Studies of ties of service and dependence in the early middle ages have traditionally been split into two main directions of enquiry: slavery and unfree labour on the one hand, and free military or political service on the other. Both, in different ways, have been the object of fierce debate. In the case of slavery, the historiography has by and large told a story of endings rather than beginnings, with most arguments relating to identifying the date by which the classical model may be said to have died out. The question of free military service has proved equally controversial, but has generally been approached more from the perspective of later developments of the central medieval period. Arguments initially focused on the birth of vassalage as an institution and on the early signs auguring the advent of ‘feudalism’ (in the Ganshofian sense of formalised military service), emphasising its novelty and uniqueness, and its heralding of a crisp new way of defining and formulating power relationships, wholly distinct from the earlier, institutionally messier late antique system of patronage and clientelism. This teleological approach was firmly refuted by Susan Reynolds, in her ground-breaking Fiefs and Vassals.Footnote 1 Concurrently, the reevaluation and discrediting, during the 1990s, of arguments positing a major turning-point around the year 1000 has also emphasised continuity between Carolingian lordship and that of later periods.Footnote 2 This has resulted in an interpretation of the Carolingian era as a period of unprecedented aristocratic growth, coupled with an acknowledgement, at least in the British historiography, that Carolingian military dependence cannot be said to constitute an origin-point for an inexorable progression towards feudalism (though, as we shall see, Carolingian ‘vassalage’ as an institution remains alive and well in much continental historiography).
Most examples of ties of dependence found in the sources have thus tended to be interpreted as falling into one of the following two categories: dependence is either characterised as fully consensual and as a mark of enhanced status, in the case of free military dependants, or as coerced and degrading, in which case it is assimilated to unfree status. Beyond the legally unfree, this latter category has been seen as extending progressively to all low-status people, technically free but in practice subjected, with dependence in all these cases linked with the constitution of a new, more diffuse and inclusive servile category (‘serfdom’ as defined by Marc Bloch). The divide is also historiographical: slavery and unfreedom have mostly been studied by social historians, while military service has tended to fall more within the remit of political and constitutional history. Partly as a result of the dearth of source material, however, social ‘history from below’ and analyses of ‘top-down’ political structures tend not to stray very far from their initial starting-point, and only rarely meet in the middle. As a result, the terms of these debates have rarely been combined effectively, except in the Marxist model, which, at its best, has perhaps been most successful at articulating the two, though it too posits a deep structural divide between them as radically different spheres of dependence, the one high-status and honourable, the other low-status and humiliating. Put in the starkest terms, these two categories could be said to represent the putative winners and losers of the ‘crucial period of aristocratic affirmation’ associated with the Carolingian period, which, Chris Wickham has recently argued, allowed all-out oppression by this aristocracy of the lower echelons of society, and of the peasantry in particular, thereby bringing to an end a period of relative autonomy enjoyed by the peasantry after the breakdown of the Roman taxation system.Footnote 3 There may, however, be some value in considering them not as natural opposites, but as two ends of a spectrum, by looking for broader underlying principles and examining in more concrete terms how and why people across the social scale entered into such ties of dependence, and what this meant to them – in short, what lords and dependants expected to get out of them, and from each other.
I
One text in particular may serve as a useful starting-point, since over the years it has successively been taken as evidence for both high-status military service and low-status dependence. It is a legal formula, no. 43 of a formulary (that is, a collection of models for legal documents) compiled in Tours at some point during the eighth century:
For someone who is putting himself under the power of another man. To the magnificent lord A, I, B. Since it is known to all that I do not have the means from which to feed and clothe myself, I therefore asked your piety, and this was my decision (et mihi decrevit voluntas), that I should put and transfer myself under your protection (me in vestrum mundoburdum tradere vel commendare), which I did in this manner, in such a way that you should provide me with help and comfort with both food and clothing, according to how much I may deserve in serving you; and for as long as I live in dependence (in caput advixero), I should pay you service and obedience (servitium et obsequium) as a free man (ingenuili ordine), and I may not have the power to withdraw myself from under your power and protection in my lifetime, but I should remain all the days of my life under your power and protection (defensio). And it was decided that if one of us wants to change this agreement for himself, he should pay n. solidi to the other, and let this agreement remain firm; and it was decided that they should make and sign for each other two documents regarding this copied with an identical content, which they did.Footnote 4
Long associated with the origins, from the eighth century onwards, of the ‘triumphal march of vassalage’ (in the words of Gerd Althoff), this text has been said to constitute the earliest surviving evidence for an embryonic ‘vassalic’ relationship between a secular lord and his dependant, and taken as a prototype for later developments.Footnote 5 It is often referred to in textbooks on feudal relationships as if it presented few or no problems of interpretation: the clear outline of such elements as the service of the dependant, the duties of the lord and the lifelong and irrevocable nature of the agreement seemed enough to interpret it as showing an example of vassalage, which may be why it has proved so enduringly attractive to historians. A closer reading, however, shows clear weaknesses in this interpretation.
The lack of any mention of ritual, such as the oath and symbolic touching of hands generally thought to have been typical of ‘vassalic’ entry into service,Footnote 6 is not in itself enough to deal a death-blow to the hypothesis that this formula involved a bond of vassalage. Rituals of this sort tend to be little emphasised in sources beyond relationships directly involving the king in a formal court setting.Footnote 7 This could of course be due only to the limits of the available evidence, but it could also imply that the symbolism of power at lower levels remained more fluid.
More worryingly for those who want Tours no. 43 to be the foundation-text of feudal relations, however, this text also makes no mention of the most important structural feature normally associated with vassalage: the performance of military service in return for land. Our Tours formula does not mention military service as the specific occupation of the new dependant; nor does it contain any reference to his being granted any land by his new lord.
The representativeness of this text may also be called into question, in view of its very limited impact on the manuscript tradition of the Tours collection. This text is a legal formula, that is, a model prepared by a scribe on the basis of an existing legal document to help him with his future work or in teaching his pupils.Footnote 8 Although some of these models were first compiled in the Merovingian period, they survive only in Carolingian manuscripts, which often give vastly different versions of the same collections. Charting the development of these texts over time can give us an impression of what models the scribes who copied our surviving manuscripts felt they would need in the course of their professional careers. The surest way to establish the general usefulness and relevance of any given formula during this period is therefore to see how many later scribes, working on the basis of the collection of which it was part, chose to copy or reuse it themselves in their own manuscripts.Footnote 9 By this standard, Tours no. 43 does not score very highly, since this text can only be found in a single manuscript, the fullest and most comprehensive out of the seven or so containing material relating to this collection: it was therefore clearly far from being a favourite of later scribes in the process of compiling their own selections. This once again points to the fragility of this text's status as the founding-stone of feudalism; perhaps a founding-pebble would be more accurate.Footnote 10
All this can only cause the gravest doubts regarding Tours no. 43 as evidence for anything that might be thought of as straightforward vassalage.Footnote 11 Indeed, the particular arrangement made in the Tours formula and the wording used to describe it may have owed less to the observance of a preexisting, standard way of entering into someone else's service ‘as a free man’ (or as a vassal), and more to negotiations and bargaining in the particular case which had formed the basis for this formula, in a similar way to those which could lead to unfree service (as in self-sales), or even, as in another formula, to a man handing over ‘half’ his free status to another in exchange for a loan:Footnote 12
Here begins a loan security concerning a man. To the magnificent lord brother A, I, B. Let it be established by this deed of security that I received from you, as I did indeed receive, a loan of n. ounces of silver. I give to you as a pledge (in loco pignoris) half my free status (statum meum medietatem), so that I will have to do whatever appropriate work (qualecumque operem legitema) you order me to do for n. days out of every seven. After n. years have elapsed, I will have to return your property, and I will recover my deed of security. And if I am slow or negligent regarding this work or in returning this property at the appointed placitum, or if I do not act according to your wish in this matter, I am to pay you back twice the amount of the loan you gave me, either to you or to whomever you will have given this deed of security to enforce.Footnote 13
As with the ‘half-freedom’ emphasised in this case, the insistence on ‘free’ service in the Tours formula may simply have been intended to ensure better terms for the dependant; it may have meant, for instance, that the lord could not inflict corporal punishment on him, as he could have on an unfree dependant.Footnote 14
The absence of most of the classic features of vassalage from the Tours formula could not but be noticed in the end, and this text has recently experienced a fall from grace in historians' eyes, if not in general textbooks, at least in new research. In his recent book on Frankish social history, Puissants et misérables, Jean-Pierre Devroey thus correctly pointed to similarities between Tours no. 43 and documents relating to unfree service. These similarities, however, led him to suggest that this example could not be considered as dealing with truly free service in any real sense.Footnote 15 This is a little strange given the text's insistence on the dependant's service ‘as a free man’ (ingenuili ordine, literally ‘as a man in the free order’); Devroey takes this as having only been a ‘pathetic’ attempt to save face by hiding the humiliation inherent in such dependence. In spite of the apparent general thrust of this text, Devroey thus associated it with what he saw as virtually unfree ‘clientelism’, by which previously free people would have been pushed, by poverty and the harshness of the times, into offering their service to a lord in exchange for protection and legal responsibility; he contrasted this with vassalage as a freely chosen and mutually beneficial association.Footnote 16 His interpretation of Tours no. 43 echoes Detlef Liebs's recent evaluation of the reasons for self-sales, which he considered a last resort in cases of extreme poverty.Footnote 17 Devroey therefore paradoxically takes this text as evidence for a very strong differentiation between truly free (or ‘vassalic’) service, in which the duty of obedience would have been only ‘symbolic’, and unfree or not-so-free service, as in the Tours formula, in which the duty of obedience would have been more ‘real’. Any resemblance between the two he views as merely superficial, which allows him to preserve the rift between free and unfree service which has remained a persistent trend in the historiography.Footnote 18
Tours no. 43 has thus so far been interpreted in two mutually exclusive ways: either as evidence for the origins of the highest-status type of dependence (vassalage), or as evidence for the lowest-status type (unfree or nearly unfree service). Either way, it has been seen as paradigmatic: for vassalage, in defining the relationship as mutually beneficial and the result of a free decision for both parties; for low-status dependence, in stressing such features as the poverty of the dependant, his willingness to give service not in exchange for land, but with food and clothing as his only reward, and his desire to secure legal protection, all of which have been seen as typical of exclusively low-status concerns. As we have seen, the first interpretation is virtually impossible to sustain, since this text manifestly differs from vassalage in several highly significant ways; whether the second one proves to be any more convincing remains to be seen. The fact that our Tours formula could be interpreted in such radically different ways indicates the need to investigate what it was, if anything, that made these two types of dependence so fundamentally distinct.
The best way to make such comparisons is to set Tours no. 43 against the background of other legal formulae produced during this period. As I argued earlier, the continued copying of some of these texts indicates that they remained broadly relevant across the alleged chronological divide between structures of lordship before and after the affirmation of aristocratic domination under the Carolingian kings: they therefore provide useful evidence for continuity as well as renewal. Since they could relate to almost any situation for which a scribe might be expected to have to write a document, their wide range of subjects means that they can effectively be seen as handbooks setting out the ways in which legal, social and economic relationships could be formulated in the Frankish kingdoms. They are also especially valuable in that they deal not only with the king and his great men, but with a much bigger slice of the social spectrum, since most scribes were only marginally involved in drawing up royal documents, and were more often called on to prepare documents for their own religious institution or for the lay community in its neighbourhood. Legal formulae also document the construction of particular relationships through explicit agreements, rather than merely referring to them as self-explanatory, as do most other sources. This allows us to focus on the relationship being expressed, rather than on the identity of the protagonists or the use of words expressing particular titles or functions (leudes, fideles, vassi and others), the precise meaning of which is always difficult to ascertain; the use of the word commendatio in sources is similarly both vague and non-systematic, so that fixing on that word does not tell us as much as we might wish either.Footnote 19 Since status itself did not follow a precisely and coherently defined system in this period, this terminology was not used by early medieval authors in as systematic or technical a manner as may have been suggested in modern analyses.Footnote 20 As Susan Reynolds has convincingly argued, deriving conclusions from the use of such terms is therefore problematic, since their own meaning is far from clear.Footnote 21 Formularies, by contrast, give room for a different, more structural, approach, and may thus offer us a better chance of understanding relationships of domination and dependence across the whole of Frankish society.
II
Let us begin with poverty as a motivation for entry into service, and the lack of free choice that this supposedly implied. Do parallels with self-sales mean that Tours no. 43 is not ‘really’ dealing with a free man? Poverty is certainly mentioned in the text, as it often is in self-sales, but is it all that it seems?
The process of comparison is complicated by the difficulty of finding such a thing as a standard, generally applicable model for entry into unfree service in our sources: the specific rights and duties which unfree service could involve seem to have themselves been eminently negotiable throughout the early medieval period.Footnote 22 Nor were self-sales necessarily the result of coercion; this was demonstrably the case only when people who had been found guilty of a crime effectively had to sell themselves to whomever agreed to pay their fine, if they could not afford to pay it themselves. Poverty is certainly often mentioned as a motivation for entry into service, in self-sales as in our Tours formula; but it is difficult to know how seriously to take such claims in either case. Take, for instance, the following formula, from a collection compiled in Angers during the Merovingian period, but preserved in a late eighth-century manuscript copy, in which a couple sold their freedom along with their land:
To the magnificent lord brother A and his wife B, we, C and his wife D. Let it be established that we sold to you (as indeed we did) our free status along with all the possessions that we may own or rent, mansus and land and vineyards, however much we are known to possess at this present day on the land of the villa of E, on the territory of the church of Angers, together with everything that we are known to have anywhere. For this we received from you the price that pleased us, worth n. solidi in gold, so that from this day the said buyers may have the free power to do whatever they want with us and our heirs.Footnote 23
The property described here is not large, but then the model documents found in this formulary hardly ever deal with large estates or with a villa in its entirety. Although this couple may not have been particularly well off, they were therefore clearly not completely destitute either, at least in comparison with the other lay people whose transactions were included in this collection. The future conditions under which they were to live as a result of this sale are not stated, but they may well subsequently have received their property back as a precarial grant, or farmed it as tenants. Whatever the reason for this couple's decision to give up their free status, they must have been expecting it to bring about an improvement on their situation as free smallholders, for instance by providing ready cash or securing enhanced protection.
Earlier evidence for this type of arrangement can be found in a famous passage of Salvian of Marseilles's vociferous tract on the state of the late Roman empire, the De gubernatione Dei, in which he shows ‘poor’ people handing themselves and their property over to more powerful persons, against the backdrop of the fifth-century invasions. Salvian considered this practice as yet another sign of the decadence of the empire and of the injustice of the late Roman taxation system, but his only explanation for this behaviour was that these people had been driven to these measures by their desire to retain their property and stay with their families; otherwise, according to Salvian, they would happily have fled to the barbarians. This certainly suggests they could keep their land under these new arrangements, and contradicts Salvian's claim that precisely these arrangements were causing such people to be despoiled. The discrepancies in his account may have arisen because the only type of right over land which he seems to have acknowledged as worth having (and, indeed, the one most heavily privileged in Roman law, though, perhaps precisely because its status was uncontroversial, it was subject to fewer regulations) was that of full ownership, rather than any right to exploit the land or retain its revenues. Entry into service in these cases could therefore similarly be interpreted as involving people who were not necessarily in financial extremities, but in search of protection and patronage.Footnote 24 Salvian in fact went on to say that many of his pauperes were ‘of not obscure birth, and educated like gentlemen’.Footnote 25
These texts reflect the tendency to use the discourse of poverty to represent weaker social and political status rather than to indicate an actual financial situation. As Karl Bosl showed over forty years ago, the word pauper did not mean ‘poor’ so much as ‘weak’, and could also refer to those without the protection of a lord.Footnote 26 Formulae describing entry into service were therefore bound to cite ‘poverty’ as a motive; this does not in itself constitute supporting evidence for nightmare scenarios postulating slavery and famine as the only two possible alternatives for the majority of early medieval people. The literary sources seldom specify the actual socio-economic situation of protagonists, but instead tend to offer shades of moral characterisation linked with status; historians are therefore left to guess at the underlying circumstances of their actions, and in so doing often reproduce the assumptions of medieval authors in equating legal subjection with complete powerlessness and financial destitution, and vice versa. This, however, reveals much more about what the rather grand persons who wrote these sources considered to be a truly free and dignified life than about social reality for the majority of the population. The self-sale formulae should therefore be seen not as a last-resort option, of which only the most desperate would ever have considered availing themselves, but as a deliberate choice (which does not, of course, imply that it was necessarily a pleasant one); the same may be said of our Tours formula.Footnote 27 Entering someone's service of one's own free will cannot, therefore, be seen as a feature exclusive to free military service; nor, conversely, can coercion be seen as necessarily typical of unfree service. In this sense at least, the opposition between these two types of dependence is not a useful one.
Let us now consider the smallness of the reward envisaged in Tours no. 43, and whether it can really be considered as typical of near-unfree service, and only of near-unfree service. The new dependant's concern, according to this text, was only to obtain material support ‘according to how much I may deserve’ in the form of food and clothing (me tam de victu quam et de vestimento, iuxta quod vobis servire et promereri potuero, adiuvare vel consolare debeas). This, however, does not fit unfree service more than it does any other type of service, since greater rewards at the discretion of the lord could also be expected by unfree dependants, who, as other formulae show us, could even hope for grants of land, whether under precarial tenure or in outright ownership. One formula from the Marculf collection thus shows us a mansus being given to an unfree military retainer (gasindus; Marculf added servus as its Latin equivalent), though this unfree status did not constitute a salient feature of the text, and the same formula could easily have been used regardless of the legal status of the recipient.Footnote 28 Another formula from the Angers collection, in which a man gave a small estate to his nutritus, displays a similar ambiguity between free and unfree.Footnote 29
It is striking that the two texts which come closest in the surviving corpus of formulae to using the classic vocabulary of a land grant in exchange for service (without, however, using the word beneficium) should concern people who seem to have counted as technically unfree: these are two ninth-century formulae dealing with grants of land by a bishop to a serviens of his church.Footnote 30 The word serviens is again vague, but it does seem, at least in formulae, to have mostly been used to refer to the unfree: servientes are thus sometimes opposed to ingenui in lists of property, and are often found in manumissions.Footnote 31
Prestaria. A, archbishop of the church of Sens. We decided, with the agreement of all my canons, that we should grant a mansus to a certain serving-man (serviens) of Saint B named C, which he had. . .to pay for his. . .riding (ad cavalicandum) and his. . .[in the pagus] of D, in the villa which is called E, for this service; which we did, in such a way that both he and his heirs [should hold] the said mansus, as he [held it] in this manner. . .,and they are to hold this mansus from this day by right of inheritance (iure hereditatis). And if they appear to be slow or negligent in this service, [let them] pay [a fine?] according to the law, and let them not lose this mansus as a result of this. And we ask and beg our successors that all. . .and that this [document] made for them should remain firm, so that they should take care to keep our [document] in their favour. I signed this document below in my own hand, and decided to have it confirmed by our canons. Explicit.
Here begins a prestaria. Thus I in God's name A, archbishop of your holy church. We decided, with the agreement and the decision of our brothers and B, the abbot (rector) of the monastery (cella) of Saint C in our gift (ex nostro dono), that we should grant in exchange for an annual rent (ad censum concedere) to a certain man, our servant (serviens) named D, an urban property within the city of E, out of the property of Saint C, and [situated] in this place; which we did, in such a way that every year, on the feast of Saint C, he should take care to pay x amount to this abbot, and he may hold and use the said urban property through the days of his life in exchange for the said annuity (census). And if he is slow or negligent in [paying] this annuity (census), let him promise [to pay it] (fidem exinde faciat), and having made his promise, let him pay [a fine?] according to the law, and let him not lose this urban property as a result of this. But after his death, let the abbot of this monastery recover it under his power and authority with all its increased and added value. And we ask our successors that [this document] in his favour may remain valid. B, abbot of this monastery, consented [to this].
The second text seems to depict a more or less straightforward precarial arrangement: the use of the property is granted in exchange for an annual payment (which had become a standard feature of precarial tenure in the Carolingian period)Footnote 32 and the property was to revert to the monastery after the death of the servant. The servant is not said to fulfil a military function. The formula preceding it is more unusual: the reference to providing for his riding service may suggest it dealt specifically with a military retainer, though this could also have referred to bearing messages. The land was here granted in full ownership, rather than as a benefice in exchange for continued service: in spite of the title describing this text as a ‘prestaria’, the servant was to hold it iure hereditatis, so that, as in the two formulae from the collections of Angers and Marculf mentioned earlier, he would have had the right to leave it to his heirs (though the land had apparently been held by the same servant as a benefice beforehand; on what terms is unclear because the text here is partly illegible). Both texts stipulate that the land could not be recovered by its original owners if these servants proved negligent in performing their duties; this mention suggests that the land could have been taken away under a different arrangement, but that this was decided against in these two cases. The coupling of these two texts in the manuscript points to an association in the mind of the scribe between these different ways of granting land as a reward for service (whether as a benefice or in full ownership), without laying much emphasis on the distinction between either military and non-military service or free and unfree dependence.
The fairly small reward for service, limited to food and clothing, envisaged in Tours no. 43 therefore does not in itself bring this text closer to unfree service, since even unfree servants could sometimes hope for more than this. Conversely, the concern to ensure provision for daily sustenance was not limited to the unfree, but seems to have been widely shared. Tours no. 43 shares features not only with self-sales and unfree service, but also, in some ways, with arrangements made within a family context, in particular those dealing with adoptions and provision for family members:
I, in God's name A, living in the villa of B, have determined in my mind to give two [thirds] of all the property which I am seen to own in this life to my son [C] by this document, so that he may have it in his possession from this day. . .in recognition of his assiduous services and kindness (pro adsidua servicia sua vel benevolentia). . .with this condition, that while I live, he should take care of me, and feed and clothe me (mihi in omnibus tam de victo quam et de vestito soniare mihi debiat).Footnote 33
The focus on reward for service and the provision of food and clothes echoes the concerns of the Tours formula. A degree of assimilation seems to have existed between the service offered by junior family members and that offered by servants in the mind of at least one scribe, who curiously merged in his collection two formulae taken from Marculf, one dealing with a man giving property to his grandchildren in exchange for their support in his old age and ‘poverty’, the other concerning entry into unfree service.Footnote 34
Gifts of land could also be made by parents to reward a son's performance of military service above and beyond what would be included in his inheritance, as in the following case provided by a formula from the Angers collection:
I, in God's name A, and my sweetest wife B, to our son C, loved by us with full affection. Since you are seen to serve us faithfully in all things and in every way, and have endured on our account many hardships and injuries in various places, and went in my place to fight the Bretons and Gascons in the service of our lords, we therefore decided to give you something from our property; which we did. Therefore we give to you in writing our small estate (mansellus) of D. . .And when anything else from our property which is not entered in this charter is left to our children, you should share it with them equally.Footnote 35
The link between ties of service and obligation and family support is nowhere more clear than in cases of adoptions, in which non-family members took on the responsibilities of sons. These, at least in formularies, always included a gift of land in exchange for material support:
If someone adopts a stranger (extranio homine) as his son. To my lord brotherFootnote 36 A, B. Since, for my sins, I have long been bereft of sons, and poverty and infirmity are seen to afflict me, I was seen to adopt you in the place of my sons, according to what was decided and agreed between us with good will, so that, while I live, you will spare and provide in sufficient quantity food, clothing, both for my back and for my bed, and shoes (victum et vestimentum tam in dorso quam in lecto seu calciamentum mihi in omnibus sufficienter inpercias et procures); and you are to receive in your power all my property, whatever I am known to have, including mansus, vineyards, meadows, cattle and every other content of my house, on the condition [that you observe] this right while I live. I therefore decided to have this document made for you, so that neither myself nor any of my heirs nor anybody may be able to change this agreement made between us, but, as mentioned above, you must provide for my necessity while I live, and all my property will remain in your power, both at present and after my death, and you will have the free power to do whatever you please with it.Footnote 37
The claim of poverty is here again undermined by the property listed immediately afterwards. This contradiction is a more or less constant feature of such documents: another formula of adoption, which made similar claims of poverty, thus transferred a fairly substantial amount of property, including a number of mansi, in exchange for the provision of food, drink, clothes and shoes.Footnote 38
This type of arrangement, offering land rather than service in exchange for material support, did not necessarily take the form of adoption as part of a family, but could also characterise relationships with a lord or a religious community: the choice between these two different forms of bond, adoption or dependence, no doubt depended on the relative social positions of the parties involved, with adoption only a possibility when comparable social status made it acceptable to both parties. In purely material terms, however, their consequences may not have been that different. One formula from the late ninth century shows us the model to follow when a man handed over all of his property to a religious house (qualibet locum) or, alternatively, ‘to any powerful man’ (vel cuiquolibet potenti viro), in exchange for material support, again ascribing his need to poverty and old age, ‘which often go together’. What this lord was supposed to provide in exchange is here described in far more detail than in most such cases: this included
two sets of clothes each year, one of wool, one of linen, and sufficient provisions, including bread, beer, vegetables and milk, and meat on feast days; and, every three years, a blanket, gloves, shoes, leg-bands, and soap and a bath-tub, which are very necessary for the sick, and the tools with which to work, as they promised me.
The arrangement was to last only if they received and treated him well (si, ut reor, benignius me voluerint habere); if they failed to do this, the man was to recover his property.Footnote 39 No reference is made to the status under which he was to live, so we can assume he would certainly have counted as free, since he was not subject to any future obligations.
Such arrangements to provide material support in exchange for land also seem to have been perceived as equivalent, or at least comparable, to land grants and precariae, generally seen as the hallmark of military or political dependence. One collection, perhaps originating from Reichenau, contains a series of formulae describing the different possible options that could be followed to establish the terms of the counter-favour expected from a monastery in exchange for a couple's gift of their entire property, with variations depending on the preferred arrangements.Footnote 40 The counter-favour could take the form of a precarial grant, but could equally well involve instead the promise to provide food and clothing for the lifetime of the party who made the gift.Footnote 41
Obtaining the basic necessities of life on the best possible terms from the persons or institutions best able to provide them, whether directly, in the form of food and clothing, or indirectly, through the granting of land allowing such sustenance, thus seems to have been a very general concern during the early medieval period, and to have remained central to a great many people's decisions and choices. The arrangements into which people could enter to achieve this were not limited to relationships of service in any strict sense, whether specifically military or not, but were framed according to whatever possibility would secure the most favourable outcome: sometimes in exchange for service, sometimes in exchange for a gift of land; sometimes within the family unit, sometimes with persons outside it, whether a lay or ecclesiastical lordship or a person of more or less equal status whom one could adopt (thereby effectively assimilating the tie to a family link). In any case, the concern to secure a regular supply of food and clothing seems to have been widespread, and not to have been restricted to the unfree or quasi-unfree, nor even to the very poor: Hincmar's De ordine palatii thus refers to food and clothing among the appropriate rewards to be given to the king's military followers.Footnote 42 A variety of people seem to have been interested in forming such relationships with more powerful persons, as long as the cost of ensuring that person's support remained proportionate to the advantage they would get out of it. Such arrangements included a great number of variables; for instance, offering one's land in exchange for material support, judging from this evidence, seems to have been a solution chosen mostly by elderly people, presumably because they were no longer able to exploit their land directly, and could not otherwise easily obtain such sustenance through service: this limited the range of what they had to offer, and may sometimes have led them to pay a more hefty price to secure their entry into a relationship of support and dependence. The size of the reward for service (or in exchange for gifts of land) thus seems to have depended much more on complex appraisals of what each party had to offer than on a strict distinction between high- and low-status types of dependence.
III
Next to obtaining basic sustenance, the need to obtain protection constituted a comparably important and generally shared concern. I will now turn to the last defining characteristic of dependence included in our Tours formula: the question of legal protection, and whether the need for it would have been limited to the powerless and the unfree. Tours no. 43 includes this aspect of relationships of dependence in its reference to mundeburdium and its Latin equivalent defensio, emphasising the legal protection afforded to the dependant by his new lord. Devroey takes such legal protection to have been an essentially low-status concern, leading vulnerable free people to abandon their liberties and hand themselves over to lords.Footnote 43 It is true that most references to mundeburdium in formulae are found in deeds of manumission and relate to freedmen. This may not mean, though, that such protection applied only to unfree or semi-free people, but rather that those were the circumstances in which it was most likely to need mentioning: since unfree persons would automatically have been under the protection and legal responsibility of their master,Footnote 44 exactly who would fulfil these functions after their manumission needed to be clarified. Sometimes freedmen were allowed to choose their own protector;Footnote 45 sometimes their choice was restricted by the terms of the manumission. An ex-master could for instance require his freedmen to choose their protector from among his heirs,Footnote 46 or, more frequently, assign them to the protection of a particular church.Footnote 47 The latter solution seems to have become widespread during the Carolingian period, at least in the Eastern Frankish kingdom, where manumission apparently grew increasingly assimilated to the transfer of unfree dependants to a church; so close was this conceptual association that a manumission which allowed a freedman to choose his own protector came to be referred to specifically as being made ‘extra ecclesiam’.Footnote 48 The transfer of dependants to a church presumably gave added value to the spiritual benefits of making a manumission. Manumission in such cases may only have implied slightly better terms for the dependant under his new ecclesiastical lord. Either way, manumissions in which a particular protector was assigned to the freedman certainly suggest that responsibility for legal protection was understood to give protectors certain rights over the beneficiary.
These rights could include the payment of dues in money or in kind: while the duty to provide wax candles or a supply of hosts for a church or for the tomb of one's ex-master, if the freedman had been assigned one of his heirs as protector, seems to have been fairly common,Footnote 49 protection was sometimes also afforded in exchange for a yearly payment or census.Footnote 50 One formula gives an example of a testament in which a couple's gifts to some freedmen whom they had entrusted to the protection of a church were automatically to go to that church after the death of these freedmen,Footnote 51 which suggests that protectors acquired a measure of control over the property held by the people assigned to them in manumissions. One way or another, such protection can therefore still be said to have constituted a tie of dependence, albeit perhaps a less demanding one than unfree service. Indeed, people seem to have been conscious, in making manumissions, of the possible abuses or over-interpretation of such ties by the freedmen's new protectors: ex-masters thus sometimes felt the need to protect the free status of their freedmen in the future by specifying that this duty of protection was to be ‘non ad adfligendum sed ad defensandum’ or ‘ad defensandum non ad inclinandum’.Footnote 52 Such distinctions suggest the existence of grey areas in what could be demanded of someone under one's protection, and of a level of anxiety on the part of masters about the risk that the spiritual impact of their gesture would effectively be annulled as a consequence.
There is therefore little doubt that legal protection could constitute a powerful tool of domination. On the other hand, acknowledging such limitations on a freedman or dependant's autonomy did not mean that these arrangements were wholly one-sided: taking someone under one's protection also involved a degree of responsibility for the welfare of the dependant, and a certain amount of give and take. In Tours no. 43, an explicit link is made between giving protection (mundeburdium) and bearing responsibility for the provision of support and sustenance for the dependant. Another formula may suggest a similar link, this time by describing what happened when such arrangements went wrong: in this case, a monastery which had previously enjoyed royal protection (mundeburdium) had later been given by the king to someone else as a favour (nos ille beneficiasti); since then, as the monks bitterly complained to the king, begging him to take them back under his own direct protection, ‘we have not received clothes, shoes, ointment, soap or food, as we used to (sicut antea fuit consuetudo)’.Footnote 53 Rights to protection, and the level of control accorded to those who provided it over those who received it, thus seem to have been intimately linked with the concerns of dependants to secure an adequate living, as discussed earlier. Although these were, by and large, highly unequal relationships, and we may assume that lords had more to gain from them than dependants (since they would have had no reason to enter into them otherwise), we should therefore not underestimate dependants' ability to define and claim the rights which the establishment of these relationships also afforded them. The particular terms of such arrangements would no doubt have been down to bargaining, depending on the relative position of strength of each party, how much they had to give and how much each thought they could reasonably expect from the other in exchange.
The variables involved in such negotiations are central to understanding how fundamentally similar ways of formulating relationships of dependence could be applied in such a wide range of circumstances. As already seen in the case of the monastery cheated of its food and soap, protection was far from being exclusively the mark of the socially weak, or of unfree or quasi-unfree service: on the contrary, it too seems to have been a very widely shared concern.
Lords' duty of protection and legal responsibility for their dependants seems to have been a commonly accepted idea in formulae documenting already established relationships as well as in those documenting the creation of new relationships through entry into service. One small cluster of formulae in the Marculf collection provides good evidence for the crucial importance of a lord's personal legal action on behalf of his dependants. One, for instance, shows the king issuing a document ensuring that no legal cases could be brought against a lord's dependants while he was away on state business:
We therefore decide and order by the present charter that, until he returns from these parts, all his legal affairs and those of his friends (amici), both those who are going with him and those who are remaining in their homes, or of anyone for whom he is legitimately answerable (undecumque ipse legitimo redebit mitio), should remain on hold, and afterwards he is to make right every legal accusation on their behalf, and in a similar manner obtain justice from others.Footnote 54
Who had legal responsibility for whom could become a fairly intricate matter, as in the following royal mandate, in which one lord was authorised to assume legal responsibility for another, who himself was responsible for his dependants:
Our fidelis A, with God's favour, having come into our presence, put it to us that due to his ignorance he could not pursue his legal cases or enter a plea before a court (mallus). He asked the clemency of our rule that the illustrious man B should take up all his legal cases on his behalf, in local as well as our royal courts, and pursue them and enter pleas regarding them before a court. He was seen to delegate them to him presently by the rod (per fistuca [sic]). Therefore we order that, since this was the will of both parties, the said man B must pursue and enter pleas regarding all of A's legal cases everywhere, and must do right concerning legal accusations (unicuique. . .de reputatis condicionibus et directum faciat) on his behalf and on that of his men (pro ipsum vel ominibus suis), and in a similar manner obtain justice from others, for as long as this is the will of both parties.Footnote 55
The vocabulary used in royal documents to refer to people under a lord's protection tends to refer to amici, pares, gasindi,Footnote 56 or homines, all of which could imply a military context, but not necessarily so: A's homines could have included both military retainers and his other dependants, including rural tenants and freedmen. The absence of lower-ranking dependants in this cluster of texts is probably due to the higher social and political sphere in which these documents were being produced, and ought not to be taken to imply that such dependants were not also meant to be included in these arrangements.
This same text also shows that legal protection could be sought by socially more prominent persons, as in the case of ‘A’ himself. He is described as a fidelis, and clearly had access to the king; B seems to have been of higher status, since, unlike A, he is described as inluster. A's plea of ignorance may or may not have been real; either way, he clearly thought that B would be better able than himself to uphold his interests. Although this arrangement gave B a level of control over A's affairs, in this case the relatively high status of both parties meant that the relationship was framed as one of patronage, in which the counter-favour was left vague or to be called in at a later date, rather than as a wider-reaching tie of dependence, in which the precise terms of the relationship tended to be defined more strictly. The practice of enlisting the support of a more powerful person through mandates, however, was not merely a feature of the royal court, and there are clear equivalents at a more local level.Footnote 57
Should legal protection at this higher level be construed as similar to mundeburdium, or did mundeburdium imply a fundamentally different, more constraining type of relationship? The majority of known cases in which kings made specific grants of their own mundeburdium do seem to relate to members of less dominant groups, such as women, freedmen or Jews.Footnote 58 It is likely, however, that royal grants of this kind arose out of a particular interest in being seen to defend such groups for reasons that were primarily ideological. In one formula reproducing the text of a letter addressed to Charlemagne at some point after his imperial coronation, a woman who did not have a man to represent her in court, and who did not yet have a charter of royal mundeburdium, wrote to the king asking for the rights such a charter would have extended to her, emphasising the protection of helpless women (in this case perhaps a widow) as a Christian virtue. Although her situation was contrasted in the title of the formula with the kind of automatic access to the king afforded to fideles in his service (the title suggests she was not technically allowed to speak directly to the king, though she could submit a written petition: ‘if you do not have the permission to speak with the king, these are the words you should send to him regarding your case’), she clearly felt she too had a legitimate claim to his protection, and made her case by explicitly tapping into the protective streak of the Christian discourse of Charlemagne's regime.Footnote 59 The fact that most surviving charters of mundeburdium should relate to people belonging to less powerful categories does not mean that protection in their case was in any way more onerous or less effective than that offered to the king's fideles: although it was given for different reasons, that is, out of Christan duty rather than as a reward for service, there is no reason to assume that royal protection in the case of such groups would have had fundamentally different effects in practice. This could explain why grants of this kind could also be made to religious houses.Footnote 60 Royal grants of mundeburdium do not, therefore, constitute evidence for a lower-level form of protection, though they may well have been used more often in the case of certain special categories (churches, widows and orphans in particular), the protection of which was, in this world of ‘little patriarchs’, as Jinty Nelson has evocatively put it, seen as a fundamental cornerstone of Christian kingship.Footnote 61 In the case of charters of this kind given to Jews, the recipients are described as ‘serving the king’ (‘partibus palatii nostri fideliter deservire’), which could suggest that mundeburdium was simply meant to extend to them in a more secure and explicit manner the same sort of protection available to other men in the king's service.Footnote 62
Legal protection, far from being an essentially low-status concern, was an almost universal factor in the workings of both society and the state. Warren Brown has confirmed in a recent article the fundamental importance of personal relationships and connections in the settlement of disputes as evidenced in Carolingian formularies, with particular reference to letters of intercession.Footnote 63 As we have seen, evidence for this is not only to be found in letters, and relationships of legal protection were also formally recognised in legal documents throughout the Merovingian and Carolingian periods: public and private power and jurisdictions complemented rather than limited each other.Footnote 64 This is at odds with Devroey's hypothesis of a fundamental opposition between service to the king and service to a patron, and of vassalage as opposed to other kinds of service:
The relationship of reciprocity initiated by royal protection did not directly threaten the dignity and freedom of his subjects. By contrast, the relationship of protection / dependence between a humble man and his patron entailed the irremediable degradation of the ‘poor’ and a rupture of the link between them and the ruler (whereas the Carolingian vassal remained among the king's faithful because he was his subject).Footnote 65
But there is actually very little in the sources to support such a distinction: ‘poor’ people would certainly not have had much direct involvement with the king, but this would have been true whether they were in the dependence of a lord or not. Whether the concept of ‘subject’ can be said to have been relevant at all in relation to early medieval kingship is debatable,Footnote 66 and there is no evidence to support the notion that low-status people who put themselves under the protection of a more powerful party were somehow removed altogether from the authority of the king as a matter of principle. Simon MacLean has shown that succession crises played a greater role in the end of the Carolingian empire than any increasingly rampant pauperisation leading free people to put themselves under ‘private’ ties of dependence, as Devroey suggests.Footnote 67
IV
In many respects, the opposition between free military service on the one hand as honourable and all other forms of dependence on the other as dishonourable thus seems somewhat artificial. Devroey's interpretation of Tours no. 43 as not ‘really’ dealing with free service rests on the assumption that there was such a thing as classic vassalage for lords and their men in this period, but that Tours no. 43 does not correspond to it. Yet dismissing this formula leaves us with very little evidence for the existence of vassalage as such a neatly separate institution, at least outside the sphere of the very highest political elite. Devroey correctly points to the inappropriateness of identifying this formula as relating to vassalage, but then tries to reduce the relationship involved here to a concealed form of humiliation: he imagines the position of free people putting themselves in the dependence of someone else as ‘irrémédiablement dégradée’, no better than that of children or slaves, and altogether precluding any claim to personal independence.Footnote 68 Such an interpretation is linked to a wide assumption in the modern historiography that dependence in the case of the less powerful sections of society would necessarily have been shameful or demeaning, and could therefore only have been the result either of coercion or of abject poverty. However, there is in fact little to suggest that people were faced with such a stark choice between complete freedom and complete lack of freedom: as formulae show us, in particular with Tours no. 43, there were many kinds of middle ground. What comparison between our Tours formula and self-sales shows is that early medieval people perceived free and unfree service as similar, not as opposites.
The flexibility and negotiability of arrangements over service could explain why even lower-status people could think that they might have something to gain from entering a relationship of dependence with a lord. Rather than forcing them into a status lower than that of independent free peasants, association with a lord may in fact have given dependants a distinct edge in competing with other members of their local communities, particularly in relation to the legal protection such a relationship involved, since it would have increased their ability to win court-cases against their neighbours. Lower-status types of service and dependence were therefore not necessarily the result of coercion, but of individual decisions based on calculations of short-term benefits.
The distorting elite viewpoint of most of our sources is what led them, and consequently has led modern historians, to turn vassalage and military service into an exception, by considering them as more honourable in principle, and allowing them a unique status as the only positive form of service. But medieval views on dependence may not have been so rigid as that, except at the very highest echelons of society which brought forth the authors of all our literary sources. It is certainly important to distinguish military service from other kinds of service in political and cultural terms, but it was still part of wider social structures: historians should perhaps be wary of following their lead entirely in this respect, and of isolating free military service from this surrounding background as a distinct and wholly self-contained type of relationship.
Einhard's letters show us that a similar interplay in the language of pressure and protection, of exploitation and the recognition of rights, as that observed earlier for low-status dependants could also be at work for higher-status dependants. In one letter, for instance, he effectively ordered one of his fideles, who was clearly not low-status, to allow his daughter to marry a vassalus of Einhard's: although the father's decision was free in principle, Einhard took it for granted that he would follow his instructions, and applied pressure by referring to the support he had given him:
I do not doubt that you recall how you commended yourself and yours to me. Since you decided to do this of your own will, it is also my wish in turn and by all means to supply you and yours, whenever the opportunity arises, with fitting support to the extent that I know how and possess the ability. Thus, let it be known to your kindness that my vassalus and your daughter desire with the Lord's approval and your permission to take each other in marriage. . .To conclude this business [all] that remains is either for you not to defer coming yourself at the present time in order to bring this matter to a close or to allow me to finish it.Footnote 69
Other letters back up Einhard's claim regarding the support he gave to his dependants, showing him helping to secure their benefices and actively interceding on their behalf in legal cases.Footnote 70 High-status dependence did not necessarily imply a less hands-on approach or a lesser degree of intereference on the part of the lord: here as in the other cases discussed in this paper, patronage implied control.
Judging from the evidence of formularies, no single feature seems to have been exclusively characteristic of either high- or low-status service, whether in terms of freedom of choice in entering service, of the types of reward on offer or of legal protection and the level of control it afforded to lords. Ties of dependence at all levels seem to have been contracted on a largely ad hoc basis, depending on the outcome of negotiations. There was no sudden break on the social scale on either side of which power relationships could be said to have worked themselves out in fundamentally distinct ways: what we see instead is a continuum. Considering military service as the only form of truly free dependence and everything else as either explicitly or virtually unfree thus obscures similarities in the basic concerns shared across the social spectrum, and in the level of control that the establishment of relationships of patronage could involve: the very ambiguity and conflicting interpretations associated with Tours no. 43 show just how similar the terms of either relationship could be. In this context, it might be more fruitful not to worry so much about whether such free service as is described in the Tours formula ‘really’ corresponded to vassalage or to unfree service, but to examine this evidence against the broader background of the various ways in which ties of obligation and dependence could be formulated and understood during this period.
A useful comparison with the Tours formula may be made in the case of a group of tenth-century charters from northern Spain, recently studied by Wendy Davies, recording gifts of land made in exchange for future support, with the expectation that the recipient would ‘do good’ (bonum facere) for the donor.Footnote 71 These charters, like our Tours formula, were for a long time interpreted as straightforward acts of formal commendation.Footnote 72 As Wendy Davies has convincingly shown, however, such agreements were far from forming part of ‘a general European post-Roman process, in which small-scale free proprietors submitted their persons and gave up their property to great landlords in order to secure assistance and protection in an increasingly unstable world’: they did not always involve all of the donor's property; they were not always meant to create a lifelong relationship; and donors in these cases could be aristocrats as well as peasants.Footnote 73 Instead, Davies places these texts in the realm of ‘light-touch’ patronage, ‘lining up and getting supporters, in an altogether looser, less autocratic, framework than the “commendation” model’, epitomised by the use of the ambiguous term gubernare, which could mean ‘to rule’, but also ‘provide’ and ‘support’.Footnote 74
The reevaluation of this Spanish evidence could in turn help us to revisit our Frankish material. If these charters are not about formal commendation, perhaps we should revise our interpretation of the Tours formula and of similar texts from the Frankish kingdoms along these lines. Susan Reynolds has already dealt a death-blow to the modern conception of vassalage as a fixed, self-evident construct; perhaps it is now time to free ourselves from another tyrannous construct, and to abandon the idea of lower-status ‘commendation’ as a rigid, formal process. If we are to understand them properly, the need to allow for a similar level of complexity and ambiguity in the formulation of power relationships in the case of ties of dependence at lower social levels is just as crucial as in the case of vassalage. The negotiations over the rights and duties owed to and by peasants and lower-status dependants, documented in formulae as well as in other sources, implies a level of recognition of their rights by lords.Footnote 75 As with most human relationships, some degree of oppression and some degree of reciprocity were both nearly always part of the deal; the real question lies in the balance between the two, and this seems to have been governed not by hard and fast rules so much as by improvisation and opportunity. These different ways in which ties of dependence and protection allowed people to obtain what they needed from each other, which overwhelmingly involved the provision of support and security, certainly point in the direction of a strong consciousness of power relationships, but not ones that were fixed in a conceptually very systematic way: a range of dependent relationships are best viewed within a single, broad, social and legal context of need and support. Strategies of patronage, support and control could take many different forms and include people from virtually all sections of society, and it is consequently very hard to find a single form of dependence which could be called typical for this period. That is why historians should be wary of attributing to lower-status dependence as an undifferentiated whole such strong connotations of humiliation and powerlessness: early medieval people were, after all, familiar with ideas of honourable dependence and empowering service.
