The Last Pagan

INTRODUCTION

This article assesses the evidence for the survival of pagan views in skaldic poetry from the period of the reign of king Óláfr Tryggvason of Norway (r. 995–1000) onwards. Based on a new typological subdivision of mythological kennings, I suggest that one group of these remained in largely unchanged use after Óláfr's time, whereas another disappeared almost completely. This division makes it possible to see much clearer lines of development than in previous research. In the following, I first briefly describe the research history and the two groups of mythological kennings, and I address the question of why one group remained while the other fell out of use. Thereafter, the study turns to the second group, the one that nearly disappeared and that therefore is likely to have diagnostic value. These kennings are rare enough in the eleventh century that they merit analysis on a case-by-case basis, and I begin by discussing four stray examples from the early decades as well as a number of occurrences that are likely to be spurious. Thereafter I turn to some particularly informative poets. These include what appears to have been a milieu of pagan poets around Eiríkr jarl (r. 1000–12), the poet Hofgarða-Refr (1030s), and Arnórr jarlaskáld (1060s). The observations are based on a comprehensive reading of the poetry dated to ca. 995–ca. 1130 in Den norskislandske skjaldedigtning, edited by Finnur Jónsson, since the new edition (Skaldic Poetry in the Scandinavian Middle Ages) is not yet complete, in tandem with Rudolf Meissner's Die Kenningar der Skalden. Ein Beitrag zur skaldischen Poetik.¹ Later critical scholarship and editions have been consulted in order to secure the readings, most notably the poetry that has so far been published in SkP; Ernst Albin Kock's Notationes Norroenae: Anteckningar till eddaoch skaldediktning; Anthony Faulkes's edition of Snorri Sturluson's Edda: Skáldskaparmál; and editions of Íslenzk fornrit.² The relevant stanzas by Refr, as well as one by Skúli, have very recently appeared in SkP, and I wish to express my gratitude to the general editors of the volume, Kari Ellen Gade and Edith Marold (editor of Refr's stanzas), as well as Kate Heslop (editor of Skúli's stanza), for allowing me to use their unpublished texts while working on this article. With regard to Refr's alhent stanza in particular, the new reading is much more plausible than the old one. Since the edition has now appeared, references are made to it below.

Because it is datable with relative security, skaldic poetry offers us a unique possibility to follow cultural and literary developments in western Scandinavia much further back than the time of the first written documents, which began to emerge in the twelfth century. The skaldic corpus takes us all the way to the ninth century and thus across the divide between Christianity and paganism. No other source group is thus equally suited for the study of diachronic changes in religious attitudes, even if the strong stylistic conventions must warn us against simplistic interpretations. Within the skaldic corpus, kennings are ubiquitous, and they are therefore a promising feature for detecting changes on the level of diction. Scholars have long realized this, and several studies of stylistic changes over time have been published, often focusing on the use of inconsistent metaphors.³ Two scholars in particular, Jan de Vries in the 1930s and Bjarne Fidjestøl in the 1990s, have focused rather on mythological kennings and studied diachronic changes in their use.⁴ They have noted a strong increase in the use of mythological kennings in the latter half of the tenth century, followed by a slump in the period ca. 1000–ca. 1150 (ca. 1200 according to Fidjestøl) and a subsequent rise. In their studies, the fall is far from absolute and the rise ca. 1150–1200 is limited or, according to Fidjestøl, even questionable. It should be noted, however, that Fidjestøl excluded much of the poetry that most clearly shows how these kennings rose again from his analysis (thus...