Approaching the Evidence

As a first step in engaging with these particular issues, let us explore some of the problems that arise from the different types of evidence that are available to us, since the nature of the evidence itself and the methodological problems associated with interpreting it can significantly determine our approach. Essentially, the evidence available to scholars of Late Antiquity divides into two kinds: material sources

A Companion to Late Antiquity. Edited by Philip Rousseau © 2009 Blackwell Publishing Ltd. ISBN: 978-1-405-11980-1

(often derived from archaeological investigation, and here inclusive of art) and textual sources (that portion of evidence, of whatever kind, that comprises some form of written record). In this respect, we can already see some of the problems that arise, since, as Brandt and Francis point out, this dichotomy has been influential in establishing barriers between traditional disciplines. Scholars of art history or archaeology are often not skilled in interpreting textual evidence, while scholars skilled in the interpretation of textual evidence are often poorly trained, if at all, in the interpretation of archaeological or material remains. Furthermore, these classifications are in fact unhelpful in themselves, since textual evidence is often a subset of material evidence as presently defined. That is, the lines between the two can blur considerably, since much epigraphy is located on buildings, rock faces, caskets, gems, or other objects of interest in themselves, while mosaic-covered floors and walls, usually of interest for their pictorial or decorative elements, can contain names, dedications, and other forms of writing. Moreover, while the term ‘‘textual’’ implies that such evidence begins and ends in its present written form, this classification can in itself be misleading, since texts such as homilies or orations originated in an oral form and the written record can preserve in fossilization any one of a number of editorial layers. Letters likewise were communicated orally, usually being read aloud to the recipient or dictated in the first instance by the sender (Ebbeler, ch. 19), but at the same time the written letter as preserved can represent only a fraction of the actual communication (see Sotinel, ch. 9). The letter-carrier, an important link in the communicative act, as often as not delivered a report by word of mouth, which might represent the real content of the communication, the letter itself containing platitudes. That in Late Antiquity literary texts, too, should be seen as ‘‘performative’’ is a point raised in this volume by Haines-Eitzen (ch. 17) and Gillett (ch. 26).

This brings us to another issue. Traditionally, when we think of‘‘textual,’’ we tend to think of a particular kind of written evidence - namely that which is literary in form, such as historiographies, saints’ lives, or philosophical treatises, with the concomitant effect that scholars who apply themselves to textual evidence often place less value on less literary textual forms, such as graffiti, epitaphs, or documentary papyri. In a similar fashion, the decorative arts - sculpture, painting, relief-carving, and the like; in reality a subset of material evidence - for historical reasons come with their own set of interpretive problems and ideological approaches (see Harley, ch. 21). As a result, the picture of Late Antiquity that we receive from any one scholar is in the first instance always going to be distorted, since it is a rare interpreter of the evidence who is skilled in every single problematic and every single kind of evidence. When we add to this the current tendency in Anglo-American scholarship to dismiss the political and administrative aspects of Late Antiquity in favor of the cultural and social (Ando, ch. 5), we (and here I mean as much ‘‘we the interpreters’’ as ‘‘we the viewers’’) need in addition to be wary both of the questions that have been asked of the evidence and of the landscape that we gaze upon in consequence.

These are simple problems, however, in relation to the difficulties that the evidence itself throws in our path, when we attempt to interpret it. A key theme in the chapters that follow is the fragility of evidence and its impermanence. There are salutary lessons to be learned in neglecting to take into account what has failed to survive when interpreting what is extant. In archaeology, bones, walls, coins, jewelry, tableware, and other durable artifacts are but a fragment of the entire picture. As Halsall (ch. 27) points out in regard to the territories that bordered the former Roman Empire, the very fact of finding Roman ware in these northern lands means that something that has left no trace in the archaeological record must have been traded in exchange, such as raw materials, slaves, or fabrics. The survival of textual evidence is equally quirky. As Choat (ch. 23) points out in regard to Coptic papyri in Egypt, the more humid climatic conditions in the Nile delta have meant that what must have been a large body of documentary evidence from a major urban center (Alexandria) has failed to survive, while the dry conditions further up the Nile insured that the documents of less significant settlements were preserved. Likewise, the simple decision to conduct an archaeological dig at Kellis (Ismant el-Kharab) in recent years, with its resultant corpus of finds, has demanded the adjustment of a range of assumptions about the rise of written communication in Egypt in Coptic. That is, before this additional body of evidence was recovered, it was assumed that there was a direct link between the use of Coptic and the rise of Egyptian monasticism. Now it can be seen that Coptic was more widely used, had connections with the domestic sphere, and was adopted by women more readily than Greek. New discoveries of doublets of already existing texts can also lead to dramatic reappraisal. In the case of the recently discovered ‘‘new’’ sermons of Augustine (Dolbeau 1996), some of those already known to us are proven now to have been severely shortened, due to editing. Dolbeau 23 is five-sixths longer than its counterpart, Sermon 374; and Dolbeau 26 contains more than 1,500 lines, compared to some sixty in Sermon 198 (Hill 2000). Much of the material that had been stripped out was topical and now helps us to understand better some aspects of the Donatist movement in North Africa. Or, as Trout (ch. 12) points out, the simple act of seeking out late antique material in a particular region (in this instance, Spain) can lead to the emendation of our picture of that territory as an epigraphic desert.

Failure to survive into the modern period, because of impermanence or fragility, or not failing to survive but existing undiscovered are, however, only two aspects of this problem. As Brandt (ch. 11) points out, the privileging of the classical era of Greece and Rome on ideological grounds often led in the past, in archaeological excavations, to the permanent destruction of late antique layers or architectural phases, in order to free that which was valued. Administrative and ideological decisions made in Late Antiquity itself can be equally responsible for the disappearance or suppression of sources. As Humfress (ch. 25) indicates in the case of Justinian’s Digest, in the process of its assembly four centuries worth of jurisprudence was condensed from ‘‘more than three million lines... into roughly 150,000 lines, one twentieth of the original mass.’’ Our reading of the contents of the Digest necessarily alters significantly when we understand, as Humfress goes on to explain, that the promulgation of the document was accompanied by the deliberate suppression of a vast quantity of purged material. Carolingian cultural imperialism, as Leyser (ch. 3) points out, is similarly responsible not only for the survival of Latin texts but also for their loss, just as the wilfulness of individual Carolingian scribes in copying nonapproved texts helped to reduce those losses. The point to be made here is that for a variety of reasons a large proportion of late antique evidence is always missing, and while sometimes parts of it will later be found, as in the examples of the dig at Kellis and the sermons of Augustine, more often than not it can never be recovered. The picture of Late Antiquity that we currently see, in consequence, is always subject to readjustment by the acknowledgment of gaps or by the appearance of new discoveries.

At the same time, gaps that appear to be gaps in the record can prove to be not gaps at all, but a failure in interpretive method. A case in point is the prevalent belief in a seventh-century ‘‘dark age’’ in the Near East as a consequence of the Islamic conquest. Where previously a theory of decline and impoverishment had been based on monumental architecture and site inscriptions, recent archaeological work in Jordan and Palestine has led to an improved coin record for the seventh to eighth centuries, which has in turn led to improved ceramic chronologies and typologies (Walmsley 2007). Where the seventh century had until recently been largely empty of pottery, that void is now being filled, with a resultant shift in the interpretation of the economic situation. This is due not only to freshly excavated material, the value of which can now be recognized, but also because the newly developed tools now permit retrospective identification. Brandt’s Example 1 points to a similar reinterpretation of this period at Rome, likewise on the basis of an improved analysis of pottery. In another example, Trout points to a recent reframing of the apparent paucity of Latin epigraphy in the sixth and seventh centuries. Rather than positing a theory of decline, the suspicion has now been raised that in these centuries inscriptions were increasingly painted rather than incised, leading to a failure not in epigraphic habit, but in survival.

Awareness of gaps in the evidence, their possible causes, and the way in which they can distort our received picture of Late Antiquity, is only one aspect of which to be aware. Another, quite different point that we need to consider when approaching the evidence that does survive is the question of layers. As often as not, different types of evidence exist in multiple sets, or consist of a number of coexisting layers. Negotiating our way through those layers requires making value judgments and applying interpretive frameworks. The simple act of editing a text from a multitude of manuscripts, for instance, implies that we can recover the author’s original version, that we can determine the precise relationship between the different versions of the text (as codified in the stemma), and even that the text has an author in the first instance. These assumptions and the resultant decisions that are made, problematic in themselves, dictate how we read the text and thus analyze the information that it contains. But, as we observed in the case of Augustine’s sermons, the manuscripts on which an edition is based do not always lead back to the ‘‘original,’’ but may represent instead an interim stage in the text. In that instance, the manuscripts known prior to the 1980s had all been copied from a version severely edited by medieval redactors, who had shaped the material to suit their own purposes. Multiple editions within the lifetime of an author are also not uncommon, leading to the coexistence of different ‘‘original’’ phases of a work, as Woods shows in the cases of the histories of Eunapius of Sardis and Malchus (ch. 24). In those instances, do we place priority on the most recent version, as being definitive, rather than privileging the original or first version as we would normally do, or do we value each version equally as having been produced at different times in the author’s life with different agenda? What do we do, moreover, with works that fail to survive in their own right, but exist reshaped within a later work, such as Photius’ ninth-century epitome of Philostorgius’ Ecclesiastical History, or the Chronicle of Pseudo-Joshua the Stylite, an important sixth-century record of local events in Syria preserved entirely within the eighth-century Chronicle of Pseudo-Dionysius (Trombley and Watt 2000)? Editors who publish a document divorced from its context make decisions that permanently shape the reader’s response. Or, what of texts such as the Jerusalem and Babylonian Talmuds, the Tosefta and Midrashim, or the Theodosian and Justinianic ‘‘law codes,’’ which constitute the collation of multiple layers accrued over time? To talk of authorship is complicated and, in the case of the first four items, dates can be assigned only within a broad range. Even within the Theodosian Code, a legal opinion ascribed to a particular ‘‘author’’ is just as likely to have been drafted by a bureaucrat before being promulgated in the ‘‘author’s’’ name. In the same way, letters that purport to be from a particular person’s hand may in fact have been composed at that person’s direction by a notarius, whose formulaic phrases, rather than the ‘‘author’s,’’ the letter conveys. When we make assumptions about authorship in those instances - that a particular rabbi spoke, that a particular consul or emperor promulgated a law, or that a particular individual wrote a letter - our approach to the texts might already be skewed.

The question of authorship itself can be problematic, and attribution, false or otherwise, can shift the context within which evidence is located and therefore interpreted. The late antique period is the golden age of pseudepigraphy, with many texts being passed down to the copyists of medieval times under false names. We have already encountered the case of Pseudo-Joshua and Pseudo-Dionysius within the genre of historiography. The homiletic genre is perhaps the most complex, but also serves to give an idea of the range of possibilities. In the case of John Chrysostom, while around 820 authentic sermons survive, another 3,000 or so are attributed to him in the manuscript tradition. In some cases (well over 1,000) the homilies contain genuine material from John’s sermons, excerpted and combined with new material to create a text that suited the needs of audiences in the seventh, eighth, and ninth centuries. In the case of another 1,000 or so, the homilies were originally authored in their entirety by someone else. The corpora of a number of lesser-known preachers have survived in this manner, such as the sixth-century Leontius, a presbyter of Constantinople, and John’s contemporary and enemy, Severian of Gabala. Similarly, in the case of Caesarius of Arles and Maximus of Turin, survival of many of their homilies has depended on attribution to Augustine. The case of Ephrem is even more complex. Works attributed to him survive in both Greek and Syriac, but the relationship between the two corpora is distant, and the biographies demonstrate two distinct personae (Griffith 1998). Identifying an ‘‘author’’ or even discussing authenticity in this instance is problematic, and so we talk of Ephrem Graecus and Ephrem Syrus to distinguish the two corpora. To further complicate matters, monks of the Greco-Syrian communities of the sixth century, in addition to transmitting the works of Ephrem Graecus in both Greek and Syriac, composed new poetic homilies and songs in his style and under his name (Griffith 1998). Gaddis (ch. 34) points to falsification of a quite different and more malicious kind in the case of conciliar Acts, particularly those of Chalcedon. There, language disparity enabled manipulation of translations of documents from the opposing sides to produce subtly altered texts that supported hostile propaganda.

In the case of art, architecture, and archaeological excavation, the impact upon interpretation of the way in which we handle layers (or phases) is just as profound. The church of Santa Maria Antiqua in the forum in Rome, which dates from the sixth century and is currently closed for study and restoration, is a case in point. The wall paintings, which date variously from the sixth to the late eighth century, present their greatest challenge on the wall to the right of the apse. There, seven separate layers of decoration can be identified: the earliest (marble tile) predates Christian use of the building (in the fourth and fifth centuries); the latest date to the time of the iconoclast controversy (Www. archeorm. arti. benicultural. it/sma/eng). Each layer is significant in itself, each is only partially extant, and each is fragile. Recovery of any one layer is at the expense of another. Whatever decisions are made during the process of restoration will determine how the program of illustration inside the church is viewed for generations to come. The very phrase ‘‘pre-Christian use of the building’’ raises an issue discussed at length by Brandt in his chapter (11) on the archaeological record for Late Antiquity - namely, that buildings go through different phases of use or construction, of which any late antique phase may be only one of many. Do we analyze the late antique phases in the context of other late antique construction and building use, or do we also view the late antique phases of a particular building within the context of that same building’s earlier and later use? Again, whatever decisions we make when analyzing these buildings do much to shape the picture that emerges. Brandt’s discussion of the Harris stratigraphic method and its implications in this respect is particularly instructive. Reuse of building materials or of spolia from the classical period, as in the case of the Constantinian arch in Rome or the statuary relocated in the Constantinopolitan hippodrome by emperors from Constantine to Theodosius II (Bassett 2004), again present us with a diverse range of possibilities for interpretation. As Harley (ch. 21) also shows, whether we privilege the classical origins of the material or the way in which it is reframed and reshaped in Late Antiquity determines whether the late antique construction is viewed as symptomatic of transformation or decline. Whether we view the late antique result as an aesthetic object (in terms of visual reception) or as a political act (or in terms of its religious, social, or cultural function), our choice shapes the way in which we respond to it. Similarly, an art historian, an archaeologist, a political scientist, an anthropologist, a military or social historian will shape in their own way the questions that they ask. Understanding the multilayered nature of late antique evidence and the multiple contexts that consequently arise is an important part of approaching Late Antiquity.

A quite different theme that recurs throughout this volume is the at times contradictory nature of different types of evidence, and the problems this can pose for the interpreter. As we saw in the case of the seventh-century Near East, the evidence of monumental architecture and site inscriptions has been interpreted as pointing to economic decline and increasing impoverishment. Evidence from coins and pottery, on the other hand, now points on the contrary to an active monetary and trade economy, which displays little sign of major economic or organizational failure. In this case, the assumed ‘‘dark age’’ is dark only if we interpret monumental architecture and inscriptions as the chief signs of economic prosperity, and if we are unaware of the other type of evidence. Brandt’s example of the tituli in Rome (where textual evidence - the Liber Pontificalis - claims that they were founded in the early second century, but the archaeological record shows no evidence of this) illustrates how conflict of this kind can lead to a variety of interpretations. It also shows the weight that tends to be placed on textual evidence, with the archaeological evidence being adduced to prove or disprove it: so, in this case, the assertion of the Liber that the tituli existed in some form is taken as reliable, even if its dating remains open to suspicion. Brandt’s point is that such confidence can lead to overinterpretation of the evidence. The comparison with Jewish evidence of community organization also demonstrates how the location of evidence within a broader context can compel reinterpretation. As a final point, Brandt raises a more important question in regard to the handling of textual evidence: whether information from a sixth-century text should be read as anything other than evidence of a sixth-century preoccupation. Marsham (ch. 32) points to a less problematic case: the late seventh-century decree that Arabic be used as the official language for tax administration in the occupied territories in the east. Documentary papyri from Egypt show, on the contrary, that Greek and Coptic persisted alongside Arabic well into the ninth century, demonstrating a gap between official policy and actual practice. As Sotinel (ch. 9) shows, similar discrepancies existed between official and unofficial communication. In those instances, the contradictory evidence does not point to a problem in interpretation, but rather encourages us to look for the inconsistencies that could and did occur within late antique society.

Far simpler problems can cause the unwary to err, and thus reshape the picture. One of the most simple, yet most profound, is the question of date. The dating of late antique texts and material culture is often itself based on a series of assumptions (as in the case of seventh-century near eastern pottery above) and can be far less secure than one realizes. Since the discovery, for example, of the ‘‘dead cities’’ of the limestone massif in northwest Syria, the same coin, pottery, and architectural evidence has been dated variously on the basis of different sets of assumptions, giving rise to substantially different interpretations (Tchalenko 1953-8; Sodini et al. 1980; Tate 1992; Magness 2003; for an overview see Foss 1995 and Magness 2003: 195-9). The resultant variation in dating for the decline of the site spans the period between the sixth and tenth centuries, with profound implications for the thesis of Syrian decline in the sixth century prior to Umayyad rule. In another example, the dating of papyrological documents, spells, or amulets, in cases where no internal evidence for date exists, relies largely on the context of the find and the style of the script in which it is written. These criteria tend to offer no greater precision than a quarter or half century and can as often result in a range of several centuries, again with implications for interpretation. In cases where the termini of an author’s career are known and provide a helpful terminus ad and post quem, our ability to pinpoint the date of texts that they are known to have authored tends to rest largely on a reading of internal evidence. In the case of Augustine’s homilies, agreement on how to interpret the same internal evidence has shifted a number of times in the past few decades, with the result that even the latest dating system (Hombert 2000, who

Challenges La Bonnardiere 1965) is now being overturned (Drobner 2000, 2003, 2004). Likewise, in the case of the largest surviving corpus of Greek homilies (John Chrysostom’s), dates that had been established over three centuries of scholarship are now being overturned (Pradels et al. 2002; Mayer 1999, 2006). When it comes to the dating of a historical event referred to in texts, even when the date of the textual evidence itself is not in question, as often as not two interpreters who derive their evidence from the same texts will fail to agree. A compelling example is the date of the translation to Constantinople in the fourth century of the relics of Saints Andrew, Luke, and Timothy. Utilizing the same evidence, a variety of dates have been offered in recent decades, the latest diverging by some twenty years (Woods 1991; Burgess 2003).

The simple fact of how evidence is presented can also have a profound effect on how we approach a particular body of artifacts or texts. One negative example is the editions through which we access texts. Since the process of producing modern scientific editions is extremely slow, we are as often reliant on text editions produced in the early 1900s or even as far back as the 1700s. The way in which a group of documents is arranged in these instances consequently reflects scholarly opinion of the time of the edition, rather than the present. In the case of the letters of John Chrysostom, the current edition was produced by Bernard de Montfaucon in 1724, and the criteria according to which he arranged and numbered the letters are far from chronological. Recent scholarship has, moreover, completely overturned the received view of the chronology of the letters (Delmaire 1991), with the result that a person who reads them in their present sequence gains a far different picture when they reread them in the sequence proposed by Delmaire. Without the benefit of a reorganized edition, significant information is obscured from the reader, including a sense of how John’s correspondence progressed as historical events related to his exile unfolded, of how different letter-bearers were utilized to deliver items to various locations in packets, and of who was writing to him and with what frequency. The positive impact of new technologies in terms of the way it presents evidence to us, on the other hand, cannot be overestimated. Three-dimensional imaging and computer graphics facilitate the reconstruction of buildings and sites, or the restoration of furnishings and interior decoration (see the computer-generated images of mosaics from Antioch and of the church at Seleucia Pierea in Kondoleon 2000), allowing the viewer to scan and interact with buildings and sites in ways that were impossible with photographs, drawings, or hand-constructed models. At the same time, the Internet is increasingly being utilized by archaeological teams to publish interim results in new ways and with unprecedented rapidity (e. g., Www. sagalassos. be, which includes 360-degree scans of the site topography and summaries of current field notes). Only a small number of scholars of Late Antiquity currently utilize these technologies (for the more advanced resources available in classical antiquity see, e. g., Www. stoa. org), but they are increasingly changing the way in which we approach and ask questions of the field, particularly in regard to material evidence.

While it may be exciting to apply novel approaches to the evidence, particularly those derived from other disciplines, it must also be recognized that each type of evidence has its limits. Asking new questions in order to expand our horizons and to look at the evidence in fresh ways may in fact produce false results (see Rebenich, ch. 6, on social Darwinism and evolutionary biology). While this should be no disincentive to approaching the evidence from new directions, since, as the other chapters show, significant reshaping of the picture of Late Antiquity can emerge, the same chapters as often as not demonstrate the problems that can occur when we push the evidence beyond its natural limits. Gillett (ch. 26), in particular, shows how the desire to read ‘‘barbarians’’ into late antique texts, driven by a late twentieth-century interest in ethnology and identity, is doomed to frustration. With only Greek and Roman texts in our possession, and in the total absence of‘‘barbarian’’ ones, ‘‘we can look through Greco-Roman eyes to see how they perceived the Other and, therefore, themselves’’; yet, no matter how much we want or try to, we cannot, via these same texts, recover ‘‘barbarian’’ self-identity. In a similar fashion, the feminist drive to recover late antique women from male-authored texts, which was prevalent in Anglo-American scholarship of the 1980s and 1990s, is now quietly being laid to rest, in recognition that such an endeavor is fraught with difficulty, if not impossible (Elizabeth Clark 1998). It remains to be seen whether the new tools currently being borrowed from other disciplines - discussion, for example, of‘‘intertextuality’’ and the ‘‘linguistic turn’’ - further our knowledge of Late Antiquity or lead to distorted readings.