Critique of Archaeological Reason
5. Excerpts and summaries

Giorgio Buccellati

Marco De Pietri – February 2014

Back to top: Giorgio Buccellati


Back to top: Giorgio Buccellati

Excerpts from Buccellati, G. 2017, A Critique of Archaeological Reason. Structural, Digital and Philosophical Aspects of the Excavated Record, Cambridge: Cambridge University Press

broken tradition(s) pp. 22-23 2.4.2 Broken Traditions
     The distinctiveness of archaeology in terms of the secondary definition rests on the notion of broken traditions. A human cultural tradition is broken when the observer cannot rely on the native competence of a carrier of the culture to which we owe the evidence, nor is there a continuous line of documentation that embodies a self-assessment of that culture (Sagan 1978 for a “reverse” broken tradition).
     The archaeological site is both a metaphor for broken traditions and the locus where these traditions are embedded. Much like how the history of a site was truncated when it began to collapse in its own physical debris, so a “broken” tradition was truncated when its carriers disappeared and its values could no longer be handed down (see also “absence” in Lucas 2010).
     The resulting disconnect is the basis as to why a distinct methodological approach is needed – the approach, precisely, of archaeology. Much as the archaeologist qua excavator must reconstruct the presumed structural integrity of a physical system now in ruins, so the archaeologist qua interpreter must reconstruct the presumed structural integrity of a cultural system that is also present to us only through its own ruins. It is not that the tradition must be “ancient.” Chronological antiquity that is linked to the culture of the investigator through a continuous tradition does not qualify as archaeological and requires a different methodology altogether (see Harrison 2011; Horning 2011; Lucas 2015). The recapturing of a broken tradition entails a kind of phenomenological epoché that proceeds along a quite independent path (see Husserl 1913: §32; Moran 2000: 146–52). It builds on a sample of data that is not filtered through conscious selection, through deliberate choices in function of a canon (more or less explicit), or through the accretion of a still growing self-awareness.
     It is not that we have an “either–or” situation – either archaeology or history. The two certainly meet. But their reciprocal levels of analysis must be kept distinct. Failure to do so is at the basis of many a misunderstandings in such fields as biblical studies, where one can easily fall prey to the danger of arguing as to the merits of a broken record, that of archaeology, from the standpoint of a continuous tradition, that of religion, and vice versa. The Bible as a central point of reference is an apt metaphor: the text is handed down through a manuscript tradition that acquires a sacral value and thus enshrines the very notion of the canon in the most rigid way possible. Analogously, the search for continuities in the ethnographic record (ethno-archaeology) can be both enlightening and deceiving.
     In this light we may understand more fully the implications behind the notion of what is commonly known as a “dead language.” A language cannot possibly be dead, because then it would not have ever been an operative system (see Heidegger 1927: §34). It is only its competent native speakers (all of them) who are dead. This is analogous for a broken tradition; it is an extinct culture, but one that can be rediscovered as a system and thus become re-embedded in human awareness – ours.
     Building on this, we will arrive (see 15.5) at a fuller understanding of the notion of archaeological reason, understood as human reason severed from the stream of cultural self-understanding. It is a significant concept that has its roots solidly in the experience of archaeology but leads beyond it to question the very essence of what hermeneutics is.”
integrated whole pp. 37-38 2.8.3 Structural archaeology
     If there is a label that applies to the kind of archaeology characterized by the approach suggested here, it is “structural” – in the sense that it builds on the principle that all data are seen as part of an integrated whole that is more than the sum of its parts (for a fuller development of the concept of structure see below, chapter 14). This integrated whole is the way in which the data are excavated and then are understood and translated into a coherent representation. Of course, any excavation report does this – i. e., it translates the site into an analogical counterpart. But for this analogical counterpart to be structural, the emphasis must be on the internal logical coherence, the “grammaticality,” of the end result.
     One might also argue that any report is structural, in the sense that for it to make sense it must have some degree of internal organization that meets known logical criteria. But such a loose understanding of the concept of structure proves soon to be inadequate. For the resulting whole can easily become a mere congeries of components, defned in an ad hoc sort of fashion and linked by external criteria which do not explain how the individual parts add up to a meaningful whole. There is no doubt that even such a loose arrangement is useful – much as a telephone book is useful. But under these conditions the usefulness is limited to a narrow number of purposes, in fact, the single purpose, in the case of a telephone book, of serving to provide name retrieval in an alphabetical order. In other words, the telephone book is undoubtedly endowed with structure, but one that is so simple as to have hardly any use for any advanced type of analysis.
     Another example would be to say that all the possible sentences of a language do constitute a grammar of that language, since they describe fully the language. Or again we might think of a linguist who, facing informants, simply asks them to speak, without posing leading questions. What kind of linguistic analysis would be possible under such conditions? Clearly, a very limited one. Analogously, the unexcavated site is, one might say, a form of report, since all the data are there – but a hopelessly opaque report, and impervious to analysis. Or similarly, and at the opposite extreme, if we were to have dozens of video cameras recording continuously every moment of the excavation from every conceivable angle, we would indeed have a record. But, again, one that is essentially impervious to analysis, hence one that could hardly be called structural. Structural archaeology, instead, aims to identify the flaments that hold together all the data as an integrated whole, and to address the very question of what this whole is as such.”
emplacement and deposition pp. 57-58 4.4 Emplacement and deposition: a basic antinomy
     Serious refection on the nature of emplacement and deposition, and on the need to consider them as distinct moments of stratigraphic analysis, is a crucial prerequisite in the search for objectivity (see Tani 1995). There is a fundamental antinomy built into the interrelationship of the two. On the one hand, we can only document contacts between elements as they are seen physically in their (static) emplacement. On the other, we cannot physically see these contacts unless we frst remove (dynamically) the elements that are defned by those very contacts. In other words, we must develop depositional inferences in order to lay bare the emplacement.
     The nature of the antinomy emerges even more sharply if we look at another aspect of the same problem. When we seek to identify types of contact between elements in the ground, we are seeking to document purely spatial relationships. But when, in order to see the contact, we must choose which element to remove, we operate on the basis not of spatial, but of temporal considerations: what came later must be removed frst. So – the determination of a temporal sequence is properly depositional in nature, but the primary aim of documentation is spatial contact. In other words, the static aspect of spatial juxtaposition cannot even be seen without a foregoing hypothesis about the dynamic temporal process that resulted in the juxtaposition. The antinomy then is that the identifcation of static juxtaposition (emplacement) depends on our understanding of the dynamic process that caused it (deposition), and yet conversely deposition can only be inferred from an observation of emplacement.
     To state the antinomy in these terms helps us to propose an answer to the basic underlying problem. We must indeed operate in a dual mode, seeking the evidence for emplacement all the while developing a strategy based on deposition. But the levels of analysis must be kept sharply distinct, and the moments through which the inferential argument develops must be clearly identifed. Obviously, the depositional inference is not created in a vacuum: it rests on initial clues, which of their very nature relate to emplacement. On the basis of these clues, one develops an excavation strategy – the frst moment of the stratigraphic argument. The second moment is a careful check as to whether the depositional inference holds up: does the emplacement record as unveiled confrm the initial inference? How should the excavation strategy be modifed, if at all? There develops a constant give and take between the observation of actual (if fragmented) data in the ground on the one hand, and on the other the choice as to which element ought to be removed next. And a close monitoring of this give and take, coupled with a constant alertness for the distinction between static spatial emplacement on the one hand, and dynamic temporal deposition on the other, are the two claims we can make for objectivity.”
pp. 75-77 5.2.5 In search of laws
     A major theoretical concern of archaeology has been the establishment of laws. But the ones proposed pertain typically to the social sciences, which are not specifcally archaeological either in scope (they do not pertain to stratigraphy) nor in their method (which is borrowed from such other felds as philosophy or sociology). What are then properly archaeological laws? In a broad sense, they are the rules that preside over the excavation process, understood as the disentangling of stratigraphy. There will be, accordingly, two major sets of laws, those that pertain to emplacement and that add up to a properly defned theory of excavation; and those that predict deposition from emplacement, i. e., depositional laws that describe time as a function of space. I will deal briefy with each in turn, in this and in the next section.
     I have already addressed the need for developing a proper theory of excavation (5.2.5; 4.4). It is, after all, the main thing for which archaeologists can claim exclusivity, since it is the task that no one but an archaeologist does. The distinction between emplacement and deposition; the defnition of what minimal constituents are; the clarifcation of the range of contacts – these are aspects of what a full-fedged theory of excavation will aim to develop. I will not here develop the topic about laws in and of itself. I am however explicitly presupposing such a theory, and I am articulating some of the principles that underlie the specifc implementation I am proposing. Beyond that, I can only point to the need for developing a greater awareness for it, as well as to the desirability of a more extensive discourse about it within the discipline.
     It should be noted that the concept of law is essentially grammatical in nature, in that it must be not only predictive, but also circumscribed within an all inclusive system. This is an important aspect that is easily underestimated. The larger whole, to be grammatical, has to subsume constituents that are not only clearly defned in themselves; they must also be described as elements of arrays, or paradigms, that are systemically linked. Thus a law will never afect constituents in isolation, but rather it will resonate across the spectrum and potentially touch every other constituent across the linked paradigms. We may refer back to the three-way law pertaining to stoplights as they afect vehicular trafc (3.4.1). If other colors were added, it would afect the meaning of the original three; it would reduce the assurance with which the diference may be perceived; it would afect the speed in the response to the stimulus; and so on. The original contrast is loaded with meaning in ways that are only made explicit as one investigates the impact that each individual element has on the others. Accordingly, defnitions and laws afecting their relationships must be established through a careful process that is deductive, i. e., one that generalizes from universal principles, but is the same time also inductive, i. e., one that draws on actual data as observed. Such a comprehensive theory of excavation is one of the important tasks ahead.

5.3 Deposition: time as function of space
5.3.1 An archaeological inference
     Deposition is a properly archaeological inference, meaning that the argument on which it is based is based exclusively on a consideration of the traces in the material record (2.3.1), i. e., on emplacement. This refers essentially to the ability to construe spatial relationships as defned by contiguity and contact in function of the temporal sequence with which elements have in fact come int contact with each other. The correlation between type of contact and inferred sequence is such that we can properly speak of archaeological or depositional “laws,” i. e., laws that explain the dynamics of how things got, over time, to where they are now, statically, in space (see also below, 15.10.4) (see also the principles in Bahn 2001: “stratigraphy”).

5.3.2 Emplacement as a clue to deposition
     The fundamental presupposition in this respect is that emplacement contains in itself an inherent clue to temporal sequence. In other words, each type of contact as explicitly and precisely defned when observed in the ground is assumed to refect a univocal depositional cause that has efected emplacement as we see it. This is why the paradigm of the types of contact (described in detail in the Grammar) is one of the most important features of an archaeological grammar, both theoretically and practically. Theoretically, it must be truly inclusive and exclusive, i. e., it must include every possible spatial confguration while it must exclude any variable that is purely allomorphic (see 3.6). Only thus can the relationship between emplacement and deposition be truly univocal, i. e., such as to “require” an inevitable consequence (on which is predicated the proper functioning of a law as such). Deposition emerges in this light as the assumed development for the embedding of elements as we discover them in the ground. And this approach gives us the key that makes possible a higher level documentation, namely the key for documenting process.
     Practically, the paradigm of the types of contact is of critical importance because it has a direct impact on strategy. We must remove frst what was last depositionally, and a proper understanding of this sequencing, as well as its documentation, is thus the indispensable presupposition to any decision reached during the course of excavation.”
conservation pp. 140-142 9.2 Conservation
     Conservation must not be viewed as a service provided by outsiders, after the excavation. Even if conservators are not present with their technical expertise during each stage of the excavation, it is important that the excavators develop sufficient sensitivity for the needs of conservation to shape their strategy accordingly (see Ashurst 2007). And it is important that the recording system be so structured as to accommodate this dimension from the start.
     In the practice of the system presented here, this means that there should be a proper categorization of the criteria that pertain to conservation (in the form of a Special Roster and a Special Lexicon, given in detail in the Grammar) and that these criteria should be applied when strategy decisions are made and de facto observations are recorded. Such a practice is common among conservators working in their traditional role as technicians whose expertise is invoked a posteriori: they naturally maintain documentary logs of initial conditions, treatment procedures and eventual disposition. My point is that this should not be, precisely, a posteriori with regard to the excavations, neither temporally nor conceptually – especially not conceptually (see Matero 2006; 2008).
     This is what I have called “extrinsicism.” If and when that is avoided, the whole conservation endeavor becomes an intrinsic part of the broader documentary effort. This is particularly important for the architectural and stratigraphic components of the record. Let us look at a concrete example. There are cases where a section illuminates relationships among architectural fragments, illustrating not only aspects of the chronological sequence, but also depositional and functional relationships (e.g., a wall collapse related to the structure to which it belongs). It is useful, in this case, for the excavation to proceed in such a way that the section, too, is preserved, and not only the walls. The logistic effort may well be minimal, and so too, therefore, the expenditure of resources. It is only the strategic decision taken from the start that makes a difference, and the level of sensitivity for the potential problems of conservation. This translates, therefore, not only in the conceptual record, but also in the physical dimension. A section saved and conserved, along with the walls, constitutes a physical ensemble that is part of the “reassembled construct” as much as the descriptive or analogical parts of the graphic and digital record.
     It is interesting to think of the value of conservation in the light of a seemingly unrelated event, namely the visit on the part of archaeologists to a site that is under excavation. The kind of explanation that is given in such cases relies fully on a field situation that does not need conservation because the excavation is current, and the architectural remains are in their pristine state. The relevance of this observation is apparent if one confronts this situation with a visit to a site after excavation: it is then well-nigh impossible not only to appreciate details, but also, in many cases, to even fully identify the architectural volumes. Conservation may never achieve the same goal of completeness that one has when visiting a site during excavation, but it should aim at retaining significant selected portions of the site so that there may be a similar confrontation with the physical record at any time after the excavation.”
restoration pp. 142-143 9.3 Restoration
     Conservation tends to privilege the piece in and of itself as extant (e.g., broken and incomplete), while restoration aims at rendering the whole. In this respect, restoration concerns the archaeologist in a very important way that may easily be overlooked: it forcefully steers attention in the direction of structure. It is a point that needs to be elaborated in our context because of what it tells us about the understanding of the archaeological primary evidence in its fragmented state. Restoration affects very concretely the perception one has of the data. Two convergent observations will help to bring out the import of a restored structural whole.
     On the one hand, the fragments emerge with a significance of their own. The view of the archaeological ruins of an architectural complex holds a fascination all on its own; it is, we may say, a romantic view, in the sense that it emerges as a scene with its own consistency, which bears no relation to the way in which the architecture actually functioned in antiquity. It is an autonomous landscape. In a different, yet analogous, way, the sherds that we collect in great numbers on an excavation acquire an autonomous identity, to which we relate as entities with an existence of their own, even though they bear no relation, qua sherds, to the perception that the original users had of the vessels whence the sherds came. In other words, the unrestored dimension, in the form of a ruined landscape or a collection of sherds, acquires an identity of its own that in some way may overlay and obliterate a concern for what the original identity in fact was. If one stops there, the structural whole one considers ends up being this particular identity of the broken element, as if on the same level of what had been the originary structural whole.
     On the other hand, there are intermediate stages, where one forms a mental perception, however inarticulate, of how the fragments would have cohered into such an originary structural whole. With regard to architecture, one may think of the impact that the loss of the roof has on the perception of a Greek Temple: a mental shape develops of the interplay between light and darkness that is radically different from the one that would have obtained given the presence of a roof: the columns of the peristasis acquire greater perceptual valence than is the case when the presence of a roof highlights the dark space framed by the columns as much as, if not more than, the bright frame represented by the columns. With regard to ceramics, the profile drawings, which are meant to serve as a shortcut, may end up being the thing in itself.
     This points to the value of restoration as a call to reflect concretely on the perception as elicited by the originary structural whole. Regardless of what limits may constrain the effort or what errors may arise in the process, the fact remains that the endeavor in itself underscores the relevance of structure. Take, for example, the restoration of walls to even just a limited height: here the limits are the lack of finishing details, of the roof, of furnishings; but the presence of the architectural volumes, instead of partial fragments of walls or even total voids, generates a heightened perception of circulation patterns, especially within structures with higher numbers of interior rooms. Or take a reassembled vessel: correlations among structural details are easier to notice when the fragments are pieced together and the vessel is present in its physical reality. Also, the restoration of a vessel may have important stratigraphic implications by bringing together sherds that were found in different locations. There is a valid concern, in restoration, about distinguishing clearly the original pieces from what is added. This may, in a small measure, detract from the full perception of the whole, in that it juxtaposes elements that are perceptually heterogeneous. But it is a small price to pay for the sake of keeping the proper distinction between the authentic and the “intrusive” elements. One interesting experiment has been very successful with flat images such as frescoes: it consists of projecting an image of the restored version upon the damaged and incomplete original image (see de Raemy et al. 1999, pp. 21–23). This way, one can alternate on the same surface between one and the other version, where the restored version shows no distinction between the two moments.
     A special form of restoration is relocation, where an ancient site is moved in its integrity to a different place. The most spectacular such project was the relocation in 1968 to higher ground of the temples of Abu Simbel in Egypt, which were going to be flooded as a result of the construction of the Aswan High Dam on the Nile river (Allais 2013).”
outer limits pp. 224-225 12.6.2 The Outer Limits
     Websites declare their starting point by using the metaphor of a “home.” This is where a website opens, and where one may find a statement of purpose. Perceptually, this functions like both the cover and the preface of a book.
     By contrast, there is, in a website, no place for a conclusion (see already 12.4.2). This is essentially due to its fluid nature: a website is generally not conceived as ever having been terminated and thus having reached a final point. In principle, an ending comes when it is no longer “maintained,” but in that case it is largely because of some exterior event, and not because of an explicit decision: it is, as it were, abandoned, without any sense or intent of closure. In other words, it is not in the nature of a website to reach an end point, and therefore there is no structural equivalent and no outward signal of any completion having taken place. This has a conceptual relevance with regard to the notion of a digital text: there is in effect no outer limit other than the beginning. It is one of the reasons that reduce the degree of awareness of a website as a “text” (see also below, 13.3.1). It is therefore a challenge to maintain the integrity of the site as a coherent whole: on the one hand, fluidity itself must be seen as a mode to maintain the unity, in an intrinsically deconstructive mode (11.3.5); on the other, one must rely more heavily on the other compositional mechanisms.
     And yet there are set moments in the life of a website which may be considered as temporary terminal points.
     The first is the dynamics of evaluation. To evaluate properly a text one should gain a sense of its wholeness. In the case of a website, this means that any evaluation episode must refer to a given window in time when the text is being considered. In other words, one must on some level consider the existence of a potential conclusion, however temporary and unexpressed; there has to be a sufficient cohesiveness to the website to make it properly a text, with its own compositional wholeness, subject to an inspection capable of properly assessing it.
      The second aspect to be considered is the way in which updates are to be identified. The de facto custom is simply to omit any identification of changes, while the current best practice is to identify each individual change in a separate roster (the system adopted by Wikipedia). This differs considerably from what happens with printed publications, where reprints or new editions are identified as such for the book as a whole at set intervals. In the case of a website, a proper citation requires therefore not the number referring to an edition, but rather the date in which the resource has been accessed – an approach that underscores the lability of the text as a compositional whole. An alternative is instead to “publish,” i.e., to open to the public, the revised website at set intervals, and to maintain online a separate copy of each previous edition. The result would be similar to what happens with an astronomical ephemeris (12.7). This approach protects the integrity of the text and would make citations easier, but it is not in current usage.”
bibliographical status pp. 230-231 12.7 Bibliographical Status
     An advantage of online publishing is that updates are easy. A corresponding drawback is that one may just as easily lose control of the referential value of the citations. In other words, what I quote today may no longer be available for inspection tomorrow. This is a well-known problem, and the safeguard that is commonly used, citing the day in which the website has been accessed at the moment of the citation, is only a palliative: the authenticity of the reference rests with the statement given, and it cannot be checked against the original if that has changed in the meantime.
     A more rigorous solution is the Wikipedia approach, which gives a complete history of all changes that are ever been made. An implicit problem is that authors are often identified only through nicknames that hide the identity of the person, and thus negate some of the key points that are associated with the concept of authorship and responsibility (see Berners-Lee et al. 2008: 96, 100). Also, individual pages are often the work of a single author who will normally and frequently come back to the same page. The full history of these changes is of little consequence, and can best be omitted.
     The ephemeris approach offers an alternative which seems better to me (a similar approach uses the term “fixed edition”, see, e.g., the website Stanford Encyclopedia of Philosophy). The published updates occur in batches, each of which is preserved as a distinct entity in a section of the website that preserves its history. Each batch is a distinct version or edition – an “ephemeris,” as in astronomy – where the term refers to tables that define the orbital movements of celestial bodies at any given point in time (see Sobel 1995: 24–27). There are two symmetrical uses of this approach.
      On the one hand, each page carries the of the version or ephemeris to which it belongs. Reference to the pertinent version establishes a framework that is consonant with normal bibliographical standards, and lends any given citation its proper “status. The penalty is that the updating may not be as frequent as one may wish. But it is more in keeping with the responsibility any publishing enterprise has towards transparent referentiality.
     On the other hand, a stored archival version (ephemeris) subsumes a batch of files that have been updated since the last version has been posted online. Each version is pegged to a given date, and it includes only the pages that have been changed since. In this respect, the concept differs from that of a new edition for paper publications, since in our case only those pages are earmarked with the new label that have changed since the previous version was closed.
     This approach allows to keep the proper balance between fluidity (12.6.2) and referentiality in a state of punctuated equilibrium (13.3.1).”