By Anatoly Liberman
At the end of the 19th century, rare “underdone” had so little currency in the British Standard (“Queen’s English”) that educated speakers labeled it as an Americanism, though language historians, Skeat among them, knew better than that. He wrote in 1868: “…a very common old English and provincial English word” (the low case o in old indicates that older periods were meant; otherwise, Skeat would have said Anglo-Saxon). About fifteen years later, the London-based periodical The Academy asked the readers to send information about their familiarity with rare and was astounded to learn that the word was in use practically everywhere. In England, the more common form is rear or rere; as a rule, rare appears in British dictionaries with the reference: see rear. This fact alone is enough to show that rare “underdone” and rare “infrequent; unusual; precious” go back to different sources; they are homonyms. Rare, rhyming with “beyond compare,” poses no problems: it is an adjective English took over from French, while in French it is from Latin. By contrast, rare “underdone” is obscure.
Throughout its history, rare has been applied not only to meat but also to eggs. Moreover, it seems that, as far as references go, eggs were roasted rare or rarely (that is, with their white still fluid) more often than meat (the pun on rare/often is unintentional). In 1868 Skeat, already an experienced editor of Middle English texts, was a novice in matters etymological and attacked some hard questions with more vigor than prudence. He cited “two cognate” Old English forms (both having long vowels)—hrere and hreaw: “from the first…comes rere, and from the second raw. There is little difference in shape, and apparently none in meaning.” Alas, the meanings, as we will see, are not fully compatible, and “the difference in shape,” far from being little, is wide as a gulf, for long e and ea cannot alternate in one and the same root of an Old English word (in older Indo-European, vowels formed series that resembled trains, each of which was allowed to move along its own track and prohibited from encroaching on the track of its neighbor). In the first edition of his dictionary (1882), Skeat said that although a connection between rear/rare and raw had been suggested, it is “very doubtful.” Also in 1882, in a letter to The Academy, he derived long e in Old Engl. hrere from long o (a legitimate procedure) and compared hror with the Germanic verb for “stir.” Its reflexes have continued into Dutch (roeren), German, and all the modern Scandinavian languages. In English, the only cognate of hrere can perhaps be seen in reremouse “bat” (with rere- from hrere-). This is a reasonable conjecture in light of German Fledermaus (as in the title of the famous opera), a cognate of Engl. flittermouse “bat,” that is, a mouse with flittering, fluttering wings. (By the way, flatter belongs here too: a flatterer “flits around” the person whose vanity is to be gratified.) Later Skeat did not venture any hypothesis on the origin of rare “underdone” and confined himself to citing the Old English form.
The reason for this diffidence is clear. The editors of the OED doubted that rere- in reremouse should be connected with the verb for stirring and looked upon Old Engl. hreremus as a folk etymological alternation of hreathemus (modernized spelling), another word for “bat.” Hreathe- has fairly respectable, but not a hundred percent solid cognates in and outside Germanic meaning “shake”; so here too the bat comes out as a creature with quickly moving wings. The last edition of Skeat’s dictionary appeared in 1910, shortly before his death. Neither at that time nor in the eighties did he follow the OED slavishly, and many of his conjectures differ from and, as I think, sound more promising than James A.H. Murray’s and Henry Bradley’s (those were the first great editors of and etymologists for the OED), but old age taught him circumspection. So he gave up his former suggestion on the origin of rear ~ rere ~ rare and left the question open. There may have been another consideration. If rare had existed only as the opposite of hard-boiled, the sense “movable, easy to stir” would have been a good bridge to “with the white still fluid.” But if the original sense of rare was “uncooked” or “underdone,” flittering, fluttering, and the rest have nothing to do with it.
There still may be some way out of the difficulty. Contrary to the verdict of the OED, Old Engl. hreremus need not be an illegitimate side-form of hreathemus. The bat, feared for its habit of flying at night and for being half-bird and half-“beast,” became the object of numerous superstitions. To avoid calling the creature by its real name, lest it hear the call and pay an unwelcome visit, various taboo words came into existence. They sometimes resembled one another, but not each of them has an acceptable etymology. For example, bat, a 16th-centry word, ousted reremouse and replaced bakke ~ backe, that is, back, evidently a noun of Scandinavian origin. The k ~ t alternation, though not unheard of, is hard to explain. In Medieval Latin, blatta, blacta, and batta have been recorded, partly borrowed from Germanic and modified. Hreremus and hreathemus may have coexisted as two independent names of the bat.
The path from eggs to meat is more difficult to visualize. Perhaps rare was first applied to eggs and later broadened its meaning, but this is guesswork, unsupported by evidence or analogy. The opposition cooked ~ uncooked lies at the foundation of civilized life, and that is why, in Indo-European (to give one example), words for “raw” usually have transparent origins. For example, raw (from hreaw) is related to Latin crudus “raw; fresh; unripe,” and Russian krov’ “blood” (both raw and krov’ have close cognates in Germanic and Slavic respectively; the alternation non-Germanic k ~ Germanic h is regular). Since in contradistinction to them, Engl. rare (from hrere) is isolated, it probably sprang up as a metaphor or an extension of some other meaning. Anyway, rare means “underdone,” not “uncooked,” but whether it developed from “moving, movable; fluid” remains unclear. A low probability of such a process should be admitted.
Two more remarks are in order. 1) Rare “underdone” competes with rare “early” (now dialectal); it is a variant of rather (compare whe’er “whether”). Rather is the comparative degree of an adjective meaning “quick”; hence the suffix -er. Despite some similarity of meaning (what is underdone has been taken off the fire too early), the two words are different. 2) Among the reference works not devoted exclusively to etymology, only the OED, The Century Dictionary, and Henry C. Wyld’s The Universal Dictionary of the English Language contain original or nontrivial information on word origins. Wyld, an excellent scholar, suggests tentatively that rare “underdone” may be related to Latin crudus. Obviously, not only Homer was given to occasional nodding. If it could be shown that crudus, which is certainly akin to raw, is akin to rare, then rare and raw would emerge as related—an unacceptable conclusion. Returning to my favorite idea (“meaning and etymology form a symbiosis”), I must say that the result of our investigation does not surprise me: a word meaning “underdone” was doomed to have a “medium rare” etymology.
Anatoly Liberman is the author of Word Origins…And How We Know Them as well as An Analytic Dictionary of English Etymology: An Introduction. His column on word origins, The Oxford Etymologist, appears here, each Wednesday. Send your etymology question to firstname.lastname@example.org; he’ll do his best to avoid responding with “origin unknown.”