ascertain which finds are contemporary and which are mixed from earlier or later layers.
Finally, radiocarbon and other scientific dating methods usable in Atlantic Scotland –
for settlements, ecclesiastical sites and burials without grave goods – typically produce
error ranges of between one and three centuries (unlike the dendrochronological dates
available elsewhere in the Viking world). Ongoing work regarding Old Scatness is
attempting to resolve this last problem by using multiple stratified dates and Bayesian
statistics (Dockrill and Batt 2004 ), but this approach cannot be applied retrospectively
to sites such as the Brough of Birsay.
A related problem is establishing when one might expect to find the first Scandinavian
settlement in Scotland. It was once assumed to have begun c. 800 (e.g. Hamilton 1956 )
based on the date of the earliest raids in the North and Irish Seas – and the assumption
that these raids were launched from bases in Atlantic Scotland (cf. Morris 1998 : 74 ;
Ó Corráin 1998 a; Sawyer 2003 ). However, early ninth-century settlement has been
questioned by those who would argue for both earlier (Myhre 1993 ) and later (Graham-
Campbell 1998 : 106 ; Barrett 2003 ; Parker Pearson et al. 2004 a: 129 ) alternatives. An
early, pre-ninth-century, option is conceivable if envisioned as a period of trade and
contact rather than migration and settlement (see below). A late, mid-ninth-century,
option is favoured by the present author. It is only at this time that the historical and
archaeological records provide explicit evidence for settlement rather than raiding in
Scotland and it is also the period when Scandinavian warbands first began to overwinter
elsewhere in the west – such as Ireland and England (see Barrett 2003 and references
within). (An even later, mid-tenth-century, option has been proposed by Parker Pearson
et al. 2004 a: 129 , but is inconsistent with the historical, burial and settlement evi-
dence.) Clearly the contemporaneity of indigenous and Scandinavian material culture in
Atlantic Scotland will depend on which of these interpretations is correct.
In sum, it is evident that one’s interpretation of Viking Age Scotland will be
dependent on which of the above bodies of evidence are favoured and what assumptions
are made regarding chronology. The potential combinations are legion, but most
published discussions can be categorised into one or more of four broad theories:
- The genocide hypothesis
- The Laithlind hypothesis
- The Myhre (or pagan reaction) hypothesis
- The earldom hypothesis
Not all of these alternatives are equally plausible, but insofar as each has been given
credence in the published scholarly literature they will be discussed in turn.
The least credible (if nevertheless tenacious) option is the genocide hypothesis –
which suggests that the indigenous populations of the Northern and Western Isles were
eradicated by Scandinavian migrants (e.g. Crawford 1981 ; Smith 2001 , 2003 ; Jennings
and Kruse 2005 ). It is typically espoused by those who place most weight on the
onomastic evidence, given the (virtual) absence of recognisable Pictish place names in
the Northern Isles and of demonstrably pre-Norse Pictish or Gaelic place names in the
Western Isles ( Jennings and Kruse 2005 : 284 ). This observation – sometimes in com-
bination with ethnographic analogy citing tragedies such as the European extermination
of the Tasmanians (e.g. Smith 2001 , 2003 ) – is explicitly or implicitly interpreted as
an indication of both language death and biological death in the early Viking Age. The
–– chapter 30 : The Norse in Scotland––