-name on the back. This has given rise to some controversy, since we should naturally expect an address to state the place to which a letter is being sent, but the case in which the place-names occur is locative.16

Before attempting to resolve this problem we must look at the other evidence from Vindolanda. One other wooden leaf-tablet (Inv.no.1022) has on the back Vindolande (= the locative form Vindolandae),17 and addresses are legible on at least five of the stilus tablets: Inv.no.689, which also reads Vindolande; Inv.no.722, Coris; Inv.no.851, Vinouis; Inv.no.836, Cataractonio and Inv.no.575, Eburaci.18 In every case the place-name comes first, at the head of the address, before the name of the recipient; in two cases (343 and Inv.no.1022) it is written diagonally across the top left-hand corner. Further examples from other sources are to be found in CEL. We are quite certain that in every case the place-name is locative.19

This evidence would seem to compound the problem. It obviously makes sense for letters found at Vindolanda to have the name Vindolanda as part of their address (as happens in four of the ten examples cited); but it is not obvious why the case should be locative rather than accusative. If the places are to be explained as those from which the letters were being sent, we can understand why letters found at Vindolanda bear the place-names Londini, Eburaci, etc.; but, apart from the objection that an address should state where a letter is going to rather than the place from which it has come, such an explanation will not do for the four letters with Vindolanda on the back. One then has to fall back on the implausible notion that such letters were drafts or copies which were never sent.20

The solution to our problem, we believe, is supplied by evidence from outside Vindolanda. Firstly, in a number of tablets from Vindonissa the address begins with the word dabis;21 this is true also of a stilus tablet from Carlisle and may occur in one of the stilus tablets from Vindolanda.22 This suggests the possibility that such a word is to be understood with all addresses, which therefore mean something like "Deliver to NN".23 It then becomes easy to understand that, where a place-name was added, the sense would be "Deliver to NN at [place]", thus accounting for the locative. Against this it may be argued that the word order is wrong - that the place-name ought in that case to follow the name of the recipient. Here we may call upon the evidence of another Carlisle stilus tablet addressed to a certain Tertius, where, after the name of the recipient, we read Lugualio [sic].24 If that does not seem conclusive, we believe the evidence of yet another Carlisle stilus tablet puts the matter beyond all reasonable doubt. The address on this tablet reads (before the name of the addressee) Trimontio aut Lugua[l]io [sic].25 This is a conclusive piece of evidence since it clearly cannot refer to the place of writing and must refer to the places to one of which the letter was to be delivered (i.e. the sender was uncertain whether the addressee was at Trimontium or Luguvalium). Theoretically one could argue that the function of these place-names was not always the same, sometimes indicating the place where the letter came from and on other occasions the place to which it was being sent;26 this in our opinion is a counsel of despair.27 We must surely try and find a solution which accommodates all examples of the occurrence of these place-names and such a solution we believe we have now given. It may be added that this solution is also consonant with the evidence of the addresses in Greek letters on papyrus. There are several examples which contain instructions for delivery, with and without the imperative ; in several of these the personal name (dative) is preceded by plus place-name (accusative).28

This leaves only the oddity that six of the ten examples of letters found at Vindolanda quoted above record a place other than Vindolanda, and the need to explain why letters written for delivery at York, London, etc. should have ended up at Vindolanda. The explanation must simply be that they were received by their addressees at York, London or wherever and that subsequently the recipients came to Vindolanda bringing their correspondence with them. In view of the fragmentation of units and the movement of personnel attested in the Vindolanda tablets and elsewhere this is not difficult to believe.

It is important to emphasise the value and importance of the evidence for the physical characteristics and appearance of the Latin letter. The papyri from the Roman east supply little such evidence and the literary letters of Cicero, Pliny and Fronto give us none. Although there is a good deal of ancient evidence for the principles and practice of epistolography,29 there is no ancient text which describes the characteristics of the layout of a letter. An important development may have been initiated by Julius Caesar who sent letters to the senate quas primum uidetur ad paginas et formam memorialis libelli conuertisse, cum antea consules et duces non nisi transuersa charta scriptas mitterent ("which he seems to have been the first to change to the form of a paginated notebook whereas previously consuls and generals had only sent despatches written with the roll rotated through 90 degrees").30 The relationship between these features of layout in our Vindolanda letters and the layout of a "literary" letter is of major interest but must be left open. What is important, as with documents, is the relative uniformity of convention; correspondents knew what to expect to see and this is a vital part of the development of a "grammar of legibility".31 Documents and accounts.

Since we now have a greater range of types of document it is perhaps not surprising that we also have a much larger number of ways in which the leaves were used. The folded diptych, which is so common in the letter format, is employed. In the most notable example (154) it is used transversa charta but it should be noted that this diptych is very much larger than any other example. Another example, also a large tablet (161), exhibits the familiar notches and tie-holes. 164, which may be a memorandum of a military nature, is also written across the grain and it is to be noted that the text is in narrative form unlike other texts written across the grain. Reports detailing the activities of groups of soldiers are written along the grain possibly only on half-leaves (155-157) and this is also consistently the format of the reports with the renuntium-heading (127-153) as well as the applications for leave (166-177).

The accounts which were found in the 1970s led us to expect that texts of this form were likely to be written in narrow columns transversa charta and there are indeed several more examples (e.g. 181, 185, 194, 196). Although there is no clear example of another concertina-form notebook like 190,32 180 may be a triptych which has been folded concertina fashion; this contains an account written transversa charta on one side and a letter or petition written along the grain on the backs of two of the three sections. There are several examples of short accounts written along the grain, perhaps on half-leaves (e.g. 178, 193). More interesting are two unusual examples written along the grain, one with an account in two columns like a letter (182), the other in three columns of which the middle one straddles the central fold (184); on both these accounts the text appears to continue on the back of the leaf. It is clear that there was no standard format to which the writers of these reports and accounts (which we assume were all produced at Vindolanda) were confined, but there is what might be described as a broad commonality of convention in the use of the wooden leaves.

Stilus tablets

Although we are here excluding the Vindolanda stilus tablets (of which there are over 100 from the 1980s) from detailed consideration, it is appropriate to mention a recent survey containing material which has appeared since 1983 and to list a few recent items of particular interest.33