We have a remarkable tendency to assume that words, rather than things, tell the truth–at least when it comes to writing the history of religion. The Indologist Gregory Schopen pointed out the most concrete problems this prejudice for texts over archaeology causes for writing the history of Indian Buddhism. Here the preference for late, edited, and idealized scriptural sources over contemporary and unedited material evidence leads to straightforward historical errors. For example, while hundreds of inscriptions and excavated objects show that Indian monks owned and donated property, scholars have continued to insist this was not possible because it was prohibited in canonical texts. Similarly, Schopen points out, while a number of cemetaries have been excavated testifying to cases where monks took elaborate care of their dead, scholars have continued to insist that we cannot be sure of their funerary practices because the scriptures say nothing about it.
But the problem of isolating and privileging words over things goes beyond making for bad history: it creates a fundamental limit in our theory itself. If ideology “has no life outside the things that give it substance,” as the archaeologist Adam Smith—following Althusser–points out, why has the study of ideology so often denied its material dimensions, removing words and ideas from their physical lives? The habit is clear even in those fields of study that would seem best able to attune us to material patterns. Famously, to understand the meaning of objects in the ancient record archaeologists have analogized them to written text. Archaeological theorists such as Ian Hodder describe “reading” assemblages of cookware or grave goods as if they are books. To reduce the essence of culture to discourse and meaning is a interpretive shortcut that cuts out precisely what is most hermeneutically powerful and distinctive about archaeology. In fact, and not coincidentally, Fredric Jameson identified a interpretive problem very much like this as the most common hermeneutical challenge which which the postmodern cultural theory of the 70s attempted to grapple–the archaeology of knowledge in Foucault’s Mots et Choses.
Yet archaeological evidence from the history of ancient religions can provide a deeper perspective on this theoretical limit. Tracing when ritual objects gain the power to act or speak in the world of the living lets us see the role of language in the material world from a viewpoint that is not idealized but historicized, and can even cast light on the historicity of our own theory.
This paper examines how the changing relationship between words, things, and death is marked in the history of Levantine funerary monuments. Throughout the second millennium BCE a ritual division of labor was broadly shared from Syria to Mesopotamia in which one made things but performed words to invoke the continuing presence of dead ancestors. By the middle of the Iron Age, the work of craftsmen and ritualists reflected a change in the means by which the dead became immanent–through inscribed language. And not coincidentally, this was a time when peoples were made immanent through being addressed in the first vernacular script-languages.
Perhaps nowhere has the fetishization of the abstract over the concrete been as prominent as in the study of ancient religions. As Gregory Schopen points out, scholars of Indian Buddhism have consistently taken the heavily edited and often unplaceable texts of Buddhist scripture as their basis for understanding the religion, even or especially when they contradict the extensive unedited material and epigraphic remains that come directly from the times and places in question. Written ideals are persistently taken as more real than physical remains, precisely because they are written ideals: material evidence is by definition incidental to the essence of a religion. This startling but actually conventional viewpoint is made explicit in the archaeology of Christianity, where “… the material remains that characterize the early Christian archaeology of North Britain cannot be, paradoxically, in any way essentially and historically Christian” because “they are independent of the Word” (Schopen 1997:10-11). Schopen points out that this view of archaeology dovetails with classic Protestant views of religious language and concludes that “the old and ongoing debate between archaeology and textual studies is not–as is frequently assumed–a debate about sources” but “rather be a debate about where religion as an object of investigation is to be located” and in this regard “It is possible, perhaps, that the Reformation is not over after all.”(1997:14).
The theoretical problem of the location of religion is a symptom of a issue broader than religious studies. In fact it is strikingly similar to a problem identified by Deleuze and Guattari and Frederic Jameson in hermeneutics and Webb Keane in anthropology: the fact that language is best grasped as a pattern with its semantics set aside. As Deleuze and Guattari put it, “No one has been able to pose the problem of language except to the extent that linguists and logicians have first eliminated meaning; and the greatest force of language was only discovered once a work was viewed as a machine, producing certain effects, …But on condition that meaning be nothing other than use.” [Anti-Oedipus 109]. Jameson shrewdly noted that all of the now-classic postmodern responses to this problem, from Foucault and Derrida to Kristeva, involved some kind of renamed hermeneutics. He himself proposed that “the ideal of an immanent analysis of the text,… amounts less to a whoesale nullification of all interpretive activity than to a demand for the construction of some new and more adequate immanent or antitranscendent hermeneutic model…” [1981:23].
A stable relationship between words, things, and death seems to have held for an extended period across a Near Eastern cultural area stretching from Syria to Mesopotamia. A comparison of archaeological remains and written evidence for funerary ritual from the late third through the second millennium bce suggests a broadly shared ritual division of labor in which one made things but performed words to commemorate the dead and index their proximity. Long-term presence was marked by objects, renewed by short-term ritual acts [for the best documented 3rd millennium Syrian evidence see Archi, for an introduction to the shifts in iconography and ritual see Bonatz and Porter].
An idealized and schematic version of this solution to the problem of death is set out in the Old Babylonian Sumerian poem The Death of Bilgames:
Men, as many as are given names,
their statues have been fashioned since days of old,
and stationed in chapels in the temples of the gods,
so that their names, being read aloud, cannot be forgotten
–(lines 298-301 in the translation of Veldhuis, following George)
In related texts and archaeologically documented patterns, the problem of Death was understood to be two-pronged: a lack of presence and memory in the living world, and starvation in the underworld. A wide range of sites from Mesopotamia and Syria to the southern Levant centered rituals of naming and feeding the dead around monuments and other funerary loci—always uninscribed [Greenfield, Jonker, Sanders, Niehr]. Language was understood as a feature of realtime performance, not a long-term physical feature of objects. During this period archives happen to document the rituals, they make clear that the major role of language was in performance, not inscription. But crucially, parallel patterns of burial and monument construction are also archaeologically documented at sites without archives such as Tel Banat. Funerary objects played varied roles as loci, monuments, or participants but were typically defined as things to which language was momentarily offered, not sites on which language was materialized.
Sometime early in the Iron Age, craftspeople and ritualists initiated a shift in this semiotic ideology…