I. Introduction
The oft-neglected loom weight is finally receiving its due. While these artefacts have long been ubiquitous in archaeological assemblages and museum collections across the Mediterranean, only recently have scholars begun to appreciate the insights they afford into daily life in antiquity. New-found interest in loom weights has spurred the formation of a thriving subfield which has introduced innovative methodologies and significantly expanded our knowledge of the warp-weighted loom.Footnote 1 This recent research has also highlighted one of the unique characteristics of the loom weight as an artefact. Weavers did not use loom weights as individuals but rather in groups and, as a result, these artefacts only assume their full analytical significance when recovered in large numbers. The multifarious quality of loom weights has meant that the set, that is, a group of loom weights used together on a single loom simultaneously, has become a central concept in the subfield of textile archaeology.Footnote 2
This reasonable focus on identifying productive sets, however, has obscured a fundamental issue in the analysis of groups of loom weights recovered in archaeological contexts. There is a reflexive tendency in publications of archaeological textile tools to label any group of weights found together within a single room or the same area of a building as a set. Once weights have been labelled in this way, experts are able to perform technical analysis and form hypotheses about the type of production performed in the structure where they were discovered. While this approach has yielded valuable results, I nevertheless argue that too little attention has been paid to clearly delineating the varied depositional processes responsible for the accumulation of weights in the archaeological record. In response, I aim to refocus our attention on these processes and demonstrate how a broader range of social practices beyond the act of weaving account for their find-spots.
This article therefore begins with a survey of archaeological contexts in which loom weights have been found in the exact location where they were used for weaving in antiquity. The survey shows that such situations are exceedingly rare and that most loom weights were deposited in the archaeological record through various processes of reuse, loss, breakage, construction and storage. In particular, close study of textile tool publications reveals that this final category of storage accounts for a far more significant proportion of archaeological loom weights than is usually recognized.
Far from devaluing these weights, however, an acknowledgement of the importance of storage opens new avenues for research. I will conclude this article by exploring three interrelated implications. First, I conceptualize storage as a dynamic process and highlight how loom weights left in such contexts likely never constituted the full range of tools owned by an ancient household. Second, I consider how storage contexts offer a window into the decisions ancient people made about what to leave behind in moments of extreme duress and what these decisions reveal about the value they accorded to their tools. Lastly, I demonstrate that implicit in the act of storage is an expectation of retrieval, which in turn sheds light on circulation patterns within households. By tracing these patterns and the interactions they provoked, I argue that other, less visible responsibilities assumed by weavers in their daily lives become discernible, such as their role in the surveillance of enslaved labour in the ancient Greek household. Altogether, these insights establish how a renewed focus on the storage of loom weights affords a more nuanced picture of the lives of weavers in antiquity.
II. Where are all the looms?
Looms in the ancient Mediterranean world were typically constructed in wood and seldom survive in the archaeological record. There are rare exceptions, such as the Workmen’s Village at Amarna, where the hot, dry Egyptian climate has preserved several wooden warp-spacers, but generally archaeologists must reconstruct ancient weaving practices through the combined study of iconographic representation, ethnography and the remains of other textile tools made in more durable materials.Footnote 3
Under certain circumstances, however, it is still possible to determine exactly where a loom once stood in antiquity, even in climates not conducive to organic preservation. When a warp-weighted loom burns during the rapid destruction of a building, its loom weights will often fall onto the surface directly below, to be found where the loom once stood, surrounded by the carbonized remains of the wooden frame. This process was first studied in detail by Ingrid Schierer, whose pioneering experimental work simulated the destruction of looms set up for various types of production, including plain weave tabbies and several types of twills.Footnote 4 These tests showed that for a typical tabby production, the weights generally fell to the ground in two distinct parallel lines (fig. 1).Footnote 5 For other types of weave, such as elaborate twills, the weights fell in a more haphazard and overlapping fashion, but still always arrayed roughly in a line running parallel to the wall which supported the loom.
Rows of loom weights excavated at Pakenham, Suffolk (photo by N. Smedley).
Considering the centrality of weaving to daily life in the ancient world, one might assume that it is relatively common to find the remains of a loom in this manner. This is not the case. The recovery of a burnt loom requires that several increasingly unlikely events occurred in succession. First, there must have been a loom in use when the structure burned down, and its charred remains must have been sealed rapidly as a result of the destruction. While weaving was an essential and time-consuming activity in ancient households, there were certainly lulls in production when the loom would be put into storage. Second, residents would need to decide against rebuilding the destroyed structure because doing so would remove the debris created by the fire. Third, they must fail to return to the structure to retrieve their belongings. Finally, even if all these events occurred, the line of weights would still need to survive any later pilfering of the structure as well as any number of other post-depositional processes.
This hypothetical reconstruction demonstrates why such situations are exceedingly rare. Indeed, a survey of the evidence from archaeological sites across Italy bears this out. Despite the vast amount of excavation conducted in Italy, Margarita Gleba’s landmark study of archaeological textile tools from the first millennium BC includes only a handful of potential examples of burnt looms.Footnote 6 At three sites, Pozzuolo del Friuli, Santorso and La Piana, carbonized wooden beams were recovered alongside concentrations of weights possibly indicating the remains of a loom, although at Santorso the excavators concluded that the loom was not in use at the time of abandonment based on the arrangement of the weights, while the publication for La Piana refers to the potential loom remains as ‘very tenuous’ (fig. 2).Footnote 7 Gleba’s study also cites three other examples, Acquarossa, Serra di Vaglio and Poggio Bacherina, where weights were recovered in a rough line along a wall within deposits which presented evidence of burning, providing a relatively firm basis for inferring locations where weaving was once performed.Footnote 8 To these can be added two more recent examples from Entella and Torre di Satriano.Footnote 9
Map of Italian sites mentioned in this article.
Fig. 2. Long description
Pozzuolo del Friuli and Santorso are located in Northern Italy; La Piana, Poggio Bacherina and Acquarossa in Central Italy; Cancellara, Torre di Satriano, Oppido Lucano, Serra di Vaglio and Herakleia in Southern Italy; Himera, Entella, Mozia and Morgantina in Sicily.
While these contexts constitute vital evidence for the locations where looms were used in ancient households, their rarity also limits the conclusions which may be drawn from them.Footnote 10 For example, lighting appears to have been a central concern in the placement of the loom at Torre di Satriano, found conspicuously close to the northern part of the structure’s monumental entrance, and at the rural farmstead of Poggio Bacherina, where the burnt loom was recovered in a small room adjoining an open courtyard.Footnote 11 At Acquarossa and Entella, however, roofed interior rooms with limited access to natural light were chosen rather than more suitable locations closer to an entrance, with weavers perhaps prioritizing shelter and comfort over natural light.Footnote 12 Whether these divergences relate to differences in preference, or whether they resulted from some other factor, such as the time of year the structure burned down, is impossible to determine at this point. The value of these examples is also curtailed by the fact that in several cases the loom weights were destroyed beyond recognition due to the very fire which preserved them in a line, rendering any further analysis of their functional properties infeasible.Footnote 13 Ultimately, these examples account for a miniscule percentage of the full range of weights available for study from Italian contexts. It is therefore necessary to broaden the analysis to assemblages of weights deposited in the archaeological record through other means.
This review of the evidence from Italy serves as a useful point of departure because the sweeping scope of Gleba’s study highlights the rarity of such discoveries particularly well. A more impressionistic survey of the broader Greek world produces the same results. Olynthus has held a particularly influential place in the study of Greek textile production ever since Nick Cahill’s reanalysis of the city’s houses.Footnote 14 Although his publication lists the locations of looms in various buildings throughout Olynthus, Cahill only cites one potential example, the poorly preserved Villa CC, where a group of weights was recovered in a line indicative of the destruction of a loom.Footnote 15 Moreover, a closer look at this case reveals that it is unlikely that the weights represent the detritus of a loom because the room in which they were found seems to have been too small for weaving. Instead, these weights probably fell from a low shelf where they had once been stored. Otherwise, there is not a single example of a burnt loom from the entire site. Similarly, at Athens few excavations have found large groups of loom weights. The one exception is Building Z from the Athenian Kerameikos, a controversial structure which has been variously identified as a textile workshop and a brothel.Footnote 16 Nevertheless, even in this much-cited building, there are no groups of weights which have been identified as having fallen directly from a loom during destruction.Footnote 17 The same pattern holds at other Greek sites where substantial numbers of domestic structures have been excavated, such as Halieis, although one interesting example of a burnt loom was recently discovered at Hellenistic Euesperides in Libya.Footnote 18 While a full survey of the entire Greek world is beyond the scope of this article, this brief discussion of some of the most commonly cited textile tool assemblages entirely conforms to the picture painted by Gleba for Italy in the first millennium BC.
III. Loom weight storage
Most loom weights are recovered as singlets during excavation. Such weights end up in the archaeological record through a variety of activities which are often impossible to fully reconstruct. Reuse must have been an important factor. Gleba proposes that single loom weights might be reused ‘for holding down roof thatching, for tying sacks, as fishnet weights, or as supports for holding a spit over fire’.Footnote 19 Many other uses could surely be suggested. Once a reused weight was broken or no longer needed, it could be discarded individually, severed from its primary function as a textile tool.
Construction also played a significant role in the deposition of loom weights. Building new structures and refurbishing old ones often requires the deposition of fills for levelling foundations and raising floors, what Michael Schiffer describes as ‘earth-moving processes’.Footnote 20 One by-product of such earth-moving endeavours is the unintentional transfer of artefacts from one location to another. As small ceramic objects, it is no surprise that loom weights often found their way into such contexts. Perhaps they were sometimes even thrown into fills intentionally, as Paolino Mingazzini suggests in relation to the fills of the Pnyx.Footnote 21 A recent publication of the House of the Two Skeletons at Morgantina provides a clear picture of this process, with several loom weights listed among the finds from levelling fills and floor contexts across the building.Footnote 22
Reuse and construction therefore account for a significant amount of the loom weights recovered on archaeological sites. To these can be added weights which were broken during production and discarded, or others which were simply lost or misplaced. As mentioned above, in a few rare situations loom weights have also been found in the exact location where they were used on a loom. Nevertheless, there is one other context in which weights are frequently discovered which deserves significantly more attention. This context is storage, and its importance in structuring the recovery of archaeological loom weights has not yet been fully appreciated.Footnote 23
It is common for archaeological reports to record the finding of a group of loom weights clustered on a floor surface (fig. 3). Publications often explicitly assert that these findings provide evidence for the location of a loom, even when such weights are not arrayed in a line and no charred organics are present. For example, the publication of the burnt loom from Torre di Satriano cites two well-known cases from Oppido Lucano and Cancellara as comparanda.Footnote 24 Both are instructive as to how this interpretive move works. The excavators at Oppido Lucano published their findings from four houses which all produced significant amounts of loom weights. In particular, a group of 21 weights was recovered in a single room in Casa A, while two other groups of 58 and 21 weights were brought to light in the corners of a room in Casa D.Footnote 25 The publication claims that each case attests to the location where a loom had been set up during the final phase of occupation.Footnote 26 Yet there is in fact no evidence to support this assertion, since each group is described as being piled up on a floor surface. Likewise, at the site of Cancellara, a similar interpretation is put forward for groups of 11 and 13 weights found in contexts dating to two different phases of a single house. The publication on these findings asserts that each pile of weights corresponds to the location where a loom was set up during its respective phase, again without any corroborating evidence.Footnote 27 The same move has been continually replicated at many other sites, such as Himera, Mozia and Herakleia di Lucania, always without any of the expected signifiers identified by Schierer.Footnote 28 In each of these cases, the most reasonable explanation is that the weights had been left in storage at the time of their abandonment.
Cache of stored loom weights excavated on Oxford Road, Thame (courtesy of Oxford Cotswold Archaeology).
I recently encountered a similar context while working on a collaborative project excavating the House of the Two Mills at Morgantina. This excavation recovered a group of weights in a relatively restricted part of a room on the eastern side of the building. While our initial reaction was to assume production had taken place in that room, upon further reflection it became clear that the arrangement of these weights in no way indicated that they had fallen from a loom. The best explanation for their arrangement was that they were initially piled in a wooden box, sat on a low shelf or had been placed in a bag made of organic material. Their container must have since decomposed, leaving behind this concentration of loom weights. Although most containers used to store weights were probably similarly ephemeral, there are at least two examples from Italy in which excavators found weights stored inside a surviving ceramic container.Footnote 29 These are exceptions, however, and archaeologists need to pay closer attention to how groups of weights stored in organic containers would appear after decomposition.
This acknowledgement of the centrality of storage in the formation of loom weight deposits has vital implications for their study. While the technical affordances of stored weights should certainly be analysed with the most advanced methods, it must also be realized that they do not provide evidence for production in an entirely straightforward manner. Yet this does not diminish the analytical value of such weights; it enhances them, as will be shown through the close analysis of storage contexts in the following sections.
i. Incomplete sets
The relationship between storage and use is always dynamic. Objects are generally placed in storage with the intention of retrieving them later. Textile production would have been performed throughout the year in ancient Mediterranean households but was likely somewhat inconsistent. Experimental work has noted that the spinning of threads was more labour-intensive than weaving itself and could conceivably have formed a bottleneck that limited production.Footnote 30 Once weaving had begun, a weaver needed to consider several factors when evaluating how many weights to employ for a specific project. The width of a textile piece woven on a warp-weighted loom is ultimately determined by the number of weights used and the thicknesses of those weights. To form a larger piece, an ancient weaver could use many weights arrayed across the entire width of their loom, or they could weave smaller strips and sew them together later. A simple tabby requires only two rows of weights, while more elaborate twills could require up to four. As such, some projects might have needed every weight in a weaver’s collection, while others might only utilize a small percentage. These considerations surely impacted the number of weights in storage at any given moment.
These interpretive issues are further compounded by the functional properties of loom weights. Experimental studies have shown that weights of different sizes are required for working with different types of thread.Footnote 31 A weaver who wanted to use both fine and coarse threads would have needed access to distinct types of loom weights. While archaeologists might prefer that ancient weavers always stored different groups of weights separately, there would have been no reason for this to be the case, since an experienced weaver could always distinguish between them. Therefore, multiple distinct types of weight could easily be stored together in a single container, complicating any straightforward reading of the assemblage.
An important technique for addressing this issue is compatibility analysis.Footnote 32 Based on a loom weight’s dimensions and mass, it is possible to calculate a range of potential thread types which could be attached to it. The results for different weights can then be compared to each other and those which overlap considered compatible. By evaluating each weight, one can discern whether an entire stored assemblage might have been used together as a solitary group or whether several groups of weights were mixed. However, it is essential to reiterate that these results are purely hypothetical. The dynamic nature of storage implies that even an assemblage of homogenous weights found together in such a context may not have constituted the full group once used by their owner. Ultimately, the fundamental implication of this analysis is that archaeologists should not expect to find full sets of compatible loom weights in storage contexts. This insight explains why the number of weights recovered together in storage contexts is highly variable across the case studies cited above.
ii. Storage and value
Other factors beyond production and storage play a role in the formation of loom weight assemblages. Recent work on loom weights from Sicily and southern Italy has highlighted the ways in which the archaeological record strongly indicates that these quotidian tools were often accorded a value which exceeded their mere utilitarian function.Footnote 33 These arguments draw on the literary and ethnographic record to reconstruct how women learned to weave. This was a process during which young women came into repeated contact with loom weights and over time these weights could become associated with specific memories of the older relatives from whom they learned their craft. Such weights might in turn take on the role of an heirloom, forming intergenerational connections through the transfer of craft knowledge. Studies on the embellishment of loom weights confirm this conjecture (fig. 4).Footnote 34
Examples of loom weights with embellishments from Hellenistic Morgantina: a) Inv. 57-2545; b) Inv. 63-461a; c) Inv. 62-1376; d) Inv. 66-748 (courtesy of the American Excavations at Morgantina).
Most excavations of ancient Greek houses yield only a few sporadic loom weights scattered through various fills. In light of their heirloom status, however, this result is unsurprising. If a household moved or their residence fell out of use, it must have been typical to simply take the family set of weights with them. A large box or bag of 50–100 weights would have been heavy but certainly movable, as were the wooden beams of a loom. Archaeologists should expect to find groups of stored weights only in contexts where houses were abandoned or destroyed abruptly. This accounts for the large numbers of stored loom weights recovered in the houses of Olynthus. It also explains the large assemblages of stored weights associated with the two major destruction events at Morgantina in 459 and 211 BC, as well as the complete lack of weights from the slow final abandonment of the city in the first century AD.
In moments of great duress, it might not be feasible to carry large boxes or bags of weights, and residents might leave them behind with the intention of retrieving them later. Nevertheless, because loom weights were often accorded a value beyond their mere material worth, it seems possible that, even in moments of duress, a weaver might grab a handful of particularly valuable weights. Those embellished with the personal emblems of previous generations may have been the first removed as valued heirlooms. Such a process is hard to observe in the archaeological record but worth keeping front of mind when analysing groups recovered in storage. Even when many weights are found together, there remains a strong possibility that a few embellished weights were removed as a sentimental gesture in a moment of upheaval.
iii. Weaving and surveillance
Stored weights do not provide evidence for where weaving was performed in households, although it seems likely that weavers would have preferred to move these heavy objects only short distances. While burnt looms do provide direct evidence for loom placement, even this evidence affords only a snapshot in time and is difficult to interpret. A previous era of scholarship often aimed to identify the gunaikeion through the analysis of artefacts related to weaving, but the prevailing wisdom today is that the use of space in ancient Greek households was flexible.Footnote 35 Seasonality would have been an important consideration in loom placement as well.Footnote 36 Therefore, it is more productive to identify a range of rooms suitable for weaving than to try to pinpoint one dedicated weaving space.
Interestingly, storage locations may have been more rigidly defined than weaving spaces. They are also much easier to find archaeologically. In this last section, it is my contention that identifying weights in storage spaces opens new interpretive possibilities for studying women’s daily lives in antiquity and that a myopic focus on locating production has obscured this potential insight. In short, storage affords valuable information about the spaces regularly frequented by those who wove textiles, almost exclusively women in the ancient Greek household.Footnote 37 Once archaeologists identify other types of objects stored alongside weights, they may be able to observe obliquely other activities these weavers performed. Moreover, locating storage rooms can shed light on circulation patterns within the household.
A preliminary analysis of the evidence from Hellenistic Morgantina will serve as a case study for this approach. Several assemblages of loom weights were recovered at the site in structures dated to the Hellenistic period, spanning roughly from the second half of the fourth century BC until the city’s destruction in 211 BC. For the most part, weights dating to this period come from contexts which do not relate to the primary occupation of the site. However, three groups were recovered on the final floor levels of houses violently destroyed in 211 BC.Footnote 38 Intriguingly, two of these groups of weights were recovered in close proximity to a millstone which still sat on the floor of the same room.Footnote 39 The room in which the third group of weights was recovered was poorly preserved since it had been subject to illegal looting prior to its full excavation, but the disturbed fill from this house did include various fragments of a millstone as well.Footnote 40
It should be stressed both that these are preliminary results and that a sample size of three is perhaps too small to assert broad conclusions. Nevertheless, it is still worth considering why loom weights might be stored near millstones, since at first glance the association between these two classes of artefact is anything but straightforward. It is highly unlikely that weaving occurred in the same spaces as milling. Milling with a rotary millstone (fig. 5), the type commonly used at Hellenistic Morgantina, would have created unpleasant sounds due to the friction between the fixed conical base (meta) and the upper stone (catillus). The mill would also have constantly thrown up particulate matter and spread it throughout the room.Footnote 41 Rotary mills are extremely heavy, and it is improbable that they would be moved into a different room for use. While it is conceivable that weaving and milling occurred in the same room at different times, it seems much more feasible that the loom weights were simply moved from this milling area to another more suitable space.
Rotary millstones from the Contrada Agnese district of Morgantina (courtesy of the American Excavations at Morgantina).
Yet there may be more significance to this recurrent association between milling installations and stored weaving equipment, especially when we consider the social structures and household organization of the Hellenistic world. Due to its back-breaking nature and terrible conditions, milling was considered to be among the least pleasant productive activities undertaken in households and was delegated to enslaved labourers whenever possible.Footnote 42 In contrast, weaving would have been performed by free and enslaved women alike, unless a household was extremely wealthy and could afford for its women not to work at all. Xenophon provides a useful picture of what a woman in a working household would have been expected to do, albeit from his particular vantage point. One of the responsibilities specifically mentioned in his account is the surveillance of enslaved labourers in the house, in particular that of an enslaved woman described as a miller (Xen. Oec. 10.10).Footnote 43 The surveillance of this activity within the household must have been an intentional strategy to discipline milling, especially considering the arduous nature of the labour. The co-occurrence of stored loom weights in spaces with dedicated milling installations perhaps confirms the picture painted in Xenophon’s account of the surveillance of enslaved labourers by free women in Greek households. The placement of these objects in proximity to each other may be interpreted as a deliberate choice to keep tools regularly used by women in the spaces where millers worked.
This observation is further reinforced when the House of the Two Mills is considered in more detail (fig. 6). The stored weights and millstone were found below a tile fall which sealed the final floor level of room 11b. A rich collection of finds was recovered at the same level, including several large pithoi and Graeco-Italic amphorae concentrated in the corners of the room.Footnote 44 Excavators also recovered three echinus bowls, several out-turned-rim plates, a mostly intact guttus in the shape of a boar’s head, a bronze grater and a few stone weights. This assemblage indicates that the room was not only used for storing weights and milling grain, but also for the storage of objects related to food preparation. It may suggest that the women involved in weaving played a role in such activities as well. Furthermore, a glance at the plan of this house shows that room 10 is the closest one to 11b suitable for weaving. If this space was used for this purpose during certain seasons, then weavers would have needed to carry their weights through room 11a, which in its final phase included a large bread oven.Footnote 45 As a result, a picture emerges of a three-room suite in which the most comfortable room was used for weaving, flanked by a room for food preparation and then a third for milling and storage. The free women of the household would only work in the first two rooms but could circulate through all three. The result was that the milling operation in this house was deliberately placed in a space which could be entered by free women of the household at any moment, thereby keeping enslaved labourers under the constant threat of surveillance. When viewed in this light, the practice of storage does not merely fulfil a utilitarian function but becomes embedded within a broader range of daily practices and gendered roles within households.
Plan of the House of the Two Mills with rooms from its final phase labelled (courtesy of the American Excavations at Morgantina; state plan by G. Filantropi).
The sample size here is very small and further analysis at other Hellenistic sites with significant domestic assemblages is necessary to confirm these results. At a minimum, however, this analysis highlights how close attention to the deposition of loom weights paints a more vibrant and dynamic picture of the daily lives of weavers. While direct evidence for production remains elusive, storage offers a window into the social strategies which guided the allocation and organization of work in the Greek household.
IV. Conclusions
Most loom weights recovered during archaeological excavations were deposited through various processes of reuse, loss, breakage and construction. A small number have been found exactly where they were used for weaving. In this article, I have argued that piles of stored weights left behind on the floors of abandoned and destroyed structures form another crucial and underappreciated category. Understanding such weights requires an awareness of the dynamic nature of storage and the range of factors which might lead to their removal in times of duress, including the value accorded to loom weights by their owners. Storage also reveals circulation patterns within households and the social strategies they deployed to surveil enslaved labour. These insights provide a more holistic picture of production in the household than an approach which focuses myopically on their use for weaving alone.
The implications of this article need not be limited to our understanding of textile tools. Archaeologists and historians are predisposed to think about storage practices when analysing large vessels, coin hoards and various artefacts related to food processing. Yet objects with primarily productive functions are often reflexively studied through the lens of their use, even when they are recovered in other sorts of context. By refocusing on those moments when tools were not in use, however, we can consider a broader range of questions and find new avenues for studying the daily lives of craftspeople in the ancient world.
Acknowledgements
I would like to thank the Editor, Lin Foxhall, and the anonymous JHS readers for their generous comments and suggestions. Alex Walthall and Alyson Melzer provided valuable feedback on earlier drafts of this article. I am also extremely grateful to all my colleagues at the American Excavations at Morgantina and the Parco Archeologico di Morgantina e della Villa Romana del Casale di Piazza Armerina.
Open access
