Steve Roskams (2001, 267–70) has challenged archaeologists to theorize excavation practices and Patricia McAnany and Ian Hodder have responded in a cogent manner. They draw the most fundamental of archaeology's field methods – stratigraphy – into the light of social theory. The product is ‘social stratigraphy’ and the authors offer an array of interpretive schemes and processes through which social stratigraphic approaches might be considered and developed. McAnany and Hodder want us to think beyond the geological facets of stratigraphy, including our section drawings, photographs, matrices, phase designations, chronologies, thin sections and artefacts. They suggest we strive to do more: to discern and interpret social meanings, patterns and practices in every deposit, cut, pit, erasure, concealment, return and episodic rebuilding. The great bulk of their discussion focuses on the study of past contexts, though in closing they address how strata and stratigraphy can be altered or used to influence present-day social and political contexts. It is here – that is, in their concluding comments – that their emphasis on memory (and forgetting), repeated like an incantation throughout the paper, comes full circle.

McAnany and Hodder's discussion will no doubt inspire new explorations of how to wed social theory to fieldwork and field methodologies, and their consideration of augmentative techniques offers vital support for the role of ‘science’ in these efforts. Many of these new efforts will focus strictly on the relevance of social stratigraphy to interpretations of ‘past’ contexts. I think the authors recognize this: they provide section drawings from their own excavations, at K'axob and Çatalhöyük, respectively; moreover, the interpretive schemata of their paper are most easily conceived for use in studying ancient patterns and especially stratigraphic configurations related to complex societies (thin or very simple layered strata probably will not respond to their programme without substantial micromorphological analyses). We can even imagine that sets of flexible general laws might emerge from the social concepts they propose, such that the ‘rapprochement’ they attempt between American and British stratigraphic schools can be more explicitly extended to the processual and postprocessual schools. This could be an important undertaking, in line with other recent efforts to resolve philosophical and practical differences in approaches to historicism and evolutionary change, the social and the cultural, antipositivism and positivism (e.g. McAnany 1995; Yoffee 2005; Trigger 2003; 2006) All this is not to say that that the authors’ observations of the intersection between past and present, and of the present context of archaeological stratigraphies, are underdeveloped or unpromising, only that these are discussed briefly and that their implications may be more complex and take longer to explore and substantiate.

In the remainder of my comments I focus on three published debates or exchanges that center on stratigraphy. In discussing these I look more or less simultaneously at two issues that I think might help to encourage further exploration of the intersection between past and present in an emerging social stratigraphy. The first of these involves technical aspects of how we write and read – record, analyse and discuss – stratigraphic data. This has obvious implications for the support of a social stratigraphy geared to past contexts, for there exist biases and differences in training, terminology and research objectives that can hamper even the most basic social interpretations of ancient and prehistoric strata. These biases and tensions also bring us into the present context as we consider how archaeologists engage one another through the medium of stratigraphy-based dialogues and debates, responses and rejoinders. Thus the second issue I want to address regards how we might conceive of the history of archaeology in terms of social (disciplinary) stratigraphies that can be read and studied: precedents practised, manipulated, remembered and forgotten; features of the intellectual and imagined landscape modified, erased and/or deposited in thick layers of varying substances, hues, textures and influences; the results of encounters and conflicts, debates and patterns of attribution, silences; barely visible gravitational-like forces, ideas and impositions that weigh and reflect but may be hard to discern without guidance and tools. What type of trowel can we use to clean and delineate the sections that cut through our disciplinary history? Would this tool respond to a type of use-wear analysis?

The intellectual dynamics and discourses of archaeology are practised, and they transpire linearly (vertically and historically) and diffuse and incorporate nonlinearly (horizontally and spatially) depending on precedents, goals and purview. The debates I address reflect discursive and field practices and centre on the Americanist tradition in archaeology and, in particular, on Mexico and Mesoamerica. The first debate, from 1960, remains well known and well remembered; it played out in the pages of American antiquity between Richard ‘Scotty’ MacNeish and Walter Taylor. As expressed in his famous book A study of archeology (1948), Taylor was the father of the ‘conjunctive approach’ and a sharp critic of pre-Second World War American archaeology. Scotty MacNeish, when he died accidentally in 2001, was known as one of the most productive and resourceful archaeologists of the 20th century and as someone who single-handedly rewrote the prehistory of the Americas. Among his major contributions was his Tehuacan project (MacNeish 1964) on the origins of agriculture in Mesoamerica; this research alone had made him a household name by the mid-1960s (Flannery 2001, 153).

Years before MacNeish rose to eminence, however, Walter Taylor (1960a) published a review of MacNeish's book on northern Mexico, Preliminary archaeological investigations in the Sierra de Tamaulipas (1958). It was not particularly acerbic as reviews go, but it clearly questioned MacNeish's worth as an archaeologist. Wheeler (1954, 59) has declared that ‘the digger must learn to read his sections, or he should be constrained from digging’. As if following this tenet, Taylor judged MacNeish based on the latter's skills with stratigraphy and phase designations. Known to this day as the ‘MacNeish–Taylor debate’, their exchanges were closely read and influenced later research on agriculture and on the development of complex societies. Owing to this wide readership, and also to tenor, the MacNeish–Taylor debate has been memorialized in the annals of American archaeology (e.g. Flannery 2001).

Taylor (1960a) noted that MacNeish had begun his excavations using outdated, New Deal-era (1930s) six-inch arbitrary levels and that he seemed to be giving each level a new phase name or cultural designation. Taylor (1960b) challenged MacNeish to justify his separation of cultural assemblages into so many (seven) different phases and also showed that some of his phases were not, as stated (MacNeish 1958), the results of his own preliminary analyses but rather that MacNeish arrived on-site (at Diablo Cave) with preconceived (borrowed) phases and phase relationships. Taylor also challenged the accuracy of MacNeish's use of terms such as ‘phase’ and ‘complex’. MacNeish was ‘appalled’ (1960, 593) at the accusations and what he saw as misinterpretations of his methodologies. It was 1948 all over again; Taylor had made another colleague – a rising star – uncomfortable. MacNeish became defensive, claiming that some of the problems with his conclusions were Taylor's fault because he had not answered MacNeish's written requests for comparative information from his (Taylor's) excavations. He also went on the offensive, challenging Taylor to publish his excavations in Coahuila and submit to similar scrutiny. Win, lose or draw for MacNeish, one thing is certain: his exchange with Taylor forced him to clarify his methods and methodologies. Taylor's critique redefined MacNeish's approach to stratigraphic recording and interpretation, and shaped the rest of his career as well as his ability to conduct and interpret archaeological operations (MacNeish 1978, 247; Flannery 2001, 152).

The MacNeish–Taylor debate draws attention to several matters of relevance to my discussion. The first has to do with practices in stratigraphic analysis and how these may vary from individual to individual, from project to project, and/or according to cultural or national tradition. I was trained in the Levant under a British-derived system of excavation techniques and stratigraphic analysis; natural/cultural, as opposed to arbitrary, levels were standard, as were Harris matrices. I was shocked to arrive at Copán, Honduras, for graduate fieldwork and discover that 20-centimetre arbitrary levels were the norm for excavations and had been the basis for virtually all work since the 1970s. Not surprisingly, perhaps, the recent 30-year phase of research there was begun by an archaeologist trained on New Deal-era excavations, namely Gordon Willey. My interests in modifying procedures were met with misgivings by more senior archaeologists, but I nevertheless found a happy – and acceptable – middle ground by excavating natural levels within arbitrary ones. Arbitrary levels, however, endure as a mainstay of many excavations in the Maya area and beyond, yet one has to wonder if it would not be useful for Americanists to discuss more directly why, how and when this practice is appropriate. Sadly, debate and confrontations in Americanist archaeology are still saddled with stigmas, in large part owing to the memory of Walter Taylor (1948; 1960a; 1960b; 1972).

The matter of designating, determining and identifying phases is another problem where we see terrific variation in strategies, everywhere in American archaeology. Especially in the Maya area and along its peripheries, for example, new sites and sequences are appearing every year, yet all too often these are approached with preconceived models and assumptions regarding assemblages, typologies and, of course, excavation techniques. Part of the problem, certainly, is the massive increase in data over the past few decades, while the publishing outlets for site reports have decreased. But there is also a general lack of willingness to engage in spirited debates over the minutiae and the basics of our fieldwork, such as differences in stratigraphic data and interpretations. Journals and their editors are only partially to blame; it also seems that Americanists are trapped in middle-range theorisms and their attending (and varied) epistemologies, with no way out. In sum, before we ask what the implications are of these problems for the social stratigraphy that McAnany and Hodder encourage, we might simply ask what we are missing in terms of more basic data and dialogue.

The second point of relevance to my discussion has to do with the implications of the MacNeish–Taylor debate for the complex social stratigraphy – history, historiography and future – of Americanist archaeology. Scotty MacNeish, partly thanks to Taylor, became a prolific scholar and hero of processual archaeology. His research and publications (his ‘deposits’) have directed the study of agricultural origins around the world, influenced dramatically the study of complex societies in Mesoamerica and elsewhere, and elaborated Mexico's pivotal role in the history (and future) of the world's maize production. MacNeish left behind a considerable cohort of friends and colleagues and was acknowledged and granted the highest awards by numerous academies and institutions. Walter Taylor (d. 1997), on the other hand, left visible marks that are more localized, read almost solely within the Americanist field of archaeology. These were in no small part ideological and psychological, and substantially less material than MacNeish's, but they diffused widely and forcefully. Taylor felt that it is an archaeologist's obligation to assail and defend theories and interpretations of data – that this was the sign of a healthy discipline. Few others at that time or since have taken this stance. Due in large measure to his willingness to critique others’ work, Taylor has been erased from certain historical profiles, misinterpreted in others, and exploited without attribution (i.e. memory) in yet others (Reyman 1999; Maca 2009). His influence remains visible in some pockets and regions (e.g. Golden and Borgstede 2004; Maca, Reyman and Folan 2009), and his gravitational force endures in the conscience of American archaeology. Taylor's ‘profiles’, however, look very different from those of MacNeish and their articulation requires different – perhaps as yet unknown – tools and theoretical strategies than we are used to using. The example offered by another debate draws this point out a bit more clearly.

Socially and politically established – that is, anchored and embedded – segments of Maya archaeology have recently begun adopting versions of Walter Taylor's conjunctive approach as a validation for past research and as a guide for future approaches (Marcus 1995; Bell, Canuto and Sharer 2004; Canuto and Fash 2004; Golden and Borgstede 2004; Sabloff 2004; Sharer and Golden 2004). A 1983 article by Joyce Marcus is one of their most-cited general foundations. Therein Marcus offers critiques of research in Maya archaeology as well as recommendations for more effective approaches. These are strikingly similar in tone and content to those of Taylor (1948), yet she never once cites him – Taylor is remembered but erased, or summoned but forgotten. It was years before Maya conjunctivists (re)attached his name to their movement (Marcus 1995). Nevertheless, in a response to Marcus's (1983) paper Norman Hammond (1984) immediately recognized its critical Taylor-like (‘outsider’) perspective. Hammond's rejoinder was written to express dissatisfaction with Joyce's incomplete knowledge of some of the topics she addressed. In particular, he highlighted her ‘lack of appreciation of the nature of archaeological, stratigraphic, evidence and the ambiguities and uncertainties that are part of its nature’ (ibid., 821). Hammond in effect ‘pulled a Taylor’ on Marcus who had ‘pulled a Taylor’ on Maya archaeology as a whole.

Hammond was especially perplexed by Marcus's understanding of the stratigraphic and radiocarbon evidence in the Maya area for the periods during and subsequent to the transition from the late Archaic period to the Formative or Preclassic, Early Village, period, ca 2000 B.C. He opens his corrective by conjuring MacNeish (naturally):

Hammond's subsequent remarks focus mainly on questions of the delineation of phases at Cuello, a site he excavated in Belize, and discuss (almost exclusively) issues related to dates, especially uncalibrated versus calibrated radiocarbon dates from the periods, phases and sites in question. His is a highly technical response, but one that we do not see often enough in the literature (and almost never in the literature on social theory); these are frequently downplayed in favour of seemingly higher-level discussions. While correcting Marcus on the finer points of his own and others’ work in Formative-period Maya archaeology, Hammond touches on matters that are fundamental to piecing together complicated stratigraphies from different sites. Ultimately we can ask, what is at stake in this sort of dialogue and what do we learn?

One of the vitally important results that derived from Hammond's (and others’) stratigraphic and phase analyses was the evidence for the in situ development of Maya civilization: it did not arise from cultural diffusion or immigration from other regions. This has, and has had, implications for living indigenous Maya peoples as well as for the nation states that govern them and their ancient sites. Exploring this development with the care it deserves, however, requires close attention to stratigraphic details and the mechanisms and interpretations we use to derive the absolute chronologies of our sequences. Clearly not all professional archaeologists have a handle on these issues or on the data in their geographical and cultural areas of expertise. One of the better-known examples of this problem can be seen in the debate regarding obsidian hydration dates and the Classic period collapse at Copán (Webster and Freter 1990; Braswell 1992; Manahan 2004). Without a good handle on the absolute (and relative) dates of stratigraphic sequences, we are at a loss piecing together phases and relationships between sites and regions, and an effective social stratigraphy is difficult to muster. On the other hand, when we have good dates and/or corroborating chronological information we can not only construct reasonable local and regional sequences but also identify gaps in our knowledge. A look at one final debate demonstrates this last point fairly well.

The third debate I will mention is more of an ongoing assault on a two-decades-old theory of ancient statecraft and political change at Copán, a UNESCO World Heritage Site in Honduras; analyses of architectural stratigraphy have proved to be key in this regard. Copán is currently the site of the longest-running ‘conjunctive approach’ anywhere in the world. While Walter Taylor has recently been hailed as the founder of the Copán approach, his general and specific original ideas are present only as vestiges. Today the conjunctive approach in the Maya area largely means multidisciplinary and social-historical research; at least this is the model encouraged at Copán. One of the leading advocates of the conjunctive approach is William Fash (Fash and Sharer 1991; Fash 1994; Canuto and Fash 2004), whose Copán Acropolis Archaeological Project has, among other things, excavated and restored a building on Copán's acropolis known as Structure 22A.

Owing to tunnel excavation into the acropolis, it is known that Structure 22A has no historical or stratigraphic precedent; it appeared anew sometime during the Late Classic period dynastic decline in the 8th century AD. Employing multiple, conjunctive lines of evidence, especially sculptural iconography and epigraphy, the excavators have interpreted Structure 22A as a popol naah or council house. Further, on the basis of what appear to be Mayan date glyphs on the building's façade, Structure 22A is thought to have appeared immediately after the death by capture of one of Copán's greatest kings (in AD 738). The presence of the council house, therefore, is thought to represent a shift to council rule during the reign of a subsequent weak king. This has been the standing theory for more than 15 years (Fash 1992; Fash et al. 1992; Larios, Stuart and Fash 1994; Stomper 2001). It has been challenged on the basis of iconographic studies of the façade sculpture; these show that the ‘council’ members’ alleged home locales are actually supernatural place names (Wagner 1998). But no one ever thought to check the architectural stratigraphy.

Shannon Plank (2004) finally did this as part of her doctoral work and discovered that there is no secure stratigraphic evidence that the building dates to the time period cited by the excavators. She combined a basic reanalysis of their stratigraphic interpretations with epigraphy, Maya calendrical information, and iconography, and arrived at a very different conclusion regarding the date and function of the building. Plank argues convincingly that it was built three decades after the death of the great king as a shrine or sleeping house for an assemblage of deities. This adds an entirely new dimension to the history of the Copán dynasty and state, but this is not all that is learned through this example.

The popol naah theory for Structure 22A has become so pervasive in Copán's museum exhibitions that a replica has been built in the town hall to signify community and democracy. More importantly, its façade images now adorn modern buildings and businesses across Honduras and its symbolism is one of the reasons that Honduran presidents are now inaugurated at Copán. A simple lesson here, following on the heels of the other cases and debates I have cited, is that before exploring social stratigraphic processes, including especially those that intersect with the present, we will have to carefully evaluate our practices of stratigraphic recording and analysis. There is much work to be done. In the case of Copán's Structure 22A, this may simply mean that we return to published accounts and/or drawings of stratigraphic relationships. In other instances, we will need to evaluate the dates, phase designations, terms and definitions, and artefact complexes associated with stratigraphic interpretations. These considerations can serve as the basis of an emerging social stratigraphy, one that includes studies of published evaluations and critiques and the influence these have on our discipline. Other types and forms of analysis may proceed from this that demonstrate how, and the extent to which, present and past intersect at numerous levels of significance.

I will close by citing a line from Patty Jo Watson's review of the book Social archaeology (Shanks and Tilley 1987). She writes (Watson 1990, 219),