Language: Its Nature, Development and Origin

XX.—§ 1. Sound and Sense.

The idea that there is a natural correspondence between sound and sense, and that words acquire their contents and value through a certain sound symbolism, has at all times been a favourite one with linguistic dilettanti, the best-known examples being found in Plato’s Kratylos. Greek and Latin grammarians indulge in the wildest hypotheses to explain the natural origin of such and such a word, as when Nigidius Figulus said that in pronouncing vos one puts forward one’s lips and sends out breath in the direction of the other person, while this is not the case with nos. With these early writers, to make guesses at sound symbolism was the only way to etymologize; no wonder, therefore, that we with our historical methods and our wider range of knowledge find most of their explanations ridiculous and absurd. But this does not justify us in rejecting any idea of sound symbolism: abusus non tollit usum!

Humboldt (Versch 79) says that “language chooses to designate objects by sounds which partly in themselves, partly in comparison with others, produce on the ear an impression resembling the effect of the object on the mind; thus stehen, stätig, starr, the impression of firmness, Sanskrit lī ‘to melt, diverge,’ that of liquidity or solution (des zerfliessenden).... In this way objects that produce similar impressions are denoted by words with essentially the same sounds, thus wehen, wind, wolke, wirren, wunsch, in all of which the vacillating, wavering motion with its confused impression on the senses is expressed through ... w.” Madvig’s objection (1842, 13 = Kl 64) that we need only compare four of the words Humboldt quotes with the corresponding words in the very nearest sister-language, Danish blæse, vind, sky, ønske, to see how wrong this is, seems to me a little cheap: Humboldt himself expressly assumes that much of primitive sound symbolism may have disappeared in course of time and warns us against making this kind of explanation a ‘constitutive principle,’ which would lead to great dangers (“so setzt man sich grossen gefahren aus und verfolgt einen in jeder rücksicht schlüpfrigen pfad”). Moreover blæse (E. blow, Lat. flare) is just as imitative as wind, vind: no one of course would pretend that there was only one way of expressing the same sense perception. Among Humboldt’s examples wolke and wunsch are doubtful, but I do not see that this affects the general truth of his contention that there is something like sound symbolism in some words.

Nyrop in his treatment of this question (Gr IV § 545 f.) repeats Madvig’s objection that the same name can denote various objects, that the same object can be called by different names, and that the significations of words are constantly changing; further, that the same group of sounds comes to mean different things according to the language in which it occurs. He finally exclaims: “How to explain [by means of sound symbolism] the difference in signification between murus, nurus, durus, purus, etc.?”

XX.—§ 2. Instinctive Feeling.

Yes, of course it would be absurd to maintain that all words at all times in all languages had a signification corresponding exactly to their sounds, each sound having a definite meaning once for all. But is there really much more logic in the opposite extreme, which denies any kind of sound symbolism[100] (apart from the small class of evident echoisms or ‘onomatopœia’) and sees in our words only a collection of wholly accidental and irrational associations of sound and meaning? It seems to me that the conclusion in this case is as false as if you were to infer that because on one occasion X told a lie, he therefore never tells the truth. The correct conclusion would be: as he has told a lie once, we cannot always trust him; we must be on our guard with him—but sometimes he may tell the truth. Thus, also, sounds may in some cases be symbolic of their sense, even if they are not so in all words. If linguistic historians are averse to admitting sound symbolism, this is a natural consequence of their being chiefly occupied with words which have undergone regular changes in sound and sense; and most of the words which form the staple of linguistic books are outside the domain of sound symbolism.

There is no denying, however, that there are words which we feel instinctively to be adequate to express the ideas they stand for, and others the sounds of which are felt to be more or less incongruous with their signification. Future linguists will have to find out in detail what domains of human thought admit, and what domains do not admit, of congruous expression through speech sounds, and further what sounds are suitable to express such and such a notion, for though it is clear—to take only a few examples—that there is little to choose between apple and pomme, or between window and fenster, as there is no sound or sound group that has any natural affinity with such thoroughly concrete and composite ideas as those expressed by these words, yet on the other hand everybody must feel that the word roll, rouler, rulle, rollen is more adequate than the corresponding Russian word katat’, katit’.

It would be an interesting task to examine in detail and systematically what ideas lend themselves to symbolic presentation and what sounds are chosen for them in different languages. That, however, could only be done on the basis of many more examples than I can find space for in this work, and I shall, therefore, only attempt to give a preliminary enumeration of the most obvious classes, with a small fraction of the examples I have collected.[101]

XX.—§ 3. Direct Imitation.

The simplest case is the direct imitation of the sound, thus clink, clank, ting, tinkle of various metallic sounds, splash, bubble, sizz, sizzle of sounds produced by water, bow-wow, bleat, roar of sounds produced by animals, and snort, sneeze, snigger, smack, whisper, grunt, grumble of sounds produced by human beings. Examples might easily be multiplied of such ‘echoisms’ or ‘onomatopœia’ proper. But, as our speech-organs are not capable of giving a perfect imitation of all ‘unarticulated’ sounds, the choice of speech-sounds is to a certain extent accidental, and different nations have chosen different combinations, more or less conventionalized, for the same sounds; thus cock-a-doodle-doo, Dan. kykeliky, Sw. kukeliku, G. kikeriki, Fr. coquelico, for the sound of a cock; and for whisper: Dan. hviske, ON. kvisa, G. flüstern, Fr. chuchoter, Sp. susurar. The continuity of a sound is frequently indicated by l or r after a stopped consonant: rattle, rumble, jingle, clatter, chatter, jabber, etc.

XX.—§ 4. Originator of the Sound.

Next, the echoic word designates the being that produces the sound, thus the birds cuckoo and peeweet (Dan. vibe, G. kibitz, Fr. pop. dix-huit).

A special subdivision of particular interest comprises those names, or nicknames, which are sometimes popularly given to nations from words continually occurring in their speech. Thus the French used to call an Englishman a god-damn (godon), and in China an English soldier is called a-says or I-says. In Java a Frenchman is called orang-deedong (orang ‘man’), in America ding-dong, and during the Napoleonic wars the French were called in Spain didones, from dis-donc; another name for the same nation is wi-wi (Australia), man-a-wiwi (in Beach-la-mar), or oui-men (New Caledonia). In Eleonore Christine’s Jammersminde 83 I read, “Ich habe zwei parle mi franço gefangen,” and correspondingly Goldsmith writes (Globe ed. 624): “Damn the French, the parle vous, and all that belongs to them. What makes the bread rising? the parle vous that devour us.” In Rovigno the surrounding Slavs are called čuje from their exclamation čuje ‘listen, I say,’ and in Hungary German visitors are called vigéc (from wie geht’s?), and customs officers vartapiszli (from wart’ a bissl). Round Panama everything native is called spiggoty, because in the early days the Panamanians, when addressed, used to reply, “No spiggoty [speak] Inglis.” In Yokohama an English or American sailor is called Damuraïsu H’to from ‘Damn your eyes’ and Japanese H’to ‘people.’[102]

XX.—§ 5. Movement.

Thirdly, as sound is always produced by some movement and is nothing but the impression which that movement makes on the ear, it is quite natural that the movement itself may be expressed by the word for its sound: the two are, in fact, inseparable. Note, for instance, such verbs as bubble, splash, clash, crack, peck. Human actions may therefore be denoted by such words as to bang the door, or (with slighter sounds) to tap or rap at a door. Hence also the substantives a tap or a rap for the action, but the substantive may also come to stand for the implement, as when from the verb to hack, ‘to cut, chop off, break up hard earth,’ we have the noun hack, ‘a mattock or large pick.’

Then we have words expressive of such movements as are not to the same extent characterized by loud sounds; thus a great many words beginning with l-combinations, fl-: flow, flag (Dan. flagre), flake, flutter, flicker, fling, flit, flurry, flirt; sl-: slide, slip, slive; gl-: glide. Hence adjectives like fleet, slippery, glib. Sound and sight may have been originally combined in such expressions for an uncertain walk as totter, dodder, dialectical teeter, titter, dither, but in cases of this kind the audible element may be wanting, and the word may come to be felt as symbolic of the movement as such. This is also the case with many expressions for the sudden, rapid movement by which we take hold of something; as a short vowel, suddenly interrupted by a stopped consonant, serves to express the sound produced by a very rapid striking movement (pat, tap, knock, etc.), similar sound combinations occur frequently for the more or less noiseless seizing of a thing (with the teeth or with the hand): snap, snack, snatch, catch, Fr. happer, attraper, gripper, E. grip, Dan. hapse, nappe, Lat. capio, Gr. kaptō, Armenian kap ‘I seize,’ Turk kapmak (mak infin. ending), etc. (I shall only mention one derivative meaning that may develop from this group: E. snack ‘a hurried meal,’ in Swift’s time called a snap (Journ. to Stella 270); cf. G. schnapps, Dan. snaps ‘glass of spirits.’) E. chase and catch are both derived from two dialectically different French forms, ultimately going back to the same late Latin verb captiare, but it is no mere accident that it was the form ‘catch’ that acquired the meaning ‘to seize,’ not found in French, for it naturally associated itself with snatch, and especially with the now obsolete verb latch ‘to seize.’

There is also a natural connexion between action and sound in the word to tickle, G. kitzeln, ON. kitla, Dan. kilde (d mute), Nubian killi-killi, and similar forms (Schuchardt, Nubisch. u. Bask. 9), Lat. titillare; cp. also the word for the kind of laughter thus produced: titter, G. kichern.

XX.—§ 6. Things and Appearances.

Further, we have the extension of symbolical designation to things; here, too, there is some more or less obvious association of what is only visible with some sound or sounds. This has been specially studied by Hilmer, to whose book (Sch) the reader is referred for numerous examples, e.g. p. 237 ff., knap ‘a thick stick, a knot of wood, a bit of food, a protuberance, a small hill;’ knop ‘a boss, stud, button, knob, a wart, pimple, the bud of a flower, a promontory,’ with the variants knob, knup.... Hilmer’s word-lists from German and English comprise 170 pages!

There is also a natural association between high tones (sounds with very rapid vibrations) and light, and inversely between low tones and darkness, as is seen in the frequent use of adjectives like ‘light’ and ‘dark’ in speaking of notes. Hence the vowel [i] is felt to be more appropriate for light, and [u] for dark, as seen most clearly in the contrast between gleam, glimmer, glitter on the one hand and gloom on the other (Zangwill somewhere writes: “The gloom of night, relieved only by the gleam from the street-lamp”); the word light itself, which has now a diphthong which is not so adequate to the meaning, used to have the vowel [i] like G. licht; for the opposite notions we have such words as G. dunkel, Dan. mulm, Gr. amolgós, skótos, Lat. obscurus, and with another ‘dark’ vowel E. murky, Dan. mörk.

XX.—§ 7. States of Mind.

From this it is no far cry to words for corresponding states of mind: to some extent the very same words are used, as gloom (Dowden writes: “The good news was needed to cast a gleam on the gloom that encompassed Shelley”); hence also glum, glumpy, glumpish, grumpy, the dumps, sulky. If E. moody and sullen have changed their significations (OE. modig ‘high-spirited,’ ME. solein ‘solitary’), sound symbolism, if I am not mistaken, counts for something in the change; the adjectives now mean exactly the same as Dan. mut, but.

If grumble comes to mean the expression of a mental state of dissatisfaction, the connexion between the sound of the word and its sense is even more direct, for the verb is imitative of the sound produced in such moods, cf. mumble and grunt, gruntle. The name of Mrs. Grundy is not badly chosen as a representative of narrow-minded conventional morality.

A long list might be given of symbolic expressions for dislike, disgust, or scorn; here a few hints only can find place. First we have the same dull or dump (back) vowels as in the last paragraph: blunder, bungle, bung, clumsy, humdrum, humbug, strum, slum, slush, slubber, sloven, muck, mud, muddle, mug (various words, but all full of contempt), juggins (a silly person), numskull (old numps, nup, nupson), dunderhead, gull, scug (at Eton a dirty or untidy boy).... Many words begin with sl- (we have already seen some): slight, slim, slack, sly, sloppy, slipslop, slubby, slattern, slut, slosh.... Initial labials are also frequent.[103] After the vowel we have very often the sound [ʃ] or [tʃ], as in trash, tosh, slosh, botch, patch; cf. also G. kitsch (bad picture, smearing), patsch(e) (mire, anything worthless), quatsch (silly nonsense), putsch (riot, political coup de main). E. bosh (nonsense) is said to be a Turkish loan-word; it has become popular for the same reason for which the French nickname boche for a German was widely used during the World War. Let me finally mention the It. derivative suffix -accio, as in poveraccio (miserable), acquaccia (bad water), and -uccio, as in cavalluccio (vile horse).

XX.—§ 8. Size and Distance.

The same vowel [i] is also symbolical of a very short time, as in the phrases in a jiff, jiffy, Sc. in a clink, Dan. i en svip; and correspondingly we have adjectives like quick, swift, vivid and others. No wonder, then, that the Germans feel their word for ‘lightning,’ blitz, singularly appropriate to the effect of light and to the shortness of duration.[104]

It has often been remarked[105] that in corresponding pronouns and adverbs the vowel i frequently indicates what is nearer, and other vowels, especially a or u, what is farther off; thus Fr. ci, là, E. here, there, G. dies, das, Low G. dit, dat, Magy. ez, emez ‘this,’ az, amaz ‘that,’ itt ‘here,’ ott ‘there,’ Malay iki ‘this,’ ika ‘that, a little removed,’ iku ‘yon, farther away.’ In Hamitic languages i symbolizes the near and u what is far away. We may here also think of the word zigzag as denoting movement in alternate turns here and there; and if in the two E. pronouns this and that the old neuter forms have prevailed (OE. m. þes, se, f. þeos, seo, n. þis, þæt) the reason (or one of the reasons) may have been that a characteristic difference of vowels in the two contrasted pronouns was thus secured.

XX.—§ 9. Length and Strength of Words and Sounds.

We see now the emotional value of some ‘mouth-filling’ words, some of which may be considered symbolical expansions of existing words (what H. Schröder terms ‘streckformen’), though others cannot be thus explained; not unfrequently the effect of length is combined with some of the phonetic effects mentioned above. Such words are, e.g., slubberdegullion ‘dirty fellow,’ rumbustious ‘boisterous,’ rumgumption, rumfustian, rumbullion (cf. rumpuncheon ‘cask of rum’ as a term of abuse in Stevenson, Treas. Isl. 48, “the cowardly son of a rum-puncheon”), rampallion ‘villain,’ rapscallion, ragamuffin; sculduddery ‘obscenity’; cantankerous ‘quarrelsome,’ U.S. also rantankerous (cf. cankerous, rancorous); skilligalee ‘miserable gruel,’ flabbergast ‘confound,’ catawampous (or -ptious) ‘fierce’ (“a high-sounding word with no very definite meaning,” NED); Fr. hurluberlu ‘crazy’ and the synonymous Dan. tummelumsk, Norw. tullerusk.

In this connexion one may mention the natural tendency to lengthen and to strengthen single sounds under the influence of strong feeling and in order to intensify the effect of the spoken word; thus, in ‘it’s very cold’ both the diphthong [ou] and the [l] may be pronounced extremely long, in ‘terribly dull’ the [l] is lengthened, in ‘extremely long’ either the vowel [ɔ] or the [ŋ] (or both) may be lengthened. In Fr. ‘c’était horrible’ the trill of the [r] becomes very long and intense (while the same effect is not generally possible in the corresponding English word, because the English [r] is not trilled, but pronounced by one flap of the tip). In some cases a lengthening due to such a psychological cause may permanently alter a word, as when Lat. totus in It. has become tutto (Fr. tout, toute goes back to the same form, while Sp. todo has preserved the form corresponding to the Lat. single consonant). An interesting collection of such cases from the Romanic tongues has been published by A. J. Carnoy (Mod. Philol. 15. 31, July 1917), who justly emphasizes the symbolic value of the change and the special character of the words in which it occurs (pet-names, children’s words, ironic or derisive words, imitative words ...). He says: “While to a phonetician the phenomenon would seem capricious, its apportionment in the vocabulary is quite natural to a psychologist. In fact, reduplication, be it of syllables or of consonants, generally has that character in languages. One finds it in perfective tenses, in intensive or frequentative verbs, in the plural, and in collectives. In most cases it is a reduplication of syllables, but a lengthening of vowels is not rare and the reinforcement of consonants is also found. In Chinook, for instance, the emotional words, both diminutive and augmentative, are expressed by increasing the stress of consonants. It is, of course, also well known that in Semitic the intensive radical of verbs is regularly formed by a reduplication of consonants. To a stem qatal, e.g., answers an intensive: Eth. qattala, Hebr. qittel. Cf. Hebr. shibbar ‘to cut in small pieces’ [cf. below], hillech ‘to walk,’ qibber ‘to bury many,’ etc. Cf. Brockelmann, Vergl. Gramm., p. 244.”

I add a few more examples from Misteli (428 f.) of this Semitic strengthening: the first vowel is lengthened to express a tendency or an attempt: qatala jaqtulu ‘kill’ (in the third person masc., the former in the prefect-aorist, the latter in the imperfect-durative, where ja, ju is the sign of the third person m.), qātala juqātilu ‘try to kill, fight’; faXara jufXaru ‘excel in fame,’ fāXara jufāXiru ‘try to excel, vie.’ Through lengthening (doubling) of a consonant an intensification of the action is denoted: Hebr. šāβar jišbōr ‘zerbrechen,’ šibbēr jẹšabbēr ‘zerschmettern,’ Arab. ḍaraba jaḍrubu ‘strike,’ ḍarraba juḍarribu ‘beat violently, or repeatedly’; sometimes the change makes a verb into a causative or transitive, etc.

I imagine that we have exactly the same kind of strengthening for psychological (symbolical) reasons in a number of verbs where Danish has pp, tt, kk by the side of b, d, g (spirantic): pippe pibe, stritte stride, snitte snide, skøtte skøde, splitte splide, skrikke skrige, lukke luge, hikke hige, sikke sige, kikke kige, prikke prige (cf. also sprække sprænge). Some of these forms are obsolete, others dialectal, but it would take us too far in this place to deal with the words in detail. It is customary to ascribe this gemination to an old n derivative (see, e.g., Brugmann VG 1. 390, Streitberg Urg pp. 135, 138, Noreen UL 154), but it does not seem necessary to conjure up an n from the dead to make it disappear again immediately, as the mere strengthening of the consonant itself to express symbolically the strengthening of the action has nothing unnatural in it. Cf. also G. placken by the side of plagen. The opposite change, a weakening, may have taken place in E. flag (cf. OFr. flaquir, to become flaccid), flabby, earlier flappy, drib from drip, slab, if from OFr. esclape, clod by the side of clot, and possibly cadge, bodge, grudge, smudge, which had all of them originally -tch. But the common modification in sense is not so easily perceived here as in the cases of strengthening.

I may here, for the curiosity of the thing, mention that in a ‘language’ coined by two English children (a vocabulary of which was communicated to me by one of the inventors through Miss I. C. Ward, of the Department of Phonetics, University College, London) there was a word bal which meant ‘place,’ but the bigger the place the longer the vowel was made, so that with three different quantities it meant ‘village,’ ‘town’ and ‘city’ respectively. The word for ‘go’ was dudu, “the greater the speed of the going, the more quickly the word was said—[dœ·dœ·] walk slowly.” Cf. Humboldt, ed. Steinthal 82: “In the southern dialect of the Guarani language the suffix of the perfect yma is pronounced more or less slowly according to the more or less remoteness of the past to be indicated.”

XX.—§ 10. General Considerations.

Sound symbolism, as we have considered it in this chapter, has a very wide range of application, from direct imitation of perceived natural sounds to such small quantitative changes of existing non-symbolic words as may be used for purely grammatical purposes. But in order to obtain a true valuation of this factor in the life of language it is of importance to keep in view the following considerations:

(1) No language utilizes sound symbolism to its full extent, but contains numerous words that are indifferent to or may even jar with symbolism. To express smallness the vowel [i] is most adequate, but it would be absurd to say that that vowel always implies smallness, or that smallness is always expressed by words containing that vowel: it is enough to mention the words big and small, or to point to the fact that thick and thin have the same vowel, to repudiate such a notion.

(2) Words that have been symbolically expressive may cease to be so in consequence of historical development, either phonetic or semantic or both. Thus the name of the bird crow is not now so good an imitation of the sound made by the bird as OE. crawe was (Dan. krage, Du. kraai). Thus, also, the verbs whine, pipe were better imitations when the vowel was still [i·] (as in Dan. hvine, pibe). But to express the sound of a small bird the latter word is still pronounced with the vowel [i] either long or short (peep, pip), the word having been constantly renewed and as it were reshaped by fresh imitation; cf. on Irish wheen and dialectal peep, XV § 8. Lat. pipio originally meant any ‘peeping bird,’ but when it came to designate one particular kind of birds, it was free to follow the usual trend of phonetic development, and so has become Fr. pigeon [piʒɔ̃], E. pigeon [pidʒin]. E. cuckoo has resisted the change from [u] to ʌ as in cut, because people have constantly heard the sound and fashioned the name of the bird from it. I once heard a Scotch lady say [kʌku·], but on my inquiry she told me that there were no cuckoos in her native place; hence the word had there been treated as any other word containing the short [u]. The same word is interesting in another way; it has resisted the old Gothonic consonant-shift, and thus has the same consonants as Skt. kōkiláḥ, Gr. kókkux, Lat. cuculus. On the general preservation of significative sounds, cf. Ch. XV § 8.

(3) On the other hand, some words have in course of time become more expressive than they were at first; we have something that may be called secondary echoism or secondary symbolism. The verb patter comes from pater (= paternoster), and at first meant to repeat that prayer, to mumble one’s prayers; but then it was associated with the homophonous verb patter ‘to make a rapid succession of pats’ and came under the influence of echoic words like prattle, chatter, jabber; it now, like these, means ‘to talk rapidly or glibly’ and is to all intents a truly symbolical word; cf. also the substantive patter ‘secret lingo, speechifying, talk.’ Husky may at first have meant only “full of husks, of the nature of a husk” (NED), but it could not possibly from that signification have arrived at the now current sense ‘dry in the throat, hoarse’ if it had not been that the sound of the adjective had reminded one of the sound of a hoarse voice. Dan. pöjt ‘poor drink, vile stuff’ is now felt as expressive of contempt, but it originates in Poitou, an innocent geographical name of a kind of wine, like Bordeaux; it is now connected with other scornful words like spröjt and döjt.

In E. little the symbolic vowel i is regularly developed from OE. y, lytel, whose y is a mutated u, as seen in OSax. luttil; u also appears in other related languages, and the word thus originally had nothing symbolical about it. But in Gothic the word is leitils (ei, sounded [i·]) and in ON. lítinn, and here the vowel is so difficult to account for on ordinary principles that the NED in despair thinks that the two words are “radically unconnected.” I have no hesitation in supposing that the vowel i is due to sound symbolism, exactly as the smaller change introduced in modern E. ‘leetle,’ with narrow instead of wide (broad) [i]. In the word for the opposite meaning, much, the phonetic development may also have been influenced by the tendency to get an adequate vowel, for normally we should expect the vowel [i] as in Sc. mickle, from OE. micel. In E. quick the vowel best adapted to the idea has prevailed instead of the one found in the old nom. forms cwucu, cucu from cwicu (inflected cwicne, cwices, etc.), while in the word widu, wudu, which is phonetically analogous, there was no such inducement, and the vowel [u] has been preserved: wood. The same prevalence of the symbolic i is noticed in the Dan. adj. kvik, MLG. quik, while the same word as subst. has become Dan. kvæg, MLG. quek, where there was no symbolism at work, as it has come to mean ‘cattle.’ I even see symbolism in the preservation of the k in the Dan. adj. (as against the fricative in kvæg), because the notion of ‘quick’ is best expressed by the short [i], interrupted by a stop; and may not the same force have been at work in this adjective at an earlier period? The second k in OE. cwicu, ON. kvikr as against Goth. qius, Lat. vivus, has not been sufficiently explained. An [i], symbolic of smallness, has been introduced in some comparatively recent E. words: tip from top, trip ‘small flock’ from troop, sip ‘drink in small quantities’ from sup, sop.

Through changes in meaning, too, some words have become symbolically more expressive than they were formerly; thus the agreement between sound and sense is of late growth in miniature, which now, on account of the i, has come to mean ‘a small picture,’ while at first it meant ‘image painted with minium or vermilion,’ and in pittance, now ‘a scanty allowance,’ formerly any pious donation, whether great or small. Cf. what has been said above of sullen, moody, catch.

XX.—§ 11. Importance of Suggestiveness.

The suggestiveness of some words as felt by present-day speakers is a fact that must be taken into account if we are to understand the realities of language. In some cases it may have existed from the very first: these words sprang thus into being because that shape at once expressed the idea the speaker wished to communicate. In other cases the suggestive element is not original: these words arose in the same way as innumerable others whose sound has never carried any suggestion. But if the sound of a word of this class was, or came to be, in some way suggestive of its signification—say, if a word containing the vowel [i] in a prominent place meant ‘small’ or something small—then the sound exerted a strong influence in gaining popular favour to the word; it was an inducement to people to choose and to prefer that particular word and to cease to use words for the same notion that were not thus favoured. Sound symbolism, we may say, makes some words more fit to survive and gives them considerable help in their struggle for existence. If we want to denote a little child by a word for some small animal, we take some word like kid, chick, kitten, rather than bat or pug or slug, though these may in themselves be smaller than the animal chosen.

It is quite true that Fr. rouler, our roll, is derived from Lat. rota ‘wheel’ + a diminutive ending -ul-, but the word would never have gained its immense popularity, extending as it does through English, Dutch, German and the Scandinavian languages, if the sound had not been eminently suggestive of the sense, so suggestive that it seems to us now the natural expression for that idea, and we have difficulty in realizing that the word has not existed from the very dawn of speech. Or let me take another example, in which the connexion between sound and sense is even more ‘fortuitous.’ About a hundred years ago a member of Congress, Felix Walker, from Buncombe County, North Carolina, made a long and tedious speech. “Many members left the hall. Very naïvely he told those who remained that they might go too; he should speak for some time, but ‘he was only talking for Buncombe,’ to please his constituents.” Now buncombe (buncome, bunkum) has become a widely used word, not only in the States, but all over the English-speaking world, for political speaking or action not resting on conviction, but on the desire of gaining the favour of electors, or for any kind of empty ‘clap-trap’ oratory; but does anybody suppose that the name of Mr. Walker’s constituency would have been thus used if he had happened to hail from Annapolis or Philadelphia, or some other place with a name incapable of tickling the popular fancy in the same way as Buncombe does? (Cf. above, p. 401 on the suggestiveness of the short u.) In a similar way hullaballoo seems to have originated from the Irish village Ballyhooly (see P. W. Joyce, English as we speak it in Ireland) and to have become popular on account of its suggestive sound.

In loan-words we can often see that they have been adopted less on account of any cultural necessity (see above, p. 209) than because their sound was in some way or other suggestive. Thus the Algonkin (Natick) word for ‘chief,’ mugquomp, is used in the United States in the form of mugwump for a ‘great man’ or ‘boss,’ and especially, in political life, for a man independent of parties and thinking himself superior to parties. Now, no one would have thought of going to an Indian language to express such a notion, had not an Indian word presented itself which from its uncouth sound lent itself to purposes of ridicule. Among other words whose adoption has been favoured by their sounds I may mention jungle (from Hindi jangal, associated more or less closely with jumble, tumble, bundle, bungle); bobbery, in slang ‘noise, squabble,’ “the Anglo-Indian colloquial representation of a common exclamation of Hindus when in surprise or grief—Bap-rē! or Bap-rē Bap ‘O Father!’” (Hobson-Jobson); amuck; and U.S. bunco ‘swindling game, to swindle,’ from It. banco.

XX.—§ 12. Ancient and Modern Times.

It will be seen that our conception of echoism and related phenomena does not carry us back to an imaginary primitive period: these forces are vital in languages as we observe them day by day. Linguistic writers, however, often assume that sound symbolism, if existing at all, must date back to the earliest times, and therefore can have no reality nowadays. Thus Benfey (Gesch 288) turns upon de Brosse, who had found rudeness in Fr. rude and gentleness in Fr. doux, and says: “As if the sounds of such words, which are distant by an infinite length of time from the time when language originated, were able to contribute ever so little to explain the original designation of things.” (But Benfey is right in saying that the impression made by those two French words may be imaginary; as examples they are not particularly well chosen.) Sütterlin (WW 14) says: “It is bold to search for such correspondence as still existing in detail in the language of our own days. For words like liebe, süss on the one hand, and zorn, hass, hart on the other, which are often alleged by dilettanti, prove nothing to the scholar, because their form is young and must have had totally different sounds in the period when language was created.”

Similarly de Saussure (LG 104) gives as one of the main principles of our science that the tie between sound and sense is arbitrary or rather motiveless (immotivé), and to those who would object that onomatopoetic words are not arbitrary he says that “they are never organic elements of a linguistic system. Besides, they are much less numerous than is generally supposed. Such words as Fr. fouet and glas may strike some ears with a suggestive ring;[106] but they have not had that character from the start, as is sufficiently proved if we go back to their Latin forms (fouet derived from fagus ‘beech,’ glas = classicum); the quality possessed by, or rather attributed to, their actual sounds is a fortuitous result of phonetic development.”

Here we see one of the characteristics of modern linguistic science: it is so preoccupied with etymology, with the origin of words, that it pays much more attention to what words have come from than to what they have come to be. If a word has not always been suggestive on account of its sound, then its actual suggestiveness is left out of account and may even be declared to be merely fanciful. I hope that this chapter contains throughout what is psychologically a more true and linguistically a more fruitful view.

Though some echo words may be very old, the great majority are not; at any rate, in looking up the earliest ascertained date of a goodly number of such words in the NED, I have been struck by the fact of so many of them being quite recent, not more than a few centuries old, and some not even that. To some extent their recent appearance in writing may be ascribed to the general character of the old literature as contrasted with our modern literature, which is less conventional, freer in many ways, more true to life with its infinite variety and more true, too, to the spoken language of every day. But that cannot account for everything, and there is every probability that this class of words is really more frequent in the spoken language of recent times than it was formerly, because people speak in a more vivid and fresh fashion than their ancestors of hundreds or thousands of years ago. The time of psychological reaction is shorter than it used to be, life moves at a more rapid rate, and people are less tied down to tradition than in former ages, consequently they are more apt to create and to adopt new words of this particular type, which are felt at once to be significant and expressive. In all languages the creation and use of echoic and symbolic words seems to have been on the increase in historical times. If to this we add the selective process through which words which have only secondarily acquired symbolical value survive at the cost of less adequate expressions, or less adequate forms of the same words, and subsequently give rise to a host of derivatives, then we may say that languages in course of time grow richer and richer in symbolic words. So far from believing in a golden primitive age, in which everything in language was expressive and immediately intelligible on account of the significative value of each group of sounds, we arrive rather, here as in other domains, at the conception of a slow progressive development towards a greater number of easy and adequate expressions—expressions in which sound and sense are united in a marriage-union closer than was ever known to our remote ancestors.