[Aiwendil]

mhagain - you make a good argument. However, I am far from convinced of this:

that they are the same word (rog/raug/arauka|rauko), and that the meaning of this word has changed since LT was written.

I do agree that the whole linguistic part of the debate hinges on this. If the name 'Rog' and the element '-rog' in 'Balrog' are identical, then the later meaning of that word ('demon') makes it inappropriate for an Eldarin lord. However, if the two elements are different, then the word 'rog' meaning 'demon' cannot be taken as evidence for the unsuitability of the name 'Rog' in FoG.

I was once of the former opinion, but prior to our completion of the FoG draft, I changed my mind.

Let me set out the linguistic evidence.

From 'Names in the Lost Tales part 2' (HoMe II):

Rog GL gives an adjective r^og, rog 'doughty, strong'. But with the Orcs' name for Egnor Beren's father, Rog the Fleet, cf. arog 'swift, rushing', and raug of the same meaning; Qenya arauka.

From this it is clear that we have, at the Lost Tales stage:

1. An element 'rog' = 'doughty, strong' (explicitly in GL)
2. An element 'rog'/'raug', Q. form 'arauka' = 'swift, rushing' (surmised by CRT based on very strong evidence).

In 'Names in the Lost Tales part 1' (I) we have:

Balrog [. . . ] Separate entries give bal 'anguish (original initial consonant mb-), balc 'cruel'; and graug 'demon'. Qenya forms are mentioned: arauke and Malkarauke.

We have also, then:

3. An element 'graug' = 'demon' (explicitly in GL).

The evidence for 1 and 3 comes from a single source, GL. The evidence for 2 is partially in GL, partially in QL, and partially in the 'Tale of Tinuviel'. It is clear, then, that, unless we posit some rather intricate and baroque developments during the writing of GL, these three elements coexisted simultaneously. We have, then, not one or even two distinct words but three.

Now, after the LT stage, elements 1 and 2 are not given in any etymological discussion. Element 3 retains its meaning but is altered slightly in form in the Etymologies (V):

RUK- demon. Q ranko demon, malarauko (*ngwalarauko, cf. NGWAL); N. rhaug, Balrog.

So Gnomish 'graug' becomes Noldorin 'rhaug'. At both stages, the 'au' diphthong appears to be resolved to an 'o' in compounds.

It is worth noting, also, that the character Rog of Gondolin still appears in the 1930 Q (IV):

Of the deeds of desperate valour there done, by the chieftains of the noble houses and their warriors, and not least by Tuor, is much told in The Fall of Gondolin; of the death of Rog without the walls; and of the battle of Ecthelion of the Fountain with Gothmog lord of Balrogs in the very square of the king . . .

Taken together, I think that all this makes it quite clear that the elements 'rog' in 'Rog' and 'graug/rhaug' in 'Balrog' were distinct from the beginning. This in itself is enough to take the force out of the argument to the effect that the name 'Rog' is impermissible because it must mean 'demon'. We can further demonstrate that, as late as 1930, the elements remained distinct; otherwise the name 'Rog' would not appear in Q.

So, all evidence points to 'rog' = strength and 'rhaug' = demon being unrelated elements; there is no indication anywhere that this situation was ever altered.