Uses of the vernacular in the acts of Welsh rulers 1120-1283

Foreword

Wales in the twelfth and thirteenth centuries was a land of both political conflict and cultural change. Politically, a number of native dynasties competed for power amongst themselves as well as with Anglo-Norman lords who succeeded in establishing Marcher lordships along the border with England and along the southern coast. In addition, kings of England sought to secure recognition of their overlordship, attempts that culminated in the final conquest of Wales by Edward I. One consequence of Anglo-Norman and English conquest and settlement was that Wales became more ethnically and culturally diverse than ever before. The Marcher lordships were peopled by substantial numbers of French- and English-speakers. In addition, the native rulers of the remaining Welsh principalities – of which the most powerful and long-lasting was Gwynedd, defended by the mountains of Snowdonia in the north-west – assimilated new influences from England and farther afield in Europe, influences reflected, for example, in castle building and the patronage of reformed religious orders of Continental origin, especially the Cistercians1.

This paper focuses on another aspect of this process of cultural assimilation and adaptation, namely the written acts issued by Welsh rulers between the early twelfth century and the final extinction of native rule by Edward I in 1282–32. All of the surviving texts are written in Latin apart from a tiny number of letters and petitions, all dating from the late thirteenth century, which are in French3. Here, I shall concentrate attention on charters4. Latin charters had been written in Wales before the arrival of the Normans from the late eleventh century onwards5. However, the charters issued on behalf of Welsh rulers in the twelfth and thirteenth centuries marked a new departure in that, unlike those produced earlier in the Middle Ages, they conformed, on the whole, to diplomatic forms used in Anglo-Norman and Plantagenet England, especially the writ charter6. This was linked to the switch in Welsh rulers’ patronage from long-established native churches to religious houses – especially Benedictine and Augustinian priories and Cistercian monasteries – introduced to Wales by the Anglo-Normans. Most of the surviving charters are for such ecclesiastical beneficiaries; and it also appears that most were drafted by the beneficiaries, who were influenced by the diplomatic of charters from previous Anglo-Norman benefactors. Of course, the picture was more complex than this brief sketch suggests, not least because beneficiaries varied in the extent to which they adhered to Anglo-Norman diplomatic forms. In particular, it is important to distinguish, on the one hand, between charters issued for religious houses that already had a strong tradition of Anglo-Norman patronage and, on the other hand, charters for houses which native Welsh rulers had either founded or of which they were the principal benefactors. In considering the production of charters it should also be emphasized that the communities of the first group of houses were exclusively or predominantly Anglo-Norman or English, whereas those of the second group were mainly or entirely Welsh. Two Cistercian monasteries, both daughter houses of Clairvaux, symbolize this distinction: Margam Abbey, in the south-east, founded in 1147 by Earl Robert of Gloucester, Norman lord of Glamorgan; and Whitland Abbey, in the south-west, which, although originally a Norman foundation, soon gravitated into the orbit of the native dynasty of that region and became the mother house of a network of monasteries founded by Welsh rulers7.

The Latin charters I wish to consider also contain words in Welsh, the Celtic vernacular spoken by the rulers in whose names the documents were issued. Usually these words consist simply of personal names and place-names, but in some cases there is more, as we shall see. Welsh was well established as a written language by the twelfth century, and had been used for the composition of a wide variety of texts, though there is no evidence that pre-Norman charters were written entirely in Welsh8. As Welsh continued to be a vigorous written language in the twelfth and thirteenth centuries, it would have been technically possible to have used it for the writing of charters9. In practice, however, Welsh intruded into the Latin charter texts only to a very limited extent: there is no parallel for the writing of vernacular charters such as we find, say, in thirteenth-century Germany10. Why this was so lies largely outside the scope of this discussion11. Likewise, I shall not attempt to provide a detailed examination of the orthography of Welsh names and other words, although the subject is obviously of great importance12. Suffice it to say that some scribes of original charters written for Cistercian monasteries in areas under Welsh control reveal a familiarity with orthographical conventions used in writing Welsh13, whereas several scribes of the Anglo-Norman foundation of Margam in Glamorgan rendered Welsh personal names and place-names according to English orthography, including the Old English letters thorn and eth14. I shall concentrate attention, rather, on the relationship between the Welsh vernacular and Latin. Let us look, first, at the treatment of Welsh personal names before moving on, second, to boundary clauses.

1. Welsh personal names

One indicator of how far Welsh personal names were Latinized is the treatment of the word for ‘son’ in the patronymic. Welsh names in this period normally included a patronymic15. On the whole this was linked to the Christian name by Latin filius, although in some cases scribes used the Welsh equivalent of filius, namely mab or its shorter form ab or ap. For example, both usages occur with respect to Morgan ap Cadwallon of Glamorgan (fl. 1191–1228), namely, Morganus filius Kaduathlan and Morganus ab Kadwathlan16. A Welsh form of the patronymic likewise appears on the legend of the seal of another member of the same dynasty, Morgan ab Owain (fl. 1183–1246): ‘+ S. MORGAN MAB OEIN’17. On the whole, however, this Welsh vernacular usage is uncommon in the intitulatio of charters or on the legends of seals, whereas it is more frequent in witness lists18. Presumably greater care was taken to use filius in the intitulatio because of the key role played by this part of the document in expressing the dignity and status of the act’s author.

The same consideration may also help to explain the Latinization of Christian names and patronymics. It is notable that Latin forms are consistently applied to the names of the rulers of Gwynedd, the most powerful Welsh principality, in the intitulatio of their acts: thus Lewelinus (or sometimes just the initial L.) is used for both Llywelyn ap Iorwerth (d. 1240) and his grandson Llywelyn ap Gruffudd (d. 1282). Likewise, it is hardly a coincidence that the princes of Gwynedd were unusual among Welsh rulers in being styled with a title19. In other words, there appears to be some correlation between the Latinization of names and the articulation of a ruler’s status. Possibly less care was taken to Latinize the names of less powerful rulers, who often lacked titles, such as the Welsh lords of upland Glamorgan: in these cases, practice was more varied. Nevertheless, it is difficult to detect clear patterns. For one thing, individual scribes had different preferences: not only were some readier to Latinize names than others, but they also devised different Latin forms for the same Welsh names.

This is illustrated by the varying treatment of the Welsh name Iorwerth. Here we see two different approaches to Latinization. The first, found fairly commonly, was to deem the Welsh name as being equivalent to the French name Gervase (Latin Gervasius): hence Llywelyn ap Iorwerth could be rendered as Lewelinus Gervasii filius20. A comparable, but not identical, phenomenon was the adoption by some Welsh clergy of a second, internationally recognizable, name in addition to their Welsh name, the most famous example being Thomas Becket’s crossbearer Llywelyn, otherwise known as Alexander21. The second, less radical, solution was simply to add Latin case endings to the Welsh name: for example, Iorwerth is rendered by Ioruerthus and related forms in the charters of the Welsh lords of Gwynllŵg or Caerleon in the south-east. Thus Morgan (d. 1158) and Iorwerth sons of Owain are styled variously as Morganus et Iereuert filii Oni, Morganus filius Oeni et Geruerdus frater meus and Morganus et Ioruerdus frater eius22. These examples also illustrate the lack of standardization: scribes devised different Latin forms for Iorwerth and in one case did not Latinize the name at all23.

Admittedly, all the forms of Iorwerth I have cited so far are extant only in copies. This raises the general methodological difficulty of determining how faithfully name forms are reproduced in acts preserved only in later copies such as English chancery enrolments or cartularies. An arresting example of the pitfalls of depending on such copies is provided by a recently rediscovered original charter issued in the early thirteenth century by Llywelyn ap Iorwerth of Gwynedd. Whereas the copy in the late fifteenth-century cartulary of Haughmond Abbey (Shropshire) renders the patronymic of one of the witnesses as Gervasius, the original has the Welsh name Iorwerth (spelt Iaruuortd).24 Quite possibly the scribe of the cartulary amended the text in the knowledge that Gervasius was a normal Latin equivalent for Iorwerth. How frequently copyists Latinized personal names in this way would merit further investigation25. In general, the trend was probably towards Latinization. One corollary of this is that the appearance of Welsh words in a copy is likely to reflect the content, if not the precise spelling, of the original. This point is relevant to the second part of my discussion, in which I propose to examine some aspects of boundary clauses, as several of the charters to which I shall be referring are extant only in copies.

2. Boundary clauses

Place-names in Welsh rulers’ charters are normally rendered in Welsh, though one charter for the Cistercian monastery of Basingwerk, an Anglo-Norman foundation near the English border in the north-east, provides an English gloss of the Welsh name for Bala Lake (« Thlin Tegit, id est Pemblemere »), while the foundation charter of Strata Marcella Abbey, established c. 1170 near Welshpool by the native ruler of southern Powys, is unusual in giving most of the Welsh place-names in Latin translation26. In the majority of charters the lands granted are described fairly briefly, in conformity with normal Anglo-Norman practice. However, a significant minority describe the lands granted by means of detailed perambulations, a practice ultimately derived from the Roman world, most often in the form of a circuit beginning and ending at the same fixed point27. This is a conservative aspect of Welsh diplomatic by comparison with royal charters in England, where perambulations tended to die out with the demise of the Anglo-Saxon diploma (whose bounds were usually recorded in the vernacular, in this case Old English), although they continued to be used in private charters, notably in northern England, and were also widespread in Scotland28. All the surviving twelfth- and thirteenth-century Welsh examples are in favour of Cistercian houses closely associated with native rulers, especially Strata Florida in west Wales and Strata Marcella in Powys; yet similar boundary clauses are also attested in summaries of Welsh rulers’ charters for the Premonstratensian house of Talley in south-west Wales29.

In most instances, the perambulations are described in Latin, but a few include connecting and other descriptive words in Welsh, as shown by the following two examples (connecting words in Welsh are italicized):

« (…) Tref Boith cum apendiciis suis et hii sunt termini eiusdem, describente Reso cum optimatibus suis et etiam Griffino filio eius: o Aber Meylyr sursum ar hit Arth usque fossam que manat de fonte Bleydud, or fannaun hir pant, pant in hit usque fossam Bileyneyt, hi foss ar hit usque Pant Gweun inter Marchidi et Brinn Llende, hi pan ar hit usque fossam que est terminus inter Marchidi et villam que vocatur Ardiscinkiwet quam optulit Gwenliant predictis monachis in perpetuam elemosinam cum consilio nostro et filiorum nostrorum, fossa vero prefata in termino est usque ad mare, mare vero usque ad ostium Arth, Arth hit in Aber Meilir, et in litore et mari ab ostio Arth usque ad ostium Ayron. Coredeu et omnem piscaturam sepedicto monasterio offerimus imperpetuum (…) »30.

« (…) omnes pasturas totius provincie que dicitur Keueilliauc infra istos terminos: scilicet a Bon Main Melin usque ad Lluin Ecrois, et inde in directum usque ad blain Nanhanauc, et inde Nanhanauc usque ad eius aber, et inde usque ad Abernant Karthbrandu, et per longitudinem ipsius rivuli usque ad suum blain, et inde directum usque ad Carnethwen, et inde usque ad Gobleitheu, et a pen Gobleitheu blain Nant Teyling usque ad suum aber, et inde Bacho usque ad aber Dengum, et inde per Dengum usque ad eius ortum, et inde usque ad Keilligogeu, et inde usque ad Reidiaul et per Reidiaul usque ad Gwrhet Kei, et inde Reidiaul iterum usque ad aber Camdwr Keueiliauc, et ab aber Camdwr Keueiliauc usque ad eius ortum, et inde in directum usque ad blain Eynniaun, et sic per Eynniaun usque ad eius aber, et inde per Deui usque ad aber Dwlas, et inde per Dwlas usque ad eius ortum, et inde in directum usque ad Kenghulf, et inde usque ad blain Lloido, et per Lloido usque ad eius aber et inde Deui usque ad aber Llewenith, et sic per Llewenith usque ad eius ortum, et inde in directum usque ad Red Pebellua super Clawedauc, et inde per Clawedauc usque ad Gwernach, et per Gwernach usque ad eius ortum, et inde sicut ducit mons superior usque ad Red Derwen, et sic per Derwen usque ad Euernoe, et inde Nant Er Heyre usque ad Lledwern, et a blain Lledwern in directum usque ad Bon Main Melin »31.

The first of the charters cited was issued by Rhys ap Gruffudd (also known as the Lord Rhys) of Deheubarth in south-west Wales for the Cistercian abbey of Strata Florida in 1184, and draws on a number of models: for example, the clause introducing the lands granted borrows from a papal bull, almost certainly that issued for the monastery by Pope Alexander III32. Yet, whereas the papal bull refers tersely to Rhys’s gift of the site of the house ‘with all its granges and appurtenances’, the lands in Rhys’s charter are described in two detailed perambulations of boundaries. The second of these boundaries, most of which are reproduced in the extract cited above, are said to have been described by Rhys with his nobles and his son Gruffudd. Likewise, the first perambulation was probably described by Rhys and three of his sons, since these are named earlier in the charter as granting all contained within the bounds to the monks in perpetual right in the presence of Rhys’s army. The way in which the text moves back and forth between Latin and Welsh suggests that the scribe of the charter was literate in both languages and, perhaps, not too worried about distinguishing sharply between them. What we seem to have, then, are examples of what sociolinguists term ‘code-switching’, a phenomenon for which there are plenty of parallels in other kinds of medieval texts33. In the case of the charters under consideration here, the process was no doubt facilitated by two factors. First, some of the descriptive words were also common components of place-names; indeed, it is sometimes unclear whether a word such as aber, meaning ‘river-mouth’, should be taken as part of a place-name, in a genitival relationship to the following river name, or merely as a descriptor.34 Second – and the same applies to other vernaculars apart from Welsh – some terms were derived from, and closely resembled, their Latin equivalents: for example, Latin fossa and Welsh ffos (a correlation illustrated in the extract cited above from Rhys ap Gruffudd’s charter, « usque fossam Bileyneyt, hi foss ar hit »).

Why did the drafters of some boundary clauses in Welsh rulers’ charters insert words and short passages in Welsh? Probably one important part of the answer is precedent. Both perambulation and a mixture of Latin and Welsh for the linking words – as well as some boundaries written entirely in Welsh – were features of early medieval charter writing in Wales, to judge especially by pre-Norman charters, belonging to what Wendy Davies has identified as a ‘Celtic’ charter tradition, preserved in the early twelfth-century Book of Llandaff or Liber Landavensis35. It is quite possible, therefore, that the twelfth- and thirteenth-century scribes who drafted the boundary clauses under discussion here had seen earlier Welsh charters written in this form. If, as is most likely, these acts were drawn up by the beneficiaries, this would provide a further instance of the way that some Welsh Cistercian houses preserved and adapted aspects of the written culture of the native, pre-Norman Church in Wales: the continuation of early medieval annals and the occasional use of Insular decoration and punctuation in thirteenth-century manuscripts are other cases in point36.

Yet, if earlier Welsh charter writing provided a model for these boundary clauses, their appearance in twelfth- and thirteenth-century acts cannot be explained simply in terms of scribal conservatism. After all, the overall diplomatic structure of most of the charters under consideration here was the Anglo-Norman writ charter, and, as already mentioned, the 1184 charter for Strata Florida was also indebted to papal diplomatic. The drafters of charters were eclectic in their selection of models to follow, and made no attempt to adhere to the form of pre-Norman, so-called ‘Celtic’, charters in all respects. The rendering of the boundary clauses as perambulations was therefore a matter of deliberate choice: more concise forms could have been – and indeed in the majority of cases were – chosen. A further incentive for recording boundaries in the form of perambulations may well have been the continuing use of perambulation as a means of marking out boundaries in twelfth- and thirteenth-century Wales. Certainly, Welsh lawbooks of the period assume that perambulation might be used by rulers and other lords to settle disputed boundaries37. As already mentioned, the 1184 charter for Strata Florida explicitly states that the bounds were described by the Lord Rhys and his sons and nobles: as, for example, in Anglo-Saxon England, perambulation gave lay donors an active role in the making, and hence the guaranteeing, of gifts. It may be that both the monks and the prince wished to record detailed perambulations in order to prevent any disputes arising over the boundaries of the lands that had been given to Strata Florida. The use of Welsh may also have derived from contemporary practice when the acts were drafted, namely that the perambulations originated in oral declarations in the vernacular by the prince and his nobles and sons, quite possibly drawing on the oral or written testimony of those with the necessary topographical knowledge. If so, however, the retention of vernacular phrases probably reflected the preference of individual scribes rather than any requirement for Welsh to be used in order to strengthen the grant’s validity. After all, most extant perambulations are written entirely in Latin, and those containing Welsh are still mainly written in Latin. In other words, the crucial issue is perambulation, not its partial expression in the vernacular.

Conclusion

I would like to conclude with two points. First, it is clear from this brief survey that the use of the vernacular in the acts of Welsh rulers was very limited: to study the appearance of Welsh in these documents is therefore to underline their overwhelmingly Latin character. This is not surprising. The adherence to Latin probably reflected above all a desire on the part of both ecclesiastical beneficiaries and native rulers to conform with wider European (and perhaps particularly English) norms of documentary production, though it may also have been influenced by the existence of an earlier tradition of charter writing in Wales whose principal language was Latin, vernacular boundary clauses notwithstanding. Yet, limited though they were (and this is my second point), the uses of the vernacular I have considered reveal much about the production of acts and hence about written culture in Wales more generally. Thus, the choices made between Latin and Welsh throw valuable light on the assumptions and cultural affinities of those who drafted the documents. The particular form adopted in rendering the name of an act’s author may be indicative of how that ruler’s status was perceived, while the intrusion of Welsh words into boundary clauses probably testifies both to the influence of earlier Welsh charter writing and to an oral vernacular context for the making of grants that is usually obscured by the Latin record. For the student of written culture, the use of the vernacular in charters not only illuminates the extent to which scribes were literate in Welsh as well as Latin but also, in the case of some boundary clauses, provides examples of breaking down the barriers between those languages through code-switching. In short, the vernacular components of the acts of Welsh rulers merit careful attention if we wish to understand the nature and significance of those documents as Latin texts.