Frieman (Reference Frieman2024) observes in her own, highly metaphorical language that one can offer an unbounded number of interpretations to explain the distribution of archaeological remains in time and space. These interpretations offer different perspectives that can inform action—in Frieman's case an explicitly feminist understanding of the past informing the present. She provides two brief examples from the literature, suggesting that each embodies present-day biases: the distribution of Bronze Age swords relative to the provenance of ornamentation sets in Denmark and Germany, and the ‘Egtved Girl’, a Bronze Age burial of a young person of undetermined sex clad in a bronze-decorated tunic, associated with jewellery and the cremated remains of a child. Interpretations previously advanced for the first example include a patrilocal residence system wherein male warriors brought to their natal homes women ornamented with objects from their own homelands; from this interpretation we hypothesise the presence of patriarchal chiefdoms. The second example, the Egtved individual, has been characterised as a foreign bride, isotope analyses suggesting an itinerant life in the months prior to death. As each interpretation lingers in the literature, it becomes a certitude on which researchers build. Alternative interpretations go unimagined. But Frieman argues for the need for multiple, culturally complex interpretations that emerge from the gaps in the evidence, or the ‘unproofs’.

We are unable to directly observe Bronze Age culture. We can describe the material remains, but those are, at best, proxies for that which we cannot see and, therefore, cannot directly describe. We are confronted with the ineffable, Frieman's unproofs, borrowing from a scene in an Ursula Le Guin novel. These unproofs—these uncertainties—offer potential for the development of new, insightful models of past events and behaviours where there is uncertainty there is space to think (Frieman Reference Frieman2024). To the extent that we can imagine alternative scenarios and that those views could influence action in the present, I am with Frieman. But I do not think that imagining without evaluation is prudent, nor do I think that is the only way we can model the ineffable. Frieman does not directly address evaluation in her article, but it is critical that we do so.

I could imagine an ecological setting populated with characters and objects, and then use that scene to explain what we think people were doing and why, accounting for the material residues of that time and place. Name me an archaeologist who has not done that. The problem lies in evaluating the plausibility and supportability of competing scenarios, because without doing so we introduce yet another ‘just-so’ story to the discipline and to society, possibly—even likely—supporting prevailing prejudices. Post-processualism permeated my graduate student years, but I entered archaeology during the height of the New Archaeology era and hypothesis testing informs everything I do, including recognising a certain probability of error which varies inversely with the quantity and quality of data.

A theory that is not testable is not a theory, it is a fantasy and, in the context of current political discourse, it could be a malignant fantasy. We must weigh alternative interpretations, otherwise every view is valid. This is not to say that only one can be right. Frieman recognises that multiple alternatives can be true, contributing to a more comprehensive view of the past, but I think she would agree that they cannot contradict one another. And they should be subject to analysis: “As soon as we map out our unknowns, a thousand new stories emerge, each ready for critical analysis and each with a different resonance in our present” (Frieman Reference Frieman2024: 1684, emphasis added). Elsewhere, I have suggested playwrighting as an analytical tool (Gibb Reference Gibb, Jameson, Finn and Ehrenhard2003, Reference Gibb, Van Dyke and Bernbeck2015, Reference Gibb, van Helden and Witcher2020), but any structured method will fit the bill.

I detected in the introductory paragraphs of Frieman's (Reference Frieman2024: 1679) article a jab at Big Data and new technologies for data collection and analysis, citing those who have remarked upon the “death knell of theory” and the “retreat to data [in] an epistemological poverty yielding an unrepresentative, singular discourse about past lifeways”. To be clear in my bias, theory is a means to understanding; it is not an end and data do not exist to serve theory. Without both we lack the ability to create those imaginings that describe the past and serve the present: without them there is no understanding. Nor is there the ability to evaluate competing interpretations.

I appreciate that there are gaps in what we know, or can even suspect, about human time and space. Call them unproofs, ineffables, or just what-the-hells. Imagination—a faculty enjoyed by every scientist, to varying degrees—is a powerful tool for identifying and exploring those gaps. It helps us to describe the indescribable, provided we do so in a structured, disciplined way and that we always establish those criteria by which we would deem our description insupportable. But there is a companion approach: collecting and analysing a broad range of data—Clifford Geertz's (Reference Geertz1973) ‘thick description’—seeking the understandings of our subjects through detailed observation and analysis. For example, Frieman briefly reports other researchers’ patrilocal-patriarchal model of Bronze Age society in one part of northern Europe. But the analysis is not deep. Some of the octagonal-hilted swords, she tells us, were recovered from burials—but not all. Presumably others were recovered from rubbish-filled pits, middens, or lack any provenance beyond site location. Does this not suggest that these swords might have been used in different contexts, the meaning changing with cultural context? They certainly entered different archaeological contexts that, in some way, reflect different contexts of use. The chiefly polities—stand-ins for homelands—of women's ornament sets are all represented by circles with identical radii on the map (Frieman Reference Frieman2024: fig. 1). Were these polities of equal size, and therefore of similar power and prestige? Did the authors whose model Frieman critiques take into consideration differences in settlement size (number and sizes of dwellings) and, to borrow from classical market geography, different orders of central places (presence and distributions of craft, trade and institutional sites)? Their model may not be wrong—I have only Frieman's summary to go by—but it is not sustainable based on the data presented. Thick description and experimentation with unproofs may be what we need before letting imaginings loose on the discipline and on the wider public.