1Ever since Georges Duby demonstrated in his thesis on the Mâconnais the riches that could be extracted from eleventh and twelfth century charters, these documents have become the principal source for the social history of this period. Commonly, they have been considered unproblematic witnesses, directly transmitting the past to us almost without interference; in the language of electronic engineering, they are assumed to have a high signal-to-noise ratio; their changing vocabulary directly reflects the usages in the ambient society, the actions they embody or report are there, as it were, in the flesh, unprocessed. Such an assumption is particularly strong when historians interpret the « legal » vocabulary of charters, directing their efforts at ascertaining the exact technical meaning of particular words. In this essay I will ask whether such an « assumption of transparency » is always valid, looking at the terms « alod » and « fief » and a commonplace transaction in which both are involved –the so-called fief de reprise.

2From time to time in cartularies we come across a document or pair of documents which assert that a person or group of persons have given or sold their « alod » to another who in turn has given it back to the donors (vendors) as a « fief ». The cartulary of the Guillems of Montpellier is particularly rich in such documents, and since A. Germain’s edition of that cartulary they have been interpreted as stages in the process by which the lords of Montpellier extended their power over independent castellans in the countryside, subjecting those castellans to higher lordship, getting them to recognize that they held their castles in fief and were obligated to serve for them rather than holding those castles « in full right » free of all obligations1.

3Here is one such story.

4In 1139, Rixendis, widow of Pons Fulco of Popian, along with her son William « gave » William VI of Montpellier everything they had in the castle of Pignan, including its lordship and command ( « tradimus... totum quod habemus... in seniorivo seu mandamento ipsius castri »), and receiving it back in fief « and in [William VI’s] service and fidelity »; Rixendis had this castle, she said, by succession to her brother William2. The scribe who drew up this particular two-way conveyance did not use the word « alod » for what Rixendis was giving, but that word often appears in other conveyances of the same kind. Taken in isolation the document seems to match the standard interpretation. When we place it in the context of other documents concerning Pignan, however, we see something very different.

5Already twenty-five years earlier, in 1114, Peter William « called Laige Infant, » the father of Rixendis and son of an earlier Rixendis had given the castle of Pignan ad alodium to William V of Montpellier, receiving 500 sol. Melg. from William’s wife Ermessendis, plus 40 sol. « for my [i.e. Peter William’s] love » (pro drudo meo). In 1133, six years before the later Rixendis’s gift, her son Raimond gave all he had in Pignan « as my uncle William and my ancestors held it » to William VI per alodium franc, receiving it back in fief along with 1500 sol. Melg. As Claudie Amado has shown, this is the same family line, and we can follow it at Pignan through the rest of the century. We could hypothesize a complex family division of this castle, whose parts the lords of Montpellier only slowly absorbed into their lordship. That would save the idea that with each conveyance some « full rights » (alods), though not the entire castle, were converted into fiefs. But it would be just as easy to hypothesize that whenever a new individual succeeded to the headship of the family –Laige Infant, then William, then Rixendis’s eldest son Raimond, then (on his death?) Rixendis herself– he or she negotiated a new exchange with the lord of Montpellier, « giving » the castle and receiving it back in fief, sometimes along with a sum of money, the exchange fixing in the land– and in everyone’s memory– the service and fidelity and the rights to the castle that would ensue. If so, the same thing most probably happened when an heir succeeded to the lordship of Montpellier3.

6That expectation, indeed, is what the rough Latin of Laige Infant’s gift to William V announces: « If William of Montpellier has not devised (dividebat) this castle [by will], then I give it to his eldest son; if his eldest son has died (moreretur) without heir, then I give it to his second son », and so forth, with more hypotheses following. This could be read as nothing more than a way to assure the donor’s continuing loyalty to the dynasty, a complement to the notice that William’s wife has given 40 sol. to the donor for his love. But another, and very different, possibility behind this strange sequence of imperfect indicative and imperfect subjunctive verbs –all regarding future events (for William V did not die until 1121)– is that on William’s death Laige Infant would give the castle again to William’s sons in the order of succession that William had already determined. That is the literal meaning of the sentence (when we allow for the scribe’s limited command of Latin tenses), and there seems no reason to reject it. That is, Laige Infant’s gift ad alodium was not a once-and-for-all event; no one imagined that it transferred a specific « bundle of rights » from the donor to William and that another, different or lesser bundle came back as a fief. Laige Infant and his successors would give it again, just as he did in 1114, and receive it back again, as he did, in fief.

7If we turn to the charters for Le Poujet in the same cartulary, such repeated giving is exactly what we find.

8Sometime late in the eleventh century, probably soon after William V succeeded to the lordship of Montpellier in 1068, a woman named Girondes, daughter of Adivenia, gave her oath of fidelity to him for this castle4. A decade earlier, in 1059, she had divided her lands and rights there among the sons from her two marriages. Then, three-quarters of a century later, in 1124, her son Bernard Raimond along with his nephew Raimond Berenger and niece Girondes, sold the « honor that belonged to [their mother and grandmother] Girondes » in le Poujet to William V’s widow Ermessendis, and she returned it to them in fief; from its description we know it was exactly the same land and rights that the older Girondes had divided in 1059 and for which she had sworn fidelity around 10685.

9Another grand-daughter of Girondis, Azalais, daughter of Hugh Peiron of Poujet, was also castellan in this village; in 1114, as William V was preparing his expedition to seize Majorca, she and her son swore fidelity for the castle to him, and ordered her husband to do the same6. When William V died in 1121 leaving the western part of his domains to his second son, William of Aumelas, Azalais swore fidelity to the young heir. On the death of William of Aumelas his rights passed to William VI, and in 1132 after Azalais’s husband died, it was to this lord that the widow and her son gave their « alod, » including control of le Poujet for three months each year ( « trado... III menses quos annuatim per alodium... cum toto alio alodio »), receiving it back in fief and again swearing fidelity for it7. Here again, the chance copying of documents into the Montpellier cartulary lets us see in action the repeated process laid out by Laige Infant in his gift to William V. Neither Azalais nor Girondes’s descendants were independent castellans « converting » their alods into fiefs; at the time of the surviving « gifts », they and their family were already in the fidelity of the lords of Montpellier for the castles they were « giving », and had been so, perhaps, for generations. The alods they were giving were already fiefs.

10Whatever « alod » meant, when a lord bought a castle or received it as a gift it did not necessarily mean he was gaining a lordship he did not have before, no matter what appears to us to be the « legal » language of the conveyance. We need only read the charter drawn up in 1158 when two Aimerys of Barbaira, father and son, sold to William VII of Montpellier « the entire honor and all the rights which Dalmacius of Castries had or should have had in the castle of Castries and its territory, in lordship, towers, houses, fortifications, fields, pastures, woods, [etc.], all of which we freely and entirely sell to you ». All the latest vocabulary from the Montpellier law school is deployed in this text, and it would seem as though this is the moment when the lords of Montpellier finally acquired this important castle so near their city. It would seem so were it not for all the other documents concerning Castries that appear in the Montpellier cartulary.

11From father to son and uncle to nephew over five generations we can follow the lordship of this castle in the family « of Castries » until it falls to an heiress, Ermessendis, whose aunt and distant cousin, represented by their husbands and a son, are the other possible claimants8. Ermessendis was married to William VII’s brother, William of Tortosa, and left the castle to him when she died in childbirth in 1157. Soon thereafter he joined the Templars and headed off to the Holy Land, leaving Castries and the rest of his inheritance to his older brother, now sole lord of Montpellier. William VII thus became lord of Castries. In the mid-1140s, however, when Ermessendis was still a child, her father, Dalmace of Castries, had specified that if all his children were to die (presumably meaning « without heirs ») the castle of Castries was to pass to his nephew, Aimery of Barbaira, the son of his sister Galberga. It was these rights that William VII purchased in 1158; though the document is drawn up as a sale, he was paying the Aimerys not to bother him with the claims they might push on the basis of Dalmace’s oral last will (and perhaps to bind more closely to himself a family that had long served the Trencavels). And indeed at the same time, Gaucelm of Claret, husband of Ermessendis’s distant cousin Agnes, agreed to forego his claims to the castle9. If the sale by the Aimerys of Barbaira was the only document in the Castries file to survive we would know nothing of all these complications. Misled by its seemingly unimpeachable legal language would we not take it at face value?

12These are not the only complications in the Castries file to call the literal interpretation of twelfth-century legal language into question. Dalmace of Castries, lord of the castle in the years around 1100, was at least the third generation of his family to control it. Yet two charters of about that time record gifts of the castle « ad alode » to him by men who are probably his fevales, both donations followed by grants of fief, both classical fiefs de reprise10. The more complex of these is actually a double gift, first from Raimond « son of Viola » and his wife Garsendis « daughter of Froiles » to their sons, and then from those sons to Dalmace. The donors are otherwise unidentifiable11.

13How shall we interpret their donations? As extensions of Dalmace’s castellany (the interpretation of A. Germain, the editor of the Guillem cartulary)? Why then the two-stage gift? Why, above all, the generality in describing what is given: « [I] psum castellum... et muros, et quod est intus muros » in one, « castellum.. et ipsas fortezas que ibi sunt » in the other? It is more plausible to imagine these as the same ritual « transfers » we have seen at Pignan and Le Poujet, but here further down the castellan hierarchy, those actually manning the castle « giving » it to their lord « as alod » and receiving it back in fief as their predecessors had done. The occasion here was the departure of Dalmace’s father, Eliziar, and uncle, Rostagnus, on crusade12. It is a novel interpretation, to be sure, but one that more fully explains the peculiarities of the text.

14These examples have all come from the region of Montpellier, but the practice was not peculiar to the lords of that city and its immediate countryside. The evidence is just as clear to the west, in the lands ruled by the Trencavels: up in the mountains of Lacaune at the stronghold of Sénégats, at Laure in the Aude plain, at Termes in the Corbières13.

15What we are witnessing in the fief de reprise, then, is a ritual of succession. When an heir or heiress succeeded to a castle or the overlord succeeded to his or her lordship, more was required than the renewal of the oath of fidelity; there had first to be the giving and giving back. The fief de reprise is but another variant of the gift [don] and countergift [guerredon] so common to that same world, the repeated ground of so many a chanson de geste and twelfth-century romance. To record that exchange, the scribes had only their standard formulae derived at great distance from the very different world of Antiquity. Those formulae were made to record the transfer of property, and so that is what they made the exchange look like. But in fact the person who received the gift did not gain something he did not have before, nor was the person who gave it reduced in wealth or status. How, then, should we understand the terms « ad alodem » and « ad fevum »? Adverbially, I would suggest, rather than substantively. They do not name « bundles of rights » but rather ways of giving. To give « ad alodem » was to hold nothing back; to give « ad fevum » was to create a permanent tie. As follower and lord handed back and forth the same castle, they fixed in the landscape the paired and inseparable values of fidelity and good lordship; the one had to be without reserve as the other had to be for life. What was realized in the exchange was what Laige Infant in 1114 so aptly called « love ».