Introduction

1By the late Middle Ages Catalonia was unusual among Mediterranean countries in having a substantial class of legally defined unfree persons. These remences, as they came to be called in the fourteenth and fifteenth centuries, comprised as much as half the rural population of Old Catalonia, the northern and eastern parts of the Principality, essentially the areas conquered from Islam during the era of Charlemagne that formed the Carolingian Spanish March2. Although not completely unknown in what came to be called New Catalonia (southern and western regions seized in the twelfth century), peasant servitude was uncommon in those districts (comarques) nearest to Old Catalonia and virtually non-existent in the regions surrounding the great cities of Tortosa to the south and Lérida to the west. Catalonia was unique among European realms in throwing off servile status by legal abolition in 1486, the result of a successful peasant insurrection allied with the royal forces against a portion of the nobility in the long Catalan Civil War.

2Investigation of servile anthroponyms in Catalonia therefore is significant not only within a comparative European or Mediterranean framework for determining possible links between social status and naming patterns but for the question of when, how (and ultimately why) a distinct class of unfree persons was identified. What in the tenth century had been a frontier society settled by free tenants and even allodial small holders was by the thirteenth century organized into a seigneurial regime that forced peasants to acknowledge a servile social condition symbolized most clearly by the payment of a manumission fine to depart from the lord's jurisdiction3. This redemption payment would give its name, redimentia, to the late medieval Catalan term for serf, remença. According to jurists and legal codifications of the late Middle Ages, remences were also liable for certain payments concerning inheritance, adultery, and other circumstances affecting the lord's interest in the properties they held. Lords also possessed certain rights to deal with their servile tenants, even in an arbitrary fashion, without the intervention of public courts or royal jurisdiction.

3One can trace the beginnings of this seigneurial regime to the political and social upheavals of the eleventh century as well as to the imposition of a more formal legal procedure for defining servile status beginning in the late twelfth century. This does not mean that those so defined in the period under consideration were necessarily constrained in the manner that ultimately provoked the late-medieval rebellion. The presence of numerous documents recording the payment of redemption in order to change lords, marry off the property, or to leave the region altogether indicate both the extent to which servitude had become routinized by the early thirteenth century and the ability of remences to manipulate the system or at least find some opportunities for movement within it4.

4These documents, in which a lord acknowledges receipt of payment from a peasant and recognizes his resulting freedom are the basis for our assessment of servile names. Other charters involving unfree persons, such as recognitions of lordship or placement into the hands of a lord, tend to show similar results but have been excluded here because at the time such records were drawn up those entering into servitude might not previously have been serfs. Recognitions of lordship are meant to demonstrate, but do not by any means prove, an already existing bond. Linked to inheritance or perhaps to the ability of lords to demand formal acknowledgement of a previously ambiguous status, recognition charters often involved peasants at a moment of transition into servile status rather than those who can more clearly be said to possess such status at the time the document was drawn up. Servile names can be assumed to apply to those making redemption payments (exiting from servitude or changing lords) with somewhat more confidence than to those making recognitions or commending themselves at the moment of entrance into servitude, persons who had perhaps been free until recently and who bore their names as free men and women before the document was drawn up.

5A sampling of unfree names has been amassed from a variety of archival and published sources for the century after redemption charters first began to appear, from 1180 until 1283. The latter date is chosen because the Catalan representative assembly or Corts meeting that year in Barcelona effectively defined servitude (in the opinion of later jurists) by prohibiting the free movement of seigneurial tenants onto royal land unless a redemption payment was made5. The records come from the regions of Old Catalonia where servile status would remain most widespread until the fifteenth-century revolts: the dioceses of Girona, Vic, Elne, parts of Barcelona and Urgell. They are supplemented by the first book of the notarial registers of Vic which deals with the years 1230-1233.I owe a great debt to the detailed analysis of the contents of this register undertaken by Rafel Ginebra i Molins6. He has listed all the redemption documents and the names of those purchasing their freedom found in this volume. I have considered these separately from other records as the brevity of notarial entries might affect how names are entered, but in fact the Vic notarial documents confirm what is found elsewhere.

6Women figure quite frequently in documents of redemption, often making rather small payments to obtain permission to marry out of the lord's jurisdiction. By 1300 in the diocese of Girona the sum for young women leaving to marry would be fixed at 2 sous 8 deniers7. However, as Catalan women had not experienced to the same degree the change towards double names, they are difficult to compare with male naming patterns which present greater variety and complexity. I have therefore concentrated on male names for which the documentation assembled provides evidence concerning approximately 200 unfree persons (including the Vic notarial entries).

7In order to determine whether or not there was any association between servile status and personal naming patterns, comparison has been made with free peasants who appear in charters from the same chronological period and archival sources other than the Vic notarial archive. In most transactions such as sales of land it is not clear what the status of the principles was — whether free or unfree or even whether or not they were peasants or landowners. As evidence for free status I have chosen documents in which landlords establish new tenants on land on conditions free of servile payments. Some of these are what would later be termed emphyteutic leases in which vacant land was to be developed in return for a fairly small annual rent (census). The tenant could alienate the property subject to the lord retaining ultimate control and receiving certain additional payments. These leases thus clearly involve peasants of free condition as opposed to transactions among substantial property holders or servile establishments. Slightly more than 150 documents have been considered in this category, all again involving male recipients.

8Finally, for purposes of chronological comparison, I have taken advantage of a recent article by Manuel Sánchez Martínez on complaints by the inhabitants of four parishes in the area of Castellfollit de la Roca (east of Girona, the comarca of La Garrotxa)8. In a document drawn up between 1331 and 1335 edited by Sánchez Martínez, 186 individuals complained to King Alfonso the Magnanimous of violent acts of extortion and confiscation of movable property by the viscounts of Bas, lords of Castellfollit. The wholesale pillaging of tenants may be, as Sánchez Martínez argues, an instance of the seigneurial right of mistreatment (ius maletractandi) by which the royal authorities were prohibited from interfering even in the unjust exercise of arbitrary lordship over unfree tenants.

9It can not be absolutely proven that the homines of Castellfollit who appear in the document were serfs, but the apparent exercise of the right of seigneurial mistreatment and also the nature of tenancy in La Garrotxa, among the most thoroughly enserfed regions of Old Catalonia, makes it more than likely that they were remences. At any rate, the purpose of the comparison will be not so much to identify servile names of a later period but to indicate some changes for peasant names generally.

Overview of Catalonian Anthroponyms

10Before the year 1000 personal names in Catalonia were in most instances composed of a single term. An analysis of tenth-century documents (mostly from Vic and Girona) has found that single names formed 89% of the anthroponyms9. The variety of names was very great as has been demonstrated by the recently published first volume of the Repertori d'antropònims catalans (RAC) compiled by Jordi Bolòs and Josep Moran10. This monumental work extracts a total of 39,159 individuals from records of the ninth and tenth centuries. There are some 4,250 names grouped into categories, so that, for example, the relatively common name Wifredus (485 examples) includes nearly 40 variations: Giffridus, Guifreda (a female name), Quifredus, Huifredus, Guidgafredus, etc.11.

11For the tenth-century documents examined by Lluís To Figueras, the ratio of individuals per name was below 1.5. Thus in any random group of 100 individuals there would on average be found at least 70 different names. There were, to be sure, some popular names that were repeated. The most common name before 1000 was Sunifredus (in 72 varieties), mentioned in 905 instances (but amounting to merely 2.3% of the total), followed by Miro (531), Wifredus (485), Suniarius (469) and Oliba (464)12. Despite the almost infinite variety of names, there are a certain number of them repeated not only within villages (possible family names) but between widely spaced settlements. Of 180 masculine names that appear in a document of 913 from the monastery of S. Joan in the Pyrenees (the future Sant Joan de les Abadesses), 33 also appear in a record of 916 from coastal Empúries at the other geographical extreme, while 28 of them are matched at Artés on the Islamic frontier in 93813.

12There was something of a fashion for Visigothic/Germanic names in the tenth century. Early ninth century documents show almost exclusively Latin names but the Cartulary of Sant Cugat shows merely 20 to 35% of names were Latin in the first decade of the tenth century. By 1000 almost all names appearing in records of the monastery were Germanic14. A similar partiality towards Germanic names is found contemporaneously in Galicia and Portugal15. None of this shift from Latin to Germanic names can be related to migration or any other ethnic movement or upheaval, a reminder of both how volatile naming patterns have been and how their shifts may be unrelated to measurable social change.

13In the eleventh century the stock of names in Catalonia was reduced and double names of various forms became common without completely displacing single names. The reduction in naming stock is first noticeable in the early eleventh century. By 1060 the ration of individuals to names was above 2.0; for every 100 persons there would be now 40 rather than 70 names. The concentration of names accelerated after 1060. The ratio was 3.2 in the late eleventh century. By the end of the twelfth century, the ratio was 9.7 (for 1176-1200)16. The names Raimundus, Guillelmus, Bernardus, Petrus, and Berengarius achieved a striking hegemony. By the end of the twelfth century, nearly 83% of all men named in the documents studied by Lluís To bore one of the nine most popular names17.

14At the same time that the variety of names was reduced, the practice of using two names becomes more common. Dual names could take several forms: 1) an alternate common name, rendered in documents as «qui vocant», thus «A qui vocant B»; 2) the nomen paternum, «A son of B»; 3) craft names (faber, texidor); 4) double names with the second either in the nominative or the genitive (the latter possibly a form of nomen paternum). As a broad generalization, the alternate common name is the first form of two element name, appearing in the late tenth century. The others become widespread more suddenly, after 1030. The presence of the second element in personal names would become very common without quite displacing altogether the single name, a development constituting what has been called the «anthroponymie revolution» beginning in the eleventh century18. Catalonia and Languedoc both experienced a rapid acceptance of dual names, which accounts for the revolutionary nature of the change19. At the same time Catalonia was slower to adopt the practice than Navarre and the Rioja, peculiarly precocious regions with significant basque influence, where double names were common by the ninth century20.

15To what extent are the reduction in the stock of names and the addition of a second naming element related? Enric Moreu-Rey asserted that the predominance of second names in the last two-thirds of the eleventh century was a consequence of the impoverishment of the anthroponymic repertoire, making distinctions necessary among the many persons named Petrus, Berengarius etc21. This view is rejected by Michel Zimmermann who argues that the decline in the variety of available names was a consequence rather than cause of the practice of adding a second name22.

16During the twelfth century (more specifically, after 1125), the use of a place-name as the second element of a double name came to predominate, replacing (although not completely effacing) the previously dominant patronymic. The single name continued to decline among males but again, without being completely eliminated23. For the last quarter of the eleventh century the single name constitutes 14.7% of masculine names while the patronymic represents 72.9%. The toponymic name is virtually unknown (1.4%). A similar distribution pertains to the early twelfth century. From 1125 to 1150, however, the single name gains slightly (22.7%), the patronymic declines notably (39.1%) and the place names jump suddenly to 19.4%. By the last quarter of the twelfth century the toponymie name would form the majority: 62.9% with the single name fading to 6.5% and the patronymic continuing its decline to 14.2%.

17Two related social forces have been adduced by way of explanation for the overall shifts in naming patterns: the so-called «feudal revolution» of the eleventh century and the triumph of lineage over extended kinship24. Both factors can be linked to names by positing changes in inheritance practices. The assertion of aristocratic power over the countryside produced a greater degree of social control which included a tighter regulation of succession practices among tenant farmers culminating in the institution of servile status that, among other things, guaranteed stable tenancy of seigneurial property or fines in the event of departure, female adultery or death without heirs. Definition of peasant families would seem to further closer seigneurial control over land and revenues. In fact, however, the identification of family names would be somewhat slower among the peasantry than the other social groups, so that the desire to identify lineage must have had greater urgency for aristocratic families25.

18The rise of the aristocracy was accompanied by (and related to) a change in family structure from an extended kinship network to a lineage with restricted ties to other branches and a consequent emphasis on patrimony. In Catalonia the equal division of inheritance among children characteristic of Visigothic Law yielded in the eleventh and twelfth centuries to the exclusion of daughters (who now received dowries as anticipatory shares of the patrimony) and a tendency toward primogeniture. The new naming patterns identified aristocratic lineages over time by means of double names, patronymic (identifying filiation) or later toponymic (identifying territory, usually a castle).

19Thus far much of what has been described is similar to other regions of Iberia: the replacement of the single name and an immense inventory of possible names by the double name and a corresponding concentration of first names. As noted Navarre, the Rioja and some neighboring regions tended to produce double names in the tenth century in contrast to the rest of the Peninsula while Castile saw an early popularity in the use of the form «A son of B»26. In the eleventh century, with the domination of the nomen paternum, the Iberian lands were most closely in tandem. With the twelfth century Catalonia became more clearly distinct within an Iberian context. While the nomen paternum persisted elsewhere, in Catalonia it declined during the twelfth century in favor of the toponym. The turn away from the nomen paternum distinguished Catalonia in the twelfth century from most of the rest of the Iberian Peninsula, especially the west (form Rioja to Portugal) where it remained dominant27.

20For the tenth to twelfth centuries, the pattern of naming in Catalonia resembles that found in Languedoc and Gascony more than the rest of Iberia28. In both Catalonia and Languedoc single Germanic names prevailed in the tenth century and would be replaced by double names in the eleventh century (although this change occurs earlier in Languedoc). The stock of names was reduced in the eleventh century and the most popular names were similar although not identical. Berengarius was popular in Catalonia but not in Languedoc while the widespread use of Pondus in Languedoc was not quite matched in Catalonia. The introduction of double names was also similar with the nomen paternum giving way to toponymic names. Greater differences between Catalonia and Languedoc would arise in the thirteenth century: the persistence of single names in Catalonia, the relative scarcity in Catalonia of personal surnames (including but limited to the nomen paternum), and in Catalonia the relatively frequent use of complex names (those with more than two terms, often combining a personal surname and toponymic designation) beyond aristocratic circles.

First Names

21Names of four different origins have been used for this study: 1) names of servile tenants throughout Old Catalonia; 2) mames of free tenants receiving leases, again for all available parts of Old Catalonia; 3) names of servile tenants mentioned as paying redemptions in the first volume of the Vic notarial records; 4) names of those who release them from servitude (thus members of the free landed elite or officials)29. The two latter are taken from the doctoral thesis of Rafael Ginebra which reports the result of an exhaustive and exemplary analysis of this source, an indication of the extraordinary depth of material contained in the early records of the Vic notarial archive. It should be pointed out that the first two lists include documents from Vic outside the notarial records.

22The fundamental result of a comparison between free and servile names is negative: there is scant correlation between naming patterns and social or legal status. The female name Barcelona and the male Domenicus are the only ones that appear to be limited to those of servile condition. Otherwise serfs did not have distinct names that marked them as unfree but rather shared most of the anthroponymic characteristics of their social superiors. There is little variation over the course of the period covered by the documentation in any of the aspects of naming patterns, including the relative absence of distinctly servile names.

23There are some nuances in the frequency and distribution of names that inform the general observation made above, particularly among the Vic notarial records. While the servile names in the Arxiu de la Cúria Fumada exhibit no significant differences from those elsewhere, those of high status who freed serfs in return for redemption payments do seem to have certain traits not shared with other groups. Joan (John), which figures among the most common names in Catalonia at the end of the twelfth century, is completely absent from the names of the Vic elite. Berengarius and Bernardus, also common across Catalonia at this time, exert a certain dominance within this category in which Ferrarius and Poncius are also common. It therefore seems to be the upper levels of society that are distinct from the ranks of both serfs and free peasant proprietors.

24The distinct feature of the servile population is not so much a stock of first names uniquely theirs as a relatively high level of name variation30. The stock of names used by serfs is at every point greater than that of the other social groups thus the repetition of names (homonymy) is relatively low. For Old Catalonia as a whole of 188 names, 31 appear only once («unique names» or «hapax»). Here too, the Vic notarial material is similar. Although the Vic serfs have a stock of fewer names, (and at 4 markedly fewer unique names), their rate of homonymy is virtually the same (0.108 for Vic, 0.114 for Old Catalonia generally). The greater variation may be less a feature of serfs than its opposite (limiting the choice of names) might be related to hereditary transmission of property among free families and a corresponding tendency to identify certain familial first names31.

25The one common first name that seems over-represented among serfs (if not amounting to an exclusive association with servile status) is Petrus. Even here, however, there is a certain vagueness in distinctiveness as Petrus is also to be found reasonably often among the higher social orders while rare among those of simply free condition. For free peasants Bernard and John are more common than for the other social levels. In Gascony, as in Catalonia, Petrus was associated with serfs, although there John and Martin also function as servile names unlike Catalonia32. For the Vic notarial documents, Arnau has a certain frequency but this is not the case for the rest of Catalonia.

Surnames

26In discussing surnames we might begin with the observation that single names persisted more among serfs than the other classes. Serfs more frequently dispensed altogether (as far as written documents are concerned, at least) with any surname. When servile status can first be identified, at the end of the twelfth and beginning of the thirteenth century, the single name survives among serfs but has been superseded among the rest of society. At this point one might consider the survival of single names simply a sign of a slower evolution of change among serfs. In the middle of the period under consideration, 1230-1260, both servile and free populations have the same fairly low percentage of single names (8% — excluding the Vic notarial documents which may be affected by the need for brevity). Later, however, serfs again exhibit a greater tendency towards single names.

27Catalonia is unusual in the rarity of rural surnames based on craft or other occupation. While such names were reasonably common among urban populations and comprise a large number of the most common contemporary Catalan surnames, they make up only 10% of surnames for our documentation and there is no variation between serf and free, except for the complete absence of craft names among the Vic notarial material. Sobriquets too were uncommon among all elements of society, slightly rarer among serfs, and again completely absent for Vic serfs. In both cases this could be due to a desire of the notary to abbreviate, except that as we will see, the toponymic surname is quite frequently employed in these records so that any general rule of abbreviation seems unlikely.

28The dominant categories of surnames are compound names and toponyms. Compound names are of course difficult to evaluate because it is often uncertain whether or not the second name refers to a parent (hence to be considered a family name albeit with a merely one-generational retrospective horizon). The clearest example of filiation is the designation of an individual as «A son of B» and this is significantly more characteristic of the servile than the free population. It declines after about 1230 but persists in complex names in which a surname (usually something approaching a family name) has additionally the name of the parents, e.g. Bernardus Thome son of Guillelma Thome; Ferrarius de Lobatera son of Petrus de Lobatera and his wife Guillelma; Petrus de Rovira son of Deuslosal de Rovira33.

29Also associated slightly more with servile than free status is a surname formed by an anthroponym in the genitive case (Guillelmus Fortoni; Berengarius Renardi)34. For the period 1205-1260 such names represent 15% of the free population, 20% for the servile population both of Catalonia generally and virtually the same (21%) for the Vic notarial records. Whether or not such names are explicit evidence of filiation, they must indicate a sense of ancestry and familial identification (nomen paternum). If we take both of the above mentioned categories to indicate some species of filiation we find that from 1200-1230 the difference between serfs and free men is marked: 36% and 21% respectively. After 1230 filiation is less common but the greater frequency among serfs is maintained (26% and 10%).

30Toponyms as surnames are quite common in the thirteenth century and not at all peculiar to serfs. They are most common among the higher order of society. Fully 84% of the notables of the region of Vic possessing any sort of surname are designated by toponyms. The comparable figure for free men outside Vic is 54% (between 1250 and 1260) and 33% for serfs outside Vic. For serfs whose redemption are preserved by the Vic notaries, 48% of those with surnames are designated by toponyms.

31By the term toponym we include names of towns, villages, castles or other specific sites. Some surnames designate features of the landscape that might be characteristic of any place: «de Plano», «de Carreria», «de Fonte». For the thirteenth century these are not rare but much less frequently to be encountered than toponymic surnames narrowly defined. Not surprisingly they are wholly absent from the notables of Vic who would be associated with a castle or settlement rather than a generic term. For the Vic serfs it is also uncommon, a mere 1.8% and 4.5% for serfs outside Vic. Among free men such terms amount to 11%. One of the striking features about later records, such as the Castellfollit document to be discussed shortly, is the much greater frequency of such landscape terms which are also among the most common modem Catalan surnames (Font, Pla, Bosc, Puig).

32What is perhaps most surprising is the absence of names of particular manses. Although serfs are sometimes described as being from a particular manse (de manso de...), their surnames do not reflect identification with tenancy35. Indeed, as we have seen, there is very little in naming patterns unique to the servile population or even especially characteristic. The «anthroponymic revolution» affected equally all levels of society well into the thirteenth century: reduction in the stock of names, decline of the single name, triumph of toponymic surnames. It would therefore be hard to argue that servile names at this stage show the impress of seigneurial domination or reflect a campaign of social control except insofar as the organization of society (and especially of inheritance) encouraged familial identification a t all social levels.

The Fourteenth Century

33The situation appears somewhat changed by the early fourteenth century. Tenants of the viscounts of Bas from four parishes in the castle-district of Castellfollit complained to the king of arbitrary and violent exactions, largely in kind, seized by the viscounts who were also lords of the dramatically situated fortress that crowns a high, almost free standing plateau in the sub-region (comarca) of La Garrotxa. This record, preserved in the Archive of the Crown of Aragón, has been edited and commented on by Manuel Sánchez Martínez who has been able to date it from between 1331 and 1335.

34The victims of this seigneurial extortion are not stated to be either of servile or free condition. By the fourteenth century the Garrotxa was predominantly an area of servile tenements36. Reinforcing the impression of servitude is the relation between the seizures of cereal, clothing, tools and other personal property and the legal recognition given to Catalan lords to exert a species of right of mistreatment (ius maletractandi) against their tenants without royal intervention. We don't know if the judicial authorities heard this case, much less what their decision was, but, as Sánchez Martínez points out, the behavior of the viscounts is probably related to the exercise of this right that, according to the enabling legislation of 1202, affected only serfs37.

35These tenants are most often identified by surnames derived from features of the landscape that were almost surely given to manses within the four parishes. Some of these names are formed as one might expect, by a common first name joined to a generic toponym: Guillelmus de Podio, Bartholomeus de Plano, Raimundus de Casadaval, Guillelmus de Fonte. A large proportion of the names in this document, however, appear in their Catalan form with the prefix en or a.n, a characteristic (now slightly archaic) indication of male names (na being the corresponding female form, even more archaic and used almost solely for obituary notices in modem Catalan). In contemporary Catalonia one can still refer to a person as En Joan or En Ramon. In the register of complaints the names preceded by en or a.n usually are single toponyms: a.n Fabrega, a.n Noguer, en Casademont, n'Oliveres. It is probable that the first name has been omitted and that these are surnames without first names38. It is difficult to determine how many of these names refer to manses which tended to have names derived from local topography. Thus «a.n Carero» could be associated with a manse of that name (carrer = road) or be a family name independent of the manse. Similarly «n.a Garriga» (stream) is a common manse name and also in the modem era a common last name. Probably a majority of such names (and certainly a substantial number) do refer to manses. From the parish of Sant Joan de la Font complaints were made by «En Planson», «N'Oulines de Saclam», and «Bartholomeo de Granoyes», names corresponding to identifiable manses of Plançó, Aulina de Ça Calm, and Granoyers39. There are two people with the surname Casadevall («lower house») and one listed as En Casadamont («upper house») which are also surviving manse names and that also appear among plaintiffs of the parish of Sant Pere de Montagut.

36According to Sánchez Martínez there are 186 names of peasant tenants in the document (allowing for the difficult cases of repeated names that might or might not refer to the same individual)40. There are 218 entries of which 128 are single surnames with the masculine or feminine prefix «En» or «Na». Virtually all of these are landscape toponyms (probable manse names) or obvious manse names. Where common first names appear (Guillelmus, Raimundus, Bernardus, etc.) they are always followed by a toponym: Berengarius de Casa de Golos, Guillelmus de Casadamont. Bartholomeus de Casadaval. There are no instances of single common first names, nor of occupational names, nor of sobriquets, nor of compound names with the nomen paternum.

37It is instructive to contrast this list with one of 1347 from Sant Andreu de Palomar in the region of Barcelona (the so-called Pla de Barcelona) that records the names of 104 individuals (all male) who gave up all rights of patronage over an altar dedicated to the Virgin Mary41. The status of these people is not immediately evident. Sant Andreu Palomar was part of the agricultural belt surrounding Barcelona and its prosperous proprietors tended to hold land on emphyteutic leases rather than as serfs. Making allowances for the variation of region (the impoverished Garrotxa versus the fertile and commercially important Plain of Barcelona) and the different forms of redacting the documents, there is a valid comparison between servile and free peasants.

38All the subscribers to this donation possessed common first names even when these appeared only once (Thomas, Vitalis, Michael). By far the most common were Bernardus and Petrus, each amounting to 26.6% of the total, thus together making up more than half the first names. The surnames show considerably more variety than those of Castellfollit. There are anthroponymic surnames given in Latin in the genitive form (Petrus Poncii, Bernardus Oliveri), occupational surnames (Bernardus Pferrarii, Berengarius Fusterii), specific toponymics referring to towns, villages or castles (Berengarius de Finestrelles, Petrus de Arenes) and more general locations (Berengarius de Curtibus, Bernardus de Ortis). One has the impression that many of these surnames are well on the way towards becoming modem family names in the sense that they do not refer to any immediate ancestor or point of origin. Thus Bernardus Cathalani or Bernardus Roure might already be the common modem Catalan surnames Català, Roure rather than specifically associated with those individuals. In nine cases men are listed by a full name but with the additional indication of their father's name, thus «Bernardus de Sabadell, filius Guillelmi Sabadelli quondam. » In all cases the surname of the father and son are the same (four are members of the Poll family).

Conclusion

39Apart from a few nuances revealed by statistical analysis of frequency distribution, there was no identifiable set of names marking servile status in Catalonia before the fourteenth century. Neither particular first names nor a style of giving surnames could be said to characterize those of unfree status. This would be an unexceptionable result except for the growing strength of servile institutions clearly visible in records that survive from the thirteenth century. More work would have to be done to understand better the impress of servitude in other parts of the western Mediterranean but its significance in Old Catalonia seems to have affected a greater percentage of the population than elsewhere and to have been vital to the maintenance of the seigneurial regime. The absence of distinct naming patterns makes it all the more interesting to consider what means of social control existed to enforce servile status, especially given the dispersed habitat and weak forms of legal procedure in contrast with England, for example, where servile status was defined along with means for proving it42. It makes instinctual sense to posit the existence of distinctly servile names in a place where servile status was important to lords and to link such definition to the extension of the seigneurial regime, but such is not the case in reality.

40How much this remained the case in the later Middle Ages deserves further study. The sources for the fourteenth and fifteenth centuries are, of course, voluminous but their very quantity has discouraged analysis, especially in view of the paucity of published archival sources from the late medieval period in contrast with what has been accomplished for the tenth to thirteenth centuries. By the fourteenth century surnames of peasants in general seem to be associated with manses, but whether or not there is any differentiation among free and servile tenants is for the time being unknown.