I’ve come across the Revd A Smythe Palmer's Folk Etymology (first published 1882) which defines itself as a
Dictionary of verbal corruption or word perverted in form or meaning by false derivation or mistaken analogy
It is an interesting reading, but is it reliable? Could it be cited as reference on ELU for instance?
I find Smythe Palmer’s dictionary almost impossible to use, in part because he presents the argument for a folk etymology in a deadpan way and only occasionally points out a fundamental flaw in that argument. Consider his dictionary’s entry for “blancmanger”:
BLANCMANGER : the latter part of this word is said to have no connexion with manger, to eat. The old spelling was blanc-mangier, and blanc-mengier, a corruption of ma-en-sire, i.e. “fowl-in-syrup,” which is the chief ingredient of the dish in old recipes, Its other names—Blanc Desire (i.e. de sire, “of syrup”), Blanc desorre, Blanc de sorry, Blanc de Surry—are of similar origin.—Kettner, Book of the Table, pp. 211–218. But where is this ma(?)-en-sire to be found?
The problem for a reader is how to tell (1) where the entry’s content (prior to the question at the end) comes from and (2) whether any of it is valid in any way. Is Kettner responsible for everything before the final question? Is anything in the first part of the entry valid? For example, is it true that the old spelling was blanc-mangier? Was “fowl-in-syrup” the chief ingredient of blancmange “in old recipes”? Are “Blanc Desire,” “Blanc de Surry,” etc., really old alternative names for blancmange?
Etymology Online offers this entry:
blancmange (n.) "jelly-like preparation in cookery," late 14c., from Old French blancmengier (13c.), literally "white eating," originally a dish of fowl minced with cream, rice, almonds, sugar, eggs, etc.; from blanc "white" (also used in Old French of white foods, such as eggs, cream, also white meats such as veal and chicken; see blank (adj.)) + mangier "to eat" (see manger). Attempts were made nativize it (Chaucer has blankemangere); French pronunciation is evident in 18c. variant blomange, and "the present spelling is a half attempt at restoring the French" [OED].
So Smythe Palmer’s observation that “the latter part of this word is said to have no connexion with manger, to eat” simply reports some unnamed person’s erroneous understanding of the word’s etymology, but the subsequent claim about the spelling blanc-mangier seems to be accurate, but the claim that blanc-mangier is itself a corruption of ma-en-sire appears to be bogus, and the remarks about “Blanc Desire” and related forms are therefore (presumably) unfounded. And we still don’t know how much of this folk etymology argument is the work of Kettner.
Ultimately, Smythe Palmer’s treatment of blancmange is not very helpful because it mixes accurate details with inaccurate ones, without specifying which are which, and because he doesn’t clearly identify the source of each etymological claim made in the body of the entry. These problems arise over and over again in his dictionary. I doubt that I've cited it more than once or twice in my many etymology-related answers at EL&U.
