ery thoroughly indeed. The scenario was that "of course" hobbits
couldn't have spoken English (the story took place far in the past; rather, they spoke their own language, called
Westron (but often referred to as the Common Speech). Tolkien "translated" this language into English, which included "rendering" all the
Common Speech place-names into the equivalent English place-names.
The object of the exercise was to produce the following effect: names
in the Common Speech (which were familiar to the hobbits) were
"rendered" into English (in which form they would be familiar to us,
the English-speaking readers); names in other languages (usually
Sindarin) were "left alone", and thus were equally unfamiliar to the
hobbits and to us. Since the story was told largely from the hobbits'
point of view, that we should share their linguistic experience is a
desirable result (especially for Tolkien, who was unusually sensitive
to such matters).In portraying the linguistic landscape of Middle-earth he carried this procedure much further. The main example was his "substitution" of Anglo-Saxon for Rohirric. The "rationale" was that the hobbits' dialect of Westron was distantly related to Rohirric; therefore, when hobbits heard Rohirric they recognised many words but the language nevertheless remained just beyond understanding (RK, 65 (V,3)). Thus, Tolkien attempted to further "duplicate" hobbit linguistic perceptions by "substituting" that language of our world (Anglo-Saxon) which has (more-or-less) the same relation to English that Rohirric had to the hobbit version of Westron.
There were many other nuances in the intricate and subtle linguistic web he devised (always, he carefully explained, in the interests of "reproducing" the linguistic map of Middle-earth in a way that could be easily assimilated by modern English-speaking readers). Thus: