Futhark International Journal of Runic Studies Containing proceedings of the Seventh International Symposium on Runes and Runic Inscriptions, Oslo, 9–14 August 2010 Part 1: Plenary lectures Main editors James E. Knirk and Henrik Williams Assistant editor Marco Bianchi Vol. 4 · 2013 Published with financial support from the Nordic Publications Committee for Humanist and Social Sciences Periodicals (NOP-HS) © Contributing authors 2013 This work is licensed under a Creative Commons Attribution-NonCommercial- NoDerivs 3.0 Unported License, http://creativecommons.org/licenses/by-nc- nd/3.0/ All articles are available free of charge at http://www.futhark-journal.com A printed version of the issue can be ordered through http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-211527 Editorial advisory board: Michael P. Barnes (University College London), Klaus Düwel (University of Göttingen), Lena Peterson (Uppsala University), Marie Stoklund (National Museum, Copenhagen) Typeset with Linux Libertine by Marco Bianchi University of Oslo Uppsala University ISSN 1892-0950 Christian Prayers and Invocations in Scandinavian Runic Inscriptions from the Viking Age and Middle Ages Kristel Zilmer Abstract Previous studies of Christian runic inscriptions have tended to deal with par­ tic ular types of inscription from defined periods of time. This article analyses all the relevant Scandinavian runic material from the Viking Age and the Middle Ages, focusing on textual features and material contexts of inscriptions that use prayers and invocations. Its main aim is to explore the dynamics of what may be termed “the runic prayer tradition” with a view to identifying potentially stable elements of this tradition as well as those that alter over time. Two main categories of prayer and invocation explored are formulations in the vernacular and in Church Latin. The results of the study reveal various possibilities of variation in the runic prayer tradition, but also suggest links and overlaps between the earlier and later vernacular prayers. The evidence further suggests some sort of a division between a monumental (or public) form of discourse in connection with rune­stones, grave monuments and church buildings — which are dominated by vernacular prayers — and that of various loose objects, where Latin prayer formulas seem to be favoured. Keywords: Viking Age, Middle Ages, Scandinavia, Christianity, Christian runic inscriptions, prayer, invocation, vernacular, Latin Many runic inscriptions from the Viking Age and Middle Ages are directly related to Christian culture — they originate from a period during which Christianity was introduced and gradually institutionalized. The emergence and increase in number of certain types of inscription have Zilmer, Kristel. “Christian Prayers and Invocations in Scandinavian Runic Inscriptions from the Viking Age and Middle Ages” Futhark: International Journal of Runic Studies 4 (2013), 129–71. © 2013 Kristel Zilmer. This is an open­access article distributed under the terms of the Creative Commons Attribution­NonCommercial­NoDerivs 3.0 Unported License. 130 • Kristel Zilmer been associated with Christian practices. There has, for example, been discussion of the role of Christianization in the rise and spread of the late Viking Age rune­stone custom. The obvious and immediate outcomes of the interaction between the native epigraphic tradition and Christianity, however, are medieval Christian runic inscriptions on grave monuments, on the walls and furnishings of stave and stone churches as well as on different types of religious object. A common approach to Christian runic inscriptions is to look at par­ ticular types of inscription, categories of artifact, regions or periods of time. This article explores some aspects of the dynamics of the overall tra dition, based upon preserved Viking Age and medieval inscriptions from Den mark, Norway and Sweden. There are a few concentrations of in scriptions in other parts of the Nordic realm where settlers of Scan­ di navian origin made use of runic script, but these date either from the Viking Age (the Isle of Man) or predominantly from the Middle Ages (Iceland and Greenland). This material does not allow for the study of the dynamics of Christian runic inscriptions in the way the material from Scandinavia proper does. The focus is on Christian prayers and in­ vocations. Other expressions of Christianity are ignored, e.g., cross orna­ mentation and pictorial evidence, which especially in the case of stone monu ments complement the textual elements. Various Christian features of rune­stones and grave monuments have been examined in previous studies — either in isolation or in combination with each other. The focus here is narrowed to the analysis of textual features and material contexts of inscriptions that use runic prayers, with the aim of distinguishing those features which are relatively stable elements of the corpus from those which alter over time. The corpus Prayers, invocations and blessings are the most common verbal ex­ pressions of Christianity in runic inscriptions from the Viking Age and Middle Ages. In addition there are Christian terms (such as kristinn ‘Christian’, kross ‘cross’, kalikr ‘chalice’), Christian personal names, and references to practices that have received a Christian interpretation — for instance, Viking Age and early medieval inscriptions may refer to the making of ‘bridges’ and clearing of paths, which was considered a pious act. In this study Christian prayers and invocations have received a broad de finition. Besides formulas that explicitly solicit help and consideration from God or some other divine agency, I have included quotations from Futhark 4 (2013) Christian Prayers and Invocations • 131 Scripture, liturgical formulas and other texts of varying content, structure and length that invoke divine assistance, blessing and protection. Also considered are inscriptions that refer to the act of praying in that they ask readers to pray for somebody’s soul. Inscriptions that consist merely of one or more sacred names — such as different names for God, names of saints, the evangelists, the archangels — qualify for inclusion in that they express a form of devotion and serve invocatory purposes. Inscriptions whose intention is unclear, and which may indeed contain no reference to the divine at all, have been omitted. Problems that arise from following the approach here outlined will be commented upon below. This study does not differentiate between the various supplicative and manipulative functions and contexts of use of Christian prayers and invocations in terms of traditional dichotomies categorized as religion versus magic. With certain types of text, like medieval charms and incan­ tations modelled upon Christian vocabulary, and certain types of object, such as Christian amulets, the manipulative and ritualistic purpose may seem evident. However, some inscriptions and artifacts hover uncertainly between the spheres of religion and magic (cf. Spurkland 2012, especially 198–200; Steen holt Olesen 2007, 38; on Christian and magic formulas in runic inscriptions, see also MacLeod and Mees 2006, especially 184–210; McKinnell, Simek, and Düwel 2004, 147–61, 172–89). Furthermore, the scholar ship of the past decades has started viewing magic as an inherent part of medieval religion and culture. Studies have concerned what can be described as “the large grey area of popular practices in Christian Europe that are not clearly magic or miracle, but lie on a spectrum in be­ tween” (Jolly 2002, 7; cf. Kieckhefer 1989; more specifically on medieval textual amulets, see Skemer 2006). Instead of operating with universal labels such as “magic” and “religion”, one needs to examine particular practices in their own contexts (Jolly 2002, 11). In the following analysis of runic prayers and invocations I will from time to time comment on such individual or general concerns and practices as the inscriptions may illu minate. Proceeding from the broad definition just outlined, and including possible fragmentary evidence, we can estimate that approximately 13 per cent of the runic material from the Viking Age and early medieval period (up to c. 1150/1200; see chart 1) contains Christian prayers and invocations. In inscriptions that date from the High and Late Middle Ages (c. 1150/1200–1500) the proportion is around 18 per cent. There are, however, a greater number of debatable cases in the high and late medi­ eval group; if we limit ourselves to the unambiguous examples, the ratio Futhark 4 (2013) 132 • Kristel Zilmer 4000 3000 2000 1000 0 Viking Age Inscriptions with High and Inscriptions with and early prayers and late medieval prayers and medieval invocations runic material invocations runic material Chart 1. Number of inscriptions containing Christian prayers and invocations in comp ar i­ son with total number of inscriptions is around 14 per cent. This division into a fundamentally Viking Age and medieval corpus and the figures presented in chart 1 (which are numbers of inscriptions) are based on the Scandinavian Runic Text Database. The data base has been used as the main source for establishing the corpus; updates and corrections as well as information on new finds have as far as possible been taken into consideration.1 The material does not indicate any drastic changes in the general pro­ portion of recorded prayers and invocations from the two periods, despite changes in the nature of the material (types of inscription, artifact, etc.). Further analysis of different chronological and regional groups of 1 I have used the latest version of the database, from 2008, with a few divergences. Certain dating principles remain debatable. There is an overlap between the two groups, which also makes the classification of some inscriptions arbitrary. Inscriptions that have been given a date range of c. 1100–1400 are categorized as high or late medieval; those from c. 1000– 1150 as Viking Age/early medieval. For information on new finds, see, e.g., Danske Rune­ indskrifter, i.e. http://runer.ku.dk/, and the reports (“Rapporter från runverksamheten”) available at http://www.raa.se/kulturarvet/arkeologi­fornlamningar­och­fynd/runstenar/ rapporter/. It is not possible to discuss all the criteria related to the inclusion or exclusion of individual inscriptions in this article; I have, however, assessed and catalogued each of them individually. Futhark 4 (2013) Christian Prayers and Invocations • 133 inscriptions would be needed to trace developments along the Viking Age and medieval timeline, but that is a matter for future studies. Two main categories of prayer and invocation can be distinguished using linguistic criteria: (1) vernacular or vernacularized formulations; (2) quotations and formulas in Church Latin.2 Besides traditional Latin texts the latter category includes words and expressions that are of Hebrew or Greek origin but were used by the Roman Church; in runic inscriptions these elements are often recorded together with Latin formulas. As an alternative, runic prayers can be categorized according to whether they are freely formulated texts, adaptations of conventional formulas, or fixed quotations. For the purposes of the current study the distinction between vernacular and Latin prayers will suffice. Besides vernacular and Latin prayers I recognise an additional group of around 45 inscriptions that are formulated in the vernacular, but also contain Latin prayer elements or record requests to pray the Pater Noster or the Ave Maria. Their main features as well as the problems related to the study of inscriptions of mixed linguistic content will be commented upon below. Determining whether a given text should be categorized as vernacular or Latin can be problematic — factors to consider are spelling, morphology, textual content and co­text (i.e. type of inscription and whether any associated or surrounding text is in the vernacular or in Latin), and the context of the artifact. What should count as explicitly Latin elements in runic inscriptions has been much discussed (see e.g. Knirk 1998, Ertl 1994, Gustav son 1994a, 1994b). Since this study is concerned primarily with formu lation, prayers and invocations are defined as vernacular or Latin based chiefly upon their content and the textual context in which they are found, although spelling variants and grammatical forms have also been taken into account. The criteria for recognising a text as “vernacular” are per haps broader than in some previous studies. In addition to clearly vernacular formulations this group includes various sacred names and Christian terms that either appear within a vernacular context or are recorded in a form that does not automatically imply direct Latin input. The Latin group consists of religious texts of various length, some of which consist of (series of) sacred names in Latin form. Inscriptions that 2 ‘Formula’ is in this study used to denote certain standard phrases — in the vernacular or in Latin — that are repeated in various inscriptions. They may follow more or less fixed patterns of wording, but they also allow for some variation in content and structure. ‘Formulation’ is a general term used to describe the wording and composition of runic inscriptions; it covers freely formulated texts as well as different types of formula and quotation. Futhark 4 (2013) 134 • Kristel Zilmer demonstrate mixed features with regard to their form and content belong in the third (combined) group, mentioned above. A different approach would be to consider the memorized quotations of Ave Maria and Pater Noster — recognisable as such because used within a predominantly vernacular context — as loan­words or vernacularized phrases. The use of the Ave Maria and the Pater Noster as conventional prayer formulas and the occurrence of the Latin titles of prayers in vernacular contexts will be dis cussed in the section on Latin prayer formulas. Contextual considerations are important when setting up different groups of inscriptions. Indeed, one and the same element may be included in either the vernacular or the Latin group depending upon the way it is used in a particular inscription. The name of the archangel Michael, Old Norse Mikjáll (recorded with a number of variant spellings in runes), can serve as an example. On the one hand, we have vernacular prayers like Mikjáll gæti ǫnd hans3 ‘May Michael watch over his spirit’, recorded on the Ängby rune­stone (U 478) or Guð hjalpi sál hans ok sankta Mikjáll ‘May God and Saint Michael help his soul’ on the Klemensker stone 4 from Bornholm (DR 402). In both inscriptions the name is recorded as mihel. According to Wessén and Jansson (SRI, 7: 298) this may reflect a pronunciation of the fricative consonant, [ç], corresponding to Latin ch in Michael. The prayer on the Klemensker stone addresses x sata x mihel, sankta Mikjáll ‘Saint Michael’ — as do some other prayers on Danish and Gotl andic rune­stones. The use of sanktus/sankta (also recorded with variant spellings) is of interest in the present context; we are dealing with a word of foreign origin but in the runic inscriptions concerned it can be understood as a vernacularized element. Indeed, on the Klemensker stone sankta modifies masculine Mikjáll, a usage that can perhaps be com­ pared to that of other loan­words signifying titles, like herra ‘master, lord’ (Jacob sen and Moltke in DR, Text, col. 706). In contrast to these examples, we find Michael (mikael) and the names of the other archangels on a medieval folded lead plate of unknown origin (A 284). The inscription contains Latin words and phrases, some of which appear corrupt. In this context we may certainly consider the name Michael as an element of a Latin inscription. Other sacred names, 3 For the sake of consistency edited texts of Scandinavian runic inscriptions are given in normalized Old West Norse, taken in the first instance from the Scandinavian Runic Text Datab ase, though also checked against corpus editions or other available publications. Trans literations too follow in general the database, with modifications such as bows for bind­runes. Also English translations are in the first instance taken from the database, but have on occasion been altered slightly or improved. Futhark 4 (2013) Christian Prayers and Invocations • 135 for example Jesus and Maria, may similarly find their place either in the ver nacular or the Latin group. Jesus used together with Kristr can be taken as a vernacularized element (Old Norse Jésús Kristr), whereas Jesus Kristus and Jesus Nazarenus can be considered Latin elements, although this would also depend upon the context they appear in. If contextual elements are lacking, the classification will simply have to rely on the recorded forms of names and, where applicable, the use of declensional endings (cf., e.g., petrus/pætrus Petrus vs. petr/pætær Pétr). As expected, vernacular prayers completely dominate the Viking Age and early medieval material (c. 450 instances). They are recorded on raised runic monuments and early Christian gravestones; in addition there are a few possible examples on loose items. Vernacular prayers also account for almost 50 per cent (c. 240 instances) of preserved high and late medieval runic prayers and invocations, which occur in inscriptions on medieval gravestones, in stave and stone churches, on church furniture, and on a variety of loose items. Even if we exclude fragmentary or question able evidence — for instance, several Gotlandic gravestone inscriptions which in their present state contain only vernacular formulations, but may originally have included Latin elements — the ratio is still around 43 per cent. The general proportion of preserved vernacular prayers in the medi­ eval material is not altered by the exclusion of inscriptions that simply refer to the act of saying prayers (most notable in Gotlandic gravestone inscriptions), which could be interpreted as implicit references to the conventional Pater Noster and Ave Maria. On the other hand, inscriptions of this type sometimes incorporate complete prayers in the vernacular or combine vernacular prayers and Latin elements. Overall, ver nacular formulations continued to form a significant part of what might be termed “the runic prayer language” even after the era of the traditional rune­ stone. Whether or not medieval vernacular prayers formed an unbroken link with the traditions of the Viking Age is less clear, and a question to which I shall return below. Latin prayers, invocations and blessings are very much a phenomenon of the High and Late Middle Ages, and they account for around 41 per cent of the material (c. 200 instances). They occur in the same contexts as inscriptions with vernacular prayers, i.e. on church buildings, gravestones, church furniture, and loose items. To the corpus we can add one relatively early example of Ave Maria, 9aue maria, recorded on a grave slab from Gretlanda (Vg 164), dated to the end of the eleventh century and thus belonging to the group of early medieval grave monuments. In the mixed group of runic prayers and invocations we find a rune­ Futhark 4 (2013) 136 • Kristel Zilmer stone from Backgården (Vg 76, c. 1100) which refers to the praying of the Pater (Noster), þat ær : re"t : h[-ær]ium : }atbiþia : bat[ær] ‘It is right for everyone to pray the Pater (Noster)’. There is a further uncertain case on a fragmentary eleventh­century stone inscription from Öland (Öl 44). The stone has a four­word vernacular prayer Guð hjalpi sálu hans and a sequence lis : iki, which has been interpreted Les Ave ‘read Hail (Mary)’ (Brate in SRI, 1: 113). The use of both vernacular and Latin elements is a phenomenon found mainly in gravestone inscriptions from Gotland and parts of mainland Sweden, but occasional examples are known from other contexts (church walls and furnishings, loose items). Vernacular prayers and invocations Vernacular prayers from the late Viking Age and the early medieval period tend to be variations and modifications of the common formula Guð hjalpi ǫnd/sál(u) hans (or: hennar/þeira) ‘May God help his (or: her/their) spirit/ soul’. The composition of the prayer has been analysed in various studies and will not be re­examined here (on structural and content­related vari­ a tions, see Zilmer 2012 with references). Previous scholarship has also tried to trace the origin of the formula through theological analysis of its com ponents and core message. There has been discussion of the extent to which the prayers recorded on rune­stones may have been influenced by elements from the Latin liturgy of the dead; although some indirect in fluence has been identified, direct borrowings are not very evident (Beskow 1994, cf. Gschwantler 1998). Certain features of runic prayers have been interpreted as evidence of popular religiosity (Gschwantler 1998; see also Williams 1996a, 1996b). The origin of Guð hjalpi ǫnd/sál(u) hans has thus far remained une x­ plained. It is conceivable that the formula stems from prayers originally designed for missionary purposes. Prayers of this type may first have been introduced into other Germanic languages as a means of spreading the Christian message and reached Scandinavia by different paths (Gschwant­ ler 1998, 738, 743 f.; cf. Segelberg 1972, 162, 170–76). Although the number of vernacular rune­stone prayers found outside the core area in central Sweden remains limited, variants of the formula can be traced in different parts of Scandinavia. This would indicate independent manifestations of the tradition, though possibly modelled on one common source. At the same time, the analysis of variation in prayer formulations and their regional distribution shows runic prayers to have been dynamic elements of vernacular tradition rather than the formalized tools of Christian mission Futhark 4 (2013)