True, the use of "comment" in "comment dit-on … en français" and even more so in "comment appelle-t-on …" or "comment s'appelle …", does not seem any more logical in French than it is in English, but there is a difference: the English noticed that—unless they were lucky and hit upon the right idea without much thinking—, and the French lived with it without thinking that it mattered much. Be that as it may, there does not exist two ways of thinking logically, one English and another, French; so, sound logic in English grammar is also sound in French grammar, and vice-versa.
Nevertheless, the apparent "usurpation" of the concept of manner, whereas one considers more to the point the idea of correspondence, finds some legitimacy in the speculative reasoning that I will expose shortly. When there is at stake the translation of a mere word, the forms "Quel est le mot anglais pour …" and "Quel mot anglais traduit …", "Qu'est-ce que l'on dit en anglais pour dire …", etc. are quite correct, and used (by a minority of speakers); therefore, there is no rejection of the "English" point of view (or more evident point of view, shall we say), and this understanding, that we can consider as unique in English, is no stranger to French.
A case of mixed uses for this specific instance of use of "appeler"
Let's remark that "Qu'appelle-t-on XXX ?" is proper for asking the definition of "XXX". We notice from this page that usage is uncertain so that quotes ("« »" and equivalently '" "') are used often enough about "XXX" but not in the majority of cases, this being due to the general lack of understanding in using the symbolism of quotes; however there is no difference, and the meaning is always that without quotes.
- On appelle penser cette activité qui …
(TLFi) [attribut de l'objet][Le verbe est constr. avec un attribut de l'obj.] Appeler qqn ou qqc. [L'attribut (subst., verbe à l'inf., etc.) de l'obj. est une dénomination classificatrice]
A point to make precise before going on : the idea of correspondence is not implemented through a relation of object of a verb but through that of object complement (as well in French the term being "attribut de l'objet").
- You call a match "allumette". ("match" is the object, '"allumette"' is the object complement)
- On appelle une allumette "match". ("allumette" is the "objet", '"match"' is the "attribut de l'objet")
(ref.) Cependant, pour faciliter les calculs, il est préférable d'utiliser ce qu'en anglais on appelle « bulk modulus » que l'on peut traduire par le module de ...
(ref.) Pour dresser un état de la recherche ( ce qu'en anglais on appelle un « State of the Art » ) , il faut d'abord s'informer
You can see from these two examples that "En anglais on appelle cette chose « bulk modulus »" answers to both the questions "Qu'appelle-t-on cette chose en anglais ?" and "En anglais comment appelle-t-on cette chose ?", and similarly for "En anglais on appelle un état de la recherche un « state of the art »".
The general shift towards a point of view based on manner could have its root in the thinking process explained next, but nothing is less sure, and just as well this initial thinking process might be that of a rough approximation. Instead of apprehending the unknown linguistic element in the lexicon as merely a linguistic form it could have been apprehended as the concept (common to both lexicons), and then the mere arrangement of letters and sounds that pertains to one or the other of the corresponding elements in the lexicons could have been seen as merely a way, a manner to word the concept. It is clear that in the speaker's mind when confronted to linguistic concepts, there is not always a clear-cut difference between the word and its referent. According to this possible perspective, in fact, in English one would say "How do you put this concept in English/French?", the proper point of view appearing now to be that of manner rather than another one. This is a perspective that takes on relevance when we are dealing with whole expressions instead of words, as then a particular way of doing things becomes of much greater concern.