Freedom and Unfreedom in Early Medieval France PDF
Freedom and Unfreedom in Early Medieval France PDF
Freedom and Unfreedom in Early Medieval France PDF
NUMBER 193
NOVEMBER 2006
CONTENTS
page
ANDREW SHERRATT: by Chris Wickham .................................
41
73
Published by
197
NUMBER 193
10
NUMBER 193
in the Aristotelian sense did not in fact end in the fth century,
but continued until around the year 1000.9 To a certain extent,
he still failed to break free from the chronological constraints of
traditional Marxism, and he retained the idea that the end of
slavery must have coincided with the beginning of feudalism.10
The essential difference was that his chronology took into account
developments typical of French historiography, and followed
Georges Duby in placing both the birth of feudalism and the
end of classical slavery after the year 1000 rather than in the
fth century. Bonnassies work also reects an important development in the denition of what constituted the classical style
of slavery. Since the 1980s, historians, Marxist or otherwise,
have moved away from the model envisaged by Marc Bloch of
chained gangs and plantation-style exploitation, for which it is
now generally recognized there is no clear evidence north of the
Alps in any period,11 and have instead focused on the social consequences of unfreedom, while the problem of economic structures
took a back seat.
This development has been paralleled by the rejection in
French- and English-speaking historiography of an altogether
different vision of unfreedom, developed from the 1950s by
Joseph Vogts Mainz school in explicit opposition to Marxist
economic models, and working instead on the basis of a purely
legalistic denition of unfreedom. This view is now mostly
9
This development has not had many adherents in English-speaking countries,
though see Carl I. Hammer, A Large-Scale Slave Society of the Early Middle Ages:
Slaves and their Families in Early Medieval Bavaria (Aldershot, 2002), reviewed by
Warren Brown, in Early Medieval Europe, xii (2003), 834. Bonnassies position
was supported by Guy Bois in his La Mutation de lan mil: Lournand, village mconnais
de lantiquit au fodalisme (Paris, 1989), now largely dismissed, though it was highly
controversial at the time of its publication: Mdivales, xxi (1991) was dedicated to
debating it. On this vexed issue, see also T. N. Bisson, The Feudal Revolution,
Past and Present, no. 142 (Feb. 1994), and the resulting debate involving Dominique Barthlemy and Stephen D. White Past and Present, no. 152 (Aug. 1996)
and Timothy Reuter and Chris Wickham, with Bissons reply Past and Present,
no. 155 (May 1997).
10
See Davies, On Servile Status, 2312.
11
See the warnings against generalizations concerning this model in Moses
I. Finley, Ancient Slavery and Modern Ideology, ed. Brent D. Shaw (Princeton,
1998); C. R. Whittaker, Circes Pigs: From Slavery to Serfdom in the Later
Roman World, in M. I. Finley (ed.), Classical Slavery special issue of Slavery
and Abolition, viii (1987). See also Wickham, Framing the Early Middle Ages,
2623.
11
12
Hermann Nehlsen, Sklavenrecht zwischen Antike und Mittelalter: germanisches
und rmisches Recht in den germanischen Rechtaufzeichnungen (Gttingen, 1972), and
recently Detlef Liebs, Sklaverei aus Not im germanisch-rmischen Recht,
Zeitschrift der Savigny-Stiftung fr Rechtsgeschichte: Romanistische Abteilung, cxviii (2001),
are both examples of German historians trained as lawyers, and their approach to
unfreedom corresponds to an essentially legalistic denition.
13
See Wickham, Framing the Early Middle Ages, 55966.
14
Orlando Pattersons Slavery and Social Death: A Comparative Study (Cambridge,
Mass., 1982) was particularly inuential in this respect.
15
Duby considered the status of the tenth-century serf to have been essentially
dened through exclusion from any kind of participation in any community or
judicial institution (lexclusion du peuple franc), arguing that the collapse of these
institutions in the eleventh century was the only thing that could do away with this
status: Georges Duby, La Socit aux XIe et XIIe sicles dans la rgion mconnaise
(Paris, 1953), 110, 210.
16
In France, the work of Barthlemy in particular has gone a long way towards
re-evaluating the notion of total social exclusion envisaged by Duby and Bonnassie:
see Dominique Barthlemy, Quest-ce que le servage, en France, au XIe sicle?,
Revue historique, cclxxxvii (1992); Dominique Barthlemy, La Mutation de lan
mil a-t-elle eu lieu? Servage et chevalerie dans la France des Xe et XIe sicles (Paris,
1997).
12
NUMBER 193
17
An example of the strength of the idea that bad treatment and lack of free
movement would have been dening elements of slavery, and of slavery alone, is
the case of the separated families in Gregory of Tours, Histories, VI. 45. Thorpe,
Gregorys translator, inserted the word serf when all that the text actually says is
that these were families from the royal estates: History of the Franks, trans. Lewis
Thorpe (Harmondsworth, 1974), 377. The further mention that some were of
good birth (meliores natu), owned property and were able to leave much of it to
the Church casts further doubt on his interpretation. Here Thorpe concluded,
simply from the way they were being treated, that these were serfs families, even
though this is not a tale of routine oppression; the event is presented as exceptional, and Gregory clearly thinks that this was an abuse of power on the part of
Chilperic, who was not his favourite Merovingian. Thorpe admittedly was not a
historian, but his translation of Gregory remains the standard one, and his reading of this passage has been inuential proof that, in English as in Latin, we
need to be careful with words: see, for instance, R. Samson, Slavery, the Roman
Legacy, in John Drinkwater and Hugh Elton (eds.), Fifth-Century Gaul: A Crisis
of Identity? (Cambridge, 1992), 221. The standard French translation of Gregory
has the non-committal domestiques: Grgoire de Tours, trans. Robert Latouche,
2 vols. (Paris, 19635).
18
Though Goetzs recent survey of the evidence for early medieval unfreedom
includes a good but necessarily brief section on legal formulae: Hans-Werner
Goetz, Serfdom and the Beginnings of a Seigneurial System in the Carolingian
Period: A Survey of the Evidence, Early Medieval Europe, ii (1993).
19
Marc Bloch, Comment et pourquoi nit lesclavage antique, Annales ESC, ii
(1947), 37; Bonnassie, Survie et extinction du rgime esclavagiste, 3289; Liebs,
Sklaverei aus Not, 30111.
13
II
UNFREEDOM IN THE LIGHT OF FORMULARY EVIDENCE
The standard edition is Formulae Merowingici et Karolini aevi, ed. Zeumer; see
also Karl Zeumer, ber die lteren frnkischen Formelsammlungen, Neues
Archiv, vi (1881); Karl Zeumer, ber die alamannischen Formelsammlungen,
Neues Archiv, viii (1883). Further editions of individual collections include
Raymond Poupardin, Fragments du recueil perdu de formules franques dites
Formulae Pithoei, Bibliothque de lcole des Chartes, lxix (1908); Bernhard
Bischoff, Salzburger Formelbcher und Briefe aus tassilonischer und karolingischer Zeit
(Munich, 1973); Marcul formularum libri duo, ed. Alf Uddholm (Uppsala, 1962).
Many of the manuscripts are admirably described in Hubert Mordek, Bibliotheca
capitularium regum Francorum manuscripta: berlieferung und Traditionszusammenhang der frnkischen Herrschererlasse (MGH, Hilfsmittel, xv, Munich, 1995). Studies
of individual formularies include Werner Bergmann, Die Formulae Andecavenses,
eine Formelsammlung auf der Grenze zwischen Antike und Mittelalter, Archiv fr
Diplomatik, xxiv (1978); Philippe Depreux, La Tradition manuscrite des Formules
de Tours et la diffusion des modles dactes aux VIIIe et XIe sicles, in Philippe
Depreux and Bruno Judic (eds.), Alcuin: de York Tours: criture, pouvoir et rseaux
dans lEurope du haut moyen ge (Rennes, 2004). The debate over the dating of
Marculf accounted for a large part of the historiography in the rst half of the
twentieth century. The classic synthesis on formulae is Heinrich Brunner, Deutsche
Rechtsgeschichte, 2nd edn, 2 vols. (Leipzig, 1906), i; and Rudolf Buchner, Deutschlands
Geschichtsquellen im Mittelalter: Vorzeit und Karolinger, suppl., Die Rechtsquellen
(Weimar, 1953), 4955, has remained the best (and most recent) overview of the
subject. Whereas formulae had been very visible in general works and textbooks of
the nineteenth and rst half of the twentieth century, they became an increasingly
obscure source as the twentieth century progressed, and they are hardly present at
all in recent textbooks.
14
NUMBER 193
wording and content, any two documents of the same type can
look very similar without necessarily being textually related. Nor
should we expect to nd an exact match between any document
and a formula: a formula would usually have required a fair
amount of adaptation to t new circumstances. Even if secure
textual links could be made, it would be in any case virtually
impossible to tell whether the formula was modelled on the document or the document on the formula.21
Equally troubling is the general uncertainty as to how long
formulae can be said to be relevant after their supposed date of
composition: in other words, their use-by date. All the surviving manuscripts were produced during the Carolingian period,
though many of the collections are thought to have been compiled earlier. As was normal at the time he was working, Karl
Zeumer, who produced the main edition of formulae, was not
primarily interested in the Carolingian end of this story, and
instead focused on reconstituting an original version.22 As a
result, his edition deliberately obscures the uidity of formulae
as texts. They clearly remained work in progress down to the
ninth and tenth centuries: individual formulae from earlier
collections were still being modied, abridged, lengthened,
reorganized and integrated into new collections during the
Carolingian period. As Warren Brown has recently suggested,
this indicates that they were still being used actively as models,
21
Many of the dates for the original compilation of collections suggested by
scholars were based on the dates of surviving documents which they considered,
rightly or wrongly, to have been the original source of a formula included in the collection. Efforts to link formulae with surviving documents include Zeumer, ber
die lteren frnkischen Formelsammlungen; Bruno Krusch, Ursprung und Text
von Marculfs Formelsammlung, Nachrichten von der Kniglichen Gesellschaft der
Wissenschaften zu Gttingen, Phil. hist. Klasse (Berlin, 1916); Wilhelm John,
Formale Beziehungen der privaten Schenkungsurkunden Italiens und des Frankenreichs und die Wirksamkeit der Formulare, Archiv fr Urkundenforschung, xiv (1936);
Heinz Zatschek, Die Benutzung der Formulae Marcul und anderer Formularsammlungen in den Privaturkunden des 8. bis 10. Jahrhunderts, Mitteilungen des
Instituts fr sterreichische Geschichtsforschung, xlii (1937). Identications are also
made in Die Urkunden der Merowinger, ed. Theo Klzer (MGH, Diplomata regum
Francorum e stirpe merovingica, 2 vols., Hannover, 2001), i; and Brigitte Kasten,
Agrarische Innovationen durch Prekarien?, in Brigitte Kasten (ed.), Ttigkeitsfelder
und Erfahrungshorizonte des lndlichen Menschen in der frhmittelalterlichen Grundherrschaft (bis ca. 1000): Festschrift fr Dieter Hgermann zum 65. Geburtstag (Stuttgart,
2006).
22
This is the common complaint of all who have used the Monumenta edition
alongside an existing manuscript: see, for instance, Brown, When Documents Are
Destroyed or Lost, 354.
15
and were not merely a vestige of earlier legal traditions.23 Evidence for continuous adaptation in the manuscript tradition is
a good sign that a formula was still alive. The idea that formulae could remain in use in the very long term is supported by
comparison with documents from the Carolingian period: a
document of Prm from 804, involving land in the region of
Angers, thus shows striking similarities with formulae from the
late sixth-century Angers collection.24
If formulae and formularies were transferable through time,
they were also transferable through space, and collections clearly
spread out beyond the region in which they had originally been
produced.25 The fact that formulae originating from a certain
area could also be used in another suggests that their content was
not very locally specic. The models would not have needed to t
local circumstances perfectly in any case, since differences could
be resolved in the process of drawing up the actual document.26
Looking to tie legal formulae down to one place and time is
therefore a futile task as well as a thankless one. In a sense it is
also reductive. The transferability of formulae should be seen
as their strength as a source, not their weakness. We should
take advantage of the fact that their raison dtre is to draw the
general out of the specic, and to transcend local contexts.
These form-documents cannot be interpreted in relation to
particular circumstances in the way that actual charters can,
and this has created their reputation as a frustrating source to
deal with. To wish that they were more like charters, however,
is to miss what makes them unique. What formulae have to
23
16
NUMBER 193
17
Pactus Legis Salicae, XIII. 8, XXV. 4 (ed. Eckhardt, 61, 94). The results of such
marriages in other early medieval codes vary widely in harshness: see, for instance,
Lex Ribuaria, LVIII. 16 and 18, ed. Franz Beyerle (MGH, Legum I, iii, 2, Hannover,
1954); Liber Iudiciorum: sive, Lex Visigothorum (Leges Visigothorum), III. 2. 3, ed.
Karl Zeumer (MGH, Legum I, i, Hannover, 1902); ibid., III. 2. 7 seems to suggest
that some masters were fraudulently marrying off their slaves to free persons in
order to acquire them as slaves as well! See also Leges Burgundionum, XXXV. 2 and
3, ed. Ludwig-Rudolf von Salis (MGH, Legum I, ii, 1, Hannover, 1892); Leges
Alamannorum, LVII, ed. Karl Lehmann (MGH, Legum I, v, 1, Hannover, 1888).
30
For example Capitularia regum Francorum, ed. Alfred Boretius (MGH, Legum
II, i, Hannover, 1883), no. 142, cap. 3, p. 292 (a. 819).
31
Formulae Andecavenses, no. 59; Cartae Senonicae, no. 6; Collectio Flaviniacensis,
no. 102 (= Marculf, II. 29); Formulae Salicae Merkelianae, no. 31; Formulae Salicae
Bignonianae, no. 11; Formulae Salicae Lindenbrogianae, no. 20; Formulae Morbacenses,
nos. 18 and 19; Formulae Augienses Coll. B, no. 41.
32
Marculf, II. 29.
18
NUMBER 193
19
One actual case of intercession in a similar situation is envisaged in the letter sent by Einhards wife Emma to Blidthrut,
written between 828 and 836. Emma asked Blidthrut, a nun at
the convent of Mosbach, to intercede with Albuin, whose servus
had married a libera femina, so as to get him to allow this woman
and any future children to remain free. Blidthruts permission
was also needed, because the servus belonged to Mosbach as
well as to Albuin. The servus is said to have ed to the nearby
church of Saints Marcellinus and Peter at Seligenstadt: the
arrangement was clearly sought after the event.36 Although this
arrangement would probably not have been to the advantage of
the servuss master, it would not have been too much of a disadvantage either, since the marriage had probably been between
people of the same local community and would therefore not
necessarily have disrupted relations between local landowners.
Different localities may have had different attitudes and
customs regarding such unions, and been more or less tolerant
of them: the Saint-Germain-des-Prs polyptych seems to show
mixed couples as quite common on some estates but entirely
absent on others. Local variation therefore needs to be emphasized in this context. Other variables could include the lords
power of enforcement. The very different situation evidenced
in the eleventh-century Book of Serfs of Marmoutier, in which
the monastery as a rule insisted that, in case of marriage between
free and unfree, the spouse and any future children should
automatically become of servile status,37 does not necessarily
indicate a change in overall practice; it may simply reect the
monasterys ability to enforce its demands. The possibility of a
discontinuous evolution is worth stressing.38
Formulae will not tell how common or how rare this solution
ever was. Nevertheless, it is surely signicant that formularies
consistently cover the case of marriages between servi and free
women occurring without adverse consequences: formulae of
this kind survive in eight different formularies in very different
forms, which implies that different models were being devised
36
20
NUMBER 193
Capitularia regum Francorum, ed. Boretius, no. 58, cap. 8, pp. 1456 (a. 80114).
This indicates that such documents were not only in use but had also caused
disputes, which in turn suggests a level of negotiability as to the fate of children in
such cases: see Janet L. Nelson, England and the Continent in the Ninth Century:
III, Rights and Rituals, Trans. Roy. Hist. Soc., 6th ser., xiv (2004), 910. On the
term franca femina in this capitulary text, see n. 54 below.
40
Wickham, Framing the Early Middle Ages, 405, 5601. Wickham has argued
convincingly that the number of freeunfree marriages in the Carolingian polyptychs reects a situation in which the gap in status would have been less keenly felt
because peasants all tended to be tenants either way, as opposed to areas where
more owned land in their own right, such as northern Italy: ibid., 405. For Italy, see
also Codice diplomatico longobardo, ed. Luigi Schiaparelli, 5 vols. (Rome, 1929
2003), i, nos. 29, 53; ii, nos. 204, 274; iii, no. 18.
41
For example Formulae Salicae Bignonianae, no. 20: mancipiis, ingenuis his
nominibus vel quicquid ibidem presens est mea dominatio (the unfree and free
tenants named XY, or whoever lives there under my jurisdiction). See also, in
Marculf, I. 2, the mention of the dues de eorum [i.e. the monasterys] hominebus
aut de ingenuis aut servientes in eorum agros conmanentes (of their men, free and
unfree, who reside in these elds), which again suggests mixed status among rural
tenants. For a detailed study of estate descriptions in legal formularies, see Dieter
Hgermann, Einige Aspekte der Grundherrschaft in den frnkischen formulae und
in den leges des Frhmittelalters, in Adriaan Verhulst (ed.), Le Grand Domaine aux
poques mrovingienne et carolingienne / Die Grundherrschaft im frhen Mittelalter
(Ghent, 1985). On the general ambiguity in the terminology of freedom and
unfreedom in early medieval sources, with social practice affecting legal meanings,
see Goetz, Serfdom and the Beginnings of a Seigneurial System, 49.
21
not have been deeply conscious of legal distinctions: if anything, the fact that such documents were produced indicates
that the parties involved understood what was at stake, and
negotiated accordingly. This type of formula therefore conrms
the theory, put forward by Emily Coleman on the basis of the
Saint-Germain-des-Prs polyptych, that marriage to free women
could have been part of a conscious bid on the part of male
servi to improve their own status as well as that of their children, something that Gurard, the rst editor of the polyptych,
had in fact suggested as early as 1844.42 The fact that such
marriages appear in formulae from as early as the sixth century
down to the ninth, by which time they are also documented in
the polyptychs, also suggests that hypergamy was a feature of
peasant life in some areas of West Francia across a very extended
period. By contrast, there is only one example in the formulae
in which the child of a free man and his ancilla obtains free status,
in what would seem to be an unusual situation (a free father
manumitting his daughter in order to make her his universal
heir).43 This is different from situations involving servi and free
women, most obviously because here the relationship is not
sanctioned by marriage. Ancillae do not seem to have had much
opportunity to marry upwards, and hypergamy is thus likely to
have only worked one way.44
Formulae therefore contradict the evidence of the law codes,
though we should not necessarily assume that mixed marriages of this kind would have been totally unthinkable even
under Roman imperial law. A major problem with Bonnassies
42
See Coleman, Medieval Marriage Characteristics; Benjamin Gurard, Prolgomnes, in his Polyptyque de lAbb Irminon (Paris, 1844), 3912.
43
Cartae Senonicae, no. 42 (= Formularum Codicis S. Emmerami Fragmenta, no. 7).
44
The fate of ancillae as opposed to servi has been the object of a debate between
Susan Mosher Stuard, Ancillary Evidence for the Decline of Medieval Slavery,
Past and Present, no. 149 (Nov. 1995), and Jean-Pierre Devroey, Men and Women
in Early Medieval Serfdom: The Ninth-Century North Frankish Evidence, Past
and Present, no. 166 (Feb. 2000). Stuard, focusing essentially on Mediterranean
areas and the later Middle Ages, argued that ancillae, unlike their male counterparts, retained the status of slaves throughout the medieval period, and she stressed
their reduced marriage prospects: Stuard, Ancillary Evidence for the Decline of
Medieval Slavery, 27. Devroey, on the other hand, argued that the meaning of the
word ancilla evolved in much the same way as that of servus, and the condition of
ancillae was equally diverse, with gender creating differences in the personal experience of unfreedom rather than in juridical status: Devroey, Men and Women in
Early Medieval Serfdom, 18. Whether or not the difference was juridical, it was
still clearly present.
22
NUMBER 193
23
Such as Codex Theodosianus, IV. 12. 1 (from the year 314), IV. 12. 6 (from 366);
or, for a more complicated situation, Novellae Valentinianae, XXXI. 1 (from 451),
included in Codex Theodosianus, ed. Mommsen and Meyer.
50
Codex Theodosianus, IV. 8. 3, model for Lex Ribuaria, LVIII. 11 and 14: generatio
eorum semper ad inferiora declinetur.
51
See Bonnassie, Survie et extinction du rgime esclavagiste, 320.
24
NUMBER 193
whatever place he wants to choose for himself under the power of Saint
E, and he is not to do the service of a freedman to any of my heirs or
indirect heirs, but, as I said, let him always be able to remain properly
free and secure.52
52
Formulae Salicae Lindenbrogianae, no. 9 (Munich Lat. 4650; see also Copenhagen
Gl. Kgl. Saml. 1943 for a slightly different version).
53
The idea that the unfree had no right to own property beyond the level of small
savings is traditionally seen as one of the crucial points of difference between freedom and unfreedom. That an unfree person could have rights over another conicts with modern denitions of slavery. There are mentions of servi owning ancillae
in Capitularia regum Francorum, ed. Boretius, no. 16, cap. 7, p. 40 (a. 758/68)
(Goetz, Serfdom and the Beginnings of a Seigneurial System, 378), although
here the situation should probably be explained more in terms of gender.
54
Francus homo and franca femina are difcult to translate; apart from ethnic
identity, itself problematic, the terms could also designate either a free person or a
particular and privileged type of dependant: J. F. Niermeyer and C. van de Kieft,
Mediae latinitatis lexicon minus, 2nd edn, revised J. W. J. Burgers, 2 vols. (Leiden,
2002), ii, 5901. See also Brunner, Deutsche Rechtsgeschichte, i, 351.
55
Formulae Senonenses recentiores, no. 3. Goetz describes this text as dealing with
the question whether a servus had the right to testify against a colonus: Goetz, Serfdom and the Beginnings of a Seigneurial System, 43. Exciting though this
would be, it seems to me that it is not actually the colonus who has ed, but the
slave. Naturally it is implied that the testimony of several free men outweighs the
testimony of one slave, but this in itself is hardly a surprise.
25
56
This is the case most clearly, for instance, in Formulae Senonenses recentiores,
no. 6. It is known in the writings of German Rechtshistoriker as a Scheinprozess: see
Fouracre, Placita and the Settlement of Disputes, 26 and n. 13.
57
Formulae Senonenses recentiores, no. 6.
58
Marculf, II. 36.
26
NUMBER 193
27
also be linked with a passage from Charlemagnes Capitula missorum, which mentions those servi who, as honorati, hold beneces
and have special duties (benecia et ministeria) or are honorati in
vassalage together with their lord, and may own horses, arms
and a shield and a wooden lance and a short sword.60 Perhaps
the recipient of the document laid out in this formula could
have qualied as what Patterson called an ultimate [i.e. elite]
slave.61 Even if this is not a typical instance, it is revealing of
the mentality of the period regarding the concept of service, a
sign that the opposition often made by modern historians
between the service of the unfree and that of vassals was not
such a signicant one in terms of mental and linguistic associations, regardless of differences in social status.62 The word
servus here clearly does not evoke a dishonourable situation:
rather one type of dependence within a general framework
based on relationships of dependence and protection. Material
rewards seem to have sometimes been granted in exchange for
unfree as well as free service, and in this context even full ownership of land was not totally beyond the reach of the unfree.
Two formulae from a ninth-century collection thus show us an
archbishop granting to a serviens of his church a mansus or an
urban property, in either case with the stipulation that the
beneciary would have the right to pass it on to his heirs, and
that the property could not be taken back from him or his
descendants, even if they were found slow or negligent in their
servitium.63
3) Self-Sale
Self-sale formulae are unusual in having received a fair amount
of attention from scholars in recent times.64 They have been
60
Capitularia regum Francorum, ed. Boretius, no. 25, cap. 4, p. 67 (a. 792 or
786).
61
Patterson, Slavery and Social Death, 299333.
62
On free and unfree service, see also n. 76 below. The vocabulary of service
is also found in very similar terms in a familial context (for instance with a wife
describing herself as her husbands ancilla in Marculf, II. 17).
63
Addenda ad Formulae Senonenses recentiores, nos. 18 and 19 (Formulae
Merowingici et Karolini aevi, ed. Zeumer, 7234).
64
This is the only type of formula that Bonnassie, Survie et extinction du rgime
esclavagiste, 3289, and Nehlsen, Sklavenrecht zwischen Antike und Mittelalter, 339,
referred to, and it has also been the subject of a more recent article along very similar lines: see Liebs, Sklaverei aus Not.
28
NUMBER 193
29
will have the right to inict corporal punishment as on your other servientes.
And when I can return your solidi, I will recover my deed of security,
without the need for a deed of annulment.71
The document seems to be issued by a third party, but comparison with the next formula shows that the third person could
71
30
NUMBER 193
be used at the same time as the rst person to designate the same
party:
As it is known that this man named A had a deed of security from this
man named B for his [free] status, whereby [A] paid [B] X ounces of
silver in order that, for Y years, [B] should perform whatever service
[A] demanded of him, and A was completely unable to nd this deed
of security, I therefore gave you this annulment by my hand and those
of good men, as if you had given me back my property: and if this deed of
security is found, let it remain void and useless, and let this deed of
annulment remain rm.74
31
76
Formulae Turonenses, no. 43. Gerd Althoff interpreted this formula as showing
a vassal entering an agreement with a lord, and as early evidence for the triumphal
march of the institution of vassalage: Gerd Althoff, Family, Friends and Followers:
Political and Social Bonds in Early Medieval Europe, trans. Christopher Carroll
(Cambridge, 2004), 104, but I see no basis for this. If anything, this formula indicates that lordship over the free was not as fundamentally different from lordship
over the unfree as Althoff suggests: ibid., 103.
77
This would explain why peasants are still so keen to negotiate and dene their
status in legal cases surviving from the Carolingian period: see Janet L. Nelson,
Dispute Settlement in Carolingian West Francia, in Davies and Fouracre (eds.),
Settlement of Disputes; Nelson, England and the Continent in the Ninth Century:
III, Rights and Rituals.
78
Formulae Andecavenses, no. 38.
32
NUMBER 193
33
Formulae Turonenses, no. 11, for instance, the fact that the infant
is intended to become a household servant is made explicit (si,
Deo presule, convaluerit, ipsum in suis servitiis ac solatiis iuxta
legis ordinem retineat).81 Formulae Andecavenses, no. 9, an act of
sale similar to the Marculf one given above,82 has the word vernaculus instead of servus. It seems highly likely that the word
vernaculus specically designates a household servant. The word
appears in sale formulae,83 as well as in many manumission formulae;84 it does not appear in any other type of document, and
never seems to be used to describe a rural tenant. In all such
cases, therefore, it makes sense to read the formula as relating
to a household slave. Domestic slaves, through constant contact with their master, would have been objects of closer attention than tenants, and that could have made them more likely
to be manumitted, as well as more liable to suffer violence.
Medieval domestic servi may therefore have been treated similarly to classical domestic servi.85 The important thing is not to
extend evidence which in all likelihood refers to domestic service to the life of rural tenants. Rights and conditions of life are
likely to have been determined by activities and circumstances,
rather than by legal status: free domestic servants, for instance
in their vulnerability to the threat of violence, would probably
have been similar to unfree domestic servants, just as unfree rural
tenants would have led similar lives to those of free rural tenants.86
The picture of early medieval unfreedom which we nd in
the formularies is therefore very diverse, and hardly one of total
social exclusion from any community or public institution. On
81
Liebs found this text particularly shocking, and thought it an example of the
lack of social solidarity in early medieval times: Liebs, Sklaverei aus Not, 311.
82
The title in the manuscript (Here begins the act of sale which one makes when
selling oneself) does not t the content of the formula.
83
Formulae Andecavenses, no. 9; Formulae Salicae Bignonianae, nos. 3 and 5.
84
Formulae Salicae Bignonianae, nos. 1 and 2; Formulae Salicae Merkelianae, nos.
13b and 43 (in which an infantulum is manumitted); Formulae Salicae Lindenbrogianae, nos. 10 and 11; Formulae Augienses Coll. B, no. 42; Formulae Codicis Laudunensis, no. 14; see also Formulae Salicae Lindenbrogianae, no. 9, quoted at n. 52 above.
85
Even in this context, however, we nd a variety of situations: the gasindus in
Marculf, II. 36 could for instance qualify distantly as a household dependant.
86
Although it is often assumed that some horror stories in our narrative sources
reect specically the violence of masters towards slaves, medieval authors rarely
bothered to make it clear whether the victims were technically free or unfree (for
instance in the case of the victimized pueri in Gregory of Tourss Histories), which
suggests that legal status was not the most important issue in practice.
34
NUMBER 193
the other hand, we still get a strong sense of its importance: that
a good proportion of all surviving formulae (over 15 per cent)
deals with issues involving servi suggests that degrees of personal freedom were of concern to both compilers and those
who needed their services. Formulae show that relationships of
dependence and rights over persons, as well as links of protection and patronage, were worked out in a more integrated and
negotiable way than tends to be allowed for, with a strong consciousness of rights on the part of our servi.87 As a result, status
and owed labour would have been all the more difcult to pin
down and reassert, and would have needed constant attention
in documents. Freedom and unfreedom were stages on a sliding scale rather than a clear-cut tiered system, or, as Susan
Reynolds has put it, the layers of [medieval] society were more
like those of a trie than a cake: its layers were blurred, and the
sherry of accepted values soaked through.88
III
UNFREEDOM AND THE LEGAL SOURCES
35
89
See at n. 73 above.
Pactus Legis Salicae, LIX. 6 (ed. Eckhardt, 223).
Again, see Formulae Salicae Lindenbrogianae, no. 20.
92
See Wormald, Making of English Law, 79: it seems that it was the general body
of Salic custom, not the actual text of the law, that judges and scribes had in
mind, and on formulae as the main point of contact between documentary practice
and written laws: ibid., 83. See also Patrick Wormald, Lex scripta and verbum regis:
Legislation and Germanic Kingship from Euric to Cnut, in P. H. Sawyer and I. N.
Wood (eds.), Early Medieval Kingship (Leeds, 1977), 122.
90
91
36
NUMBER 193
37
38
NUMBER 193
39
40
NUMBER 193
Alice Rio