Break and Brake

By Anatoly Liberman

Like a few other essays I have written in the past, this one has been inspired by a question too long for inclusion in the “gleanings.” Are break and brake related? Yes, they are, but the nature of their relationship deserves a detailed explanation. Break is an ancient word. It has cognates in all the Germanic languages, and Latin frango, whose root shows up in the borrowed words fragile, fragment, and refract, is believed to be allied to it (the infix n may be disregarded for reconstructing the protoform).

The grammatical system of the Germanic (and other Indo-European) languages depended on vowel alternations of the type we still have in break ~ broke, rise ~ rose ~ risen, drink ~ drank ~ drunk, give ~ gave, and so forth. Vowels were arranged in non-intersecting sets and resembled parallel railway tracks. Occasional shunting was allowed, but each move required special dispensation. The principal parts of break in Old English were brecan (infinitive), bræc (preterit singular; æ, as in Modern Engl. man), and brocen (past participle). All the highlighted vowels were short. At that time, verbs like break (so-called strong verbs, which displayed such alternations) had four principal parts, because the preterit singular differed from the preterit plural (the modern language has retained this distinction only in be ~ was ~ were ~ been), but three will suffice for comparing break and brake. In the history of English, vowels have been shortened and lengthened so often, and so many later changes have interfered with the ancient system that the original state is hard to observe from the perspective of the modern language. The vowel of the infinitive underwent lengthening and diphthongization; this accounts for today’s sound shape of break. The past plural form has disappeared altogether, and the extant form broke has the vowel of the past participle, also lengthened and diphthongized.

While I am at it, I may mention that in Middle English the ending -en was usually shed (compare English and German infinitives: break versus brechen), but after a good deal of vacillation past participles retained it (so in broken, spoken, given, and so forth, though we have come ~ came ~ come, as opposed to German kommen ~ kam ~ gekommen). Yet when we are bankrupt, we go broke. From an etymological point of view, broke is the same word as broken. Also, those who can afford it wear bespoke suits; recently, bespoke has spread to computer technology. Bespoke is a variant of bespoken.

The system of vowel alternations, as in Old Engl. brecan ~ bræc ~ brocen, that is, e ~ æ ~ o, is called ablaut (a term coined by Jacob Grimm, the elder and the more famous of the two brothers, who did many things in addition to collecting folk tales), and each vowel represents what is technically called a grade of ablaut. If bræc had not been lost, the modern past tense of break would, most probably, have been brake (compare spake, the archaic preterit of speak). No reflex of bræc has survived, but several words formed on that grade of ablaut are still with us. Some of the nouns spelled brake are among them. The problem is that brake has a bewildering number of homophones. This is what we find in the OED (I will reproduce only the most important senses, with the dates of the first known occurrence in parentheses): “bracken” (ca. 1325), “thicket” (ca. 1440), “a toothed instrument for braking (note the spelling) flax or hemp, a baker’s kneading machine” (ca. 1450), “a lever or handle for working a machine” (1380; now obsolete, but “the handle of a pump” continued into the 19th century), “a bridle or curb” (1430; seemingly not recorded after 1753), “a cage of iron or wooden bars; a trap” (1529; this sense is obsolete, but the word is extant as designating various frames), and finally, brake or break (!) “an apparatus for retarding the motion of a wheel” (1772). Only the last of them will occupy us at the moment. The OED says that both the spelling and the etymology of this word are uncertain. Today brake is spelled only so. The verb brake “to apply brakes” is given with one spelling, and no examples of it predating 1869 have been found.

We can see that, except for brake “an apparatus for retarding the motion of the wheel,” all the other nouns spelled so showed up between 1325 and 1529. At least two of them may have been borrowed from French, while the others were presumably coined at the epoch when ablaut was still productive, that is, when one could form a new word by resorting to the vowel alternation outlined above. For example, we have sing and song. They represent the same root on different grades of ablaut. A drunk is someone who drinks too much. The past of wring is now wrung. Wrong and wrangle are its cognates; here too the root is the same, but the vowels differ.

We can no longer form words according to this model. It will not occur to anyone to call someone who rings a bell a rung. The origin of the word rote (“to learn something by rote”) is obscure (Romance: allied to route ~ routine or rotation?), but we would be surprised if it turned out that it is akin to rite, though rote occurred in the 14th century (and in a remote way has something to do with a recurring ritual) and though we take the alternation write ~ wrote and ride ~ rode for granted. Even when ablaut is acceptable to us on chronological grounds but its grade looks aberrant, we have a hard time explaining the result. Why is Shrovetide not called Shrivetide? The past tense of shrive is indeed shrove, but Shrovetide should have been formed with the vowel of the present. Equally puzzling is spokesman, a 16th-century noun, coined as the name of a “speaksman.” However, false, or irregular, ablaut is still productive, only that vowels governed by it alternate without much system: compare tit (as in tit for tat), tid(bit), tad, tot, toddle, and tut-tut, all of which suggest smallness but which are not related the way get and got are. In our consciousness, pet and pot are not connected (despite the get ~ got pair); yet from time to time some atavistic mechanism begins to work and people produce wodge from wedge, brolly from umbrella (these are real words!) and frosh from fresh(man). Are there more such nouns, and is the game limited to nouns? I ran into wodge and brolly by chance.

Brake (as in a car) surfaced at the end of the 18th century, when old ablaut was, as it still is, tolerated in inherited words (sing ~ song, write ~ wrote, drink ~ drunk) but no longer productive. It follows that the word could not be created on the sing ~ song model. The name of some other brake must have been transferred to that device. Several candidates suggest themselves: “an instrument for braking flax or hemp,” “a lever or handle for working a machine,” and “bridle or curb” (the last of them is especially cogent). But one cannot discuss the name of a tool without knowing everything about the tool’s history. When and where did it come into use? Was it imported? What did it look like? The Oxford Dictionary of English Etymology summarizes the situation so. Middle Dutch braeke, the name of various breaking or crushing instruments, seems to have been adopted in English at different periods with different applications related to “breaking.” If, as the OED comments, such was the history of brake “an apparatus of diminishing speed,” it is a product of folk etymology, for the apparatus “broke” the motion.

Old Engl. brecan and Middle Dutch braeke had different stressed vowels: e sounded like e in Engl. bet, whereas Dutch ae designated long a, as in Engl. father. Later, e was lengthened and produced a mid-open vowel; hence the spelling ea. Eventually it should have yielded the same vowel as in speak, but for the reason that has never been explained to everybody’s satisfaction, break and steak (to cite two most frequent words) avoided the trend and are pronounced with the vowel of make. As a result, break and brake are now homophones. Nothing would have happened if they were spelled alike (ideally, brake), but who expects logic from English spelling? As is known, Cato the Elder used to end each of his speeches with the phrase “Carthago delenda est” (“Carthage should be destroyed”). I have a similar attitude to Modern English spelling.

Perhaps someone is also interested in the origin of breaker “billow.” It is indeed break + the suffix -er (the wave breaks on the shore). I will say more about the verb break next week.

Anatoly_libermanAnatoly 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; he’ll do his best to avoid responding with “origin unknown.”

One Response to “Break and Brake
  1. [...] is the promised continuation of the previous post.  As I said last week, break is an old word.  In the foggy days of Proto-Indo-European, it may have begun with the [...]

Leave a Reply