The site has long been considered the fundamental unit of archaeological analysis. There is, however, little consensus on what a site actually is and, perhaps more importantly, on the way in which we as archaeologists should think about and approach them. With wholesale paradigm shifts in archaeological thought, theory, and practice there are conceptions of “ site” that span the spectrum from the conservative (Renfrew 1973, for survey archaeology in particular) to the post-modern (Shanks 2004; Hodder 2003). Like so many attempts to complicate a seemingly simple term, we start with a dictionary definition. The closest thing to an archaeological dictionary is Renfrew and Bahn’s Archaeology: theories, methods, and practice (2006), in which they define a site as: A distinct spatial clustering of artifacts, features, structures, and organic and environmental remains—the residue of human activity
The terms of this definition are twofold: spatial distinction and features (defined broadly), which are important in that they imply human activity. While there are several points for critical unpacking, we will focus on complicating two major aspects of this conception of what signifies an archaeological “site.” First we address the difficulty of identifying discrete boundaries—both physical and cultural, and secondly we examine the tradition of assigning singular functions and uses to unquestionably complex places.
SITES as spatially discrete?
The first part of the definition by Renfrew and Bahn highlights the spatial distinctiveness of a site. While they make allowances for disparate sizes of study area based on the research questions and methods, there is still an implicit understanding of a site as something that has a discrete boundary. This conceptual framework is well illustrated by the diagrams at different scales of early farming settlements on the Rhine River. The scales increasingly pan in from the entire continental region to an individual structure. In the text associated with these diagrams they do suggest that “an archaeologist might study—at a broader scale—the interesting association between sites…At the smaller scale the association—established by excavation—of houses with other houses (55).” While this perspective reflects developments in archaeological discourse toward exploring issues of trade and exchange both between and within sites, the implication of this hierarchy is that the site of study—at any scale—is still a discrete entity that interacts with other entities similarly defined. However, people in antiquity, as today, were not necessarily bound by the constraints of their inhabited space, and therefore spheres of interaction could be broadly construed—both spatially and culturally.
The boundary of a site was not necessarily (and we would contend, not usually) coextensive with the extent of the “distinct spatial clustering of artifacts and features which denote human activity.” Further, the boundaries of these areas were permeable resulting in occupation and use of areas beyond what is archaeologically definable as a “site.” What is missing from this traditional viewpoint is attention to the ways in which boundaries were flexible such that people moved into, out of, and in between these spatially distinct entities. Two examples will provide illustration for the need to complicate our understanding of sites as things with discrete borders.
The poleis of Archaic and Classical Greece were often defined by their walls. The space contained within the walls was called the city (polis), and the space beyond the walls was called countryside (chora). It had been long held within archaeological scholarship that the polis and chora were to be considered in diametric opposition to each other—that the city represented civilization and order and the countryside was a symbol of barbarism and chaos. However, with his work on the Greek polis De Polignac (1995) has shown that the walls of the city did not delimit the extent of human activity, but rather that there was a close relationship between the city and the countryside. Furthermore, the extra-urban sanctuaries located in between poleis served as gathering places for spatially disparate communities. In this case, important places for ritual action were located outside the walls of the polis. Although spatially removed from the urban center, these sanctuaries were central to the social life within the communities that were defined (in the scholarship) by city walls.
Debates within the sub-field of island archaeology provide another example for our challenge to the notion of bounded sites. The earliest practitioners of the sub-discipline, coming from a tradition of Darwinism and biogeography, lauded the study of islands as an opportunity to investigate sites in a controlled environment—like a laboratory (Evans 1973, Held 1993). The control, they claimed, came from the fact that islands are isolated, so contact and exchange were either limited or non-existent. From this perspective, the coasts of islands were held to be discrete boundaries. Islands were the unparalleled perfect unit of study. However, after emerging from this tradition many scholars see contact, trade, and exchange between islands and to the mainland as a fundamental feature of island life (Rainbird 1999, Eriksen 1993). This result has been a rallying cry to discard altogether the study of islands as discrete units in exchange for a conception of the sea as the mode of interaction between an interconnected network of islands—or a study of “seascapes” (Broodbank 2000). The boundary between units that were once conceived to be spatially distinct are now held to be flexible in their limits—both physically (because they are considered to extend into the water that surrounds them) and culturally (because they were in contact with a multiplicity of other people and cultures).
No site is an island, entire of itself—there are always connections that extend beyond the locus of study. Both in the physical space that is defined by a “distinct spatial clustering of artifacts, features, structures, and organic and environmental remains”, and in the networks of trade and exchange, any unit defined as a “site” will extend beyond its borders. As a theoretical exercise, it is fruitful to critique the constrained spatial and relational definition of a site. However, when it comes to practical matters it would be impossible to define and follow all of the ways in which a “site” extends beyond its’ “borders.” One way to complicate our understanding(s) of the sites we investigate is to acknowledge their multiple and polyvalent uses and meanings—both in antiquity and in the present.
SITES as functionally discrete?
While the above dealt primarily with definitions of physical boundaries, critiquing how archaeology tends to delimit what are in reality very fuzzy borders, what follows will assess definitions of a different nature. Accompanying physical boundaries are often functional boundaries, or specific associations between particular locations and particular activities. Labels such as site, we have seen, can be useful analytical tools, helping the scholar towards coherent interpretations of the past. We must ask, however, whether such approaches do more to distort the variability of the past than they do to accurately portray past activities. An account of sites on this level allows for a great range of site sizes, from buildings to territories. What will be questioned below is the validity of creating a characterization of the past based on only one type of activity attributed to, for example, a temple, a market, a house, or even part of the span of a wall. While archaeology tends to define sites based on one functional attribute, many types of functions were performed at sites, suggesting a reevaluation of such a definition.
Archaeological concern for site function is not isolated, with examples form anthropology also articulating a functionalist approach to the relationship between people and things (e.g. Rappaport 1971). Such studies, and many related archaeological ones (e.g. Demarrais et al. 1996), look for the unique function of an object to investigate what amounts to a behavioral study of human social development (cf. Hodder and Hutson 2003:22ff.). While Hodder and Huston (2003:27) note that modern determinations of function are always the products of assumed knowledge of past meaning for artifacts, a similar argument can be made for the so-called places or containers of activity. When a site is determined to be ritual, for instance, this categorization has the result of raising certain expectations for the types of associated practices as well as significantly limiting the range of possible associated practices. Function, instead, should be viewed as variable and context dependent, assuming different meanings for different individuals (Alcock 2003; Appadurai 1986). To illustrate this point I present four brief examples that highlight the mutability of function within a single site. The examples portray the simultaneous presence of different functions, with a view towards how this changes our temporal definitions of site at the end.
Houses in Pompeii, for instance, constitute one type of ‘archaeological site’ that served multiple functions. The typical atrium style house, like the House of Sallust, contains a mixture of living quarters (cubicula), working quarters (frontal workshops and kitchen areas near the peristyle), business quarters (the fauces for greeting clients and the tabularium for hosting them), leisure quarters (peristyle), and entertainment quarters (triclinia). Recognition of this has led to scholarly discussion of an encompassing Roman house that serves simultaneously as a private dwelling and a show of public display (Wallace-Hadrill 1994). In a recent book that investigates the connection between artifact location and the determination of function for the houses of Pompeii, Penelope Allison has shown that modern definitions played far too prevalent of a role in the assignation of room type (1999). The actual usage of rooms in antiquity, it seems, was far more negotiable than archaeologists were willing to allow. Objects such as cooking vessels that were found in unexpected rooms were deemed ‘misplaced’ and subsequently moved to the ‘correct’ location. It seems, therefore, that ancient conceptions of the relationship between location and function were far more flexible than we are willing to acknowledge. More importantly, however, is how the multi-functional aspect of a Pompeian or Roman house necessitates a relative and situated understanding of what a house ‘is’ in 1st century AD Roman Italy. The definition of such a site as ‘house’ has both breadth and specificity to it.
The tombs of Mycenaean Heroes have similar depth to them. The products of post-Mycenaean reverence for the Greek past, long tomb chambers cut out for burial during the Late Bronze Age were unburied and pointed to as physical connections to the idealized age of the heroes. At Mycenae, for example, partially collapsed chamber tombs dot the sloping countryside outside of the city walls, indicating uncontested signs of post-inhumation looting throughout the city’s territory. Nearer to the city walls is the large, domed tomb of Agamemnon, with its monumental ashlar masonry cutting a channel through a hill that has been hollowed out and supported by smaller stone blocks. This tomb, which never actually held the remains of Agamemnon – it was constructed before the age of the heroes – remains part of the discussion on the patterning of mortuary landscapes in the Greek world. That further evidence suggests this chamber may have acted as a treasury complicates any process of labeling the room according to a pure functional component. As with the Pompeian house, the strict adherence to properly located activities is representative of ideals rather than real world situations.
The third example has to do with sites in an archaeological survey context, where oftentimes diffuse surface scatters are interpreted either as ‘site’ or ‘off-site’ by archaeologists. While there are pertinent issues here for understanding the difficulties in constructing physical boundaries of sites, the categorization of sites into functional types is equally problematic. A typical survey project for the Roman period of Italy, for example, will utilize a labeling system that includes cities, towns, villas, large and small farmsteads, and sanctuaries (cf. Barker 1995). The reasons behind such choices are to enable deeper understanding of the social organization of a given region, the purview of much archaeological work. Survey archaeology, however, is subject to the same potential problems as excavations are in terms of limiting the extent of potential interactions with the landscape through the application of strict functional labels. A sanctuary at Campochiaro, for instance, served more than just a religious purpose in the Hellenistic and Samnite periods of eastern Adriatic Italy. The site also brought people from the dispersed Samnite communities together, hosting trade, negotiation, and knowledge exchange among many other things. Such nuances of the social landscape tend to get blurred by categories that emphasize one particular aspect of a sites usage.
The final brief example of issues confronting functional categories of sites in archaeology focuses on this notion of knowledge exchange at any site. In a phenomenological study of Neolithic Britain, Edmonds (1999) intersperses his archaeological study with narratives that highlight the experiential component of sites. Things like sight, sound, smell, and temperature, which are undetectable through traditional archaeological methods, are brought to the surface in this way. Of particular interest for this discussion, however, are the interpersonal communications that accompany every example of a site in his book. Detection of an abundance of chipped stone in the present day may compel the archaeologist to label the location as a lithic production site, resulting in the loss of significant detail. In Edmonds’s (1999) narrative, the site is a yearly meeting ground for an extended family. There, younger family members received education about proper lithic production techniques, but they also share stories about the extended landscape: which areas are more fertile this year than in years past; what news is there from other groups of people encountered throughout the year; who is marrying into which groups. The site, in this example, serves to consolidate familial relations, to pass down knowledge to new generations, and to extend a collective sense of the landscape throughout the group. This site, in a sense, is both multi-functional and without limits.
Conclusion
What exactly is a site, then? Perhaps the best answer is that sites and their boundaries are constructed through discourse. This discourse runs the gamut of turns in archaeological theory, much of which focuses on how sites are defined and how they are used. In this paper we have focused attention on the ways such definitions and boundaries can break down when pushed. Physical boundaries are rarely, if ever, discrete, and the associated functional boundaries are rarely, if ever, singular. Nevertheless, the discovery of and definition of sites remains one of the salient aspects of archaeological research today. While the term is useful in how it synthesizes the wealth of data that are available for any given location, an accompanying critical awareness of the results of these syntheses should be accounted for in the literature as well.
Bibliography
Alcock, S.E. (2003) Archaeologies of the Greek Past: landscape, monuments, and memories. Cambridge; New York: Cambridge University Press. 57-77.
Allison, P. (1999) “Lables for Ladles: Interpreting the material culture of Roman households,” in P. Allison (ed.) The Archaeology of Household Activities. London; New York: Routledge.
Appadurai, A. (1986) “Introduction: commodities and the politics of value,” in A. Appadurai (ed) The Social Life of Things: Commodities in cultural perspective. Cambridge; New York: Cambridge University Press. 3-63.
Barker, G.W.W. (1995) A Mediterranean Valley: landscape archaeology and annales history in the Biferno Valley. London; New York: Leicester University Press. Broodbank, C. (2000) An Island Archaeology of the Early Cyclades, Cambridge University Press, Cambridge
Demarrais, E. Castillo, L.J., and Earle, T. (1996) “Ideology, materialization, and power strategies,” in Current Anthropology 37, 15-31.
De Polignac, F. (1995) Cults, territory, and the origins of the Greek city-state. Chicago : University of Chicago Press
Edmonds, M. (1999) Ancestral geographies of the Neolithic: landscapes, monuments and memory. London; New York: Routledge.
Eriksen, T. H. (1993) “In which sense do cultural islands exist?” Social Anthropology 1 (1993) 133-147.
Evans, J. D. (1973) “Islands as laboratories for the study of cultural process”. In C. Renfrew (ed.), The Explanation of Culture Change: Models in Prehistory, Duckworth, London. Pp. 517-20.
Held, S. O. (1993) “Insularity as a modifier of culture change: the case of prehistoric Cyprus”. Bulletin of the American Schools of Oriental Research 292 (November 1993) 25-32.
Hodder, I. (2003) “Always momentary, fluid, and flexible: towards a reflexive excavation methodology,” in Archaeology Beyond Dialogue. Salt Lake City: University of Utah Press. 31-42.
Hodder, I. and Hutson, S. (2003) Reading the Past: current approaches to Interpretation in Archaeology. Cambridge; New York: Cambridge University Press.
Rainbird, P. (1999) “Islands out of time: towards a critique of island archaeology”. Journal of Mediterranean Archaeology 12.2 (1999) 216-234.
Rappaport, R. (1971) "Ritual, Sanctity and Cybernetics," in American Anthropologist 73: 59-71.
Renfrew, C. and P. Bahn. (2006) Archaeology: theories, methods, and practice
Shanks, M. (2004) “Three rooms: archaeology and performance,” Journal of Social Archaeology 4: 147-180.
Wallace-Hadrill, A. (1994) Houses and society in Pompeii and Herculaneum. Princeton: Princeton University Press.
