Oxford University Press's
Academic Insights for the Thinking World

Which witch?

By Anatoly Liberman


To some people which and witch are homophones. Others, who differentiate between w and wh, distinguish them. This rather insignificant phenomenon is tackled in all books on English pronunciation and occasionally rises to the surface of “political discourse.” In the thirties of the past century, an irritated correspondent wrote to the editor about “the abuse of such forms as what, when, which, wheel, and others”: “Dictionaries in vain lay down the law that the h should be heard in such words. If heard at all it will probably come from the lips of Scotsmen, as they do give full value to the h. In this way the difference of a nationality can, as a rule, be detected. Long ago I had to be present at King’s College when the prizes were given away. A Mr. Wheeler was a winner of the Elocution prize; but he was called out as Mr. Weeler by, save the mark, the Professor of Elocution himself.” We’ll save the mark and go on.

In Old English, many words began with hl-, hn-, hr-, and hw-. In the beginning, the letter h stood for ch, as in Scots loch or gh as in the family name McLaughlin. Later it was weakened to h and lost. The same change occurred in the other Germanic languages, except Icelandic and, if I am not mistaken, Faroese. Sounds seldom disappear without a trace. Thus, when h was shed, it devoiced the consonant after it. In Icelandic, voiceless l, n, and r can easily be heard, but elsewhere they merged with l, n, and r in other positions. Only hw developed differently. It either stayed in some form or devoiced w.

It has never been explained why consonants tend to disappear before l, n, r, and w. A classic example of this process, not related to the subject being discussed here, is the fate of kn- and gn-, as in knock and gnaw. One can of course say that such groups are rare and inconvenient for pronunciation. But such an explanation is illusory, because it presents the result of the change as its cause. Outside English, kn- and gn- cause speakers no trouble. Besides, the loss of k- and g- happened at a certain time. Why did it “suddenly” become inconvenient to articulate the groups that had not bothered the previous generations? We will accept the history of hw as we find it and leave it to others to account for the change.

The reverse spelling (wh- for hw-) goes back to Middle English and can only confuse those who believe that modern spelling is a good guide to etymology. The letter writer, whose displeasure with dictionaries we have just witnessed, made no mistake. The speakers of London, where in the late Middle Ages the Modern English norm was being forged, lost h before w and accepted voiced w (this happened as early as the end of the fourteenth century), while northern England, Scotland, Ireland, and, to some extent, American English have either hw or voiceless w.

Yet some authorities who taught as late as the first half of the eighteenth century insisted on the necessity to enunciate h before w. They may have trusted the written image of the words in question. In 1654 and the subsequent decades, such opinions could no longer be heard. After voiced w had won the victory in southern speech, the “true” (historical) pronunciation was often recommended as correct and returned to solemn recitation and sometimes even to everyday speech. Such cases are not too rare. Consider the pronunciation often and fore-head, which owe their existence to modern spelling. Some people believe that the more “letters” they pronounce, the more educated they will sound. “Ofen” and “forid,” rhyming with soften and horrid, strike them as slipshod.

It is instructive to look at some Modern English words beginning with wh-. Quite a few, including when, where, what, and why, did once have hw- at the beginning. As a result, southerners have homophones like which ~ witch, when ~ wen, whither ~ wither, whale ~ wail, and so forth. (Shakespeare could not know that woe and wail are related, but his ear and instinct made him write the unforgettable alliterating line in Sonnet 30: “And with old woes new wail my dear time’s waste.”)

The pernicious habit of writing wh, sometimes for no obvious reason, resulted in the creation of several unetymological spellings. Whore, from Old English hore (a common Germanic noun), is akin to Latin carus “dear” (Italian caro, etc.). The Old English for whole was hal (with a long vowel). According to the OED, the spelling with wh-, corresponding to a widespread dialectal pronunciation with w, appeared in the sixteenth century. But why should this dialectal pronunciation have prevailed to such an extent that the spelling of an old and very common word was affected? Home also has a dialectal variant whoam, but, luckily, we still stay at home, rather than at whome. Equally puzzling is whelk (from weolc); here the influence of welk “pimple” has been pressed into service. Whig traces, though in a circuitous way, to a verb meaning “to drive”. Its wh- has no justification in history. Naturally, whim was bound to cause trouble, the more so as its earliest attested meaning is “pun”; no record of whim predates the seventeenth century. Then there is whiffler “an attendant armed with a weapon to keep the way clear for a procession,” from wifle “javelin” (Od Engl. wifel).

The consonant group hw- must always have made people think of blowing and light sweeping motions. Whistle, whisper, and whisk are rather obvious sound-imitating words (which does not mean that whisky ~ whiskey, from Gaelic, should have wh-; whisker, however, is derived for whisk, and its original sense was “brush”). Whir and whirl seem to belong with other onomatopoeic formations. Whew, an exclamation of astonishment, is an onomatopoeia pure and simple. Wheedle is late and has an obscure history.

Inglewhite, Lancashire.  (Cowfield. Grazing south of Langley Lane. Photo by Chris Shaw. CC BY-SA 2.0 via Wikimedia Commons)
Inglewhite, Lancashire. (Cowfield. Grazing south of Langley Lane. Photo by Chris Shaw. CC BY-SA 2.0 via Wikimedia Commons.)

By way of conclusion, I may mention several thw- words in which thw- once alternated with hw-. Today we remember only the verb thwart, but the adjective thwart “obstinate, perverse” also existed, and over-hwart has been attested. Another archaic word thwite “to cut” is a cognate of whittle. Thwack and whack used to alternate, and thwack is a synonym of dialectal thack. Apparently, thw- too had a sound-imitative value. In the place name Inglewhite (Lancashire), the second element was thwaite “meadow.” The last name Applewhite goes back to the place name Applethwaite in Cumberland. The change of thwaite to white is a product of folk etymology.

All this is very interesting, except that wh- is often an unnecessary embellishment. For the benefit of those who like learned words I may say that this group is sometimes otiose.

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 on the OUPblog each Wednesday. Send your etymology question to him care of blog@oup.com; he’ll do his best to avoid responding with “origin unknown.” Subscribe to Anatoly Liberman’s weekly etymology articles via email or RSS.

Subscribe to the OUPblog via email or RSS.
Subscribe to only language articles on the OUPblog via email or RSS.

Recent Comments

  1. Jeffrey Winter

    Is the opening “h” sound as described here a relic of the PIE “laryngeal” consonant?

  2. Jeffrey Winter

    is the opening “h” sound here a relic of the PIE “laryngeal” consonant?

  3. mollymooly

    I have heard the T in sofTen sounded. (By a Scot.)

    As you say, the H in English wHiskey owes nothing to the Irish. Whereas Irish English sounds both W and H, in Irish Gaelic even the W is now silent (or rather the U that matches the English W). The E in whiskEy is silent but vitally important.

Leave a Comment

Your email address will not be published. Required fields are marked *