Price and Jaffe (Reference Price and Jaffe2023) develop a compelling argument that archaeologists have under-theorised the role of failure in past human societies. The authors contend that we must adopt a flexible approach to failure and recognise that power asymmetries, distributed agency and the temporalities of outcomes all play a critical, if variable, role in the success or breakdown of a technology, cultural practice, or institution.

Foregrounding problems of failure brings into high relief the difficulty of disentangling agency, structure and intentionality in interpreting historical process and social change from archaeological datasets. The Annales School, Deleuzian-inspired assemblage theory, Actor Network Theory and even old-fashioned systems theory—remember negative and positive feedbacks?—all grapple with similar problems. The daunting objective remains to explain why certain traditions endure, transform or disappear and it proves even more challenging to determine whether such processes failed or succeeded according to universal criteria of what actually constitutes ‘failure’. As a minor criticism of the article, to my mind, failure as a broader concept differs from notions of a ‘mistake’, ‘error’ and even ‘malfunction’. Certainly, to err is human but blunders do not necessarily signify that culturally established procedures or traditional ways of making (i.e. technology) ‘failed’ (or were failures). In the example of the ruined pot of cheese from Bronze Age Denmark, the process on the whole was likely successful and producers may have expected the inevitability of a few botched batches. The common occurrence of wasters near kiln sites presents a similar scenario; the remains of successfully fired pots most often outnumber the aborted specimens. Therefore, the charred vessel may not fairly reflect incompetence or carelessness but is instead the accepted by-product of the larger crafting process. I agree with Price and Jaffe that ignoring errors in past societies wrongly evacuates the humanity of the subjects we study. However, we should also proceed with caution in judging actions preserved in the material record according to unquestioned benchmarks of ‘perfection’, to avoid imposing modernist values and sensibilities.

In this vein, I also argue that the identification of small-f failures can only realistically proceed if the archaeologist can establish a standard of accepted techniques and actions for a specified production process. For instance, was the faulty Minoan flooring and drainage system at Palaikastro an isolated example that deviated from the higher standards of construction documented at Knossos? Alternatively, was it more widespread in the Bronze Age Aegean and did it function adequately enough during episodes of light rain? If the flooring did depart from an established and recurring norm, then I would agree with the authors in labelling this a failure. Such a finding would also support the argument that the building in question was inhabited by lower-status individuals emulating the houses of urban elites without access to the same skilled architects. If the examples were widespread and enduring, however, then perhaps we err in calling this a failure.

Archaeologists, then, should add the identification of historically specific ‘standards’ or norms to the analytical framework of scale, time and intentionality in analysing the outcomes and performance of a specific technology or social endeavour. Price and Jaffe might dismiss my criticism as an old-fashion adherence to cultural relativity, but I think there is no avoiding the truth that, to cite the authors, failure ‘lies in the eye of the beholder’. If poorly built homes endured for 1000 years (temporal scale) in a particular culture, despite the flaws we might recognise, and with the existence of superior alternatives during the time period in question (intentionality), then I am not entirely convinced that small-f failures would adequately describe this tradition. The materiality and spatiality of such houses also constituted integral components of the ‘structure’ that conditioned intent and constrained or enabled deliberated actions. Acknowledging this fact serves as a reminder that we never actually research the culturally relative (an admittedly passé term these days); instead, we analyse decidedly real (material), if historically specific, practices. Along the north coast of Peru, certain Moche temples (AD 300–800) were built expediently with the expectation that they would require constant and ritually sanctioned renovations and upkeep (Gamboa Reference Gamboa Velásquez2015; Swenson Reference Swenson, Jennings and Swenson2018). In this instance, what we might identify as substandard workmanship is best explained in terms of Moche ontologies of place and concepts of time and impermanence.

A dedicated archaeology of failure also needs to consider the emic end of the problem to explore how past societies defined and attempted to rectify human errors and externally driven stress. Such a focus would improve interpretation of how intentional actions exacerbated or remedied failures. The correction of the drain at Palaikastro demonstrates such a response and thus supports Price and Jaffe's argument that the initial construction failed and was recognised as a failure. On the north coast of Peru, ancient farmers built redundant and parallel irrigation canals in anticipation of both dune encroachment and El Niño-induced flooding (Dillehay & Kolata Reference Dillehay and Kolata2004). This evidence indicates that coastal communities learned from previous environmental perturbations that led to systematic destruction of agricultural infrastructure.

The ‘old chestnut’ of ritual also provides an unrivalled window onto culturally mediated reactions to precarity and calamity. As Price and Jaffe rightly note, communities often intensify ritual in the face of impending failure and such rites can profoundly shape the material record. Across many societies, foundation sacrifices under dwellings or within bridge pylons commonly served to protect structures from collapse. Sacrifices in the Moche culture of the Andes often coincided with environmental catastrophes such as devastating rains and floods, and the Inca Capacoha sacrifice of elite children occurred at times of grave uncertainty and anticipated danger (Cobo Reference Cobo and Hamilton1990; Bourget Reference Bourget and Pillsbury2001; Reinhard & Ceruti Reference Reinhard and Ceruti2010). Of course, the ethnographic record documents ritual's central role in ensuring that gardens grow, boats stay afloat, commerce prospers and houses remain standing. Therefore, ritual provided a prime form of insurance against failure, and studying the former should ideally shed light on the latter. In Evans-Pritchard's (Reference Evans-Pritchard1976: 22–24) famed discussion, the Azande of central Africa were perfectly aware that termites caused the collapse of an infested granary. If the edifice fell with people inside, however, the Azande would attribute the ultimate cause to witchcraft. Indeed, in many cultures, ontological others (e.g. spirits, gods, ancestors) determine in the final instance what succeeds or fails, and both mundane technological routines and grandiose ceremonial events (such as Ongka's big moka) require complex ritual interventions. In this light, archaeologists could fruitfully investigate both the predictable patterning of ritual contexts and the exceptional deviations from such trends in order to interpret how past people rationalised and reacted to different scales of (potential) failure.

Price and Jaffe's article has proved very useful in rethinking some perennial problems in archaeology and I applaud the authors for initiating a needed conversation on how to identify and investigate failure in the archaeological record.