I congratulate Duncan Garrow on his very engaging history of the concept of structured deposition, although I find it slightly terrifying that this history now extends over nearly 30 years. I find much to agree with in his account, notably the distressing point that what was originally intended as a heuristic has sometimes become an end in itself: the identification of a class of deposits that are ‘structured’. Thus, for instance, Bishop, Church and Rowley-Conwy argue (2009, 82) that pits in Neolithic Scotland may have represented ‘places of structured deposition rather than domestic settlements’, and that therefore the plant remains contained within them should be regarded as unrepresentative. Here, structured deposits take on the abject character that used to be afforded to ‘ritual’ phenomena in archaeology: having been identified as irrational and abnormal, their interpretation is considered beyond archaeological competence, and they are not subjected to further analysis.

Having said that I go along with the greater part of Garrow's argument, there is one issue that bothers me. This is the categorical distinction that he draws between ‘odd deposits’ and ‘material culture patterning’. While he affirms that these represent the two extremes of a continuum or a ‘sliding scale’, it seems to me that he repeatedly slips into the position where an absolute dichotomy is established between the two. As an example, he lauds ‘studies which have maintained an interpretive distinction between these two “types” of deposit’ (p. 107). On the basis of this division, he sets up a series of further oppositions: between odd and everyday, meaningful and meaningless, ritualized and non-ritualized, and so on. When Garrow talks of a form of ‘material culture patterning’ that is untainted by ‘oddness’ but characterized by an ‘everyday’ signature, it seems uncomfortably close to the early New Archaeology's optimistic belief that past lifeways could be directly ‘read off’ from material residues (e.g. Hill 1968, 135).

It is on the basis of this odd/everyday dichotomy that Garrow objects to the way that ‘odd deposits’ have been used to support the argument that material culture patterning can be meaningful. Such a complaint only really makes sense if we accept that ‘odd deposits’ and ‘material culture patterning’ are actually opposed categories, rather than different parts of a single phenomenon, which I would want to call ‘depositional practice’. Now, if I were to be writing a history of the investigation of depositional practice, I would want to argue that these ‘odd deposits’ were like the visible portion of an iceberg, showing above the surface of the ocean. It was all the pits lined with potsherds, unbroken stone axes, placed animal skulls and general weirdness that could not easily be overlooked that first alerted archaeologists to the likelihood that ‘something was going on’ in the domain of deposition. Without this strangeness, it is much less probable that the more subtle forms of patterning would have been identified as a problem to be addressed, in other than functional terms. So when Garrow argues that the more formal and elaborate kinds of deposition have been overemphasized at the expense of the everyday, this is fair enough, but it neglects the point that ‘odd deposits’ were our point of entry into the complicated world of prehistoric depositional practice. It was from this bridgehead that it proved possible to address the more mundane forms of deposition, not least in Garrow's own excellent work. Consequentially, I am wary of separating the odd from the everyday in any categorical way.

Garrow is quite correct when he argues that the patterning that we can identify in material assemblages need not have been intentionally created, and that random processes do generate uneven patterning. But again, he seems to be imagining an either/or scenario, in which those deposits that are not characterized by ‘oddness’ represent the outcome of everyday, non-symbolic acts. The reality is likely to be more complex than this. Garrow says that he finds it hard to accept that all material culture patterning is symbolically meaningful or intentionally created, but these are not the same thing. Indeed, he notes that it is possible for patterning to be unintended, but still symbolically meaningful. However, he passes over this point rather too quickly, and I think it would have been helpful to give more consideration to residues that may have been generated through habitual and unconsidered adherence to cultural conventions. As he notes, Henrietta Moore's ethnographic study of rubbish disposal amongst the Endo of Kenya was a major inspiration for early studies of structured deposition, and it was concerned with the organization of material in domestic compounds and contexts that were anything but exclusively ritual. However, the patterns that she described were not ‘meaningless’, even if they might have been described as everyday. The Endo do not mix ash, chaff and animal dung, because each of these is placed in a part of the compound that has distinct connotations of age and gender (Moore 1986, 102). Moreover, these locations are judged appropriate for the burial of men or women, young or old.

So, in the Endo case, deposition is guided by a set of connections between places, age, gender and death, and while it is far from random, the throwing of ash onto a heap behind a house, or the sweeping of dung over the edge of the compound, may not always be a fully intentional process. Much of this activity is likely to be habitual and unevaluated, simply repeating a pattern of acts without deliberation. The point is that in the Endo example mundane, ‘everyday’ activity does embody and reproduce a structure of symbolic meaning. Yet rather than carrying round in their heads a massively complex cosmological scheme, Endo people are simply conducting themselves in an appropriate way, cued by the architecture, topography and materials that surround them. They operate in the grey area between Garrow's two poles of conspicuous oddness and meaningless mundanity, where the two overlap and merge.

What I take from this is the need to complement a concern with the observed pattern of material in the archaeological record with a focus on deposition as a social practice, and I find this underemphasized in Garrow's paper. Garrow rightly stresses the importance of ideology and belief in early postprocessual archaeology, but he does not say enough about practice, agency and intentionality. In this context we might draw attention to Giddens's (1984, 41) distinction between discursive and practical consciousness, Heidegger's (1962, 67) ‘presence-at-hand’ and ‘readiness-to-hand’, and Bourdieu's (1977, 78) concern with the habitus. In each case, what is being referred to is the way that humans do not always operate in a state of explicit and calculative awareness. On the contrary, people often conduct themselves without deliberation, in an ‘instinctual’ and unconsidered fashion. But as Bourdieu in particular stresses, our habitual practices are learned ones, in which the arbitrary and conventional takes on the character of a ‘second nature’. In practice, then, people may reproduce symbolic orders or conceptual schemes without having to think them through. It is my understanding that much of what we refer to as ‘structured deposition’ is the outcome of this kind of action: where people have placed or dropped particular classes of material in particular locations, or where they have kept specific substances separate from each other, simply because ‘that is how it is done’. The archaeological signature of these practices may not be as clear-cut as the ‘odd deposits’ that emerge from the deliberate selection and arrangement of materials in a pit or ditch-butt, but neither are they randomly generated. Moreover, they will tend to mesh and interdigitate with processes that are more random, some of which may be attributable to non-human agencies. It is an open question to what extent any depositional activity on the part of human beings will ever be entirely free from the influence of inherited cultural conventions.

It is in relation to ‘practical consciousness’ and the habitus that we can turn to the issue of ritual. The most important point to make about ritual is that rather than representing an entirely separate and elevated sphere of activity, it is actually a mode of conduct that people can slip into and out of in the course of a normal day. It does not always employ a separate rationality from everyday life, but it does often provide a context in which the guiding principles of that life can fall into a sharper focus. Ritual can be defined as expressive performance that is prescribed by appropriateness, and is sanctioned by tradition (Lewis 1980, 8). In ritual, people are alerted to the specialness of what is going on by distinct kinds of behaviour, often involving formality of utterance, body posture and movement. This awareness of the particularity of the event generally leads to an enhancement of attentiveness and a heightening of sensation. Ritual is therefore discursive rather than exclusively habitual, although it is a context in which the form of discursive action is highly circumscribed: one is intensely aware of what one is doing, even while one's focus is on ‘doing it just right’. In the original Durrington Walls paper, Colin Richards and I suggested that there might be a relationship between the kind of deposition that was conducted inside the henge and the less spectacular but nonetheless structured material identified in the Neolithic pits at Fengate (Richards and Thomas 1984, 215). This argument was less contradictory than Garrow seems to suggest. For what we sought to imply was that divisions and associations that were established in the explicit context of ritual might be learned and then reproduced in other less formal circumstances, such as the filling of pits related to domestic occupation.

Some 28 years later, I am not sure how far this idea stands up. It is clear that deposition in henges is rather more complicated than we imagined in 1984, not least in that much of it appears to have been commemorative in character, representing a means of consigning the activities that had taken place in the monument to memory as much it was as a part of those activities. However, what I think it does indicate is that research on prehistoric depositional practices should not now develop a single-minded concern with the meaningless and the mundane. Instead, we need to consider the range of shades of grey between the odd and the unremarkable, and the variety of processes and agencies that interacted in the formation of any particular deposit. It may be that it is the unintentional generation of patterned deposits that we need to consider at greater length, and the way that these habitual cultural practices mesh with the more random processes that Garrow seeks to emphasize.