This paper examines some of the basic premises of archaeology as reflected in those of the round table. Simply asking ‘why do we excavate?’ is already revealing either of a will to provoke discussion about something which many might take to be self-evident, or of an academic attempt to shift the emphasis away from praxis to theory, or even of a failure to recognize how all such premises are culturally contingent.

The first point I would like to raise is the supposed ‘widespread failure of archaeologists to ensure that the results of excavation are published and accessible’ (editorial, this issue, p. 1). This is not a ‘failure’ archaeologists face alone, and more a problem for archivists. Similar problems are being faced – and solved – by chemists and biologists who make their results available online while protecting corporate secrets, and archaeologists should learn from other disciplines.

The point is that this ‘failure’ seems to contradict the idea that ‘the material recovered only adds to an existing body of data on types of site, periods or regions we already know well’ (editorial, this issue, p. 1). I will not argue that we have not done enough with the data we have, but this seems contradictory: how can we ‘know everything’ if we cannot access ‘grey literature’? The real problem is not just a matter of not having access to information: even when we have it, we do not always learn from it, possibly because our natural habitat is not the library, museum or archive. I would like to develop this further, and ask just how effective we can be at understanding the material that is the main focus of what we do study if we are not very good at understanding the history of our own discipline. As Cherry noted (this issue), this ‘why?’ debate has taken place already, and it may be that we keep returning to these same questions because we need to. And I would like to suggest, in light of a study of the importance fieldwork has for geologists as a ‘liminal experience’ (Rudwick 1996), that maybe we have to have these periodic debates in the same way that we need to excavate in order to be initiated into archaeology.

The degree to which we fail to learn from others or from the past can perhaps be indicated by considering another of the points from the session abstract, the idea that – at least in the public mind – ‘archaeology is synonymous with excavation’ (editorial, this issue, p. 1). I would like to suggest that some members of the public equate archaeology with treasure (King Tut) and adventure (Indiana Jones and Lara Croft), ley lines, Stonehenge and ‘ancient astronauts’, maybe even dinosaurs. Some people might even associate us with boring museums, or think that archaeologists ‘kill jobs’ and delay ‘progress’ by adding costs to construction projects (and it is useful to contrast the public perception of environmental and heritage protection).

What we define as being ‘archaeology’ (or even ‘the public’, for that matter) is culturally contingent. To illustrate this point, I often list definitions of the word ‘archaeology’, or contrast English-language archaeology with the divisions between Ur- und Frühgeschichte/Vor- und Frühgeschichte and Archäologie in German-speaking countries (cf. Carver 2009), partly because ‘it is widely recognized that German archaeologists dig with a different method’ (Hodder 1999, 9). The significance of these distinctions can be illustrated by attempting to translate into German Binford's famous adage that ‘archaeology is anthropology or it is nothing’. Bernbeck (1997, 37), Veit (1998, 122) and Kümmel (1998, 122) do not even try. While Binford (1962, 217) may be right about American archaeology (where archaeology is often taught as a subdiscipline of anthropology), in Britain ‘it is concerned more, but not exclusively, with early and prehistoric phases than with those illustrated by written documents’ (Crawford 1960, 15). To confuse Binford even further, German-language texts refer to Winckelmann as the father of an Archäologie which focuses mainly on classical antiquity, and German Anthropologie usually only refers to what Anglo-Americans know as physical anthropology (although the situation was different until the beginning of the 20th century – cf. Boas 1902; Fetten 2000, 159–69). How can we claim to know what the public thinks about us when we are not even sure who we are, or what we do? Which ‘public’ associates us with excavation, and which ‘archaeology’ does this ‘public’ have in mind?

Such problems relate to my work on archaeological documentation methods, definitions, epistemology and so on. Although this might seem abstract and theoretical – armchair archaeology – I try to tell myself I am on the ‘field’ side. Which suggests that we might be having this debate again because we have yet to balance practice and theory – armchair versus field archaeology – where the differences might be as significant as those between English- and German-language or between ‘commercial’ and ‘academic’ or ‘scientific’ archaeology (Demoule, this issue), or between archaeologists’ perceptions of archaeology and how archaeologists think the public perceives us. If we stay out of the archives and do not learn from history because archaeology is fieldwork (Crawford 1960, 232), we risk repeating ourselves.

Dissecting the discipline's origin myth

If ‘archaeology is synonymous with excavation’, then this should show up in the historical record. Discussing the evolution from antiquarianism, Redman (1999, 49) wrote,

Another variant lists the ‘three great conceptual advances – the antiquity of humankind, Darwin's principle of evolution, and the Three Age System’, which ‘at last offered a framework for studying the past’ (Renfrew and Bahn 2000, 25, original emphasis), while Audouze and Leroi-Gourhan (1981, 170) traced ‘French prehistoric archaeology’ to ‘geology and palaeontology . . . cultural anthropology . . . stratigraphy . . . and evolutionism’. Comparing such genealogies reveals recurring elements and an emphasis on theoretical issues which may reflect a shared derivation from Glyn Daniel's ‘antiquarian revolution’ (Daniel 1975, 32), and a contrast with Crawford's schema (1932, 173), where archaeology leads to the recognition of ‘the antiquity of man’ and ‘evolution’ is in a separate lineage. But excavation . . .?

Ploughs and coins

From a historical perspective, archaeology used to be synonymous with ‘accidental discoveries’ in the ‘antiquarian way’ (Pegge 1789, 84), traditionally with a plough or bought and later brought to the attention of some local authority (teacher, priest, antiquary; cf. Schlanger 2010). Thus Wordsworth (1994, 275) wrote of how ‘The unlettered ploughboy pities’ Romulus and Remus (depicted on a coin) ‘when he wins / The casual treasure from the furrowed soil’. Although remote sensing offers ‘a plethora of other methods’, there have always been alternatives to excavation. ‘The purpose of an excavation is to acquire data’ (Carver 1990, 77), and we excavate now because excavation is a more efficient means to do this.

Notice the shift from ‘facts’ (i.e. ‘artefacts’) to ‘data’. We acquire data partly so that others can use it. Francis Bacon argued that making data available to a critical audience should lead to better science (Quinton 1980, 31), and we see traces of his arguments combined with concerns for efficiency in some of the debates on ‘big data’ related to computers and data-mining large databases, since such data, ‘produced at great effort and expense, are only as useful as researchers’ ability to locate, integrate and access them’ (Howe et al. 2008, 47). So if we excavate to acquire data and want to share our information, then we need to be concerned with our ‘widespread failure . . . to ensure that the results of excavation are published and accessible’ (editorial, this issue, p. 1), as this shifts the emphasis to questions of accuracy, raising issues like data quality that were unimportant in the days of accidental discovery by ‘unlettered ploughboys’. Otherwise: garbage in, garbage out.

Before we can ‘acquire’ or ‘produce’ data – and do it efficiently – ‘it is incumbent upon us clearly to comprehend the nature of our mission and the limits of our field’ (Kemble 1849, 2), if only so that we can distinguish between data and ‘irrelevant information’ (Collis 2001, 1). Like history and geology, archaeology is a study of the past. Whereas history focuses on written documents, archaeology is primarily concerned with ‘material culture’ or ‘human antiquities’, often but not exclusively ‘as revealed by excavation’ (OED 1997). Like written documents, archaeological remains are often incomplete, but generally skewed more towards everyday life and away from the lives of the literate elites.

Otherwise Kemble was not concerned with excavation:

Particularly noteworthy is that Kemble's reference to archaeologists pre-dates Daniel's ‘antiquarian revolution’ (the histories tell us he should have been writing about antiquaries). This can be ‘explained away’ by tricks of semantics: since many archaeologists pursued ‘antiquarian studies’ and antiquaries did ‘archaeology’, the terminology had not yet been settled and has since evolved, or Kemble (despite having, like Winckelmann, ‘a framework for studying the past’) was part of a pre-‘scientific’ archaeology and so on. The OED includes an 1824 reference to ‘English historical archæologists’ and a statement from 1880 to the effect that ‘The archæologists have raised the study of antiquities to the rank of a science’, while ‘“Archaeology” itself’ was

The most important point is not the fact that the words changed, or that their meaning changed over time, but rather that the disciplinary space they described evolved to reflect the shift from antiquarianism to scientific archaeology: as the linguistic component split off to become philology, as the role played by coins and medals declined from the point where numismatics could be labelled ‘an important branch of archaeology’ (Pettigrew 1848, 8), as the accidental discovery by some ‘unlettered ploughboy’ gave way to deliberate excavation, and as ‘archaeology’ expanded to include work done by Foucault (i.e. The archaeology of knowledge, The order of things. An archaeology of the human sciences and so on), with the result that archaeology now resembles Lévi-Strauss's bricolage more than it does the scientific ideal.

Archaeology and autopsy

As the session abstract notes (and as my ongoing work with soil colours and so on shows), even when data has been recorded it is not necessarily accessible; even then, having been recorded in a different theoretical paradigm, it may as well have been written in a different language. Kemble separated himself and his contemporaries – ‘modern antiquaries’ (Hutchinson 1887, 470–71) – from their ‘predecessors’ in the same way that Daniel and other historians of the discipline separate ‘modern’ archaeologists from our predecessors, the ‘antiquaries’. For Kemble, the deciding factor was the fact that his generation of archaeologists ‘group . . . facts by a system, class them as it were in genera and families’. In addition to the obvious reference to Linnaeus, it might be tempting to read this as an allusion to the three-age system. It is more likely that Kemble was referring not to the system we know (i.e. Thomsen and Worsaae's Stone, Bronze and Iron Ages), but one comprising Celtic (or ‘Druidic’), Roman, and Saxon (or Danish) periods (cf. Way 1844, 2; Briggs 2007, 251), given that the three-age system had not yet found wide acceptance among the British archaeologists of 1849.

The point is this: Kemble was arguing that his ‘archaeology’ was a science simply because its data could be classified, whereas we would now say that science comes with trying to explain such ‘arrangements’. In Kemble's eyes, it was simply ‘enough’ that ‘any one fact should be a fact of the past’ for it to be archaeology; for it to be science, it was simply ‘enough’ that ‘one fact should be capable of arrangement and comparison with any one similar fact’. For Kemble, totally oblivious to considerations of the context from which these ‘facts’ derived, it was ‘enough’ that this ‘fact’ could be arranged and compared with other ‘facts’, and

Included in this prescriptive assemblage of examples of what would have been of interest to British ‘archaeologists’ of 1849 are items which would not now normally be considered ‘archaeology’. Kemble's pots and pans also provide a response to accusations of Romanticism levelled against antiquaries by their detractors, accusations which fail to account for their obsession with ‘facts’.

Kemble's ‘cabinet of medals’ – now an apparently minor detail – actually reflects the success early archaeologists had studying medals (or medallions) and the coins found by Wordsworth's ‘unlettered ploughboy’. These antiquaries were so successful that geologists wished for something analogous. Robert Hooke is cited as lamenting the fact that geologists had nothing comparable to ‘Monuments or Medals’ (Schneer 1954, 266), recognizing the value of equating fossils with artefacts. Later geologists made the links between fossils and antiquities in general on the one hand and medals in particular on the other more explicit. William Smith (1816, p. i) wrote that ‘organised fossils . . . might be called the antiquities of Nature’; Gideon Mantell (1844) named one of his books The medals of Creation, while Lyell (1991, 47) supported a comment regarding the importance of testacea by stating that ‘they are the medals which nature has chiefly selected to record the history of the former changes of the globe’. Recognizing that ‘The comparison between fossils and medals has frequently been made and fossils have well been styled the “Medals of Creation”’, John Edward Marr (1898, 40–41) expanded upon this theme by making specific reference to the three-age system. It was only after Daniel's ‘antiquarian revolution’ of 1859 that people like Lubbock (1865, 336) began reversing the analogy.

The foregoing contrasts with the ‘cardboard-history’ depiction of antiquaries as sterile scholastics. If anything, early archaeologists opposed scholastic adherence to the written authority of classical authors and later commentaries:

The concept is actually much older – ‘when men first began to philosophise crudely, they used the evidence of their senses, which the Greeks call autopsía, seeing for oneself’ (Vico 2001, 204 [499]) – and one might consider the derivation of ‘the original Greek term for “historian,” which means an “eyewitness,”’ as ‘the one who obtains truth about what happened not merely by repeating “what they say” . . . but rather by examination of witnesses and through enquiry into the actual causes of what happened’ (Mali 2002, 214). What was revolutionary to an extent now difficult to recognize was the way antiquaries deliberately sought material culture to help address shortcomings of written text. As a literal ‘rebirth’, the Renaissance – ‘when rediscovery of classical knowledge became the primary goal of scholarship (Gould 2000, 148, added emphasis) – was largely text-based:

Cutler refers to the ‘dirty hands’ of Nicholas Steno (saint, geologist and dissector). The idea of autopsıa is useful, given the relatively common metaphor of excavation as ‘careful archaeological dissection of the earth’ (Brown and Harris 1993, 10), of archaeological excavation and legal autopsy being both methodologically analogous and comparably irreproducible experiments (cf. Oebbecke 1998, 218). Excavation, by accessing primary sources (‘archaeological’ and textual), can overcome errors caused and compounded through what Richard Dawkins (2006, 194–95; cf. Dawkins 2004, 133) called ‘copying-fidelity’, and through transmission from one medium or (as in the Binford example) from one language to another (cf. McLuhan 2003, 230; Hodder 1992, 12).

As archaeology, though, the present study is intended to reverse general trends by focusing on textual problems which spurred initial interest in material culture as an object of study: ‘antiquarianism was originally text-centered, focused on the reading of inscriptions on monuments and coins, marble, and metal’ (Burke 2003, 273). Some of the problems with textual interpretation include the fact that text can be biased, contradictory, unclear and/or incomplete (cf. Taylor 1948, 31). When T.G. Bonney (1866, 6), for example, noted that ‘there is no mention whatever of stone circles in any of the Roman accounts of Britain’, the question arises, how did the Romans miss Stonehenge? Discussing ‘the manuscript sources for the topography of Roman Britain as Stukeley would have known them’, Piggott points out that, ‘Apart from scattered references in the Greek and Roman geographers and historians . . . the essential documents are five in number’ (Piggott 1985, 134). The lack of written references to Stonehenge may only reflect sampling error, given the relatively small number of Roman sources now extant. Lyell notes how the few classical references to the destruction of Pompeii and Herculaneum enabled an Italian Wernerian to contend ‘that neither were the cities destroyed in the year 79, nor by a volcanic eruption, but purely by the agency of water charged with transported matter’. As with much Wernerian geology, despite seeming too ridiculous to be believed, ‘His arguments were partly of an historical nature, derived from the silence of contemporary historians, respecting the fate of the cities’ (Lyell 1990, 351). And in the sense that Daniel's ‘antiquarian revolution’ opened up an archaeology that was not text-based, it represented a great leap forward, except that – by leaping straight into prehistory – it contravened uniformitarian principles of moving from the known (historic or ethnographic) to the unknown (cf. Woolley 1961, 54). More important, though, is the fact that written text can be as incomplete as the archaeological record due to analogous problems of preservation. Recognition of this fact inspired Lyell to use an extended metaphor comparing the geological record to texts written in an ancient language (Lyell 1990, 461–62), inspiring Darwin in turn (1859, 310–11).

The point is that text provided an accepted analogy long before postprocessualists rediscovered hermeneutics, as British geologist Henry Thomas de la Beche (1835, 6) made clear by describing strata as ‘when rocks are divided into beds like the leaves of a book’. The strength of this analogy, then, reflected more than scientific transparency, or even an appeal to common sense; since it was tied to the common metaphor of reading geology – stratigraphy, specifically, but the past in general – as a text, any problems then became those of textual interpretation already familiar to the humanist and/or scholastic traditions, and the various forms of the metaphor of the archaeological record being like a ‘text’ were strengthened by the extent to which the archaeological record is as incomplete as written history:

Gemsege's historical context is evidenced by the fact that his fragment has no archaeological context other than ‘a brazier's shop at Canterbury’. This is not archaeology, not because Gemsege lacked the three-age system, uniformitarianism and evolution and so on, but because this ‘monument’ was not found in the ground. Despite this lack of provenance, artefacts – i.e. ‘non-verbal documents’ (Taylor 1948, 43) – had a long tradition of being valued over texts:

Archaeology's eventual evolution as a science resulted from contact not only with images on coins but with provenienced finds of all kinds (cf. Edgeworth 2003, 53; Petrie 1904, 51).

Historically, archaeologists excavate because excavation is a more efficient and systematic means for gaining data than such alternatives as the ‘antiquarian way’. Not only can archaeology be systematic – a coherent science – if it does not rely on Wordsworth's ‘unlettered ploughboys’, but the data itself is of potentially higher quality (although still subject to our own biases: paradigms and various potentially idiosyncratic research interests and so on).