« Double Vision: Imagines, Simulacra, Replicas | Main | Archaeology through the Lens of Sherlock Holmes »

Archaeological Description and Doubt

Posted by Christopher Witmore

I wrote this paper for a session at the 2011 Meeting of the American Association of Anthropology in Montreal called Science and Technology Studies (STS) and Anthropology: What is the status of our descriptions?

It is about time I posted it here. (Note 1)

Archaeological description is rather peculiar. As we work at describing old things it is often the case that we simultaneously participate in their utter oblivion. Excavation, in particular, has often been taken, not only as synonymous with destruction, but also as a kind of unrepeatable experiment, if experiment is even the right metaphor. One can never return to those deposits and cuts that have been removed. One can never engage those features and things in the way that they were prior to an intervention by archaeologists. We accept these losses as a trade off for those gains made in the descriptive process. But what happens if a description becomes suspect?


This paper shares some preliminary considerations that focus on the relationship between archaeological description and doubt. It takes as its ground a series of examples from the 2011 excavations of the erstwhile Roman fort at Binchester, UK –these are divided into three sections: 1) accuracy; 2) association; and 3) definition. More specifically, this paper seeks to understand what the question of doubt reveals about the adequacy of an archaeological description. Even more perplexing is the impossibility of full verification under the weight of scrutiny. Against a standard notion that once skepticism sets in the validity of the statements are called into question or the data become irrevocably suspect, I wish to suggest that the status of archaeological descriptions should always be approached with a spirit of caution, rooted in a deep awareness of just how risky and tentative is the act of reaching an adequate archaeological description.


Let us begin with an excerpt from Jack’s notebook, a student excavator whose real name has been omitted, from the 2011 excavation season.

July 27, 2011: “ Today we finished up clearing the cobbles in 415 (Note 2) and afterwards Natalie and I began to plan. I’ve been interested in doing this since day one – I was actually the project cartographer in Belize. Here, they do things very differently [from Belize]. That is to be expected since we work in huge trenches [at Binchester]. Planning goes a lot faster, but there are countless cobbles to draw. Using a 1x1 meter grid [a planning square or frame with taunt strings crossing every ten centimeters to form a grid] allows the cartographer to work by themselves. I found that particularly useful. The map is made much quicker, but it felt less precise. I mapped 17 squares, a feat that would not have been possible in one day without the grid. Mapping this way was a valuable experience.”

July 28, 2011: “Today I continued working on my plan. Work went by a little slower today because I was moving into an area with a much higher frequency of cobbles. After lunch, a student from ‘University X’ began to map the region adjacent to mine. She seemed a nice girl until I noticed that her grid was off by several centimeters. I commented on it, suggesting that she fix her grid. She just said “oh well” and continued drawing. It shouldn’t have bothered me, but it really did. Archaeology is a destructive science and the maps we make can sometimes be the only evidence we have of the ephemeral materials encountered in excavation. I couldn’t believe someone with such a disgustingly sloppy work ethic had not only been accepted to ‘University X’ but also to this excavation . . .”

Jack’s notebook speaks of a hard lesson that every archaeologist, aspiring or otherwise, must face – namely, not everyone holds to the same standards in the descriptive process. Jack’s dilemma centers on the question of ‘accuracy’ in planning, which in this case means fidelity to the things shown as they appear to an observer hunched over a gridded-meter frame. Precision is of great importance to Jack. First, the planning frame, which speeds up the process, seems less exact than the cartographic methods used in Belize. (Note 3) Second, the insincerity on the part of Jack’s colleague means that from now on, anything that she does will be suspect as far as Jack is concerned. Credibility and trust are found lacking and, for Jack, both planning frame and commitment on the part of the student from University X are roots of doubt. (Note 4)

Questions of a faithful portrayal count all the way down to decisions made in the course of drawing a cobbled surface. Fortunately a few centimeters are not always such a critical issue on a context extending over an area of 50 square meters or more. This is not to wish away the dilemma as Jack sees it. To the contrary, it is a constant worry on a training excavation like Binchester, which is full of such differences of care and commitment (for those sticklers among us, the inconsistency was picked up during the verification process where student from University X’s plan did not match up to the grid for the larger plan).

As a teaching excavation, these worries are encompassed in its design. (Note 5) Rather than being backed into the corner of establishing criteria for faithful witnessing and judgment, these issues have relied more on a combination of developing trust and a loose set of guidelines as to immanent discrimination in the description process. As we delve further into these checks and balances, let us also cover some basics regarding the descriptive process.


July 11, 2011, 10:07 am: Joseph and Jamie are having a conversation concerning the phasing of a series of contexts on the north edge of Trench 2, which is a c. 43 x 20 m area encompassing Dere Street and the vicus (an ad hoc civilian settlement adjacent to a Roman military fortification) immediately to the east of the former entrance of the fort. The discussion centers on the proper phasing of some discrete areas of rubble, a potential wall line, and a couple of fill layers. Because of subsequent changes in soil and their supposed relationship to the lines of stone at the edge of a possible pit, they have to adjust the phasing order present on their context sheets. 5172 is replaced by 5146 as being below 5181.

The single context system requires one to excavate in phase – one does not put off the question of what something is or how it relates to other features around it. The sequencing must be worked out during excavation. Phase translates into what archaeologists call a ‘relative temporality’. In the context sheet seen here, for example, 5184 is above 5147 and below 5074. This translates into 5184 being earlier that 5074 but later than 5147.

Centered on the sequence of statigraphic relations between different contexts, Binchester, like many British archaeological projects, enrolls the single context system. Contexts are instantiations of events – what remains of past events – that can be differentiated from other events on the basis of various observable “traits”. Contexts are manifest either as “cuts” or “deposits” and traits vary with the object observed. Jamie, for example, guides Joseph on through the descriptive procedure required by the pro-forma context sheet describing context 5184 which is interpreted as a “cluster of tightly packed stones.”

The “traits” of the context are to be specified within a series of boxes. Is the context wet or dry (weather impacts one’s ability to trace a surface, wet surfaces reveal changes in soil better than dry)? What is its texture (e.g. “sticky”, “smooth”, “slightly gritty” etc)? Compaction (e.g. “cemented”, “loose”, “firm”, “solid”, “friable” etc)? What is its composition and color? Does it have inclusions? Is it truncated horizontally (cut by plowing)? Have animals disturbed the context; or is it waterlogged?

Through the pro-forma sheet, the deposit is translated as a sketch map, a diagrammatic sequence of contexts (Harris matrices), narrative and boxes pertaining to the descriptive details that Joseph seeks to understand. A guarantor of dutifully followed protocols, the context sheet fulfills the simultaneous role of reference grid by providing continuity to other contexts, features, finds, media (Black &White, Color Slide and digital photo, video, plans), interlocutors (Joseph as the excavator, Jamie as the supervisor, JCB, mattock, trowels and so on) and other sundry details of the excavation process (whether the context was dug well, with sufficient time and in what weather conditions, and whether the supervisor is confident about the diagrammatic relation in the matrix).

The context sheets are designed to not only accommodate necessary standards (with an aim towards compatibility) and protocols, but also to avoid pre-determinations of terminology/classification to the extent that it is possible. Cognizant of the dangers that one’s actions are constrained to the observation of ordered, information sets as specified on the context sheet, supervisors like Jamie take a much more open and pragmatic line on excavation as an on-going and developing. Hence, with Jamie and Joseph we see the modification of record in response to further trowel work tracing differences between various contexts. This redefinition is part of a larger process of “coming into existence” that needs to be further scrutinized – not because of any destabilizing effects that opening these things up to questioning would have, but precisely because this instability is key to their realization as contexts.


July 5, 2011, 2:48 pm: Peter, Jamie and David discuss the next steps in excavating a “robber trench” which was cut along the middle section of the eastern wall line of a 4 x 12 m rectangular foundation for a building in Trench 2 (here I speak with a certainty that they still lack at this point). The three of them stand atop the southern wall line while they survey the robber trench along the western wall to the north.

Peter: It is not very deep is it?

Jamie: That is the problem that I’ve got with it. I don’t know that we have bottomed it . . . because of the fact that if you look at the end of the wall exposed in the section (of a later ditch running east/west that cuts through the whole building) it has two courses below the level of the bottom of the trench. And it is not there . . . I do believe it is a robber trench . . .

Peter: I don’t know about the back half of it; I didn’t look at it. But I did check the front half while they excavated it and it was a really nice smooth bottom [the terminus of the cut], which is very good. (Peter pauses)

Is this building too small to have a big foundation at either end and just a small . . .

Jamie: . . . er rubbish course of stonework in the middle . . . I don’t know.

Peter: Like isle buildings would have.

David: That wall on the other side looks like it has a single course through the middle. (David, points at the middle of the western wall line.)

Jamie: It does, doesn’t it. (Peter, Jamie and David move over to the opposite side of the far wall to have a better look at the course.)

There are at least two courses here, because you have one course and another sitting on top of it where Chris’ foot was. (They nod in agreement with David’s observation.)

Peter: So maybe they took all the walling out. (Peter points back along the robber trench.) I tell you what . . . the back end of that edge is rubbish as well because it looks as if they left some clumps in.

I will get Kat or Janet to have one more look at it as well [meaning inspection by trowel], just to make sure that it is finished as it is. (They all move on to discuss the sequencing of other contexts to the east).

How a context or feature comes to be defined is a disjointed process. Peter and his colleagues have returned to an area excavated during the previous season in order to verify the definition (and, hence, description) of a cut as a robber trench. Agreements need to established before the descriptive process can move on and here we encounter this critical moment. Excavation cannot continue until they are confident about their description and there is no returning to this moment once they continue digging in phase. Why is the trench so shallow? Has it been bottomed? If doubt is present then the decision making process expands to consider other material. The line of a wall and missing courses, grey-matter recall and changes in the traits of soils, experience and the foundations of other ruined buildings, into the situation follows other entities and their rapports.

Working through the description of any context requires this space for hesitation. Either the trench is not ‘bottomed’ or the wall courses do not extend beyond a single course with respect to what remains; those present side with the latter. Everyone leaves this exchange modified – Peter, David, Jamie, the remnants of building walls, and the robber trench. And even at this point, another competent excavator will be called in to inspect by trowel.

Deviate from this path slightly and the description will be in doubt. If Peter passes judgment on the basis of an a priori story into which he fits the site rather than remaining open; if an excavator appeals to hierarchy as a basis for a decision rather than listening to the arguments of others, other possibilities will be missed. If an excavator describes a feature out of prejudice then his competency will always be suspect. Trust will be lacking among his colleagues and skepticism will haunt everything he or she publishes.

Description and doubt

Archaeology, like other sciences, is a delicate business, whose truth and objectivity lies in the piecemeal transformation of the material past into descriptions. If truth and certainty are performances, then so are both falsity and doubt. All are modified as new information comes on the scene – leading to an increasing confidence or an increasing lack thereof.

Truth, according to William James, “is simply a collective name for verification processes” (1978, 104). Truth is a condition of the passage between what precedes and what follows our descriptions. In pedology one can return to the original area in the Amazon forest in which a sample was taken. In biology one can track down other examples of Dendrolagus pulcherrimus in the montane forests of northern New Guinea. In chemistry one can replicate a new antibacterial chemical compound. In anthropology one can often reestablish contact with an informant. (Note 6) But the possibility of verifiability through a return to a precedent situation is absent for entities encountered in the course of excavation. The material past is held in place but for an instant, a transient moment, only to fall back into oblivion, but this time for good! (Note 7) Archaeologists cannot return to a cobbled surface, the relations between contexts, or the cut of a robber trench as they were. The path of verifiability thus lacks a critical gap necessary for ‘direct’ or ‘actual’ verification. In retracing our chains of transformations archaeologists smack up against the wall of media. The sum of these descriptive media add up to more than the sum of their parts, but equally, they also count for less than that which they translate.

What then is the status of archaeological descriptions?

For the old things described here, plans, context sheets, and video diaries are all that remain of them. Even so, perpetual doubt is not our lot. It is here that we may situate doubt as a spirit of caution that is not quite the steadfast skepticism and distrust characteristic of the ordinary empiricism. Rife with indeterminacy, archaeological work doesn’t fail by proposing a possible description for archaeological materials; it fails by insisting on a necessary one.

While there are many nuances that remain ambiguous in this short paper, archaeology’s peculiar situation with respect to the possibility of retroactive verification and thus truth requires a very different approach with respect to the question of adequate description. In closing I underline several key propositions: 1) the burden of proof demands we open up every step of this process to the possibility of public scrutiny; 2) manifesting material entities requires a dual path towards translating the local into easily transferrable (“global”) terms and simultaneously articulating something of locality, specificity and the ineffable qualities of things; 3) the verification and thus truth of a description should, as Whitehead suggests, be sought “in its general success and not in the peculiar certainty, or initial clarity, of its first principles” (P&R, 8); 4) the grounds for an archaeological proposition should not be that of certainty, but rather that of openness both in approach and articulation. Archaeologists may say something is viable, plausible, possible, likely or even probable, and while all of these are just short of certain, they nonetheless are more faithful to the realities of those things that play a role in the production of the past.


1) Thanks to Tim Webmoor, Helene Ratner and Malte Ziewitz for the kind invitation to take part in the session.

2 415 refers to a layer of soil and small stones covering a stone and cobble surface to the east of the remnants of a rectangular in Trench 1, a c. 37m by 26m area. and located in the north‐east corner of the fort.

3) Here Jack is referring to the potential for corruption in copying a scene by eye while leaning over in a ten-by-ten grid of ten-by-ten centimeter squares versus direct measurement of a point with triangulation and tapes in a trench.

4) Jack’s entry also speaks to issues of morality and commitment, accountability and authority (Jack’s identity is tied to is experience in cartography).

5) The project provided training in archaeological fieldwork techniques to students from Durham, Stanford and Texas Tech Universities and to members of the Architectural and Archaeological Society of Durham and Northumberland.

6) It would, I hasten to add, be impossible to ever return to any original situation – no one can account for every variable, however the differences between pedology, biology, and chemistry on the one hand and archaeology on the other are striking.

7) As an archaeologist, I do not ‘discover’ of the past, as it is often held. Rather, I work with what becomes of what was, which is coextensive with me and such work requires great care and creativity. In the end, archaeologists never find the past as it was; they, as a matter of fact, are caught up in the articulation of something entirely novel.

8) But we may note that they strength of a description increases when weighed in light of the chain of labor, associations, and other interlocutors that lie behind this work.


James, W. [1907] 1978: Pragmatism and The Meaning of Truth, Cambridge, Mass.: Harvard University Press.

Whitehead, A.N. [1929] 1978: Process and Reality: An Essay in Cosmology. New York: The Free Press.