Roman Failure: Privilege and Precarity at Early Imperial Podere Marzuolo, Tuscany

Abstract The case of the early imperial small rural settlement of Marzuolo, in south-central Etruria, paints a micro-history of arrested developments: a couple of decades into the site's existence, an abandoned wine-production facility was converted into a blacksmithing workshop, which in turn burnt down and was abandoned soon after. But were both these endings failures? This article uses the concept of failure as an epistemic lens to examine inequality: who could fail in the Roman world, and for whom was failure not an option? It argues that failure was tied up with particular notions of the future, which were not equally distributed. Yet in contrast to modern paradigms, in the Roman world even the privileged seem not to have embraced failure as a stepping-stone towards growth.


I WHY FAILURE MATTERS
Risk and disaster set the stage for the Roman world: risk as a structural vulnerability of pre-industrial economies grafted on a Mediterranean constituted by thickly but loosely connected micro-ecologies;1 disasters such as earthquakes, droughts and pandemics as the causal movers that tipped the fragile balance of stability. 2Gambling, magic and other strategies offered tools for wide cross-sections of Roman society to turn intransigent uncertainty into malleable risk, and even the most privileged were often bound to embrace a windfall rather than project systematic gains. 3This emphasis on the coast (Fig. 1).Marzuolo epitomises the transitional nature of this historical locale.Politically, it sat at the very edge between the territories of Roselle, to which it probably belonged administratively, and Chiusi.In terms of connectivity, Marzuolo was equidistant from the two main north-south roads, the Via Aurelia along the coast and the Via Cassia inland.While the River Orcia, which skirted the site to its north before owing into the Ombrone, was not navigable, its valley would have facilitated travel.Ecologically, Marzuolo bridges the at coastal landscape characterised by villa economies and the hilly interior, cleaved by seasonal streams and home to a mixed economy of small farmers. 11he site of Marzuolo was rst recorded during a eld survey in 2006-7, and was subsequently subject to two short excavation seasons under the auspices of the Roman Peasant Project (RPP) in 2012−13. 12Since 2016, the Marzuolo Archaeological Project has carried out intensive excavations and studies, which give a rare high-resolution glimpse of the complexity of such 'minor' rural sites and their histories.The lack of later building activity in the north-western area of the site facilitated the detailed reconstruction of a rapid sequence of events, all within the span of around half a century.Around the turn of the millennium, a large building was constructed in opus reticulatum masonry, quite likely as an agricultural processing installation, as evidenced by two cylindrical tanks added shortly after, one of which contained residues of tartaric acid indicative of wine production (Fig. 2, phase 1). 13At most a few decades after construction, the tanks went out of use and were partially lled in with dumped materials.Room C and part of Room D in the opus reticulatum building became home to a blacksmithing workshop (Fig. 2, phase 2). 14In the middle of the rst century A.D., much of the building's northern wing was destroyed in a re, and the blacksmithing workshop was covered by collapse material and abandoned (Fig. 3).Reoccupation, still in the rst century A.D., does not seem to have touched any rooms west of Room D.
This ten-by-ten-metre area of a small, rural Roman settlement in inland Tuscany thus packs a dense and eventful history of endings, including a wine business that was abandoned soon after its initiation and a blacksmithing workshop left behind after a disastrous re.But were both of these events instances of 'failure'?At what cost?And FIG. 1. Marzuolo in its regional context.Red: terra sigillata production sites; orange: towns; black: villas.
(Map by author) for whom?The following two sections will unpack the stories of the vintner and the blacksmith, using the prism of failure to reect on issues of inequality in imperial Roman (rural) society.

Design and planning
The opus reticulatum building at Marzuolo was constructed in the later rst century B.C. or the very early rst century A.D. It was probably inserted into what was at the time a nascent small-scale rural settlement with a row of rectangular, wide-entry strip buildings possibly oriented around a road or trackway.The strip buildings had a socle consisting of a few courses of unmodied river stones bounded with earth and placed either directly on the walking surface or in a very shallow foundation trench (Fig. 4). 15These stone bases carried an earthen elevationpossibly piséand shielded it from any moisture rising from the ground.The technique was common in the area, both on earlier nearby sites such as Podere Cannicci16 and on contemporary sites excavated by the RPP, from dwellings to temporary stables.17This combination of stone base and earthen elevation made the most of the locally available materials in a region that was poor in stone but rich in clay.Whoever designed, built and nanced the opus reticulatum building at Marzuolo wanted to do things differently, though.Its size exceeds that of other buildings at the site by many orders of magnitude (Fig. 2).Its northern wing was at least thirty metres long (and continues beyond the excavation limits).What really set the building apart, however, was its construction technique.Deep foundation trenches were cut in the clay-rich substrate, meticulously measured out to wrap tightly around the foundations.At 80−100 cm deep, these foundations were carefully assembled using roughly cut river stones bonded and layered with white mortar.Near the base of the walls, the foundations transitioned into more regularly cut rectangular stones secured with the same white mortar.The actual walls were built in concrete with a rubble core and a facing of river stones cut into a rough diamond shape, once more with white mortar as a bonding agent. 18ach of the cubiliathe individual diamond-shaped stones that make up the double facing of the wallswas carved individually, a difcult and time-intensive feat given the hardness of the calcareous river stones.The stones were probably fetched from the River Orcia, which ows immediately downslope from the site of Marzuolo.This contrasts with the softer tuff with which opus reticulatum construction is usually associated, used in Rome and its volcanic environs and highly amenable to carving and shaping. 19The result of the harder material used at Marzuolo was stones with a frontal face that only ever approximated the regular diamond shape of true opus reticulatum (Fig. 5).Therefore, whether it is appropriate to call the Marzuolo masonry opus reticulatum or opus quasi-reticulatum is a moot point, that might erroneously ascribe chronological weight to what was in reality the product of available materials. 20More interesting is how the imperfectly diamond-shaped stones made the task of assembling the walls into a tall order, thereby confounding any explanation that sees opus reticulatum masonry as a strategy of efciency. 21Whereas cutting soft volcanic tuff stones to create a ready-made building kit to be assembled by unskilled labour speaks to rationalisation, carving recalcitrant rounded calcareous river stones to generate a giant puzzle dees any means-ends logic.Something other than efciency and labour management seems to have been at stake here.Marzuolo's opus reticulatum speaks above all about the desire to execute a design.In doing so, the performance was at least as important as the result.Its construction paraded a true spectacle of labourlabour for cutting the stones, and labour for assembling themthat was local, extended in time, and both highly visible and audible.It is impossible to pin down the nature of that labour, which could involve slaves, but is likely also to have drawn on the local farming and crafting population of Marzuolo and its environs, whether through forms of payment, social obligation, or economic dependency. 22Whatever the source of labour, construction of this building was conspicuously wasteful of it.
This spectacle of labour staked a claim to the future: building in opus reticulatum was building for the future.This future envisaged a particular function for (at least part of) the building.Room D in the northern wing, entirely paved in opus signinum (whitish mortar containing orange-to-pink ceramic inclusions), contained two cylindrical tanks at its back, installed shortly after a rst use-level consisting of a beaten earth oor.The tanks were sunken below oor level and carefully nished with an opus signinum waterproong, with partially preserved tile rims providing a lip for a possible lid or cover.Residue analysis of the mortar lining of the easternmost Tank 1 yielded tartaric acid indicative of wine production; samples from Tank 2 were not conclusive. 23The two tanks amounted to a total liquid capacity of 1500−1750 l for wine fermentation, or a modest 0.3−0.4hectares of vines. 24No press was present in the excavated areas, and it is possible that the oor of Room D served as a manual treading oor.The opus reticulatum building was set up with the intention of engaging in wine production, even if perhaps not at a scale warranting investment in a press.
Other features that were part of the building's original design similarly speak to an intended function.A channel ran parallel behind the entire length of the northern wing's back wall.It consisted of two low walls with foundations, c. 60 cm apart, employing the same ne white mortar as used in the opus reticulatum walls, and probably collected rainwater runoff.A sturdy pavement with large stones, lled in with ceramic and tile fragments, was in place in front of the northern wing, suggesting a preparation for heavy-duty work and the reception of heavy loads, possibly including the grapes that would have been carted into Room D for pressing and fermentation and the wine that would have exited.
In sum, the future conjured by the opus reticulatum building was programmatic and geared towards a specic function and goal.Yet, at the same time, this futural orientation exceeded functionalism in that the entire structure was over-engineered for its purpose.

Between promise and potentiality
Investment, integrated project design and the spectacle of labour all conspired to create, colonise and tame the future as an abstract space of promise and possibility, a time-space rife with expectation. 25Rebecca Bryant and Daniel Knight coin 'expectation' as a horizon dening the futurality of the future.In other words, when we are able to expect, we can almost 'see' the future spread out in front of us. 26Expectationand especially the expectation of expectationallows one to experience the full thickness of the present: actions in the present bring a sense of future promise.The promise of expectation inhabits an emptied, decontextualised future: 'the future is simply there, an empty space waiting to be lled with our desire …, holding out the promise that it can be what we want it to be'. 27et the initial plan as materialised by the opus reticulatum building at Marzuolo was ultimately unsuccessful.Its rst use-phase was very short-lived indeed, perhaps of the order of a decade or even less (Fig. 2, phase 2).The cylindrical tanks went out of use and were partially lled in. 28It is impossible to tell whether this repurposing entailed a shift in ownership: in other words, whether the building changed hands between its use as agricultural processing site and as craft centre.Perhaps the original owner died, and the estate was inherited or put up for sale. 29As will be discussed below, there is circumstantial evidence that hints at some form of leasing for the second use-phase, which would have meant the (old or new) owner opted to switch to a more risk-averse mode of exploitation. 30Whatever the exact scenario, abandonment of wine production and possibly also of the associated vineyards, a large-scale, long-term investmentand repurposing of the structure meant failure for the original owner/investor.
It is precisely because the project of the opus reticulatum building and its wine production had been so carefully designed and so meticulously executed that its lack of success can be labelled 'failure'.Failure presupposes the presence of and condence in the future as a 'structurally occupied space', sensu Brent Shaw in these pages, 31 or what Bryant and Knight call a 'horizon of expectation'. 32Failure happens when something falls short of someone's expectations.The futural quality of expectation resides just as much in the fact that its substance can indeed be expected to happen as in the possibility, always present, that it may not materialise.Bryant and Knight describe this as the tension between promise and potentiality, which are both qualities or avours of expectation. 33This tension is where the possibility of failure resides.In other words, expectationthe orientation that sees the future as a horizon, empty and decontextualisedis a prerequisite of failure.The possibility of failure, then, is a privilege that comes with the ability to imagine the future and its promise.
Whether the wine-making business at Marzuolo lasted a few years, a decade, or two decades, its demise was swift.No attempt was made at maintaining it, improving processes in some way, or changing the product or its marketing.Instead, activities inside the building moved away from wine-making altogether.The associated infrastructure was completely abandoned, without regard for the care, energy and labour that had gone into its establishment.Whether the building itself, and perhaps any land or resources that came with it, were sold, inherited, or repurposed, change was sweeping.

Modications and dwelling
As the wine production in Room D came to a sudden halt, Room C and the portico area were converted into a blacksmithing workshop.A domestic assemblage including cooking wares and faunal remains indicates that a blacksmith took up residence inside Room C, possibly accompanied by a family or dependents.But most exceptional was the nd of a complete assemblage of tools for blacksmithing and associated woodworking, objects in the process of repair or working, and recyclables.
The toolset encompasses a blacksmith's essential instruments. 34This includes at least two hammers, a large striking hammer with a at surface, and a much smaller typical smithing hammer.The large hammer is so heavy as to require both hands for holding it, especially when in active use, which implies a team of at least two people: a master-artisan holding the object to be worked and indicating the place to be hit with a smaller smithing hammer, and an assistant or aid striking the blows.Several sizes and types of tongs, used to hold objects while forging or hammering, were also found inside the workshop, indexing a broad and versatile task-and skill-set.Objects worked or repaired range from small (e.g.nails) to large (e.g. a ploughshare).An offcut from an iron bar points to the import of fresh iron, but there is also evidence suggestive of recycling.A woodworking toolsetincluding saws of various types and a drawing knife or billhookindicates that complementary woodworking took place in the same workshop, to t handles to tools, shares to ploughs, etc.While the scale of output is difcult to estimate without representative waste deposits, the overall picture is of a workshop comprising c. two to four people, including at least one permanent, specialist blacksmith, and servicing the local rural economy and its need for new products as well as repairs.
The blacksmith modied the space (Fig. 2, phase 2).They used a drain on top of the pavement of large-scale stones that was left in its original position in the portico in front of Room C. Inside Room C, the blacksmith cut pits, one in the back of the room lled with a deposit including large ceramic fragments and a stone, and another, wider one in the west of the room, lled with an assemblage of ceramics, animal bones, glass and slag, possibly the remnants of an interim cleaning up of the workshop oor.
Several workstations combined hearths and anvils or hammering supports (Fig. 6).The archaeological footprint of the latter was traced by mapping the concentrations of hammerscale, the small, ferromagnetic lamellar particles that escape during the hammering of hot iron.Smithing hearths can be quite rudimentary, consisting of no more than a at area, sometimes raised or slightly sunken, with clay walls or another low barrier to shield the blacksmith from the heat.At Marzuolo, the footprints of ve hearths were found inside the workshop, three of which were associated with hammerscale concentrations.Judging from in situ plano-convex slag that collects at the bottom of a smithing hearth, and a lining of tile and stones, two hearths had been in operation recently.The various hearths could represent multiple smiths working simultaneously, different workstations for different objects and/or the repeated abandonment and setting up of workstations.Several quenching basins consisting of broken amphorae are likely to have contained water for cooling off or hardening tools and objects.
The portico area was divided and shielded by the erection of two wattle-and-daub screens, which collapsed sideways (Fig. 3). 35The imprints on the remaining earth chunks show segments of a supporting wooden framework of at beams and of the tapestry of twigs that lled in the framework and on which the earthen mix was applied (Fig. 7).Traces of white plaster survive on some of the fragments.The walls were installed at either end of the portico in front of Room C and seem to have functioned to shield activities elsewhere in the complex from the smoke and other nuisances stemming from blacksmithing, while possibly also adjusting the airow in proximity to the smithing hearths.In addition, the screens could have served to claim space by dividing up the portico area and reshaping circulation patterns.36The blacksmith made the space their own and looked forward, creating a thriving workshop servicing the local community.But while the conversion relied on skill, it did not require radical structural interventions or large investment of labour or resources.Where building in opus reticulatum aspired to be building for eternity, the features employed by the Marzuolo blacksmith embodied a different futural orientation.Vitruvius writes that wattle-and-daub walls are prone to cracks in their plaster cover, 37 and excavations in London and Verulamium in Roman Britain have shown that timber and wattle-and-daub walls were frequently rebuilt, as often as every two years, and generally lasted between two and ten years. 38Preservation and duration of wattle-and-daub walls is highly context dependent, varying with (among other variables) humidity, moisture, technique and use; but while they could, in some cases, last for a lengthy period, they could just as well need repeated maintenance.Several drains in the north-western area of the site attest that humidity in particular was a recurring concern at Marzuolo, probably compromising the longevity of wattle-and-daub.Masonry or concrete structures could transform into heritable assets because their material properties lasted for the long term, thereby opening a multi-generational horizon of expectation. 39attle-and-daub walls, by contrast, harboured their own cyclicity, as did clay hearths and movable workstations.Through such shorter-term (or open-term) features, the blacksmith was not only looking forward to the success of the workshop, but also attending to the uncertainties of the near future.They were, in other words, anticipating.
While expectation looks ahead to a distant horizon and the abstract space of possibilities and promise it describes, anticipation is already creating, intervening in and grappling with that future, which can be here any minute now. 40Anticipation is a self-conscious preparation for, or reaction to, a future that is not yet actualised but already experienced as real.Anticipating happens when the future cannot be kept at bay: '[t]he future is always possible now, at this moment, viscerally present in the act of anticipation'. 41Where the  (Guyer 2007; 2019). 41Bryant and Knight 2019: 35.While anticipating prepares for a future that is already at the doorstep, it can also be seen as an attempt to delay that futureto keep it on the doorstep just a little longer.
expected future is at once abstract and distant, creating space for exploration and planning, the anticipated future presses down on the present, demanding specic responses.

Pre-empting failure
The anticipating to which clay hearths, wattle-and-daub walls and temporary workstations compelled its users at Marzuolo nds its clearest illustration in the workshop's end.Indeed, the very reason why the wattle-and-daub screens at Marzuolo were preserved is because they were accidentally baked in a re in the mid-rst century A.D. that destroyed the entire blacksmithing workshop and much of the portico area (Fig. 3).What should give us pause is not that the re happened (after all, blacksmithing was a pyrotechnic craft) or that the wattle-and-daub screens were unable to withstand it (Vitruvius's mockery of this material as a re hazard has some basis), 42 but the response to the re and its destruction.Instead of rebuilding, repurposing, or at least salvaging, this particular area of the site remained untouched, despite continued activity in other zones.This lack of salvaging attempts is all the more unusual since the destruction levels buried the assemblage of blacksmithing and woodworking tools described above.Indeed, it is most exceptional to nd a complete set of blacksmithing tools in situ, in their workshop context. 43Most archaeological nds of Roman-period tools occur instead in deliberate deposits or in secondary contexts.A dump in an abandoned silo at the site of Ruscino (France) contained a varied set of agricultural implements. 44An assemblage of ceramic pots and iron tools stored in a cave or cellar at the rural site of Côté de Contrexéville (France) was sealed by re debris. 45Among a large number of smithing tools retrieved at the vicus of Lousonna (Switzerland), none appeared in workshop contexts, and some might have been deliberately stowed away. 46Ironworking tools also frequently show up in hoards, whether their purpose was functional or ritual. 47ne possible explanation for why the tools were not retrieved and why the deposits in Room C were left untouched 48 could be found in the symbolic connotations of blacksmithing.A rich ethnography of blacksmithing and blacksmiths has detailed how the making and working of iron conjures a recurrent imaginary of transformation and regeneration. 49The state transformations of worked ironbetween uidity and soliditythe skilled management of the re and the embodied and closely guarded knowledge of reading immaterial and intangible signs such as the colour of heated iron add up to making the blacksmith at once a venerated gure of authority and a marginalised, feared character.This ambivalence extends to the ironworker's tools: their placement in graves, hoards and structured deposits might indicateat least in parta process of purication, neutralisation, or curation of these objects' potent associations. 50If the Marzuolo blacksmith accidentally set the workshop on re, then both the context and its assemblageincluding, and perhaps especially, the toolsetmight have been considered tainted and to be left alone.
That blacksmithing at Marzuolo might have occupied a similarly ambivalent social role, as is attested in other times and places, is plausible, albeit not testable.But to ascribe the lack of retrieval of the Marzuolo toolset to symbolic or ritual preoccupations is to pair such other-than-rational behaviour with the subaltern.As an explanatory strategy, it locates the subalternin this case the Marzuolo blacksmithin an ahistorical timespace, bereft of the historicity and futurity that make them human. 51What could an alternative framework look like?
The tools left behind at Marzuolo were valuable for their iron contentthe large hammer alone contained 4 kg of iron that could have been reforged. 52Cato recommends selling off 'old tools' ( ferramenta vetera) when taking over a farm, indicating that tools were part of the inventory and would be acquired and sold as the owner saw t. 53But for a craftsperson, tools also materialise a form of embodied knowledge.Craftspeople often make their own tools, or customise them to their body, hands, movements, as well as to the task at hand. 54More than extensions of body and skill, tools come to stand in a metonymic relation with their craftsperson: tool and craftsperson become indistinguishable in terms of both identity and disposition. 55articipatory ethnography of blacksmithing in the United States of the twentieth century by Charles M. Keller paid close attention to the importance of the workshop setting and layout, including the arrangement of tools. 56Of note is how tools are, once more, more than a simple extension of one's skill; instead, they constitute this skill.Tools carry knowledge by activating a rich imagery: One knows about one's tools, where they are kept, how they are shaped, the past productions in which they have been utilized to effective or ineffective ends, from whom they were purchased or how they were made.This imagery is rich and detailed, its comprehensive quality often seemingly excessive, yet the information represented allows the smith to search his stock of knowledge with respect to the task at hand. 57ing presented with a task, especially a new or challenging one, the blacksmith will glance over their toolset to map out possibilities and forge new connections.The tools set in motion problem-solving networks at the interface between the mental and the somatic.It is thus easy to understand how loss of one's trusted tools equals loss of skill, insofar as the craftsperson's body knows and remembers through its seeing and handling of the tool, and how the singularity of both the setting and the toolset are of crucial importance.
The Marzuolo toolset thus represented a signicant capital investment, but was also a core constituent of someone's human capital.Why, then, was the toolset left behind after a re destroyed the workshop, even when other areas of the site continued to be occupied? 58Examples at Autun and Blesseyboth in Roman Gaulshow that even when blacksmithing workshops were destroyed by re, craftspeople tended to return to retrieve their tools. 59At least in these cases, any concerns with the symbolic domain of reproduction and transformation did not prevent people from re-opening the contexts and recuperating its assemblages.Why, then, did not one of the presumably two to eight people (blacksmith, apprentices or aids, and family) associated with the Marzuolo workshop dig through the collapse deposits?Why did none of the inhabitants of Marzuolo aware of the existence of the workshop attempt to salvage or claim the tools?
The explanation I propose is that the exceptional in situ presence of a complete toolset in the Marzuolo blacksmithing workshop hints at a misalignment between the ownership of capital (in this case the workshop and the tools) and human capital (the embodied skills of the blacksmith or blacksmiths).Perhaps the master-blacksmith succumbed in the re, a scenario that is impossible to rule out, although no human bone material was found in relation to the otherwise well-preserved workshop space.But even then, anyone involved in the blacksmithing at Marzuolo would have interacted intimately with the toolstweaking, adjusting and slowly aligning mind, body and tooland would have understood the importance of retrieving them.And any of the other inhabitants of Marzuolo without the buffer of an ample asset portfolio, aware of the presence of the now-buried workshop, would have salvaged at least the bigger iron objects for their recycling and resale value.That instead the collapse layers and the tools they shielded remained untouched suggests that whoever owned them denied access, for whatever reason.It therefore seems unlikely that the blacksmith was the owner either of the workshop or of the toolset that was left behind, given how their skill development would have taken a signicant step backwards without it.
The possibility that the owner of the tools and building installed a blacksmithing workshop operated by slave labour seems unlikely on similar grounds.Indeed, in the case of a slave-artisan, capital and human capital coalesce in the skills of the slave, and optimal use of this investment would once again have relied on the tools. 60If a slave-owner decided to cease activity at Marzuolo and reinstall a blacksmithing workshop elsewhere, the tools would have been worth retrieving.If they relinquished blacksmithing altogether, the slave(s)-blacksmith(s) would have fetched a higher resale price as skilled specialists with tools.
The most likely scenario, although ultimately impossible to prove, sees a blacksmith (free or freed) lease the workshop space and associated tools from the building's owner, who may or may not have been the same person who designed and funded the construction of the opus reticulatum building in the rst instance, and who may or may not have died in the re. 61After the disastrous re this owner (or their heir) was, perhaps once again, privileged enough to swallow another failure and simply abandon the investment altogether.The blacksmith and their probable aide(s) insteadif still alivehad no choice but to search for other opportunities, albeit without their tools and thus with a signicant dent in their skillset.
The schism between capital and human capital would have curbed innovation and constrained the possibility of Schumpeterian growth, at least in the Roman rural economy.The blacksmith was forced to have recourse to versatility: they might have moved in an attempt to nd employment as a blacksmith elsewhere, or they might have ventured into other activities, relying on complementary skills, wage labour, or family ties.One can assume that the skillset of a blacksmitheven, or perhaps especially, one of the 'generalists' who produced such things as nails and made various ad hoc repairs, as did the Marzuolo blacksmithwas sought after in various contexts: rural estates, minor centres, or even cities such as Roselle or Chiusi.The blacksmith could thus have found a new leasing contract or some other arrangement allowing them to continue practising their craft and earning a living.Even this possibility, however, would have come at signicant cost: establishing new supply networks, new customer relations, new needs, a new workshop space and, indeed, a new toolset.Perhaps the largest cost, however, was the human toll of forced change, possibly including displacement, uprooting one's life, friendships, families, landscapes.In this scenario, resilience was not a choice for the blacksmith. 62he Marzuolo blacksmith, however, had anticipated this outcome.However detrimental and costly its impact on their life, the destruction of the workshop by re did not represent failure for them, precisely because they never really entertained the promise of expectation.Put differently, for the Marzuolo blacksmith failure was not futural; instead, it was always already 'there' and, therefore, anticipated.They did not project plans into an abstract futurity, but were constantly responding to, and preparing for, specic concerns and demands.We cannot tell what the duration of a leasing-contract for the blacksmithing workshop might have been: the Roman jurists tend to accept as standard a short contract of the order of ve years. 63Regardless of any length of time stipulated by a locatio-conductio contract, the future as a horizon of expectation, a distant, abstract space of possibility, would not have existed for the Marzuolo blacksmith in this scenario.Their futural orientation was one based on imminence, one that was always already in responsive mode, even before a response was needed.Where expectation can lead to failure, and indeed lends its futural quality to the potential of failure, anticipation inverts the temporal and causal arrow: it pre-empts failure in order to deny its potential.Here the wattle-and-daub screens, the clay hearths and the other mutable features come in again, as both product and tool of anticipation.The short term of Roman leasing-contracts parallels the lifecycle of wattle-and-daub walls as reconstructed in excavations in Roman Britain, of the order of two to ten years. 64or the Marzuolo blacksmith, building in wattle-and-daub meant building in anticipatory mode.It meant preparing for a future that would entail change, uncertainty, a future that was right at the doorstep, right around the corner.A small apotropaic stone phallus afxed to the workshop's front materialises this anticipatory impulse: warding off evil before it manifests itself (here and now, not in the future).With the future already on the doorstep, it becomes impossible to make the leap 'from more proximate needs to more distant aspirational worlds'; instead, one remains caught up in the pressing demands of proximate needs. 65Much like the other modications and the smithing hearths themselves, the wattle-and-daub screens were relatively low cost and easily reversible.They minimised the loss of labour and capital tied to a facility that would have had to be left behind at some pointwhen the lease ran out, at the owner's whimwhile maximising the return in the form of a functional space, skilfully laid out and adapted to its new role as a forge.
In laying the groundwork for the future, anticipation draws on past experience. 66Such experience offers a repertoire of actions and modalities with which to respond to the perceived future-at-the-doorstep.This should not be seen solely as a conscious and strategic process, nor as an automatic recycling of tradition.Much anticipatory action takes place subconsciously, in a choreography of calibration between experience and need.As an anticipatory tool, wattle-and-daub was deeply anchored in local and regional practices at Marzuolo.Its constituent materials were widely available and 62 Diprose 2014. 63On short-term lease contracts as the norm for Roman jurists, see Kehoe 1997: 155-6; 165; 201-7.Cf.Columella, Rust.1.7.1−2.See also Foxhall 1990: 99 and 101.On long-term lease contracts and bargaining power of tenants, see Kehoe 1997: 183-209. 64 Perring 2002: 81;Perring and Roskams 1991: 81. 65 Appadurai 2013: 213. 66 Bryant and Knight 2019: 28, 69.intimately known: earth as a bonding agent; cuttings or twigs probably from local shrubs which could have doubled as kindling for craft and domestic ring processes; wooden beamscut locally or on the slopes of Monte Amiataalso used in construction of roofs and even for complementary woodworking inside the blacksmithing workshop. 67oreover, building in wattle-and-daub had a long history in the wider area of Marzuolo.This local history was selectively activated at Marzuolo, for the specic anticipatory action of a blacksmith (re)furbishing a workshop space that could not be expected to remain theirs into the future.

V FAILURE AND ITS LIMITS
The Roman world was unequal: Roman elites extracted comparably more of the available wealth than their classical Athenian counterparts. 68What this meant for the lives of people on various rungs of the steep Roman socio-economic ladder is less clear, however, despite some pioneering quantitative modelling. 69Several recent studies describe the gap that separates GDP, Gini coefcients and other numerical abstracts from lived experience: 70 botanical remains from a Herculaneum sewer demonstrate that non-elites at Herculaneum consumed a varied diet, including sh and even pepper; 71 the size and layout of Pompeian houses refuse to be divided neatly into rich and poor; 72 and Pompeian grafti documenting expenditures on foodstuffs record the highly individual consumption choices of various (presumably non-elite) people and households. 73t is impossible to quantify the absolute level of material wealth of the Marzuolo blacksmith and to gage the number of mouths fed by the workshop activities.The domestic assemblage of cooking implements and dinner wares does not stand out as either particularly rich or deprived.Compared to some of the farmers, labourers and traders that would have circulated in and through the site of Marzuolo, the blacksmith had the advantage of possessing a specialised skillset.The owner-investor of the opus reticulatum building, in contrast, clearly occupied a very different socio-economic register, able to nance a large capital investment.An examination of the half-century of start-stop development in Marzuolo's north-western corner can texture this stark socio-economic inequality.Paradoxically, the privilege of the owner-investor was constituted by failure: the ability to fail was dependent on the ability to conjure a future horizon of expectation in which things were possible, and therefore could also fail.Conversely, what marked the blacksmith as disadvantaged and their position as precarious was already to anticipate and in some way experience loss as real even if it was not (yet) actual.The subjectivity of precarity was thus shaped by a pre-emptive 'tyranny of tomorrow' 74 which ruled out failure, or rather, displaced failure from the future to the now.
Shaw's conclusion that the Romans had no sense of an abstract future as a 'structurally occupied space' 75 hides profound inequalities: some privileged people could indeed embrace expectation, and thus the possibility of failure, while (presumably many) others were trapped in anticipatory mode.These different futurities had affective resonances. 76ondence is needed to expect, to let the future arrive and unfold: it considers, and parks, the potential that things might not turn out as planned.Failure happens when the potentiality of expectation overtakes its promise.Conversely, the inability to expect, and therefore the inability to fail, causes stress.A sense of control emerges as a key variable of stress from sociological research on relative poverty, 77 and indeed this is one way in which tenancy agreements and similar forms of dependency in the Roman world served to lock in part of the population.This stress is not acuteit is different from any stress which Marzuolo's opus reticulatum investor might have felt when their investment in wine production failed to pay off as expected, and it is different from the sudden rupture of crisis. 78The stress of precarity, instead, is chronic: it is the constant background hum of anticipation. 79he owner-investor at Marzuolo could probably swallow failure.Yet even they did not abide by the narratives of growth-through-failure that have become normative today.When the wine business failed, no attempt was made to learn from mistakes, to re-invest, tweak or otherwise improve.Instead, the undertaking was abandoned altogether.Whether or not the property changed ownership at that stage, a more risk-averse strategy was adopted, if the above reconstruction of leased workshop spaces inside the opus reticulatum building is correct.Even for the owner-investor of the opus reticulatum building, then, the privilege of failure had real limits: the horizon of expectation crafted after the wine business's failure was radically reoriented, and decidedly more cautious and restricted in nature.
Endings and abandonments abound in the archaeological record, much of which stems from pre-industrial worlds coloured by risk and uncertainty.By texturing experiences of privilege and precarity, the story of the vintner and the blacksmith at Marzuolo opens new methodological and interpretive avenues for the archaeological study of such pauses, ruptures and ssures.These phenomena have long been read mechanically, through the lens of post-depositional processes at the micro-scale 80 or of grand causal movers at the macro-scale. 81More recent studies emphasise the work of place-making, between phenomenology, mnemonics and politics. 82But the unit of analysis tends to be either an undifferentiated human agent or an equally undifferentiated group.This article shows the value of interrogating the very nature of such endings and not just their causes or consequences.Both the arrested wine production and the abandoned blacksmithing tools represent endings, but while one can be labelled a failure, the other was an anticipated reorientation, where the human and economic cost lay as much in the stress of constantly preparing for the inevitable as in its consequences.
Finally, the case of the vintner and the blacksmith at Marzuolo highlights the need to write micro-histories of Roman social and economic dynamics.At a coarser scale of analysis, the data presented here might have appeared as a typical pattern of mixed agricultural and craft activities, characterised by gradual developments and slow changes of orientation.Instead, on closer examination, the smooth curves of macro-economic models and the articial continuity of survey evidence hide a story of repeated boom and bust. 83Sustained excavation and high-resolution analyses at small-scale sites such as Marzuolo are needed to texture the social and economic ramications of such processes of boom and bust.Repeated endings, some of them failures, may well have been at the very core of the fabric of the Roman world, rather than merely constituting the sequel of its fall.

67
Wattle in Roman Britain often used '[h]azel and birch from short rotation coppicing': Perring 2002: 83.The most dominant tree cover in Marzuolo's vicinity was mixed oak wood; part of the landscape would have been open with shrubs and pasture: Rattighieri et al. 2013. 68Bagnall 1992; Scheidel 2016; Taylor 2017: 89. 69Scheidel and Friesen 2009. 70Stiglitz et al. 2010. 71Rowan 2016. 72Flohr 2016. 73Bowes 2021b. 74After Jérôme Bindé's concept of the 'tyranny of emergency' of the last decades: Bindé 2000, as cited in Appadurai 2013: 161.The tyranny of emergency leaves no space for anticipation and pre-emptive response; its future is not on the doorstep but already past.