PLACE-NAMES. Ancient historians often speculated upon the meaning of place-names, inventing persons who are presumed to have been the first founders of the country and to have left their mark on its nomenclature. But it is only within the last half century that attempts have been made to lay the foundations of a science of place-names. Forstemann in Germany, Rygh in Nor way, Noreen in Sweden, Skeat in England, Longnon in France, are perhaps the best-known pioneers in this work. The main founda tions upon which they built were the recognition (I) that no place-name can be interpreted in the light of its present-day form alone, but must be traced back to its earliest recorded form, (2) that no explanation of a place-name, however convincing as philology, can hold good if inconsistent with the known history and topography of the site, and, conversely that no explanation based upon legend or topography, is of value if inconsistent with philology.
The ordinary word is to some extent controlled by the written or printed book. But the transmission of most place names is still very largely oral, and as the name is merely a name, and rarely has any significance in itself, at least to its modern user, there is little or nothing to check the process of change. The natural result is that the modern form of the name is often misleading. Cenwulfesleah has become Knowsley, Tidwulfestreo has become Elstree; and Cholmondeley, Cirences ter, Auchinleck, and Wymondham, while in spelling they retain something like their original form, in pronunciation have been reduced to Chumley, Sissiter, Aleck and Wyndham. (It should be noted that the spread of the printed word and universal primary education are now tending to reverse the process and to substitute spelling-pronunciations for the genuine local ones.) The lack of stress upon the final element together with certain common sound assimilations have obscured the history of many names. Thus Embleton really should end in don, "hill," Hallington is really "holy dene" (the site of the battle of "Heavenfield"), Warden-on Tyne is a "watch don" or "hill" at the junction of north and south Tyne. So, also, nearly all the place-names in -hall, such as Willen hall, really end in hale, "nook" or "corner of land." The proximity of some similar sounding name may lead to its ultimate substitu tion as when Combe-in-Teignhead replaces Combe-in-Tenhide be cause the place stands on a headland at the mouth of the Teign.
When we come across a name which in itself conveys no meaning we are apt to turn it into something which conveys some sort of meaning, however inapposite. Thus all the places once called Fivehide have now become Fivehead or Fyfield, since we are no longer aware what a "hide" of land is. Highenovre (lit. "high bank") becomes High and Over. Folk etymology has led to the formation of a series of "ghost" place names, commonly spoken of as "back-formations." Cambridge is assumed to be the "bridge over the Cam" and the river is hence forward so called, though in reality Cambridge is an Anglo-Norman corruption of an earlier Grantabrycg and the correct name of the river is Granta. This process is specially common with river names. The Sussex Arun and Rother are back-formations from Arundel and Rotherfield, while the Adur is a back-formation from the false identification of Portia Adurni with Shoreham. The old names for the rivers, in this case Tarrant, Limen, Bramber, were early lost and new names had to be found.
Early forms of certain names may be recorded by early geographers, historians or chroniclers but for most we have to turn elsewhere, to national surveys such as Domesday Book or the Danish Landbook of King Valdemar, to ancient charters, which of ten give not only the name of the land granted but also the names of the land-marks on its boundaries, royal and episcopal documents of every kind and, in later days, surveys, extents, ter riers, maps. Here fresh difficulties arise. We find a large num ber of faulty and fantastic spellings, as when Knighton (O.E. cnihtatun) commonly appears in Domesday as Chenistetone, or Gilling (O.E. Gillingas) is disguised as Gwyllingues. Such forms, if they appeared repeatedly in official documents, might actually be recognized as the correct names even locally. Thus the Ciren of Cirencester is the Norman-French spelling for the native English Churn, the river on which that town stands, and Durham is the result of the Norman's attempt to pronounce Dunholm, in which the sound-combinations nh and /m proved too difficult for him.