Our knowledge of Old English dialects derives from a multitude of sources: literary texts, legal documents, coins, inscriptions, and place-names. The extent of the Old English dialect areas is bound up with the settlements of the tribes that invaded Britain in the fifth and sixth centuries. In the formative stages of Old English dialects isolative mountain ranges, fenlands, vast forests and rivers with adjoining marshes determined the divergencies between speech groups. The boundaries generally coincide with the boundaries of the kingdoms. The regional differences that developed during the early Old English period were probably not radically modified by later political and military changes, the reason being that such changes directly affected the speech of only a minority of the population.

That Old English was divided into four basic dialects, Northumbrian, Mercian, West Saxon, and Kentish, can be deduced from the preserved localised texts and from Bede’s account of the Anglo-Saxon settlement. But the picture obtained in this way is incomplete and imprecise. A.H. Smith (1956) therefore drew a dialect map ‘which shows in outline the areas which these original tribes and dialects may be supposed to have occupied’. In addition to the four basic dialect areas, the map also determined the putative areas ofthe Middle Angles, East Angles, and East Saxons. Smith’s map was based on the evidence of history, archaeology, and place-names. That such an approach is necessary if we are to get a closer knowledge of Old English dialects has long been recognised, and was emphasised by Ekwall as early as 1913. A dialect like East Anglian, for instance, is not represented by any literary texts at all, and when Fisiak (1988) tried to unravel the problem ofOld East Anglian, he had recourse to the evidence of place- and personal names as they are represented in Domesday Book and later sources.

Fisiak was – as always – on the right track, but the variegated picture of the appearance of OE y (< WGerm u-+ i, j) that he tentatively presented may be biased by unreliability of the sources. A basic tenet in historical dialectology is to rely only on such forms as were taken down locally by people living at the place and are preserved in originals or copies very close to the originals.