English place-names have frequently been used in recent decades by researchers in different historical fields. In the case of the study of linguistic evolution, the publication in 1913 of Eilert Ekwall's innovative article "Die Ortsnamenforschung cin Hilfsmittel fur das Studium der englischen Sprachgeschichte" opened up a new branch of investigation that has been especially fruitful. However, the differing interpretations of this type of evidence for the reconstruction of Old and Middle English dialects that have been proposed since the publication of the above mentioned article provide a clear illustration of the problematical nature of this material.
In spite of this, it is clear that the onomastic evidence must not be categorically excluded from the materials used by the linguist engaged in the painstaking task of reconstructing the linguistic features of Old and Middle English. Given the unreliability of other types of evidence (such as literary records), place-names should be regarded as the most reliable source available to us of the phonemic reality of previous periods in the history of the language; furthermore, Old and Middle English phonemes reconstructed through the study of onomastic material tend to show regular patterns of geographical spread, while the diffusion of the written forms recorded by the scribes in their copies of literary texts is usually subject to social and cultural processes. (1)
Taking J. Fisiak's (1984, 1985, 1990) and G. Kristensson's (1967, 1986) previous works on Domesday Book (1086) (hereinafter: DB) and thirteenth and fourteenth century Lay Subsidy Rolls (hereinafter LSR) as my model, I am proposing a diachronic approach to the onomastic material recorded in these documents, which accounts for some of the numerous phonological changes experienced by the English language in the period immediately after the Norman Conquest of 1066. Both documents have proven extremely useful for the historical dialectologist in recent times and, in spite of the methodological differences in the treatment of the place-names recorded in each one of them, I would like to focus on their complementary character as informants of processes of linguistic change in early Middle English.
In this paper, I am going to focus on the phonetic process traditionally known as second fronting and its significance for the study of the early Middle English dialect of the South West Midlands. According to classical accounts on this change, the palatalization of West Germanic a into OE ae and its later development into e took place in the dialects spoken in the Mercian dialects of the West Midland area and in Kentish. However, the quality of this sound, which is usually represented by the grapheme in the Vespasian Psalter and other Mercian texts, was slightly different from the one derived from Germanic e, as can be seen from the subsequent levelling of West-Saxon ae and Mercian e (from West Germanic a) during the twelfth century (Campbell 1959: 62-64).
Early Middle English texts produced in the South-West Midlands in the years immediately after the beginning of the thirteenth century show a great deal of inconsistency in their use of the spellings , (both for West-Saxon ae) and (rendering Mercian e) in this group of words. As I have pointed out elsewhere (Diaz 1994: 450 and Appendix B), the distribution of the three spellings corresponds to the following patterns:
(a) Copies of West-Saxon documentary and literary texts produced in the South-West Midlands between 1200 and 1300 show high degrees of retention of . Scribes systematically use only when dealing with copies of Old Mercian texts (such as St Chad), or when consciously holding on to the orthographic tradition of their own region, as in the case of the famous "Worcester tremulous hand" of the early thirteenth century (Franzen 1991).
(b) However, contemporary copies of literary texts produced in this same region show a prevalence of the grapheme , being the copyist of the Caligula manuscript of Lazamon the only one who alternates this spelling with ; further, the double graph is sporadically found in the group of texts copied in the AB-language. …