There is a folk etymologist in all of us. We expect sever, severe, and severy (“compartment of a roof”) to be cognates and resent the fact that they are not. Poets and punsters connect such words, while etymologists, these dry-as-dust killjoys, ruin native intuition. It is a pity that rhyme has fallen into desuetude, for rhyme is a great uniter. Consider the following. English love is gentle, almost heavenly, because love rhymes with above, dove, and, to a certain extent, move (we will disregard shove for the sake of argument). French love never ends, for otherwise, why should amour have rhymed with toujours “always”? Russians are a different matter, as follows from the triad liubov’ “love,” krov’ “blood,” and vnov’ “again.” German young people of both sexes (sexes, not genders, for German has three genders) have been a model of propriety since the beginning of creation: consider the time-honored rhyme Jugend “youth”/ Tugend “virtue.” A good deal has been written about such convergences, but it is equally interesting to observe how easily language drives a wedge between closely related words. One small phonetic change suffices to obliterate the ties between related words. A few examples come to mind.
Batch, to quote Skeat, is as much as is baked at once; hence, a quantity. This is right; yet we no longer associate batch and bake. In Modern English, the consonants designated in spelling by k and ch respectively alternate with some regularity. Not infrequently, the k-forms are northern, and the ch-forms southern: compare kirk and church, mickle ~ muckel (both regional) and much. Occasionally both forms belong to the Standard, and the gap between them is not too great. For instance, we are ready to agree that seek is a cognate of beseech. The two verbs have strayed apart, but not as far as bake and batch, even though their preterits are different: sought versus beseeched (besought also existed but in the speech of the majority gave way to the more productive type). Another pair whose affinity has been destroyed in our consciousness by the alternation k ~ ch is wake and watch. The etymological meaning of watch is “to be awake.” Today watch means “to look carefully, to observe,” but those who watch for an opportunity must be wakeful, that is, wide awake. The difference in vowels (short in watch, long in wake) also contributed to the rift between those doublets.
A striking example of a how a minor change of a vowel can make the origin of a word impenetrable is trade. Trade, which was borrowed in this form from Middle Low (that is, northern) German in the 14th century, meant “a path, a track,” hence “a beaten track; regular business; buying and selling.” It is related to the verb tread. Middle English had trede “a tread, a step” and trod “a track.” Not only do we dissociate tread from trade despite their material closeness; the etymology of trade comes to the uninitiated as a surprise.
Sometimes grammar, in conjunction with phonetics, plays havoc with “family ties.” Today hardly anyone realizes that truce is, from a historical point of view, the plural of true. Yet what can be more obvious? In the 13th century, when truce first turned up in English texts, it was spelled trewes, trews, and trues, the plural of trew “pledge, promise.” The meaning of the ending was forgotten or disregarded (compare the modern names of sciences like physics : physics is), the pronunciation trues changed to truce, and the form we now know emerged. Similar adventures have been recorded elsewhere. Sometimes final –s has been misinterpreted in foreign words as a plural ending. The anthologized examples are Engl. pea, which evolved from pease (Old Engl. pise, late Latin pisa), cherry (compare French cerise), skate, abstracted from Dutch schaats (whose plural in Dutch is schaatsen), and (the most outrageous of them all) Chinee, from Chinese. However, the history of truce has parallels. Pence (a variant of pennies in compounds like twopence) is still plural, but it is neither spelled nor pronounced like pens. Dice should have been a homophone of dies, but the ending, again in a collective noun, was devoiced (z changed to s). Bodice is a disguised spelling of bodies, with the ending devoiced, as in pence, and body meaning “part of a woman’s dress above the waist” (compare corset, a diminutive of Old French cors “body”). Nowadays, when even bras have been discarded along with other appurtenances of bourgeois priggishness, the word bodice is hardly ever used, but the antiquated phrase a pair of bodies “bodice” and the still familiar (from literature) a pair of stays make the derivation of bodice from bodies sustainable, as at the moment everybody says instead of satisfactory, usable, acceptable. Some etymological solutions look like circus stunts, but no less often the indubitable etymology is right there, for anyone to see, and only a slight quirk dims our vision.
What else is there to say? The ultimate source of sever is Latin separare, so that sever and separate are doublets. Severe goes back (via Old French) to the Latin adjective severus. Severy, a word revived in the 19th century, is a doublet of ciborium “canopy; a vessel for the Eucharistic bread.” Truant is not related to true or truce, cherry and cherish are not congeners, and pea jacket has nothing to do with peas. Peanut, however, is indeed pea + nut.
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.”