Reading and Interpreting Household Inventories: A Methodological Essay

The Documentary Archaeology of Late Medieval Europe

Reading and Interpreting Household Inventories: A Methodological Essay
D. Smail
Version 1 (19 February 2018)
(download PDF)

Overview: As the DALME collection grows, it will include objects listed in many different kinds of sources, including legacies made in household or probate inventories, testaments, dowry items, and lists of objects seized during the course of debt recovery. This essay addresses the methodological challenges associated with using partial or total inventories of all kinds, with a focus on complete household inventories, one of the most important type of source for the study of medieval European material culture.


As is the case with every historical document, household inventories do not reveal the past in some simple and uncomplicated way. The shafts of light that they appear to shed on the material culture of past societies need to be interpreted carefully and in light of the context of their own production. The context was composed of several facets, for the making of any given inventory was strictly governed by legal processes. In addition, the contents of inventories were subtly shaped by cultural or epistemological norms and by the competence of those responsible for redacting or translating a given inventory. Legal norms and customs varied from one region of Europe to the next and across time as well. Equally variable were the ways in which people from different regions valued things or how they “knew” them or chose to “see” them. This essay reviews the arguments that have been proposed in the literature regarding the caution that one must exercise before interpreting any single inventory or collection of inventories.

At heart, every inventory or list of objects aims to describe an ensemble of objects that is, or may soon be, in motion. In the case of post-mortem inventories, the most common type from the later Middle Ages, the objects in question were represented as belonging to a single individual, the decedent. We catch a glimpse of them at the moment where they are being transferred into the possession of the heir or heirs, even if many transfers didn’t necessarily involve any physical move. The post-mortem inventory is an example of what we may call a total inventory, a type that purports to provide a complete accounting of the possessions of a person or a corporation. There are other kinds of total inventories. Some were produced by law courts during proceedings for insolvency; in such instances, the goods of insolvent debtors were on their way to creditors. At certain moments during criminal inquest proceedings, the court might impound the goods of defendants, especially those who were contumacious or who had been banished for political crimes. The goods were inventoried because the court had the hope and expectation of soon being able to transfer the abandoned estate to the treasury.

In addition to total inventories, there are several kinds of partial inventories, such as the lists of objects found in records of auctions or debt collection as well as dowry acts and last wills and testaments. In the case of prosecutions for theft, thieves were sometimes invited to provide a reckoning of all the things they had stolen, including the names of victims and the names of the unwitting individuals to whom the thief had subsequently sold the things. Once again, we see the goods as they were on the move somewhere. Finally, we should consider a different kind of inventory, namely, those that listed the items belonging to churches, monastic or episcopal treasuries, royal wardrobes, and the like. These too are partial inventories, in the sense that they were designed to record only a certain set of possessions belonging to a corporation. The objects itemized in this way were less obviously on the move, although even here it is possible to understand that the objects were moving from the past into the present, and that the point of making the inventory was to anticipate a move into the future.

For their workings, total inventories, and to a certain degree partial inventories as well, depend upon a rather audacious claim, namely, that it was possible to imagine that a given ensemble of things “belonged” to a single (typically male) individual or to a corporate entity. The audacity here arises from the fact that “belonging” can be a fully legal relationship (our house and its contents belongs to me and my spouse) or it can just be a habit of thought or a kind of classification (the bicycle in the garage belongs to our young daughter). According to medieval European law, everything that pertained to a family belonged, at least in theory, to the head of the family, usually although not always a man. At the same time, however, everyone seems to have understood that the clothes that a wife “wore on her back,” a phrase that appears from time to time, belonged to her.[1] In addition, everything that came to her either in her dowry or, later, through legacies, inheritances, and profits from running a business, nominally belonged to her husband, but at the same time more or less belonged to her as well. Medieval law solved this by treating every wife (or widow) as the creditor of her husband’s estate for a certain sum of money. That sum was the total of the cash that came to her by way of dowry, legacy, or other means, and the estimated value of the tangible things that came to her via similar pathways. Her husband might choose to sell off a plot of land listed in her dowry or pawn a necklace given to her by an aunt, but she was still owed the value of the items in question and could ask to be compensated for that value upon her husband’s death or bankruptcy. The supposedly clear indications of ownership assumed by inventories in reality masked the existence of these complex rights.

In addition to this, inventories relied on another claim, namely, that it was possible to determine which object belonged to whom in any given household. The problem is that medieval households were sometimes complex, including not only extended family members but also servants, slaves, hired hands, and apprentices. Here, the problem is that ordinary household objects never carried signs or seals indicating their owners. Since few objects were purchased ready-made at a store, moreover, there were no receipts and little in the way of a paper trail that could be used to determine the ownership of objects. With land it was different. By the later Middle Ages, at least in Western Europe, the growing presence of legal or governmental registries made it possible for contemporaries to identify the ownership of parcels of real estate by means of some written document, such as a notarial contract or a registry of deeds. But with some notable exceptions, including the objects listed in a woman’s dowry, this was not the case with ordinary household objects. This raises an interesting question of how the very concept of ownership developed in later medieval Europe and whether the growing presence of the law, which required lists of objects for so many of its operations, played a role in changing how people thought about ownership. This topic has been at least partially addressed by Samuel Kline Cohn, but more work needs to be done on it.[2]

Although the fiction of ownership was necessary for the operations of the law, it seems as if everyone understood its limitations. This has consequences for what we see and do not see in household inventories. By way of example, it is not unlikely that the everyday costume worn by members of the family of a decedent were normally not included in total inventories. Shoes or clogs, to consider only the most obvious case in the DALME collection, are astonishingly absent from the inventories from Marseille. A possible explanation for this is that members of the household owned only one pair of shoes at a time, and, by wearing them constantly, asserted a kind of de facto ownership that trumped the demands of the law.

As a result of this situation, a surviving household inventory from medieval Europe may exclude certain kinds of objects that technically belonged to the principal party, and at the same time it may include objects that did not belong to the principal party. The act of considering the presence and absence of objects invites us to draw up two parallel columns where we list the inventory’s exclusions and inclusions. (This is necessarily an act of imagination since, with rare exceptions, we cannot derive any conclusions from the inventory itself.) Once we have our two columns set up, we can then draw two rows that cut across the columns, and label the rows according to the intention of the individual responsible for redacting the inventory.[3] What results from this exercise are the following four fields: 1a) objects-that-belong that were inadvertently excluded; 1b) objects-that-belong that were deliberately excluded; 2a) objects-that-don’t-belong that were inadvertently included; 2b) objects-that-don’t-belong that were deliberately included. To return to the observation made at the outset of these remarks, we need to be aware of the legal, epistemological, and ontological contexts of inventory production if we hope to understand, and make due allowance for, these potential exclusions and inclusions. Let us deal with each of the four fields in turn.

Inadvertent exclusions[4]

 Exclusions for reasons of thingness. Anything listed in an inventory had to be a thing that was sufficiently distinct from some other thing. The most notable examples of objects that were on the wrong side of the minimum threshold of thingness consist of elements of the built environment or “fixed furnishings,” such as fireplaces and mantelpieces, tiled floors, frescoes, staircases, and window treatments.[5] Similarly, inventories do not seem to have systematically included certain components of other objects that could, at least in theory, be understand as autonomous. To take an example, we see silver buttons from time to time on garments.

Item unam cotam domine albam cum 40ta condulis parvis argenti in manicis
Next, a woman’s cotte, white, with forty small silver buttons on the sleeves
(Object ID 13257)

It is not clear why things such as buttons were noted; perhaps it was a function of their value, but it could have resulted from the desire to identify the garment in question. There is no guarantee, however, that every redactor “saw” these buttons as being separate objects and noted them down systematically. Inexpensive metal dress ornaments made of alloys of copper, lead, and tin show up frequently in archaeological sites. In inventories, we see them from time to time as supplies stored in boxes or sachets. Clearly, they had lives that were distinct from the clothing onto which they were sewn or riveted. As Marie Casset has pointed out, descriptions of clothing in inventories rarely includes these dress accessories.[6]

Exclusions for reasons of ownability. To appear in an inventory, a given item had to be capable of being owned. Wine, to take an example, was ownable, but water, it seems, was not: inventories rarely or never include lines such as “a barrel full of water.” More importantly, the redactors of inventoires seem to have felt that dogs, cats, and pet birds (such as parrots) did not merit inclusion in inventories. It is possible that pets were not considered ownable because they were seen as entities without value (unlike chickens, donkeys, horses, and other such animals, which did have value and show up frequently). But it is also possible that pets were somehow thought to own themselves and were not included for that reason. We can sometimes determine the presence of pet-like animals in houses indirectly through the material culturel of petdom.

Item unam tristegam psitaci
Next, a parrot’s cage
(Object ID 11240)

But from this record alone, we cannot make any conclusions about whether there was a parrot in the cage. Equally interesting although perhaps less easily explained is the exclusion of musical instruments, even though instruments are found from time to time in the archaeological record. It is conceivable that flutes, lutes, percussion instruments, and other instruments were, like pets, somehow thought to belong to the house itself or to the family in general.[7] Finally, as noted earlier, total inventories seem to have excluded the garments that individuals were wearing at the time of the redaction of the inventory; in this case, the ownership of the items in question seems to have been deemed to be ambiguous.

Exclusions determined by monetary value. In theory, inventories only took note of things that possessed a certain minimum degree of value.[8] Amulets, charms, game pieces, and other trinkets appear frequently in archaeological sites but generally escaped the inventory process; presumably because they were deemed to be valueless.[9] The same holds for simple tools such as sickles.[10] It is common to discover a line item in an inventory where small objects of a similar type were all bundled together under a generic label such as in this example.

Item unum candelabrum, unum crusibolum sive calelh et de ferramenta minuta
Next, a candelabrum, an oil lamp or “calelh,” and some small ironwares
(Object ID 15196)

Individually, it seems, the small bits of iron tools or fittings were worth almost nothing, and only had value as a small ensemble. Here, an important question concerns the probability that the threshold of value or importance at which a given object became visible seems to have varied from region to region and from person to person. By way of example, Garcia and Ferrand have observed that items made of earthenware or ceramic are identified in Provençal inventories but are found only rarely elsewhere.[11] The common assumption, as they suggest, is that ceramics were simply too inexpensive to be worth enumerating, and that for some reason, the Provençaux felt differently. There are other possibilities, of course, including the slight possibility that ceramics were simply uncommon outside of Provence.

Exclusions of absent things. Items that belonged to the decedent or owner but were not found in the house at the time of the redaction of the inventory were in theory more likely to be accidentally omitted. The most common of these consists of objects belonging to the principal party that were being lent to a friend or acting as security for a loan and therefore were out in pledge with a local pawnbroker or a friend or neighbor.

Item i descargadoyra quam habet magister Ludovicus Martini faber
Next, a vat which is being held by master Lois Martin, the smith
(Object ID 10374)

Very occasionally, a marginal notice might indicate that such an item had been discovered and added after the fact. In Florence, newly discovered rights or objects could be added to an inventory months after the fact simply by appearing before an notary and declaring the addition.[12] In keeping with this practice, the preambles of inventories sometimes indicate that the redactors were concerned about the perception that missing things might have been deliberately omitted, a point that will be discussed further below.

Inadvertent inclusions

Inclusions of pledges. In the same way that pledges currently outside the house might be missed by the redactor of an inventory, pledges brought into the house in exchange for a loan did not necessarily carry marks of their “real” owner and might therefore be included inadvertently. Here, it is important to be aware that the lending on pledges was widespread in later medieval Europe and individuals did not have to be professional pawnbrokers to lend money to friends, neighbors, patrons, and clients.

Inclusions of items belonging to other members of the household. Items belonging to other members of the household could be inadvertently swept up into the general inventory.

Deliberate exclusions

 Exclusions to evade creditors. As Thomas Kuehn has argued forcefully and convincingly, heirs and their guardians had a strong interest in hiding the assets of an estate so as to minimize the estate’s liability to the claims of creditors.[13] Throughout the Italian peninsula in the later Middle Ages and early Renaisance, statutes and other legislative actions reveal a profound suspicion of inventories and the motives of those who made them. Kuehn has argued on the basis of the ubiquity of this legislation that the practice using inventories to defraud creditors must have been common. By the sixteenth century, when the certain jurisdictions began to impose death taxes of various kinds, the state itself became one of the creditors who could be defrauded by this clever use of inventories.

The possibility for fraud arose from the fact that the making of an inventory was the responsibility of the heirs themselves or their guardians. It is true that the law required the presence of witnesses, including the creditors, at the redaction of the inventory, but even so, it was still possible for the family to deliberately undercount the decedent’s possessions and thus minimize the assets that could be seized by those same creditors. Given what has been said above about the uncertainty of ownership—it was rarely the case that ownership of any object could be ascertained by written proof—few creditors would have been in a position to counter the family’s claim that this fancy silver goblet or that well-appointed bed did not actually belong to the decedent. With real estate it was a different matter. But here, it is exceedingly important to be aware that a substantial proportion of a family’s assets might be tied up in movable goods rather than real estate. In addition, there were several ways to assign the ownership of real estate to someone other than the head of the family.

In part to minimize the extent of the potential fraud, legislation in Florence and elsewhere in Italy granted benefit of inventory only to heirs who were minors, on the theory that very young children lacking full knowledge of an estate’s assets and debts ought not be required to accept the potential burden. Curiously, this opened the door to a potentially fraudulent practice whereby an adult heir, who was not given benefit of inventory, would repudiate an inheritance so that the inheritance would devolve instead to a minor heir. As Kuehn has shown, this would allow the family to draw up an inventory which in turn could be used to defraud creditors.

As noted above, late medieval post-mortem inventories in many jurisdictions were typically redacted by a member of the family and only later copied out by a notary, a procedure that made it easy for the family to remove valuable objects. Later, by the sixteenth century and beyond, public notaries or other legal officials were charged with the task of redacting an inventory. This is especially the case with the English probate process. Even in this case, however, officials necessarily relied on family members to identify family goods, and the family could quietly hide things away or assign ownership to someone else in the family. In the case of total inventories generated during criminal inquest procedures, the defendant’s family also had an interest in hiding away as much as possible.

Exclusions of items owned by others. The redactors of inventories deliberately excluded items that they knew belonged in a fully legal sense to other members of the household such as servants, apprentices, and above all spouses.[14] What this means, in practice, is that one should be careful about assuming that the contents of a given inventory are identical to the contents of a given house. If there are surprising absences—if, for example, a household inventory does not list a bed, a set of iron tools, or a cooking pot—the explanation could lie in the possibility that the objects in question belonged to someone else in the household.

Exclusions of legacies. The items that were given away by testamentary legacy prior to the making of a post-mortem inventory were not included in the inventory. In addition, as Philippe Wolff has pointed out, it is likely that any item or items that had to be sold to pay off one or more debts would also be excluded from the inventory.[15] In practice, what this means is that very special objects, things we might call heirlooms, are likely to be systematically under-represented in inventories.  It was the custom in many regions of later medieval Europe to emancipate a domestic slave upon the death of the slave’s master. For this reason, inventories systematically under-count or ignore entirely the population of enslaved persons.

Deliberate inclusions

 Inclusions to assert ownership. In instances where there was an ongoing dispute about the ownership of a given object, it might pay to slip an object unnoticed into an inventory in the hopes of creating a paper trail that could be subsequently used to assert ownership. In several rare but interesting cases in our collection of inventories, the actual line item includes an acknowledgment of the current state of ambiguity.

Item una tina ad balneum que dicitur esse dicti Blacayroni de Monteolivo
Next, a bathtub which is said to belong to Blacayron de Montoliu
(Object ID 11890)

Since inventories did create a paper trail concerning ownership, it is at least theoretically possible that a given inventory might include some items belonging to another individual who was seeking to hide them from someone else, such as a creditor or the criminal court.

˜Taken together, these observations lead us to formulate the following principle of interpretation:

Principle 1
In no case can we assume that a given total inventory represents
a) the entire and exact contents of the principal party’s possessions, or
b) the entire and exact contents of a given house.

Given this principle, it becomes very important to ask the question of what, exactly, an inventory can teach us. Some of the answers to this question will appear in the section below that deals not with the presence or absence of objects but instead with their attributes. In many cases, however, the arguments we might want to make depend upon the presence or absence of objects. By way of example, devotional objects are relatively uncommon in inventories from the later Middle Ages. It matters a great deal whether these objects were uncommon because they really were uncommon, or whether they were uncommonly listed because the contexts associated with the making of inventories conspired against their inclusion.

To address this issue, we first have to complicate our mental diagram by adding a third dimension which addresses the question of whether the exclusion or inclusion of certain objects was systematic or contingent. By way of example, we know from independent lines of evidence that late medieval houses sometimes or often had fireplaces and tile floors, and because we know this, we also know that elements of the built environment such as these were systematically excluded from inventories. In this case the exclusion was inadvertent, in the sense that it doesn’t seem as if redactors were especially conscious of “excluding” the built environment. Legacies and possessions of other household members were also systematically excluded; in this case, the exclusion was deliberate. This matters because legacies were not made at random; they were especially likely to include beds, clothing, and, as noted above, heirlooms. We should therefore assume that these kinds of things are somewhat undercounted in inventories. Other kinds of objects would have been excluded on a purely contingent basis; in other words, maybe this silver vessel or that bed was falsely represented to the notary as belonging to someone else in the household so as to minimize liability to creditors or avoid taxes. It is hard to imagine inclusions that may have been systematic, but surely there were some contingent inclusions that were either unintentional or intentional.

This rather straight-forward exercise is valuable insofar as it invites us to formulate another principle of interpretation with regard to objects that appear to be missing in inventories:

Principle 2
All arguments about inventories should consider the possibility that a  given type of object
has been systematically excluded from inventories for reasons arising from the context of production.

By way of example, let us imagine that devotional objects are commonly found in inventories from region A and uncommon in inventories from region B. It might be the case that the people living in region A possessed a robust material culture of devotion whereas those in B did not. But it might also be the case that the context of inventory production in region B caused many devotional objects to be deliberately or inadvertently overlooked. For example, perhaps people in region B thought it was sacrilegious to list devotional objects. There is a third possibility, which is that the material culture of devotion in region A centered on objects like paternoster beads that were made of valuable materials such as silver, coral, or amber, whereas the material culture of devotion in region B centered on cheap things made of lead such as medals or badges. In such a case, it may be that the context of inventory production was exactly the same in regions A and B, but that the application of a shared rule—that inventories should list only those objects whose value exceeded a certain threshold—has caused B’s material culture of devotion to vanish from inventories.

The second principle above covers only those objects that were systematically excluded from inventories. What about the objects that are absent or present in a given inventory on a purely random basis? In most cases, we cannot know anything about these possibilities. In practice, this brings us back to the first axiom, which holds that the relationship between a list of objects in a total inventory and the actual belongings of a given individual is inexact. By the same token, we cannot argue with certainty that the objects listed in a total inventory represent the entire contents of a house. No argument should ever be based on the accuracy of a single inventory.

There is, however, an important corollary to this, for unless there is reason to believe that certain kinds of objects are much more likely to be deliberately excluded, it is legitimate to make arguments on the basis of a collection of inventories. As long as the dataset is large enough, the occasional contingent inclusion or exclusion will simply amount to a kind of “noise” that will be washed out by the plenitude of data.

By way of example, let us suppose that a scholar who is working on the revolution in domestic lighting that took place between the eighteenth and nineteenth centuries decides that it is important to understand systems of lighting in the later Middle Ages. Accordingly, this scholar wants to know what kinds of lighting utensils, including oil lamps, candelabra, and torches, were found in poor or middling households, and is also interested in the rooms in which the lighting utensils were found. It is not unlikely that any single inventory has overlooked something, such as a silver candelabrum that is out at the pawnbrokers or a broken ceramic oil lamp. By the same token it is likely that every now and again, a portable candelabrum that was in reality normally used in the dining hall was left in the bedroom and therefore inventoried there. In other words, there is going to be some noise in the data. Across a sample of several hundred inventories, however, the impact of these omissions will be minimized. The only exception might lie in the possibility that certain kinds of lighting, such as silver candelabra (which were valuable) or ceramic lamps (which were cheap) were more likely to have been omitted for some reason; our scholar would have to be attentive to this possibility.

From this, we can frame another principle of interpretation:

Principle 3

As long as we take account of the possibility of systematic exclusions, it is legitimate to
draw conclusions about the presence or absence of things from relatively large datasets.

This principle, refreshingly, is distinct from the first two insofar as it is enabling rather than disabling.

Let us change gears now and consider other uses of inventories. Hitherto, we have been thinking about the kinds of evidence that inventories provide about the objects types found, or sometimes not found, in inventories. We have been paying close attention to the important question of whether the objects listed in the inventory match the principal party’s possession in some exact or one-to-one way. But there is second order of information that can be extracted from inventories. This lies in the words used to describe objects as well as the attributes or descriptive phrases that are assigned to them.

Any given line item in an inventory typically identifies a count, then an object, and then a list of attributes. The count is typically “one” but it can also extend to several or even dozens. The higher counts include sheets, tablecloths, bowls, trenchers, shoe lasts, and lots of other things that can come in multiple units. The object is a noun, typically (when in Latin) in the accusative. The attributes, finally, consist of a set of adjectival phrases that identify various features of the object or objects. There are some variations on this pattern, not least in situations where in a given a line item the redactor has chosen to group a set of objects that are linked or paired in some way. Examples include a mortar and pestle or a bedset identified by a long lists of its components. But let us set aside the complications and consider a typical example of a line item consisting of a 1) count; 2) an object; and 3) a set of attributes.  This comes from an inventory from Marseille in the year 1405.

Item, unum parvum bassinum de cupro fractum modici valoris
Next, a small basin made of copper, broken, of little value
(Object ID 11054)

In this case, the basin has four separate attributes: size, material, condition, and value. On the basis of this and other examples, we have identified twelve attributes that can describe almost every single type of descriptive phrase : Material, Size, Style, Color, Shape, Value, Condition, Additions, Contents, Gender, Function, and History. These attributes can have many facets. In the case of color, for example, the set of facets encompasses all of the color terms, several dozen in all, that are found in the acts. In some cases, such as the attribute Gender, an object can have only one facet, either male or female. In other cases, including color and material, an object can have more than one facet, as in a parti-colored garment or a silver necklace with pearls.

As the reader should expect to know by now, none of the attributes identified in the example can be understood to reflect the “real” basin in some uncomplicated way. Even the meaning of the word “basin” is up for grabs. For the sake of simplicity, let us assume that the redactor of the inventory really did see an object on that day in 1405. What happened next is that a word, presumably in the vernacular, sprang to mind, and this word ended up being rendered, in Latin, as “bassinum.” But what does the word mean? Many objects found today in archaeological stores or museum collections are bowl- or basin-like but the linguistic connection that once tied a given noun to a given object-type has been at least partially lost. By the early modern period, pictorial dictionaries were beginning to emerge, and from this evidence we can make some careful extrapolations back to medieval contexts. But we can never really know for sure.

The point here is that the variable object types that contemporaries chose to group under the umbrella of a single noun, for reasons of linguistic convenience, do not constitute a category that has discrete and easily recognizable edges. We don’t know where exactly where bowls (scutelle) shaded off into basins. It is possible that every bowl-like thing made of copper was more likely to acquire the word “basin” than a bowl-like thing made of wood or ceramic; this is a testable hypothesis. But it is just as likely that size and form determined the point at which contemporary speakers of a language would shift from “bowl” to “basin.” In a number of entries, in fact, we can see redactors who are confronting an object that lies on the border between two noun-categories. They solved the puzzle by the simple expedient of listing both words and joining them with the word “or” (sive, seu, or vel):

Item sex potos stangni tam magnos quam parvos cum duobus
meysarobis sive aygaderiis stagni

Next, six tin pitchers both large and small along with two tin cruets or ewers
(Object ID 9676)

Similarly, from time to time a redactor or notary seems to have realized that the word chosen was not exact, and scratched it out so as to supply another.[16] But we cannot assume that everyone was so scrupulous in the face of ambiguity. Furthermore, we cannot assume that there was common agreement about where the border between noun-categories lay: one person’s cruet may well have been another person’s ewer. From these considerations, we can formulate this principle:

Principle 4

The nouns used to describe objects are linguistic approximations of real-world objects.

Let us now turn to the attributes themselves, beginning, in the example above, with copper. We can probably take it for granted that the redactor, contemplating the bowl or basin, was competent insofar as he or she could distinguish between metal, ceramic, and wood, and we are justified in assuming that the redactor was capable of distinguishing between copper, tin, iron, lead, silver, and gold, the five metals that appear most commonly in these records. But we cannot assume that the redactor was capable of accurately distinguishing between copper and its alloys of tin, zinc, and other metals, and ditto for other alloys.

2 bacini de rame seu de letona
Two bowls made from bronze or from brass
(Object ID 402)

It is likely that redactors could distinguish between the various woods used for making furniture or siding, including pine, walnut, and poplar, but again, we cannot be sure of this. From this, we arrive at another principle.

Principle 5

The objects and attributes identified in inventories are only
as accurate as the expertise or competence of the redactors.

Note that this expertise is not only a function of material competence—an ability to recognize alloys, woods, specialized objects, and so on—but also linguistic competence.

In one of the examples provided earlier, an object phrase describing a basin concluded with an observation regarding the basin’s condition as being “broken.” Attributes related to condition are relatively common in inventories and lists of all kinds. Their presence is not easy to explain. In this instance, it is unlikely that they were included for purposes of disambiguation, which elsewhere may help explain why attributes such as color were used. In post-mortem inventories, condition phrases could help absolve guardians from accusations of dilapidating the objects in question when the heir came of age and took possession of the goods, but this does not explain why they appear in records of debt collection or other partial inventories. In all likelihood, these phrases responded to a requirement implicit in many or most inventories, whether total or partial, that the object in question be located somewhere on the spectrum of value characteristic of the object-type.

Note the implications of this. If the guardian himself or herself was responsible for redacting a post-mortem inventory, the guardian would have had an incentive to maximize the shabbiness of the objects in question. Similarly, a sergeant responsible for debt collection might have had an incentive to maximize the shabbiness of a given object so as to avoid the charge that he had been responsible for damaging it. In other words, although it is likely that the basin in question was fractured and not worth much, we need to be aware that attributes related to shabbiness could have been systematically favored. Finally, it is important to note that copper had a value independent of the basin. It is possible or indeed likely that the object in question was “of little value” as a basin, but of some value as a store of copper. Thus, we could qualify the sixth principle by noting that attributes were shaped not only by the expertise of redactors but also by their hidden interests.

This discussion has raised the question of why attributes were listed at all, a question that is harder to answer than it might first seem. There are two possible reasons use of attributes: disambiguation and value definition. With this in hand, let us consider the patterns that seem to have governed the process of assigning attributes. As a rule of thumb, charismatic things such as clothing were more descript—that is to say, they attracted more attributes—than bland things such as tools. Objects in wealthy households, in part for this reason, seem to be more descript than objects in poorer households. Items that have silk in them are more descript than items that are only made of linen. And so on. It may be possible to devise a statistical test that could identify the “fingerprint” of a given inventory; this would be based on the subtle patterns that formed the redactor’s deployment of attributes. Similarly, it may be possible to identify a generic fingerprint characteristic of all or most of the inventories from a given time and place.

Although this test is not yet ready, a simple and intuitive reading of a set of inventories both total and partial suggests the following conclusion: the attributes assigned to objects vary very little from one type of inventory to another. In other words, the fingerprint characteristic of object-descriptions in post-mortem inventories does not seem to be very different from the fingerprint characteristic of other kinds of total inventories as well as lists of stolen objects, goods held in pawnshops, records of debt collection, and other partial inventories, even though the legal requirements of these acts varied considerably from one to the next. In addition, the presence of many thoroughly non-descript objects suggests that redactors were not in fact all that concerned about disambiguation. Ultimately, although these two factors—disambiguation and value definition—may have played a role in determining the structure and content of attributive phrases, it seems more like that there was a broadly shared practice of description that transcended the particular legal context or legal requirements of various kinds of lists. This conclusion is more interesting than it seems, because it means that we can study the fingerprint of attributive phrases from the perspective of cultural history, in this case asking questions about how people viewed objects or related to them.[17]

In a sense, this was a foregone conclusion, because any written document provides information about the thought-world of the person responsible for its creation. Do the attributes themselves point to actual qualities of real things? Subject to the provisos above regarding the inattention and a lack of expertise on the part of redactors, the answer to this must be yes. The reason is simple. In their own day, inventories were designed precisely to do this: identify things that existed in the real world. As Mazzi has pointed out, a post-mortem inventory comes into being precisely because the decendent is concerned that members of the family might spirit away things that ought to go to the decedent’s heir.[18] Although features of the context of inventory production may have led to inclusions and exclusions, and although attributes were necessarily approximate, the system itself was designed to say something about the real world of tangible things. Our own interests and concerns in this regard parallel the interests and concerns of all the many parties who may have been affected by inventories inasmuch as inventories were legal documents that had force. This is especially true in the case of post-mortem inventories. Families as a whole may have collectively connived against the exactions of the tax collector, but members of the family, individually, would have been attentive to any attempts on the part of guardians to defraud the young heirs.

Where attributes in particular are concerned, then, we can generate a principle of interpretation similar to the 3rd principle described above:

Principle 6

As long as we acknowledge the inconsistent usage of attributes
and the
fact that words are proxies for real-world qualities, it is legitimate to draw conclusions from the attributes of objects in relatively large datasets

˜Following up on the principles of interpretation reviewed in this essay, what kinds of arguments can be legitimately made on the basis of household inventories? The three suggestions described below by no means exhaust the possibilities; they are intended only to suggest certain approaches that are facilitated especially well by the DALME database.

Interpretations based on location and propinquity. Although customs varied from one legal culture to the next, the redactors of total inventories allowed location and propinquity to determine the order in which they listed household items. In some cases, inventories list the rooms within which the items were found. At the simplest level, therefore, it is possible to use inventories of this type to identify the likely contents of rooms, and although there are bound to be errors and omissions, it is unlikely that any of these were systematic. The first item listed in a given room is sometimes an item intimately associated with the room’s function, such as a dining table in the hall or a bed in the bedroom. After that, in the case of a number of inventories from our sample, it seems likely that the redactors of inventories simply listed objects that came next to view. In other words, no effort was made to group together items of the same type, such as sheets or tunics, that may have been found in the same room. In the case of chests and cupboards, an inventory will often list the container and then list the contents before proceeding on to the next object. Following Glennie, Evans describes these assemblages of objects of different types as “material combinations,” the suggestion being that a study of these combinations can be revealing.[19]

Interpretations based on frequency. Unless the researcher has concluded that certain categories of objects have been systematically excluded from a set of inventories, it is possible, on the basis of that set, to make arguments that are based on the frequency of one or more objects. To continue with an example discussed earlier in this essay, a scholar interested in medieval lighting can use a sample of inventories to measure the average ratio of candelabra, oil lamps, and torches. Ratios such as this can then be compared across time and space. The relative frequency of all lighting utensils or items such as books can be measured according to the total number of items in a given inventory. The relative frequency of objects can demonstrate certain trends over time, such as the growing presence of cushions or bolsters on beds. The measure can also be used for other kinds of analyses. To take an example, Noël Coulet and Louis Stouff have argued that the social status of households can be deduced from the appearance of turnspits in the kitchen, since well-to-do households roasted their meats whereas the poorer households boiled theirs.[20]

Arguments from relative frequency will not work all the time. By way of example, we don’t think it is possible to draw conclusions concerning the proportion of items made of bronze to items made of brass, given the fact discussed above that there are no clear-cut categories. We can, however, probably draw conclusions regarding the proportions of items made of copper alloys, tin alloys, and so on. Here, the interpretive problems don’t necessarily arise from the accuracy of the attributes. They arise instead from the fact that we have to be cautious about assuming that that the material attributes were uniformly assigned.

Let us consider this more closely. When an inventory lists a stack of tablecloths or a collection of hoes, we can be fairly certain that the tablecloths were made of linen and that the hoes consisted of an iron blade and a wooden shaft or handle. But although inventories often list stacks of bowls and trenchers, they do not consistently identify the material, which, in this case, could be ceramic, tin, or wood. As Bozzolo and Ornato have pointed out, there is internal variation in the degree to which attributes were assigned.[21] The problem here, as they discuss, is that we cannot be sure whether attributes were randomly overlooked or systematically overlooked. By way of example, it could be that contemporaries expected bowls to be made of ceramic and, similarly, expected trenchers to be made of wood, and therefore didn’t bother to write down the respective attributes. But perhaps they consistently noted down wooden bowls or tin trenchers precisely because the materials used were somewhat unexpected. Similarly, since tin was more valuable than copper in the Middle Ages, at least in southern Europe, it is possible that tin alloys were noted more often than copper alloys.

Interpretations based on patterns of co-occurrence. In any household or collection of objects in any era, certain items or attributes may appear regularly in the presence of other items or attributes. In medieval households, to take an example, mortars regularly co-occur with pestles, quite often within the same object phrase. So regular is this expression that it might even be possible to decide that the object phrase itself describes a single item: a mortar-and-pestle. Among workers in the building and smithy trades, certain kinds of tools co-occur, constituting an ensemble that we can call a toolkit.[22] In some cases, researchers might expect to find these relationships; thus, if a compass, meaning a tool used to draw circles or to measure, appears in a post-mortem inventory, it is likely that the decedent was a carpenter, and one should therefore expect to find saws and adzes.

Item 1a compas… Item quatuor serre… Item ii aysse
Next, a compass… Next, four saws… Next, 2 adzes
(Object IDs 7739, 7781, 7784)

In other cases, large data sets might reveal patterns of co-occurrence that are not obvious. The color red, for example, might co-occur with the presence of items made of silk.

The patterns of co-occurrence are not limited to objects or attributes; indeed, some of the most interesting patterns describe relationships between objects and certain rooms or locations. In inventories from Marseille, for example, curtains appear almost exclusively in bedrooms, which means that the word cortina refers to bed canopies rather than window dressings. Armor and weapons are found primarily in dining halls, a storage habit which can help explain why it is that certain kinds of shields, perhaps stored on hooks on the walls, gradually became used for decorative purposes.

˜As these remarks have suggested, it is no doubt the case that the legal, epistemological, and ontological contexts that surrounded the production of inventories of all descriptions played some role in distorting the signal conveyed by inventories from the later Middle Ages. Yet we side with those scholars who believe that it is possible to use inventories for purposes that are not purely representational.[23] One should always take care in reading such documents. But as long as users of inventories pay attention to these contexts and formulate their questions with care, it is possible to use inventories for archaeological ends.


Bibliography of Works Cited

Ago, Renata. Il gusto delle cose: una storia degli oggetti nella Roma del Seicento. Saggi. Roma: Donzelli, 2006.

Bernardi, Philippe. “Objets Légués, Objets Loués, Objets Donnés: Les Inventaires d’outils En Provence Au XIVe Siècle,” n.d.

Bozzolo, Carla, and Ezio Ornato. “Les inventaires médiévaux et leur exploitation quantitative.” In Libri, lettori, biblioteche Italia medievale (sec. IX-XV): fonti, testi, utilizzazione del libro, edited by Giuseppe Lombardi and Donatella Nebbiai-Dalla Guarda, 115–28. Rome, 2001: Istituto centrale per il catalogo unico delle biblioteche italiane, 2001.

Casset, Marie. “Inventaires après décès et archéologie des séjours nobles.” In Des châteaux et des sources: archéologie et histoire dans la Normandie médiévale: mélanges en l’honneure d’Anne-Marie Flambard Héricher, edited by Élisabeth Lalou, Bruno Lepeuple, and Jean-Louis Roch, 341–52. Mont-Saint-Aignan: Publications des Universités de Rouen et du Havre, 2008.

Cohn, Samuel Kline. “Renaissance Attachment to Things: Material Culture in Last Wills and Testaments.” The Economic History Review 65 (2012): 984–1004.

Coulet, Noël. “La cuisine dans la maison aixoise du XVe siècle (1402-1453).” In Du manuscrit à la table. Essais sur la cuisine au Moyen Âge et répertoire des manuscrits médiévaux contenant des recettes culinaires, 163–72. Montreal: Presses de l’Université de Montréal, 1992.

———. “L’équipement de la cuisine à Aix-en-Provence au XVe siècle.” Annales du midi 103 (1991): 5–17.

Courtemanche, Andrée. La richesse des femmes. Patrimoines et gestion à Manosque au XIVe siècle. Montreal: Bellamin, 1993.

Evans, Adrian B. “Enlivening the Archive: Glimpsing Embodied Consumption Practices in Probate Inventories of Household Possessions.” Historical Geography 36 (June 2008): 40–72.

Garcia, Jean-Pierre, and Guilhem Ferrand. “Le vin dans la maison: ce que les inventaires après décès disent de la consommation du vin à Dijon de la fin du Moyen Âge à l’époque moderne.” Cahiers d’histoire de la vigne et du vin, no. 13 (2015): 23–43.

Godding, Philippe. Le droit privé dans les Pays-Bas méridionaux du 12e au 18e siècle. Vol. 14. Mémoires de la Classe des lettres 2. Bruxelles: Académie royale de Belgique, 1987.

Howell, Martha C. Women, Production, and Patriarchy in Late Medieval Cities. Women in Culture and Society. Chicago: University of Chicago Press, 1986.

Kaplan, Marion A., ed. The Marriage Bargain: Women and Dowries in European History. New York: Harrington Park Press, 1985.

Kirshner, Julius. “Wives’ Claims against Insolvent Husbands in Late Medieval Italy.” In Women of the Medieval World: Essays in Honor of John H. Mundy, edited by Julius Kirshner and Suzanne F. Wemple, 256–303. Oxford: Blackwell, 1985.

Kuehn, Thomas. “Estate Inventories as Legal Instruments of Credit in Renaissance Italy.” Forthcoming, n.d.

———. Heirs, Kin, and Creditors in Renaissance Florence. Cambridge: Cambridge University Press, 2008.

Lanaro, Paola. “La restituzione della dote. Il gioco ambiguo della stima tra beni mobili e beni immobili (Venezia tra Cinque e Setecento).” Quaderni storici 45 (2010): 753–78.

Marandet, Marie-Claude. “L’équipement de la cuisine en Toulousain à la fin du Moyen Âge d’après les inventaires et les testaments.” Archéologie du Midi médiéval 15, no. 1 (1997): 269–86. https://doi.org/10.3406/amime.1997.1329.

Mazzi, Maria Serena. “Gli inventari dei beni: Storia di oggetti e storia di uomini.” Società e storia 7 (1980): 203–14.

Orlin, Lena Cowen. “Fictions of the Early Modern English Probate Inventory.” In The Culture of Capital: Property, Cities, and Knowledge in Early Modern England, edited by Henry S. Turner, 51–83. New York: Routledge, 2002.

Piponnier, Françoise. “L’outillage agricole en Bourgogne à la fin du Moyen Âge.” In Études bourguignonnes, 131–45. Actes du Congrès National des Sociétés Savantes. Section d’histoire médiévale et de philologie / 109, 2, 1987.

Stouff, Louis. La table provençale: boire et manger en Provence à la fin du Moyen Âge. Avignon: Éditions A. Barthélemy, 1996.

Wolff, Philippe. “Inventaires villageois du Toulousain (XIVe-XVe siècles).” Bulletin philologique et historique, 1966, 481–544.

[1] Kirshner, “Wives’ Claims against Insolvent Husbands in Late Medieval Italy”; Kaplan, The Marriage Bargain; Howell, Women, Production, and Patriarchy in Late Medieval Cities; Courtemanche, La richesse des femmes. Patrimoines et gestion à Manosque au XIVe siècle; Lanaro, “La restituzione della dote.”

[2] Cohn, “Renaissance Attachment to Things.”

[3] The intention was an explicit concern of the law courts. In Florence, guardians who inadvertently forgot to list items were treated more leniently than guardians who hid them for reasons of malice. See Kuehn, “Estate Inventories as Legal Instruments of Credit in Renaissance Italy.”

[4] In general, see Orlin, “Fictions of the Early Modern English Probate Inventory.”

[5] Evans, “Enlivening the Archive,” 44. For some additional thoughts on this see Marandet, “L’équipement de la cuisine en Toulousain à la fin du Moyen Âge d’après les inventaires et les testaments,” 270.

[6] Casset, “Inventaires après décès et archéologie des séjours nobles.”

[7] Casset.

[8] Mazzi, “Gli inventari dei beni: Storia di oggetti e storia di uomini,” 211; Kuehn, “Estate Inventories as Legal Instruments of Credit in Renaissance Italy.”

[9] Casset, “Inventaires après décès et archéologie des séjours nobles,” 345.

[10] Piponnier, “L’outillage agricole en Bourgogne à la fin du Moyen Âge.”

[11] Garcia and Ferrand, “Le vin dans la maison,” 27.

[12] Kuehn, “Estate Inventories as Legal Instruments of Credit in Renaissance Italy.”

[13] Kuehn, Heirs, Kin, and Creditors in Renaissance Florence; Kuehn, “Estate Inventories as Legal Instruments of Credit in Renaissance Italy.” See also (check) Godding, Le droit privé dans les Pays-Bas méridionaux du 12e au 18e siècle.

[14] Ago, Il gusto delle cose.

[15] Wolff, “Inventaires villageois du Toulousain (XIVe-XVe siècles),” 482. See also Marandet, “L’équipement de la cuisine en Toulousain à la fin du Moyen Âge d’après les inventaires et les testaments,” 280.

[16] Garcia and Ferrand, “Le vin dans la maison,” 25.

[17] Evans, “Enlivening the Archive,” 45.

[18] Mazzi, “Gli inventari dei beni: Storia di oggetti e storia di uomini,” 208.

[19] Evans, “Enlivening the Archive,” 45. In general, see pp. 55ff.

[20] Coulet, “L’équipement de la cuisine à Aix-en-Provence au XVe siècle”; Coulet, “La cuisine dans la maison aixoise”; Stouff, La table provençale.

[21] Bozzolo and Ornato, “Les inventaires médiévaux et leur exploitation quantitative.”

[22] Bernardi, “Objets Légués, Objets Loués, Objets Donnés: Les Inventaires d’outils En Provence Au XIVe Siècle.”

[23] Evans, “Enlivening the Archive,” 46.