The Romance of Words (4th ed.)

It is simply a translation of Fr. main de gloire. But the French expression is a popular corruption of mandragore, from Lat. mandragora, the mandragore, or mandrake, to the forked roots of which a similar virtue was attributed, especially if the plant were obtained from the foot of the gallows.

Akin to folk-etymology is contamination, i.e., the welding of two words into one. This can often be noticed in children, whose linguistic instincts are those of primitive races. I have heard a child, on her first visit to the Zoo, express great eagerness to see the canimals (camels × animals), which, by the way, turned out to be the giraffes. A small boy who learnt English and German simultaneously evolved, at the age of two, the word spam (sponge × Ger. Schwamm). In a college in the English midlands, a student named Constantine, who sat next to a student named Turpin, once heard himself startlingly addressed by a lecturer as Turpentine. People who inhabit the frontier of two languages, and in fact all who are in any degree bilingual, must inevitably form such composites occasionally. The h aspirate of Fr. haut, Lat. altus, high, can only be explained by the influence of Old High Ger. hōh (hoch). The poetic word glaive cannot be derived from Lat. gladius, sword, which has given Fr. glai, an archaic name for the gladiolus. We must invoke the help of a Gaulish word cladebo, sword, which is related to Gaelic clay-more, big sword. It has been said that in this word the swords of Cæsar and Vercingetorix still cross each other. In Old French we find oreste, a storm, combined from orage and tempeste (tempête). Fr. orteil, toe, represents the mixture of Lat. articulus, a little joint, with Gaulish ordag. A battledore was in Mid. English a washing beetle, which is in Provençal batedor, lit. beater. Hence it seems that this is one of the very few Provençal words which passed directly into English during the period of our occupation of Guienne. It has been contaminated by the cognate beetle.

Cannibal is from Span. canibal, earlier caribal, i.e. Carib, the n being perhaps due to contamination with Span. canino, canine, voracious. It can hardly be doubted that this word suggested Shakespeare's Caliban. Seraglio is due to confusion between the Turkish word serai, a palace, and Ital. serraglio, "an inclosure, a close, a padocke, a parke, a cloister or secluse" (Florio), which belongs to Lat. sera, a bolt or bar.

Anecdotage is a deliberate coinage ascribed to John Wilkes—

In some cases it is impossible to estimate the different elements in a word. Arbour certainly owes its modern spelling to Lat. arbor, a tree, but it represents also Mid. Eng. herbere, erbere, which comes, through French, from Lat. *herbarium. But this can only mean herb-garden, so that the sense development of the word must have been affected by harbour, properly "army-shelter," ultimately identical with Fr. auberge (p. 164). When Dryden wrote—

he was expressing a composite idea made up from the verb seal, Old Fr. seeler (sceller), Lat. sigillare, and seel, Old Fr. ciller, Vulgar Lat. *ciliare, from cilium, eye-brow. The latter verb, meaning to sew together the eyelids of a young falcon, was once a common word—

The verb fret is Anglo-Sax. fretan, to eat away (cf. Ger. fressen). Fret is also used of interlaced bars in heraldry, in which sense it corresponds to Fr. frette with the same meaning; for this word, which also means ferrule, a Vulgar Lat. *ferritta (ferrum, iron) has been suggested. When Hamlet speaks of—

is he thinking of frets in heraldry, or of fretwork, or are these two of one origin? Why should fret, in this sense, not come from fret, to eat away, since fretwork may be described as the "eating away" of part of the material? Cf. etch, which comes, through Dutch, from Ger. ätzen, the factitive of essen, to eat. But the German for fretwork is durchbrochene Arbeit, "broken-through" work, and Old Fr. fret or frait, Lat. fractus, means "broken." Who shall decide how much our fretwork owes to each of these possible etymons?

That form of taxation called excise, which dates from the time of Charles I., has always been unpopular. Andrew Marvell says that Excise—

Dr Johnson defines it as "a hateful tax levied upon commodities, and adjudged not by the common judges of property, but wretches hired by those to whom excise is paid," an outburst which Lord Mansfield considered "actionable." The name, like the tax, came from the Netherlands, where it was called accijs—

In modern Dutch it has become accijns, through confusion with cijns, tax (Lat. census; cf. Ger. Zins, interest). But the Dutch word is from Fr. accise, which appears in medieval Latin as accisia, as though connected with "cutting" (cf. tallage, from Fr. tailler, to cut), or with the "incidence" of the tax. It is perhaps a perversion of Ital. assisa, "an imposition, or taxe, or assesment" (Torriano); but there is also an Old Fr. aceis which must be related to Latin census.

When folk-etymology and contamination work together, the result is sometimes bewildering. Thus equerry represents an older querry or quirry, still usual in the 18th century. Among my books is—

The modern spelling is due to popular association with Lat. equus. But this querry is identical with French écurie, stable, just as in Scottish the post often means the postman. And écurie, older escurie, is from Old High Ger. scura[100] (Scheuer, barn). The word used in modern French in the sense of our equerry is écuyer, older escuier, Lat. scutarius, shield-bearer, whence our word esquire. This écuyer is in French naturally confused with écurie, so that Cotgrave defines escuyrie as "the stable of a prince, or nobleman; also, a querry-ship; or the duties, or offices belonging thereto; also (in old authors) a squire's place; or, the dignity, title, estate of an esquire."

Ignorance of the true meaning of a word often leads to pleonasm. Thus greyhound means hound-hound, the first syllable representing Icel. grey, a dog. Peajacket is explanatory of Du. pij, earlier pye, "py-gown, or rough gown, as souldiers and seamen wear" (Hexham). On Greenhow Hill means "on green hill hill," and Buckhurst Holt Wood means "beech wood wood wood," an explanatory word being added as its predecessor became obsolete. The second part of salt-cellar is not the same word as in wine-cellar. It comes from Fr. salière, "a salt-seller" (Cotgrave), so that the salt is unnecessary. We speak pleonastically of "dishevelled hair," while Old Fr. deschevelé, lit. dis-haired, now replaced by échevelé, can only be applied to a person, e.g., une femme toute deschevelée, "discheveled, with all her haire disorderly falling about her eares" (Cotgrave). The word cheer meant in Mid. English "face." Its French original chère scarcely survives except in the phrase faire bonne chère, lit. "make a good face," a meaning preserved in "to be of good cheer." In both languages the meaning has been transferred to the more substantial blessings which the pleasant countenance seems to promise, and also to the felicity resulting from good treatment. The true meaning of the word is so lost that we can speak of a "cheerful face," i.e., a face full of face.

But there are many words whose changes of form cannot be altogether explained by any of the influences that have been discussed in this and the preceding chapters. Why should cervelas, "a large kind of sausage, well season'd, and eaten cold in slices" (Kersey's Eng. Dict., 1720), now be saveloy? We might invoke the initial letters of sausage to account for part of the change, but the oy remains a mystery. Cervelas, earlier cervelat, comes through French from Ital. cervellato, "a kinde of dry sausage" (Florio), said to have been originally made from pig's brains. For hatchment we find in the 16th century achement, and even achievement. It is archaic Fr. hachement, the ornamental crest of a helmet, etc., probably derived from Old Fr. achemer, variant of acesmer, to adorn. Hence both the French and English forms have an unexplained h-, the earlier achement being nearer the original. French omelette has a bewildering history, but we can trace it almost to its present form. To begin with, an omelet, in spite of proverbs, is not necessarily associated with eggs. The origin is to be found in Lat. lamella, a thin plate,[101] which gave Old Fr. lamelle. Then la lamelle was taken as l'alamelle, and the new alamelle or alemelle became, with change of suffix, alemette. By metathesis (see p. 59) this gave amelette, still in dialect use, for which modern French has substituted omelette. The o then remains unexplained, unless we admit the influence of the old form œuf-mollet, a product of folk-etymology.

Counterpane represents Old Fr. coute-pointe, now corruptly courte-pointe, from Lat. culcita puncta, lit. "stitched quilt"; cf. Ger. Steppdecke, counterpane, from steppen, to stitch. In Old French we also find the corrupt form contrepointe which gave Eng. counterpoint—

in modern English replaced by counterpane. Mid. English has also the more correct form quilt-point, from the Old Norman cuilte (pur)pointe, which occurs in a 12th-century poem on St Thomas of Canterbury. The hooped petticoat called a farthingale was spelt by Shakespeare fardingale and by Cotgrave vardingall. This is Old Fr. verdugalle, of Spanish origin and derived from Span. verdugo, a (green) wand, because the circumference was stiffened with flexible switches before the application of whalebone or steel to this purpose. The crinoline, as its name implies, was originally strengthened with horse-hair, Lat. crinis, hair. To return to the farthingale, the insertion of an n before g is common in English (see p. 84, n. 2), but the change of the initial consonant is baffling. The modern Fr. vertugadin is also a corrupt form. Isinglass seems to be an arbitrary perversion of obsolete Du. huyzenblas (huisblad), sturgeon bladder; cf. the cognate Ger. Hausenblase.

Few words have suffered so many distortions as liquorice. The original is Greco-Lat. glycyrrhiza, lit. "sweet root," corrupted into late Lat. liquiritia, whence Fr. réglisse, Ital. legorizia, regolizia, and Ger. Lakritze. The Mid. English form licoris would appear to have been influenced by orris, a plant which also has a sweet root, while the modern spelling is perhaps due to liquor.

CHAPTER X

DOUBLETS

The largest class of doublets is formed by those words of Latin origin which have been introduced into the language in two forms, the popular form through Anglo-Saxon or Old French, and the learned through modern French or directly from Latin. Obvious examples are caitiff, captive; chieftain, captain; frail, fragile. Lat. discus, a plate, quoit, gave Anglo-Sax. disc, whence Eng. dish. In Old French it became deis (dais), Eng. dais, and in Ital. desco, "a deske, a table, a boord, a counting boord" (Florio), whence our desk. We have also the learned disc or disk, so that the one Latin word has supplied us with four vocables, differentiated in meaning, but each having the fundamental sense of a flat surface.

Dainty, from Old Fr. deintié, is a doublet of dignity. Ague is properly an adjective equivalent to acute, as in Fr. fièvre aigue. The paladins were the twelve peers of Charlemagne's palace, and a Count Palatine is a later name for something of the same kind. One of the most famous bearers of the title, Prince Rupert, is usually called in contemporary records the Palsgrave, from Ger. Pfalzgraf, lit. palace count, Ger. Pfalz being a very early loan from Lat. palatium. Trivet, Lat. tripes, triped-, dates back to Anglo-Saxon, its "rightness" being due to the fact that a three-legged stool stands firm on any surface. In the learned doublets tripod and tripos we have the Greek form. Spice, Old Fr. espice (épice), is a doublet of species. The medieval merchants recognised four "kinds" of spice, viz., saffron, cloves, cinnamon, nutmegs.

Coffin is the learned doublet of coffer, Fr. coffre, from Lat. cophinus. It was originally used of a basket or case of any kind, and even of a pie-crust—

Its present meaning is an attempt at avoiding the mention of the inevitable, a natural human weakness which has popularised in America the horrible word casket in this sense. The Greeks, fearing death less than do the moderns, called a coffin plainly σαρκοφάγος, flesh-eater, whence indirectly Fr. cercueil and Ger. Sarg.

The homely mangle, which comes to us from Dutch, is a doublet of the warlike engine called a mangonel—

which is Old French. The source is Greco-Lat. manganum, apparatus, whence Ital. mangano, with both meanings. The verb mangle, to mutilate, is unrelated.

Sullen, earlier soleyn, is a popular doublet of solemn, in its secondary meaning of glum or morose. In the early Latin-English dictionaries solemn, soleyn, and sullen are used indifferently to explain such words as acerbus, agelastus, vultuosus. Shakespeare speaks of "customary suits of solemn black" (Hamlet, i. 2), but makes Bolingbroke say—

while the "solemn curfew" (Tempest, v. 1) is described by Milton as "swinging slow with sullen roar" (Penseroso, l. 76). The meaning of antic, a doublet of antique, has changed considerably, but the process is easy to follow. From meaning simply ancient it acquired the sense of quaint or odd, and was applied to grotesque[102] work in art or to a fantastic disguise. Then it came to mean buffoon, in which sense Shakespeare applies it to grim death—

and lastly the meaning was transferred to the capers of the buffoon. From Old High Ger. faltan (falten), to fold, and stuol (Stuhl), chair, we get Fr. fauteuil. Medieval Latin constructed the compound faldestolium, whence our ecclesiastical faldstool, a litany desk. Revel is from Old Fr. reveler, Lat. rebellare, so that it is a doublet of rebel. Holyoak's Latin Dictionary (1612) has revells or routs, "concursus populi illegitimus." Its sense development, from a riotous concourse to a festive gathering, has perhaps been affected by Fr. réveiller, to wake, whence réveillon, a Christmas Eve supper, or "wake." Cf. Ital. vegghia, "a watch, a wake, a revelling a nights" (Florio).

The very important word money has acquired its meaning by one of those accidents which are so common in word-history. The Roman mint was attached to the temple of Juno Moneta, i.e., the admonisher, from monēre, and this name was transferred to the building. The Romans introduced moneta, in the course of their conquests, into French (monnaie), German (Münze), and English (mint). The French and German words still have three meanings, viz., mint, coin, change. We have borrowed the French word and given it the general sense represented in French by argent, lit. silver. The Ger. Geld, money, has no connection with gold, but is cognate with Eng. yield, as in "the yield of an investment," of which we preserve the old form in wergild, payment for having killed a man (Anglo-Sax. wer). To return to moneta, we have a third form of the word in moidore—

from Port. moeda de ouro, money of gold.

Sometimes the same word reaches us through different languages. Thus charge is French and cargo is Spanish, both belonging to a Vulgar Lat. *carricare from carrus, vehicle. In old commercial records we often find the Anglo-Norman form cark, a load, burden, which survives now only in a metaphorical sense, e.g. carking, i.e. burdensome, care. Lat. domina has given us through French both dame and dam,[103] and through Spanish duenna; while Ital. donna occurs in the compound madonna and the donah of the East End costermonger. Lat. datum, given, becomes Fr. dé and Eng. die (plural dice). Its Italian doublet is dado, originally cubical pedestal, hence part of wall representing continuous pedestal. Scrimmage and skirmish are variant spellings of Fr. escarmouche, from Ital. scaramuccia, of German origin (see p. 64, n.). But we have also, more immediately from Italian, the form scaramouch. Blount's Glossographia (1674) mentions Scaramoche, "a famous Italian Zani (see p. 45), or mimick, who acted here in England, 1673." Scaramouch was one of the stock characters of the old Italian comedy, which still exists as the harlequinade of the Christmas pantomime, and of which some traces survive in the Punch and Judy show. He was represented as a cowardly braggart dressed in black. The golfer's stance is a doublet of the poet's stanza, both of them belonging to Lat. stare, to stand. Stance is Old French and stanza is Italian, "a stance or staffe of verses or songs" (Florio). A stanza is then properly a pause or resting place, just as a verse, Lat. versus, is a "turning" to the beginning of the next line.

Different French dialects have supplied us with many doublets. Old Fr. chacier (chasser), Vulgar Lat. *captiare, for captare, a frequentative of capere, to take, was in Picard cachier. This has given Eng. catch, which is thus a doublet of chase. In cater (see p. 63) we have the Picard form of Fr. acheter, but the true French form survives in the family name Chater.[104] In late Latin the neuter adjective capitale, capital, was used of property. This has given, through Old Fr. chatel, our chattel, while the doublet catel has given cattle, now limited to what was once the most important form of property. Fr. cheptel is still used of cattle farmed out on a kind of profit-sharing system. This restriction of the meaning of cattle is paralleled by Scot. avers, farm beasts, from Old Fr. aver[105] (avoir), property, goods. The history of the word fee, Anglo-Sax. feoh, cattle, cognate with Lat. pecus, whence pecunia, money, also takes us back to the times when a man's wealth was estimated by his flocks and herds; but, in this case, the sense development is exactly reversed.

Fr. jumeau, twin, was earlier gemeau, still used by Corneille, and earlier still gemel, Lat. gemellus, diminutive of geminus, twin. From one form we have the gimbals, or twin pivots, which keep the compass horizontal. Shakespeare uses it of clockwork—

and also speaks of a gimmal bit (Henry V., iv. 2). In the 17th century we find numerous allusions to gimmal rings (variously spelt). The toothsome jumble, known to the Midlands as "brandy-snap," is the same word, this delicacy having apparently at one time been made in links. We may compare the obsolete Ital. stortelli, lit. "little twists," explained by Torriano as "winding simnels, wreathed jumbals."

An accident of spelling may disguise the origin and meaning of a word. Tret is Fr. trait, in Old French also tret, Lat. tractus, pull (of the scale). It was usually an allowance of four pounds in a hundred and four, which was supposed to be equal to the sum of the "turns of the scale" which would be in the purchaser's favour if the goods were weighed in small quantities. Trait is still so used in modern French.

A difference in spelling, originally accidental, but perpetuated by an apparent difference of meaning, is seen in flour, flower; metal, mettle. Flour is the flower, i.e. the finest part, of meal, Fr. fleur de farine, "flower, or the finest meale" (Cotgrave). In the Nottingham Guardian (29th Aug. 1911) I read that—

It might be added that this author is in the excellent company of Shakespeare—

There is no more etymological difference between metal and mettle than between the "temper" of a cook and that of a sword-blade.

Parson is a doublet of person, the priest perhaps being taken as "representing" the Church, for Lat. persona, an actor's mask, from per, through, and sonare, to sound,[106] was also used of a costumed character or dramatis persona. Mask, which ultimately belongs to an Arabic word meaning buffoon, has had a sense development exactly opposite to that of person, its modern meaning corresponding to the Lat. persona from which the latter started. Parson shows the popular pronunciation of er, now modified by the influence of traditional spelling. We still have it in Berkeley, clerk, Derby, sergeant, as we formerly did in merchant. Proper names, in which the orthography depends on the "taste and fancy of the speller," or the phonetic theories of the old parish clerk, are often more in accordance with the pronunciation, e.g., Barclay, Clark, Darby, Sargent, Marchant. Posy, in both its senses, is a contraction of poesy, the flowers of a nosegay expressing by their arrangement a sentiment like that engraved on a ring. The latter use is perhaps obsolete—

The poetic word glamour is the same as grammar, which had in the Middle Ages the sense of mysterious learning. From the same source we have the French corruption grimoire, "a booke of conjuring" (Cotgrave). Glamour and gramarye were both revived by Scott—

For the change of r to l we have the parallel of flounce for older frounce (p. 60). Quire is the same word as quair, in the "King's Quair" i.e. book. Its Mid. English form is quayer, Old Fr. quaer, caer (cahier), Vulgar Lat. *quaternum, for quaternio, "a quier with foure sheetes" (Cooper).

Oriental words have sometimes come into the language by very diverse routes. Sirup, or syrup, sherbet, and (rum)-shrub are of identical origin, ultimately Arabic. Sirup, which comes through Spanish and French, was once used, like treacle (p. 75), of medicinal compounds—

Sherbet and shrub are directly borrowed through the medium of travellers—

Sepoy, used of Indian soldiers in the English service, is the same as spahi, the French name for the Algerian cavalry. Both come ultimately from a Persian adjective meaning "military," and the French form was at one time used also in English in speaking of Oriental soldiery—

Tulip is from Fr. tulipe, formerly tulipan, "the delicate flower called a tulipa, tulipie, or Dalmatian cap" (Cotgrave). It is a doublet of turban. The German Tulpe was also earlier Tulipan.

The humblest of medieval coins was the maravedi, which came from Spain at an early date, though not early enough for Robin Hood to have said to Isaac of York—

The name is due to the Moorish dynasty of the Almaravides or Marabouts. This Arabic name, which means hermit, was given also to a kind of stork, the marabout, on account of the solitary and sober habits which have earned in India for a somewhat similar bird the name adjutant (p. 34).

Cipher and zero do not look like doublets, but both of them come from the same Arabic word. The medieval Lat. zephyrum connects the two forms. Crimson and carmine, both of them ultimately from Old Spanish, are not quite doublets, but both belong to kermes, the cochineal insect, of Arabic origin.

The relationship between cipher and zero is perhaps better disguised than that between furnish and veneer, though this is by no means obvious. Veneer, spelt fineer by Smollett, is Ger. fournieren, borrowed from Fr. fournir[107] and specialised in meaning. Ebers' German Dict. (1796) has furnieren, "to inlay with several sorts of wood, to veneer."

The doublets selected for discussion among the hundreds which exist in the language reveal many etymological relationships which would hardly be suspected at first sight. Many other words might be quoted which are almost doublets. Thus sergeant, Fr. sergent, Lat. serviens, servient-, is almost a doublet of servant, the present participle of Fr. servir. The fabric called drill or drilling is from Ger. Drillich, "tick, linnen-cloth woven of three threads" (Ludwig). This is an adaptation of Lat. trilix, trilic-, which, through Fr. treillis, has given Eng. trellis. We may compare the older twill, of Anglo-Saxon origin, cognate with Ger. Zwilch or Zwillich, "linnen woven with a double thread" (Ludwig). Robe, from French, is cognate with rob, and with Ger. Raub, booty, the conqueror decking himself in the spoils of the conquered. Musk is a doublet of meg in nutmeg, Fr. noix muscade. In Mid. English we find note-mugge, and Cotgrave has the diminutive muguette, "a nutmeg"; cf. modern Fr. muguet, the lily of the valley. Fr. dîner and déjeuner both represent Vulgar Lat. *dis-junare, to break fast, from jejunus, fasting. The difference of form is due to the shifting of the accent in the Latin conjugation, e.g., dis-junáre gives Old Fr. disner (dîner), while dis-júnat gives Old Fr. desjune (déjeune).

Admiral, earlier amiral, comes through French from the Arab. amir, an emir. Its Old French forms are numerous, and the one which has survived in English may be taken as an abbreviation of Arab. amir al bahr emir on the sea. Greco-Lat. pandura, a stringed instrument, has produced an extraordinary number of corruptions, among which some philologists rank mandoline. Eng. bandore, now obsolete, was once a fairly common word, and from it, or from some cognate Romance form, comes the negro corruption banjo—

Florio has pandora, pandura, "a musical instrument with three strings, a kit, a croude,[108] a rebecke." Kit, used by Dickens—

seems to be a clipped form from Old French dialect quiterne, for guiterne, Greco-Lat. cithara. Cotgrave explains mandore as a "kitt, small gitterne." The doublet guitar is from Spanish.

The two pretty words dimity and samite—

are both connected with Gk. μίτος, thread. Dimity is the plural, dimiti, of Ital. dimito, "a kind of course cotton or flanell" (Florio), from Greco-Lat. dimitus, double thread (cf. twill, p. 148). Samite, Old Fr. samit, whence Ger. Samt, velvet, is in medieval Latin hexamitus, six-thread; this is Byzantine Gk. ἑξάμιτον, whence also Old Slavonic aksamitu. The Italian form is sciamito, "a kind of sleave, feret, or filosello silke" (Florio). The word feret used here by Florio is from Ital. fioretto, little flower. It was also called floret silk. Florio explains the plural fioretti as "a kind of course silke called f[l]oret or ferret silke," and Cotgrave has fleuret, "course silke, floret silke." This doublet of floweret is not obsolete in the sense of tape—