Many an oath in stone

Vikings, it is commonly assumed, were illiterate. They did not employ the more southern techniques of record: manuscripts and writing tablets. Instead they inscribed on stone, metal, bone and wood, producing the runic texts that are our only written evidence of their activity. Surprisingly, historians have used this material sparingly until recently. One of a new generation who exploit and interpret the clues preserved on Viking rune-stones, Birgit Sawyer shows what they have missed.

This is not an easy book to read. Its last section is a comprehensive catalogue of 2,308 inscribed stones scattered through the Scandinavian mainlands, with details of design, date, size, condition and ornamentation that can only be for the specialist. The opening chapters discuss the distribution of these inscriptions, describe their general features, examine problems of chronology, and define the nature of the texts, which are then discussed thematically. It is here that the greatest interest will lie for most readers.

In general, rune-stone inscriptions are memorials and so have, at first glance, limited range: "X raised this stone (or had this stone raised) in memory of Y." There may be more than one raiser (called in this book "sponsor"), and more than one commemorated. There is often additional wording defining the relationship between sponsor(s) and deceased, detailing property owned, circumstances of death, and so on. With such details Sawyer discerns the material of social history. She sees rune-stones not simply as memorials but as public documents defining family, political or financial relationships. They are declarations of inheritance, records of joint land ownership, assertions of rank or social status. They serve some of the purposes that charters, writs and wills did later. Some record the effects of religious conversion or imply the development of new forms of political control. Their uneven distribution through Scandinavia (168 in medieval Denmark, 51 in Norway, the rest in Sweden, including a staggering 1,016 in Uppland province) is, she argues, a sign of different religious and political developments in the lands.

Patterns of inheritance in a society of sudden death and high infant mortality could be complex. Sawyer notes the order in which sponsors are named, and what is not said as well as what is. Though men are far more likely to be commemorated than women, the latter are not absent. One important aspect of Sawyer's deductions affects the position of women in Viking society, their right to own property and to inherit. Sometimes this is clearly stated, as in the Hillersjõ, Sweden, rock inscription, which defines a series of marriage and maternal relationships ending with mother inheriting from both a daughter and son. More often it is implied; and here there is the danger of the presupposition affecting the resultant deduction. Sometimes two or more inscriptions, perhaps from different find-sites, may cast light on each other. Here the extended nature of Sawyer's study yields results. From the total evidence, Sawyer suggests rather different systems of inheritance in south and west Scandinavia and in east Scandinavia, and links them to different social patterns in the two, perhaps the effect of differing degrees of royal encroachment.

A philologist might complain there is little linguistic scepticism here. Dr Johnson observed that "in lapidary inscriptions a man is not upon oath". Epitaphs use standard, often archaic forms of sentence and lexis. It is wise not to assume as truth all that is recorded on rune-stones. I cite Sawyer's discussion of the word good in inscriptions. This, she suggests, "reinforces the status indicated by the title and rank of the people in question" or perhaps "marks status". Possibly, but epitaphs from other periods and places suggest that few of the commemorated are "bad". "Good" may be a fill-in word to be expected in a memorial. Nor does Sawyer consider whether sentence form may affect word choice; something particularly significant where the text breaks into verse, for verse lexis may differ from prose, influenced by metre, alliteration or rhyme. This is a stimulating book, challenging accepted interpretations and suggesting new sources for Viking Age social history.

R. I. Page is emeritus professor of Anglo-Saxon, University of Cambridge.