Reconstruction

The issue of the restoration of inscriptions and papyri is bound up with the much broader question of how inscriptions are used to further historical arguments. There are no simple rules that can govern all situations. Sometimes it is sufficient to work from within a specific corpus of texts. More often than not, however, inscriptions cannot be allowed to bear the brunt of a case without reference to other categories of evidence, as we shall see in reviewing some of the (excellent) work that has been done in the areas of the Roman family and Roman public administration.

In 1984 Brent Shaw and Richard Saller published a study of family structure in the western empire based on the evidence provided by an extremely large sample of funerary epitaphs (12-13,000 out of 25,000 read) (Saller and Shaw 1984: 125 n. 8). In this survey they divided the sample between texts commemorating relationships within the nuclear family and relationships outside of it (non-kin, heirs, and the extended family). In civilian populations they showed that in 75-90 percent of the texts that listed commemorators, the commemorators came from the nuclear family of the deceased. They compared these results with military populations in the western empire, noting that in some areas (especially North Africa, the Pannonias, and Spain), the military pattern conformed to the civilian, a fact that could be explained for North Africa and Spain by the fact that these legions tended to be stationary and not to engage in major wars. In north-western Europe, however, the pattern was quite different, suggesting, quite reasonably, that the commemorative evidence revealed different patterns of recruitment in the army (Saller and Shaw 1984: 145). They concluded that ‘‘though on the narrowest view these tombstone inscriptions tell us only who fulfilled the duty of providing a memorial to the deceased, there are strong reasons for believing that fulfillment of this duty was closely related to transmission of property, to a sense of familial duty and feelings of affection’’ (Saller and Shaw 1984: 145). This is reasonable as far as it goes. But they went beyond this to argue that the nuclear family was the main form of familial organization in many regions of western Europe as early as the Roman Empire (Saller and Shaw 1984: 146), and Saller went on to assert that Romans ‘‘drew a conceptual circle around the mother-father-children triad and made it the center of primary obligations’’ (Saller 1984: 355).

There are plainly two very different questions here, and neither of them can really be answered on the basis of inscriptions alone: first, what is the relationship between commemorative pattern and household organization?; second, is household organization prescriptive of affectionate patterns? In 1996, Dale Martin published a valuable study of funerary inscriptions from Asia Minor in which he showed that there was considerable regional variation in commemorative patterns, and questioned whether the dichotomy between the nuclear and extended family was useful, suggesting that a better model would be to think in terms of degrees of intimacy within the household (D. Martin 1996: 57-8). While Martin’s approach is thoroughly reasonable and will be, as we shall see, supported by a significant body of alternative evidence, his discussion failed to answer the question of why the commemorative pattern in the west should be so overwhelmingly that of the nuclear family. Is the reason ideological? This very issue was raised by Greg Woolf in the same volume of the Journal of Roman Studies that contains Martin’s essay. Woolf noted an issue that neither Sailer and Shaw nor Martin had observed, but which is pertinent to both of their studies. This is the fact that many tombstones are not inscribed at all. In some parts of the empire tombstones tend to contain a relief that is not accompanied by a text (Woolf 1996: 27). He suggested instead that we need to think of all monuments as seeking to establish false claims of permanency in an ever-changing world (Woolf 1996: 39). In this article, Woolf was expanding upon a study by Elizabeth Meyer, who had suggested that the practice of inscription should be linked not so much with family relationship as with status. In their study, Saller and Shaw paid no attention to the issue of date (something that is likewise true of Martin’s study). In Meyer’s view, there was a massive upsurge in the epigraphic habit in the western empire during the second century ce. This was a period in which many people were acquiring the privileged status of Roman citizenship, even though Roman citizens arguably remained a minority in any provincial population. As she correctly noted, the distinction between public and private is hard to maintain in any epigraphic context since, by their very nature, inscriptions are public. She suggested instead that epitaphic commemoration reached a peak in the later second century, before caracalla’s grant of citizenship to almost all free inhabitants of the empire in 212 meant that there was no longer any point to advertising this status on tombstones. If that is true, then it logically follows that the nature of commemoration in epitaphs is related to Roman rules of inheritance, which left the primary responsibility for commemoration with the heirs (Meyer 1990: 95-6). Although there are problems with Meyer’s style of analysis, the most important being that our ability to date funerary epitaphs is vastly less accurate than she asserts (Bodel 2001: 38), it is nevertheless true that, on the basis of inscriptions alone, we cannot be certain whether the evidence they provide is essentially a reflection of ideology or of actual family structure. This point is not affected by the dating problem: the style can still reflect ideology rather than family structure, if people are copying a familiar local style that is understood to reflect status, and that need have nothing to do with caracalla.

To answer the question of whether inscriptions from the Latin west are likely to provide an accurate guide to family structure, we need another kind of evidence altogether. In the case of family structure the necessary body of evidence comes from papyri. In 1994, Roger Bagnall and Bruce Frier published the results of their analysis of census returns from Roman Egypt (Bagnall and Frier 1994). On the basis of their analysis of the returns, Bagnall and Frier show that nuclear households make up a plurality rather than a majority of Egyptian households (Bagnall and Frier 1994: 602, 67). They further show that, while there is a bias in the evidence towards metropolitan households, which show a bare majority of nuclear families, these represent a far smaller proportion of the overall population. The advantage of the census returns is obvious. Although they present snapshots of a household at a given moment, they are not liable to reveal artificially permanent relationships in the same way that inscriptions might. Instead, what the returns reveal is that actual households alternate ‘‘erratically over generations between simpler and more complex forms’’ (Bagnall and Frier 1994: 64). For the historian using inscriptions, these conclusions show that the inclinations of Meyer and Woolf are more likely to be correct than those of Saller and Shaw, that even ‘‘private’’ texts are the result of a desire to publicize an ideologically defined state of affairs.