Debate: why are there so few etymologically Celtic place-names in England? Celtic language decline in England

map of place-names between firth of forth , river tees: in green, names containing brittonic elements; in red , orange, names containing old english elements -ham , -ingaham respectively. brittonic names lie north of lammermuir , moorfoot hills.



bilingual welsh-english town boundary sign, on modern welsh/english border

place-names traditionally seen important evidence history of language in post-roman britain 3 main reasons:

post-roman place-names in england begin attested around 670, pre-eminently in anglo-saxon charters; have been intensively surveyed english , scottish place-name societies.

except in cornwall, vast majority of place-names in england etymologised old english (or old norse, due later viking influence), demonstrating dominance of english across post-roman england. has traditionally been seen evidence cataclysmic cultural , demographic shift @ end of roman period, in not brittonic , latin languages, brittonic , latin place-names, , brittonic- , latin-speakers, swept away.

in recent decades, research on celtic toponymy, driven development of celtic studies , particularly andrew breeze , richard coates, has complicated picture: more names in england , southern scotland have brittonic, or latin, etymologies once thought. earlier scholars did not notice because unfamiliar celtic languages. example, leatherhead once etymologised old english lēod-rida, meaning place people [can] ride [across river] . lēod has never been discovered in place-names before or since, , *ride place suitable riding merely speculation. coates showed brittonic lēd-rïd grey ford more plausible. in particular, there clusters of cumbric place-names in northern cumbria , north of lammermuir hills. so, clear brittonic , latin place-names in eastern half of england extremely rare, , although noticeably more common in western half, still tiny minority─2% in cheshire, example.

likewise, entirely old english names explicitly point roman structures, using latin loan-words, or presence of brittonic-speakers. names wickham denoted kind of roman settlement known in latin vicus, , others end in elements denoting roman features, such -caster, denoting castra ( forts ). there substantial body of names along lines of walton/walcot/walsall, many of must include old english word wealh in sense celtic-speaker , , comberton, many of must include old english cumbre britons . these have been names enclaves of brittonic-speakers─but again not numerous.

in last decade, however, scholars have stressed welsh , cornish place-names roman period seem no more survive english ones: name loss romano-british phenomenon, not 1 associated anglo-saxon incomers . therefore, other explanations replacement of roman period place-names allow less cataclysmic shift english naming include:

adaptation rather replacement. names came coined old english may come roman-period ones. example, city of york s connection original romanised celtic name eburacum not obvious. knowledge of intermediate forms of name, old english *eburwīc (later eoforwīc) , norse jórvík, makes connection clearer.
invisible multilingualism. place-names survive in old english form have had brittonic counterparts long periods without being recorded. example, welsh name of york, efrog, derives independently roman eboracum.
later evidence place-names may not indicative of naming in immediate post-roman period once assumed. in names attested 731, 26% etymologically partly non-english, , 31% have since fallen use. settlements , land tenure may have been relatively unstable in post-roman period, leading high natural rate of place-name replacement, enabling names coined in increasingly dominant english language replace names inherited roman period relatively swiftly.

thus place-names important showing swift spread of english across england, , provide important glimpses details of history of brittonic , latin in region. not demand single or simple model explaining spread of english.