Palaeography: The Palaeographical Background

From Alan Bowman and David Thomas, Vindolanda: the Latin writing tablets London: Society for the Promotion of Roman Studies, 1983, pp. 55-60

Before we proceed to an analysis of the script of the tablets we must glance briefly at the present state of scholarship on the subject of Latin writing in the Roman imperial period. We can ignore nineteenth-century works which were written before the important papyrus discoveries at the end of this century and later. These discoveries led to two substantial books on early Latin writing, by Van Hoesen and by Schiaparelli16; but these too now have little relevance for our present purpose. All earlier work has been superseded and the whole subject has been put on a fresh basis in the last thirty years. The pioneer in the field is Jean Mallon, whose work, beginning with articles as far back as 1938 and culminating in his book published in 1952, has been universally recognised as revolutionising the study of early Latin writing17. Mallon began by insisting that the Latin papyri 'remettaient en question les théories et la terminologie établies avant leur découverte'18,and from this he proceeded to a fundamental reappraisal of the capital script ('capitale romaine') and its development into ORC (a script which he preferred to designate 'l'écriture commune')19. He went on to consider the relationship of ORC to NRC and insisted on three points: (1) it is impossible for NRC to have developed directly out of ORC; there is in fact an unbridgeable gulf (Mallon's word is 'fossé') between the two scripts; (2) the change is to be explained by a change in the angle of writing, and (3) this change in the angle does not have its origin in the cursive script ('l'écriture commune' in Mallon's terminology) but in the bookhand. Mallon does not deny any relationship ('parenté') between ORC and NRC; but he does deny that the latter is a direct descendant of the former20.

This is naturally a much more controversial matter, and Mallon has here opened a debate which as yet shows no signs of being closed. Almost at the same time as his book appeared, Cencetti published a very important study of early Latin papyri (i. e. those of the first three centuries A.D.)21.His views coincided only in part with those of Mallon. He also believed that some of the letter-forms found in NRC cannot have been derived from the forms found in ORC as we know it from papyri of the second and third centuries22; and he claimed to have detected in second-century papyri a new style of cursive writing, depending at least in part on a change of angle. Where he differed from Mallon was, first, in insisting that the development to NRC must be looked for in cursive writing, not in bookhand23. Second, since he accepted that this development cannot be traced in the material we possess, he put forward the hypothesis that the new style of cursive writing which he saw appearing in the second century and which became dominant in the third was the 'official' form, alongside which there existed a second 'unofficial' form. This official form, he suggested, became quasi-canonised in the third century in a direction which led it further and further away from any possible development into NRC. It is therefore to the second, 'unofficial' form that we must look for the style of writing and the letter-forms from which NRC began to develop. To the objection that this second, 'unofficial', form is not known in the third century, he replied that this is due to the hazard of survival, that all papyri surviving from this century are, as it so happens, from military or other official sources24.

At about the same time as Mallon and Cencetti were putting forward their different views on the relationship of ORC to NRC, Marichal began publishing his views, which were much akin to those of Mallon25. He also insisted that there can have been no direct development of ORC into NRC, and he agreed with Mallon that the change which took place must be looked for in bookhand, not cursive. Later Marichal developed a modified form of this view, in which he saw as crucial for the development of NRC the existence, as early as the date of the graffiti from Pompeii, of 'baroque' forms of the letters. He continued to insist that these forms were closely associated with bookhands since they were used by the literate members of the community; and he also continued to believe in the importance of the change in the angle of writing, which he associated with the change from the roll to the codex26. Tjäder's views are very different. He expressed himself sceptical of the emphasis placed on the change in the angle of writing27, and was firmly of the opinion that the development into NRC should be looked for in the cursive not the bookhand28. In general, we may fairly say that he endorsed the views of Cencetti, and he is especially relevant for our discussion in that he believed that forms of the letters used in NRC are already to be found in ORC in the second century29. However, he was less eager than Mallon and others to reject any possibility of a genuine development of ORC into NRC30.

In the last few years two further substantial contributions have been made to the debate . The first, by Casamassima and Staraz, is exclusively concerned with the evidence from the papyri31. The upshot of their very long, detailed consideration of the evidence was to reinforce the views of Cencetti. They agreed with him in thinking that more than one style of writing developed within the general category of ORC; in particular there developed an 'official' style of writing which tended to choose the more cursive forms of the letters. By the third century this had become 'canonised' in a direction which led away from letter-forms resembling those to appear in NRC. Alongside this there developed a script which they call 'scrittura normale' or 'commune'32; this script adopted what they described as 'posed' forms of the letters, and this they regarded as the milieu out of which the forms of NRC developed33. In particular, from a detailed examination of the letters which they regarded as crucial34, they contended that these two styles (they spoke of 'bifurcations' or of 'two branches of the same tree') can be seen in existence in Latin papyri throughout the first two centuries A.D35. Most recently, in 1980, Hornshöj-Möller published an article entitled 'Die Beziehung zwischen der ältern and der jüngeren römischen Kursivschrift'36. As his subtitle, 'Versuch einer Kulturhistorischen Deutung', makes clear, he is more concerned with the 'why' of the change than the 'how', and his work is consequently less important for our present purposes. Apart from his useful survey of earlier views, his article is principally of interest for us because he is apparently prepared to contend that the traditional view, that NRC developed out of ORC without further qualification, is still defensible; note in particular the statement: 'Ich glaube, anhand einer graphischen Analyse festgestellt zu haben, dass die jüngeren römische Kursivschrift sich hinter einer gewissen Berucksichtigung der griechischen Schrift organisch aus der älteren entwickelt hat'37.

Mallon regarded the letters in NRC which could not have developed out of ORC as A, B, E, N, P38. Some of these look more convincing than others (cf. FIG. 10), and other palaeographers have insisted only on the impossibility of deriving NRC a and b from ORC A and B39. Indeed, the letter b is something of a shibboleth and has evoked a voluminous literature. As can be seen from FIG. 10, the standard form of ORC has what the French school calls the 'panse à gauche', whereas the form of NRC has the 'panse' (loop or bow) at the right. Mallon was the first to explain the way in which the form of B used in ORC (the form with 'panse à gauche') had derived from capital B40.The correctness of his explanation would today be generally accepted. The problem is to explain how this form of B can have developed into the b of NRC by following normal palaeographical 'laws' (as the traditional view assumed). Mallon, indeed, was confident that the form in ORC could never have led to the form found in NRC and believed that we must look to the bookhand for an explanation. Cencetti on the other hand thought that it ought to be possible in principle to find in ORC of the first two centuries A. D. a form from which b with 'panse' at the right could have derived41; and Petrucci believed he had demonstrated the correctness of this by an analysis of the pottery graffiti from Contodamagos, some of which are dated as early as the first century; the writers of these graffiti were, he thought, semi-literates who had never learnt B with 'panse' at the left at school and were deriving their b in a form very much like that normally found in NRC or indeed like a modern lower-case b, from a clumsy attempt to reproduce the capital form42. Marichal, in an important study of the letter43, claimed to have found a special variant of B in graffiti from Pompeii (therefore antedating A.D. 79) in a form which he designated 'baroque'. This form he regarded as the one from which NRC b later derived. In subsequent articles Marichal was strongly critical of Petrucci, and in particular he insisted that his 'baroque' form of B is the work of literates familiar with bookhands44. At one time Tjäder sought to demonstrate that ORC B could have developed graphically into NRC b; but he subsequently abandoned this view in favour of that of Petrucci45. Finally, Casamassima and Staraz, while agreeing that B is crucial for the question of 'continuità o cesura' between the two forms of cursive, contended, in conformity with their general views, that ORC knew more than one form of B and that one of these forms was taken and developed into the b of NRC46.

This brief summary of the controversy surrounding the origin of NRC, which we trust is not in any way seriously misleading, cannot hope to reflect all the subtle nuances in the views of the different scholars. It is not our intention to discuss further the various merits of these views. For the most part this would not be strictly relevant to our purpose, since all would agree that the critical period for the changeover from ORC to NRC lay in the third century, a hundred years or more after the date of the Vindolanda tablets. It is, however, important that we should bear in mind this whole controversy while analysing the script of the tablets, for the following reasons. The thesis of Cencetti, especially as it has been developed by Casamassima and Staraz, asserts that Latin writing of the first two centuries A.D. should contain forms of letters which were critical for the development of NRC47.The question we need to ask ourselves, therefore, is whether any examples of such letter-forms are to be seen in our tablets.

Fig. 10 Letter-forms in Old Roman Cursive and New Roman Cursive Image ownership: © Society for the Promotion of Roman Studies

There is another respect in which the tablets may throw light on theses advanced by palaeographers. Mallon contended that the change in angle, on which he laid much stress, was to be seen in second-century papyri, citing as his earliest example a papyrus of 13148. Cencetti, while laying less stress on the change in angle, nevertheless considered that second-century papyri showed the development of a new type of ORC, which he described as 'maiuscola inclinata', right-sloping and characterised by more use of ligatures, and in its general aspect more 'leggero, minuto, sottile, slanciato' and distinguished by 'il traggiamento quasi del tutto strisciato'. This new type he found first exemplified in a papyrus of A.D.10349. Although it is not easy to translate Cencetti's description into precise English terms (the crucial word 'strisciato' is particularly intractable), we believe not only that there is substance in what he said, but that the reader who will take the trouble to compare reproductions of the texts to which Cencetti refers, both for the earlier form of ORC and for that which developed in the second century, will clearly see the point that he was making50. The importance for our present purposes is that this new form is first attested, according to Cencetti, at precisely the period to which the tablets from Vindolanda belong, and it might therefore be expected to make an appearance in them.

Finally, we need to note that in papyri dating from the reign of Augustus and shortly afterwards (to about A.D. 50) a style of writing is commonly found which belongs to the category of ORC, but which is much closer to the capital script than the form of ORC which is normally employed in papyri from the middle of the first century onwards51. This distinction is not quite the same as a further distinction which Cencetti seeks to make (with reference especially to the second half of the first century) between a 'stiffer' form of ORC and a freer, more cursive form. He regards the 'stiffer' form as essentially the script of wax tablets, which developed into the freer form as it became more customary to write on papyrus52. This 'wax-tablet' type of script is clearly related to the very early form of ORC mentioned at the start of this paragraph, although not identical with it. What we need to look for in analysing the Vindolanda tablets is any trace of this early type of ORC. We shall return to a discussion of all these points in Section G of this chapter, after we have examined the script of the Vindolanda tablets in some detail.