A Village That Left Almost Nothing Behind: What Am Frevelgraben Can Tell Us About Medieval Abandonment

admin · August 30, 2026
Am Frevelgraben Halberstadt Kühlingen medieval archaeology deserted villages Wüstung settlement archaeology radiocarbon dating Saxony-Anhalt abandonment archaeology

At Am Frevelgraben, on the eastern edge of Halberstadt in Saxony-Anhalt, the most conspicuous archaeological evidence is a mismatch. The settlement itself survives as a dense and intelligible pattern of cut features: rows of sunken-featured buildings, associated pits, two wells, plot divisions and ditches that were altered as the occupied area changed. Portable objects, by contrast, are scarce. So far the excavators report only scattered pottery and a small group of metal finds, including a bronze tap, fittings and belt buckles. There is no destruction layer and no evidence for a settlement-wide fire in the excavated area.

This combination has encouraged the Landesamt für Denkmalpflege und Archäologie Sachsen-Anhalt to describe the settlement as having been abandoned in an unusually orderly manner. It is an attractive interpretation, and quite possibly the right one. But the archaeological interest lies precisely in the distance between that interpretation and what the deposits can actually demonstrate. Few abandoned objects can indicate that people had time to remove useful possessions. They can also result from salvaging, systematic cleaning, recycling, later disturbance, differential preservation, or from excavating parts of a settlement where portable material was never abundant. The absence of burning excludes one dramatic scenario; it does not by itself tell us whether leaving was voluntary, gradual, coerced, economically compelled or simply the last stage in a longer contraction.

A settlement known before the excavation

The 2026 fieldwork did not begin with the accidental discovery of an unknown village. Written and cartographic evidence had long suggested a deserted settlement in this part of Halberstadt's hinterland, and municipal planning documents record geophysical investigations commissioned by the city in August 2022. Those surveys confirmed an extensive medieval settlement with numerous sunken structures and other archaeological remains. The current excavations, undertaken in advance of further industrial development and running since March 2026, are therefore the destructive but much more informative second stage: testing geophysical patterns against excavated features, establishing stratigraphic relationships, recovering datable material and documenting the site before construction proceeds.

That distinction matters. Geophysical prospection can reveal coherent buried plans without exposing them, but it does not automatically establish the date, function or sequence of every anomaly. Excavation at Am Frevelgraben has already shown that the apparent plan was not a single frozen village. Intercutting features and closely adjacent sunken buildings establish at least two settlement phases. Houses or ancillary structures were rebuilt in somewhat shifted positions, while ditch systems inside and around the settlement were repeatedly changed and extended. What looks from above like a plan is therefore a palimpsest, the accumulated geometry of a community reorganising its space over time.

The clearly traceable plot divisions may eventually prove particularly informative, because repeated boundaries can expose how space was allocated between neighbouring units and how stable those units remained. Yet a plot is not automatically a household, still less proof of a particular legal system of landholding. The same caution applies to the German term Grubenhaus. A sunken-featured building is a morphological category, not a secure statement of function. Across early and medieval Europe such structures could be used for craft production, storage, shelter or domestic activities, and their role has to be argued from internal features, residues and associated finds rather than from the sunken floor alone. Studies elsewhere show precisely this functional variability.

When absence becomes evidence

The sparse artefact assemblage is therefore not a curiosity to be treated separately from settlement formation; it is part of the evidence for how the site ceased to be occupied. Archaeologists have long distinguished ordinary rubbish from what Michael Schiffer termed de facto refuse: usable things left behind when an activity area or settlement is abandoned. In principle, a low level of such material can fit a planned departure in which people curate and transport what remains useful. It can also reflect later salvage by neighbours or by people who continued to use the locality after residential occupation ended. This is why abandonment archaeology resists the tempting assumption that the final surface of a building is a complete inventory of the last household to live there.

Am Frevelgraben is a good case in which to apply that caution. The contrast reported by the excavators is real: dense and well-preserved structural evidence, but very little portable material. That makes deliberate removal of possessions plausible. What it cannot yet establish is the tempo of departure. An entire village might be vacated over a short but well-prepared interval; it might dwindle household by household over years; buildings could be abandoned while adjacent plots remained in use; or useful materials could be stripped after occupation had ended. The site's two-phase history already shows that construction and relocation were normal parts of its life. Final abandonment may likewise have been a process rather than an event.

Nor does the absence of a burn layer equate to evidence of peace. It is positive evidence against a major conflagration affecting the exposed settlement, and the lack of a general destruction horizon makes a fiery catastrophe a poor explanation for the deposits presently known. But medieval communities could be displaced by taxation, lordship, debt, demographic decline, loss of access to land, military insecurity or environmental deterioration without their houses being burned. Even violence does not necessarily leave a village-wide destruction layer. The safest conclusion at present is narrower: the excavated settlement was not terminated by an archaeologically visible episode of widespread burning or destruction, and much of its movable material did not remain where people had used it.

The chronology is the hinge

Before asking why the settlement was deserted, archaeologists need to know when. A February 2025 municipal planning document described the presumed abandoned village as probably late medieval, approximately AD 1250–1500, and suggested the historical names Kühlingen or Frevel. That was a planning-stage attribution, not a securely dated terminal horizon from the present excavation. The LDA's August 2026 report is appropriately more cautious: the reason for abandonment remains unknown, and selected animal bones are to be radiocarbon dated in order to refine the end of settlement. Excavation is scheduled to continue until mid-December 2026.

Radiocarbon dating may sharpen the chronology, but here too the relationship between measurement and historical event needs care. A radiocarbon determination dates the death of the sampled animal, not the moment when a village was abandoned. Its value depends on archaeological context. Bone sealed in a final-use deposit or a closely defined terminal fill may constrain the last occupation far better than residual bone redeposited from an earlier pit. With a site that has at least two phases and many intersecting features, several well-chosen dates tied to a stratigraphic sequence will be more informative than a single number. The scarcity of pottery makes this especially important, because ceramic typology appears unlikely, at least from the finds presently reported, to supply a dense independent dating framework.

The site's historical name is also best treated as probable rather than demonstrated. Around 1900, maps of deserted settlements placed a place called Kühlingen more than a kilometre west of the excavation, while nineteenth- and twentieth-century Messtischblatt maps put the name immediately north of the excavated area. The LDA therefore considers it likely that the remains belong to Kühlingen. Earlier municipal documentation allowed Kühlingen or Frevel. Place-names can outlive settlements, shift across mapped landscapes or become attached to fields after buildings have disappeared, so the convergence is useful but not equivalent to an inscription naming the excavated village.

One deserted village among many

Am Frevelgraben belongs to a much larger archaeological problem. Medieval Europe contains thousands of Wüstungen, settlements that were wholly or partly abandoned, and the history of their interpretation is a warning against single-cause explanations. Epidemic disease, warfare, agrarian crisis and climatic deterioration have all played roles in particular places and periods. More recent settlement archaeology has emphasised that desertion can also emerge from long-term changes in land use, village organisation, markets and regional settlement hierarchies. Rainer Schreg's work on southern Germany, for example, argues that the late medieval wave of desertion is better approached within longer transformations of rural landscapes than as an automatic consequence of plague or crisis. Werner Rösener has likewise examined climatic stress as one factor within a broader late medieval process rather than a universal key.

That wider literature changes the question to be asked at Halberstadt. If the final occupation really falls in the later thirteenth to fifteenth centuries, then demographic contraction, changing agrarian organisation, the pull of larger settlements, seigneurial restructuring, environmental pressures and conflict all become historically reasonable avenues of research. None is presently demonstrated. If radiocarbon results shift the final phase substantially earlier or later, the relevant historical landscape changes with it. A cause should therefore not be chosen first and matched to the excavation afterwards.

A regional comparison makes the point. At Marsleben, near Quedlinburg, the LDA excavated another medieval deserted settlement during road construction. Written and archaeological evidence indicates growth into the fourteenth century and abandonment around 1400 for reasons that remain uncertain. Yet Marsleben produced a very different archaeological signature: more than sixty stone cellars and abundant material from medieval daily life, while one repeatedly rebuilt cellar appears to have burned and partly collapsed. Both places ended as Wüstungen, but the deposits left by their endings are not interchangeable. Desertion is an outcome, not a single archaeological process with one diagnostic signature.

What the next months could resolve

The most useful next results may come less from finding a spectacular object than from tightening the relationship between chronology, space and final deposits. If artefact densities can be compared plot by plot and phase by phase, archaeologists may be able to see whether material depletion affected the whole settlement at once or whether some units were already being stripped while others remained active. Careful study of the fills of the sunken buildings and wells could distinguish routine rubbish disposal from material introduced after structures had gone out of use. The wells are potentially valuable environmental archives if waterlogged plant remains, wood, insects or other organic material survive, although no such analyses have yet been announced.

The animal remains selected for radiocarbon dating can do more than supply dates if the assemblage is large and well contextualised. Species representation, slaughter ages, butchery patterns and disposal could contribute to reconstruction of husbandry and food practices across the settlement's phases. Likewise, sediments from the Frevelgraben margin might eventually help test whether changing drainage or local hydrology affected occupation, but that is at present a research possibility, not a reported result. The important point is methodological: competing explanations for abandonment become useful only when they generate different expectations that can be tested against the site's stratigraphy, environmental record and spatial patterning.

For now, Am Frevelgraben offers a rare combination of clarity and restraint. Archaeologists can already say that the settlement was organised into recognisable plots, contained numerous sunken-featured buildings and at least two wells, shifted through at least two phases, and repeatedly altered its ditch system. They can also say that the excavated deposits have yielded unexpectedly few portable objects and no general destruction or burn layer. They cannot yet say exactly when the settlement ended, whether Kühlingen is certainly its historical name, why people left, or whether the sparse final assemblage records a single coordinated move rather than gradual contraction, salvage and site-formation processes.

That unresolved distinction is what makes the site worth following. The most interesting question is not whether the inhabitants mysteriously vanished. It is whether the final archaeological pattern can be resolved finely enough to show how a medieval community dismantled itself: plot by plot, household by household, or in a comparatively short and organised departure. The radiocarbon programme and the final-phase contexts will determine how much of that question the soil can answer.

Selected Sources and Further Reading

Comments 0

Log in to leave a comment.