VIII

Language as a Historical Product: Phonetic Law

I have preferred to take up in some detail the analysis of our hesitation in using a locution like “Whom did you see?” and to point to some of the English drifts, particular and general, that are implied by this hesitation than to discuss linguistic change in the abstract. What is true of the particular idiom that we started with is true of everything else in language. Nothing is perfectly static. Every word, every grammatical element, every locution, every sound and accent is a slowly changing configuration, molded by the invisible and impersonal drift that is the life of language. The evidence is overwhelming that this drift has a certain consistent direction. Its speed varies enormously according to circumstances that it is not always easy to define. We have already seen that Lithuanian is to-day nearer its Indo-European prototype than was the hypothetical Germanic mother-tongue five hundred or a thousand years before Christ. German has moved more slowly than English; in some respects it stands roughly midway between English and Anglo-Saxon, in others it has of course diverged from the Anglo-Saxon line. When I pointed out in the preceding chapter that dialects formed because a language broken up into local segments could not move along the same drift in all of these segments, I meant of course that it could not move along identically the same drift. The general drift of a language has its depths. At the surface the current is relatively fast. In certain features dialects drift apart rapidly. By that very fact these features betray themselves as less fundamental to the genius of the language than the more slowly modifiable features in which the dialects keep together long after they have grown to be mutually alien forms of speech. But this is not all. The momentum of the more fundamental, the pre-dialectic, drift is often such that languages long disconnected will pass through the same or strikingly similar phases. In many such cases it is perfectly clear that there could have been no dialectic interinfluencing.

These parallelisms in drift may operate in the phonetic as well as in the morphological sphere, or they may affect both at the same time. Here is an interesting example. The English type of plural represented by foot: feet, mouse: mice is strictly parallel to the German Fuss: Füsse, Maus: Mäuse. One would be inclined to surmise that these dialectic forms go back to old Germanic or West-Germanic alternations of the same type. But the documentary evidence shows conclusively that there could have been no plurals of this type in primitive Germanic. There is no trace of such vocalic mutation (“umlaut”) in Gothic, our most archaic Germanic language. More significant still is the fact that it does not appear in our oldest Old High German texts and begins to develop only at the very end of the Old High German period (circa 1000 A.D.). In the Middle High German period the mutation was carried through in all dialects. The typical Old High German forms are singular fuoss, plural fuossi;[145] singular mus, plural musi. The corresponding Middle High German forms are fuoss, füesse; mus, müse. Modern German Fuss: Füsse, Maus: Mäuse are the regular developments of these medieval forms. Turning to Anglo-Saxon, we find that our modern English forms correspond to fot, fet; mus, mys.[146] These forms are already in use in the earliest English monuments that we possess, dating from the eighth century, and thus antedate the Middle High German forms by three hundred years or more. In other words, on this particular point it took German at least three hundred years to catch up with a phonetic-morphological drift[147] that had long been under way in English. The mere fact that the affected vowels of related words (Old High German uo, Anglo-Saxon o) are not always the same shows that the affection took place at different periods in German and English.[148] There was evidently some general tendency or group of tendencies at work in early Germanic, long before English and German had developed as such, that eventually drove both of these dialects along closely parallel paths.

How did such strikingly individual alternations as fot: fet, fuoss: füesse develop? We have now reached what is probably the most central problem in linguistic history, gradual phonetic change. “Phonetic laws” make up a large and fundamental share of the subject-matter of linguistics. Their influence reaches far beyond the proper sphere of phonetics and invades that of morphology, as we shall see. A drift that begins as a slight phonetic readjustment or unsettlement may in the course of millennia bring about the most profound structural changes. The mere fact, for instance, that there is a growing tendency to throw the stress automatically on the first syllable of a word may eventually change the fundamental type of the language, reducing its final syllables to zero and driving it to the use of more and more analytical or symbolic[149] methods. The English phonetic laws involved in the rise of the words foot, feet, mouse and mice from their early West-Germanic prototypes fot, foti, mus, musi[150] may be briefly summarized as follows:

1. In foti “feet” the long o was colored by the following i to long ö, that is, o kept its lip-rounded quality and its middle height of tongue position but anticipated the front tongue position of the i; ö is the resulting compromise. This assimilatory change was regular, i.e., every accented long o followed by an i in the following syllable automatically developed to long ö; hence tothi “teeth” became töthi, fodian “to feed” became födian. At first there is no doubt the alternation between o and ö was not felt as intrinsically significant. It could only have been an unconscious mechanical adjustment such as may be observed in the speech of many to-day who modify the “oo” sound of words like you and few in the direction of German ü without, however, actually departing far enough from the “oo” vowel to prevent their acceptance of who and you as satisfactory rhyming words. Later on the quality of the ö vowel must have departed widely enough from that of o to enable ö to rise in consciousness[151] as a neatly distinct vowel. As soon as this happened, the expression of plurality in föti, töthi, and analogous words became symbolic and fusional, not merely fusional.
2. In musi “mice” the long u was colored by the following i to long ü. This change also was regular; lusi “lice” became lüsi, kui “cows” became küi (later simplified to kü; still preserved as ki- in kine), fulian “to make foul” became fülian (still preserved as -file in defile). The psychology of this phonetic law is entirely analogous to that of 1.
3. The old drift toward reducing final syllables, a rhythmic consequence of the strong Germanic stress on the first syllable, now manifested itself. The final -i, originally an important functional element, had long lost a great share of its value, transferred as that was to the symbolic vowel change (o: ö). It had little power of resistance, therefore, to the drift. It became dulled to a colorless -e; föti became föte.
4. The weak -e finally disappeared. Probably the forms föte and föt long coexisted as prosodic variants according to the rhythmic requirements of the sentence, very much as Füsse and Füss’ now coexist in German.
5. The ö of föt became “unrounded” to long e (our present a of fade). The alternation of fot: foti, transitionally fot: föti, föte, föt, now appears as fot: fet. Analogously, töth appears as teth, födian as fedian, later fedan. The new long e-vowel “fell together” with the older e-vowel already existent (e.g., her “here,” he “he”). Henceforward the two are merged and their later history is in common. Thus our present he has the same vowel as feet, teeth, and feed. In other words, the old sound pattern o, e, after an interim of o, ö, e, reappeared as o, e, except that now the e had greater “weight” than before.
6. Fot: fet, mus: müs (written mys) are the typical forms of Anglo-Saxon literature. At the very end of the Anglo-Saxon period, say about 1050 to 1100 A.D., the ü, whether long or short, became unrounded to i. Mys was then pronounced mis with long i (rhyming with present niece). The change is analogous to 5, but takes place several centuries later.
7. In Chaucer’s day (circa 1350-1400 A.D.) the forms were still fot: fet (written foot, feet) and mus: mis (written very variably, but mous, myse are typical). About 1500 all the long i-vowels, whether original (as in write, ride, wine) or unrounded from Anglo-Saxon ü (as in hide, bride, mice, defile), became diphthongized to ei (i.e., e of met + short i). Shakespeare pronounced mice as meis (almost the same as the present Cockney pronunciation of mace).
8. About the same time the long u-vowels were diphthongized to ou (i.e., o of present Scotch not + u of full). The Chaucerian mus: mis now appears as the Shakespearean mous: meis. This change may have manifested itself somewhat later than 7; all English dialects have diphthongized old Germanic long i,[152] but the long undiphthongized u is still preserved in Lowland Scotch, in which house and mouse rhyme with our loose. 7 and 8 are analogous developments, as were 5 and 6; 8 apparently lags behind 7 as 6, centuries earlier, lagged behind 7.
9. Some time before 1550 the long e of fet (written feet) took the position that had been vacated by the old long i, now diphthongized (see 7), i.e., e took the higher tongue position of i. Our (and Shakespeare’s) “long e” is, then, phonetically the same as the old long i. Feet now rhymed with the old write and the present beat.
10. About the same time the long o of fot (written foot) took the position that had been vacated by the old long u, now diphthongized (see 8), i.e., o took the higher tongue position of u. Our (and Shakespeare’s) “long oo” is phonetically the same as the old long u. Foot now rhymed with the old out and the present boot. To summarize 7 to 10, Shakespeare pronounced meis, mous, fit, fut, of which meis and mous would affect our ears as a rather “mincing” rendering of our present mice and mouse, fit would sound practically identical with (but probably a bit more “drawled” than) our present feet, while foot, rhyming with boot, would now be set down as “broad Scotch.”
11. Gradually the first vowel of the diphthong in mice (see 7) was retracted and lowered in position. The resulting diphthong now varies in different English dialects, but ai (i.e., a of father, but shorter, + short i) may be taken as a fairly accurate rendering of its average quality.[153] What we now call the “long i” (of words like ride, bite, mice) is, of course, an ai-diphthong. Mice is now pronounced mais.
12. Analogously to 11, the first vowel of the diphthong in mouse (see 8) was unrounded and lowered in position. The resulting diphthong may be phonetically rendered au, though it too varies considerably according to dialect. Mouse, then, is now pronounced maus.
13. The vowel of foot (see 10) became “open” in quality and shorter in quantity, i.e., it fell together with the old short u-vowel of words like full, wolf, wool. This change has taken place in a number of words with an originally long u (Chaucerian long close o), such as forsook, hook, book, look, rook, shook, all of which formerly had the vowel of boot. The older vowel, however, is still preserved in most words of this class, such as fool, moon, spool, stoop. It is highly significant of the nature of the slow spread of a “phonetic law” that there is local vacillation at present in several words. One hears roof, soot, and hoop, for instance, both with the “long” vowel of boot and the “short” of foot. It is impossible now, in other words, to state in a definitive manner what is the “phonetic law” that regulated the change of the older foot (rhyming with boot) to the present foot. We know that there is a strong drift towards the short, open vowel of foot, but whether or not all the old “long oo” words will eventually be affected we cannot presume to say. If they all, or practically all, are taken by the drift, phonetic law 13 will be as “regular,” as sweeping, as most of the twelve that have preceded it. If not, it may eventually be possible, if past experience is a safe guide, to show that the modified words form a natural phonetic group, that is, that the “law” will have operated under certain definable limiting conditions, e.g., that all words ending in a voiceless consonant (such as p, t, k, f) were affected (e.g., hoof, foot, look, roof), but that all words ending in the oo-vowel or in a voiced consonant remained unaffected (e.g., do, food, move, fool). Whatever the upshot, we may be reasonably certain that when the “phonetic law” has run its course, the distribution of “long” and “short” vowels in the old oo-words will not seem quite as erratic as at the present transitional moment.[154] We learn, incidentally, the fundamental fact that phonetic laws do not work with spontaneous automatism, that they are simply a formula for a consummated drift that sets in at a psychologically exposed point and gradually worms its way through a gamut of phonetically analogous forms.

It will be instructive to set down a table of form sequences, a kind of gross history of the words foot, feet, mouse, mice for the last 1500 years:[155]

1. fot: foti; mus: musi (West Germanic)
2. fot: föti; mus: müsi
3. fot: föte; mus: müse
4. fot: föt; mus: müs
5. fot: fet; mus: müs (Anglo-Saxon)
6. fot: fet; mus: mis(Chaucer)
7. fot: fet; mous: meis
8. fut (rhymes with boot): fit; mous: meis (Shakespeare)
9. fut: fit; maus: mais
10. fut (rhymes with put): fit; maus: mais (English of 1900)

It will not be necessary to list the phonetic laws that gradually differentiated the modern German equivalents of the original West Germanic forms from their English cognates. The following table gives a rough idea of the form sequences in German:[156]

1. fot: foti; mus: musi (West Germanic)
2. foss:[157] fossi; mus: musi
3. fuoss: fuossi; mus: musi (Old High German)
4. fuoss: füessi; mus: müsi
5. fuoss: füesse; mus: müse (Middle High German)
6. fuoss: füesse; mus: müze[158]
7. fuos: füese; mus: müze
8. fuos: füese; mous: möüze
9. fus: füse; mous: möüze (Luther)
10. fus: füse; maus: moize (German of 1900)

We cannot even begin to ferret out and discuss all the psychological problems that are concealed behind these bland tables. Their general parallelism is obvious. Indeed we might say that to-day the English and German forms resemble each other more than does either set the West Germanic prototypes from which each is independently derived. Each table illustrates the tendency to reduction of unaccented syllables, the vocalic modification of the radical element under the influence of the following vowel, the rise in tongue position of the long middle vowels (English o to u, e to i; German o to uo to u, üe to ü), the diphthongizing of the old high vowels (English i to ei to ai; English and German u to ou to au; German ü to öü to oi). These dialectic parallels cannot be accidental. They are rooted in a common, pre-dialectic drift.

Phonetic changes are “regular.” All but one (English table, X.), and that as yet uncompleted, of the particular phonetic laws represented in our tables affect all examples of the sound in question or, if the phonetic change is conditional, all examples of the same sound that are analogously circumstanced.[159] An example of the first type of change is the passage in English of all old long i-vowels to diphthongal ai via ei. The passage could hardly have been sudden or automatic, but it was rapid enough to prevent an irregularity of development due to cross drifts. The second type of change is illustrated in the development of Anglo-Saxon long o to long e, via ö, under the influence of a following i. In the first case we may say that au mechanically replaced long u, in the second that the old long o “split” into two sounds—long o, eventually u, and long e, eventually i. The former type of change did no violence to the old phonetic pattern, the formal distribution of sounds into groups; the latter type rearranged the pattern somewhat. If neither of the two sounds into which an old one “splits” is a new sound, it means that there has been a phonetic leveling, that two groups of words, each with a distinct sound or sound combination, have fallen together into one group. This kind of leveling is quite frequent in the history of language. In English, for instance, we have seen that all the old long ü-vowels, after they had become unrounded, were indistinguishable from the mass of long i-vowels. This meant that the long i-vowel became a more heavily weighted point of the phonetic pattern than before. It is curious to observe how often languages have striven to drive originally distinct sounds into certain favorite positions, regardless of resulting confusions.[160] In Modern Greek, for instance, the vowel i is the historical resultant of no less than ten etymologically distinct vowels (long and short) and diphthongs of the classical speech of Athens. There is, then, good evidence to show that there are general phonetic drifts toward particular sounds.

More often the phonetic drift is of a more general character. It is not so much a movement toward a particular set of sounds as toward particular types of articulation. The vowels tend to become higher or lower, the diphthongs tend to coalesce into monophthongs, the voiceless consonants tend to become voiced, stops tend to become spirants. As a matter of fact, practically all the phonetic laws enumerated in the two tables are but specific instances of such far-reaching phonetic drifts. The raising of English long o to u and of long e to i, for instance, was part of a general tendency to raise the position of the long vowels, just as the change of t to ss in Old High German was part of a general tendency to make voiceless spirants of the old voiceless stopped consonants. A single sound change, even if there is no phonetic leveling, generally threatens to upset the old phonetic pattern because it brings about a disharmony in the grouping of sounds. To reëstablish the old pattern without going back on the drift the only possible method is to have the other sounds of the series shift in analogous fashion. If, for some reason or other, p becomes shifted to its voiced correspondent b, the old series p, t, k appears in the unsymmetrical form b, t, k. Such a series is, in phonetic effect, not the equivalent of the old series, however it may answer to it in etymology. The general phonetic pattern is impaired to that extent. But if t and k are also shifted to their voiced correspondents d and g, the old series is reëstablished in a new form: b, d, g. The pattern as such is preserved, or restored. Provided that the new series b, d, g does not become confused with an old series b, d, g of distinct historical antecedents. If there is no such older series, the creation of a b, d, g series causes no difficulties. If there is, the old patterning of sounds can be kept intact only by shifting the old b, d, g sounds in some way. They may become aspirated to bh, dh, gh or spirantized or nasalized or they may develop any other peculiarity that keeps them intact as a series and serves to differentiate them from other series. And this sort of shifting about without loss of pattern, or with a minimum loss of it, is probably the most important tendency in the history of speech sounds. Phonetic leveling and “splitting” counteract it to some extent but, on the whole, it remains the central unconscious regulator of the course and speed of sound changes.

The desire to hold on to a pattern, the tendency to “correct” a disturbance by an elaborate chain of supplementary changes, often spread over centuries or even millennia—these psychic undercurrents of language are exceedingly difficult to understand in terms of individual psychology, though there can be no denial of their historical reality. What is the primary cause of the unsettling of a phonetic pattern and what is the cumulative force that selects these or those particular variations of the individual on which to float the pattern readjustments we hardly know. Many linguistic students have made the fatal error of thinking of sound change as a quasi-physiological instead of as a strictly psychological phenomenon, or they have tried to dispose of the problem by bandying such catchwords as “the tendency to increased ease of articulation” or “the cumulative result of faulty perception” (on the part of children, say, in learning to speak). These easy explanations will not do. “Ease of articulation” may enter in as a factor, but it is a rather subjective concept at best. Indians find hopelessly difficult sounds and sound combinations that are simple to us; one language encourages a phonetic drift that another does everything to fight. “Faulty perception” does not explain that impressive drift in speech sounds which I have insisted upon. It is much better to admit that we do not yet understand the primary cause or causes of the slow drift in phonetics, though we can frequently point to contributing factors. It is likely that we shall not advance seriously until we study the intuitional bases of speech. How can we understand the nature of the drift that frays and reforms phonetic patterns when we have never thought of studying sound patterning as such and the “weights” and psychic relations of the single elements (the individual sounds) in these patterns?

Every linguist knows that phonetic change is frequently followed by morphological rearrangements, but he is apt to assume that morphology exercises little or no influence on the course of phonetic history. I am inclined to believe that our present tendency to isolate phonetics and grammar as mutually irrelevant linguistic provinces is unfortunate. There are likely to be fundamental relations between them and their respective histories that we do not yet fully grasp. After all, if speech sounds exist merely because they are the symbolic carriers of significant concepts and groupings of concepts, why may not a strong drift or a permanent feature in the conceptual sphere exercise a furthering or retarding influence on the phonetic drift? I believe that such influences may be demonstrated and that they deserve far more careful study than they have received.

This brings us back to our unanswered question: How is it that both English and German developed the curious alternation of unmodified vowel in the singular (foot, Fuss) and modified vowel in the plural (feet, Füsse)? Was the pre-Anglo-Saxon alternation of fot and föti an absolutely mechanical matter, without other than incidental morphological interest? It is always so represented, and, indeed, all the external facts support such a view. The change from o to ö, later e, is by no means peculiar to the plural. It is found also in the dative singular (fet), for it too goes back to an older foti. Moreover, fet of the plural applies only to the nominative and accusative; the genitive has fota, the dative fotum. Only centuries later was the alternation of o and e reinterpreted as a means of distinguishing number; o was generalized for the singular, e for the plural. Only when this reassortment of forms took place[161] was the modern symbolic value of the foot: feet alternation clearly established. Again, we must not forget that o was modified to ö (e) in all manner of other grammatical and derivative formations. Thus, a pre-Anglo-Saxon hohan (later hon) “to hang” corresponded to a höhith, hehith (later hehth) “hangs”; to dom “doom,” blod “blood,” and fod “food” corresponded the verbal derivatives dömian (later deman) “to deem,” blödian (later bledan) “to bleed,” and födian (later fedan) “to feed.” All this seems to point to the purely mechanical nature of the modification of o to ö to e. So many unrelated functions were ultimately served by the vocalic change that we cannot believe that it was motivated by any one of them.

The German facts are entirely analogous. Only later in the history of the language was the vocalic alternation made significant for number. And yet consider the following facts. The change of foti to föti antedated that of föti to föte, föt. This may be looked upon as a “lucky accident,” for if foti had become fote, fot before the -i had had the chance to exert a retroactive influence on the o, there would have been no difference between the singular and the plural. This would have been anomalous in Anglo-Saxon for a masculine noun. But was the sequence of phonetic changes an “accident”? Consider two further facts. All the Germanic languages were familiar with vocalic change as possessed of functional significance. Alternations like sing, sang, sung (Anglo-Saxon singan, sang, sungen) were ingrained in the linguistic consciousness. Further, the tendency toward the weakening of final syllables was very strong even then and had been manifesting itself in one way and another for centuries. I believe that these further facts help us to understand the actual sequence of phonetic changes. We may go so far as to say that the o (and u) could afford to stay the change to ö (and ü) until the destructive drift had advanced to the point where failure to modify the vowel would soon result in morphological embarrassment. At a certain moment the -i ending of the plural (and analogous endings with i in other formations) was felt to be too weak to quite bear its functional burden. The unconscious Anglo-Saxon mind, if I may be allowed a somewhat summary way of putting the complex facts, was glad of the opportunity afforded by certain individual variations, until then automatically canceled out, to have some share of the burden thrown on them. These particular variations won through because they so beautifully allowed the general phonetic drift to take its course without unsettling the morphological contours of the language. And the presence of symbolic variation (sing, sang, sung) acted as an attracting force on the rise of a new variation of similar character. All these factors were equally true of the German vocalic shift. Owing to the fact that the destructive phonetic drift was proceeding at a slower rate in German than in English, the preservative change of uo to üe (u to ü) did not need to set in until 300 years or more after the analogous English change. Nor did it. And this is to my mind a highly significant fact. Phonetic changes may sometimes be unconsciously encouraged in order to keep intact the psychological spaces between words and word forms. The general drift seizes upon those individual sound variations that help to preserve the morphological balance or to lead to the new balance that the language is striving for.

I would suggest, then, that phonetic change is compacted of at least three basic strands: (1) A general drift in one direction, concerning the nature of which we know almost nothing but which may be suspected to be of prevailingly dynamic character (tendencies, e.g., to greater or less stress, greater or less voicing of elements); (2) A readjusting tendency which aims to preserve or restore the fundamental phonetic pattern of the language; (3) A preservative tendency which sets in when a too serious morphological unsettlement is threatened by the main drift. I do not imagine for a moment that it is always possible to separate these strands or that this purely schematic statement does justice to the complex forces that guide the phonetic drift. The phonetic pattern of a language is not invariable, but it changes far less readily than the sounds that compose it. Every phonetic element that it possesses may change radically and yet the pattern remain unaffected. It would be absurd to claim that our present English pattern is identical with the old Indo-European one, yet it is impressive to note that even at this late day the English series of initial consonants:

corresponds point for point to the Sanskrit series:

The relation between phonetic pattern and individual sound is roughly parallel to that which obtains between the morphologic type of a language and one of its specific morphological features. Both phonetic pattern and fundamental type are exceedingly conservative, all superficial appearances to the contrary notwithstanding. Which is more so we cannot say. I suspect that they hang together in a way that we cannot at present quite understand.

If all the phonetic changes brought about by the phonetic drift were allowed to stand, it is probable that most languages would present such irregularities of morphological contour as to lose touch with their formal ground-plan. Sound changes work mechanically. Hence they are likely to affect a whole morphological group here—this does not matter—, only part of a morphological group there—and this may be disturbing. Thus, the old Anglo-Saxon paradigm:

could not long stand unmodified. The o—e alternation was welcome in so far as it roughly distinguished the singular from the plural. The dative singular fet, however, though justified historically, was soon felt to be an intrusive feature. The analogy of simpler and more numerously represented paradigms created the form fote (compare, e.g., fisc “fish,” dative singular fisce). Fet as a dative becomes obsolete. The singular now had o throughout. But this very fact made the genitive and dative o-forms of the plural seem out of place. The nominative and accusative fet was naturally far more frequently in use than were the corresponding forms of the genitive and dative. These, in the end, could not but follow the analogy of fet. At the very beginning of the Middle English period, therefore, we find that the old paradigm has yielded to a more regular one:

The starred forms are the old nucleus around which the new paradigm is built. The unstarred forms are not genealogical kin of their formal prototypes. They are analogical replacements.

The history of the English language teems with such levelings or extensions. Elder and eldest were at one time the only possible comparative and superlative forms of old (compare German alt, älter, der älteste; the vowel following the old-, alt- was originally an i, which modified the quality of the stem vowel). The general analogy of the vast majority of English adjectives, however, has caused the replacement of the forms elder and eldest by the forms with unmodified vowel, older and oldest. Elder and eldest survive only as somewhat archaic terms for the older and oldest brother or sister. This illustrates the tendency for words that are psychologically disconnected from their etymological or formal group to preserve traces of phonetic laws that have otherwise left no recognizable trace or to preserve a vestige of a morphological process that has long lost its vitality. A careful study of these survivals or atrophied forms is not without value for the reconstruction of the earlier history of a language or for suggestive hints as to its remoter affiliations.

Analogy may not only refashion forms within the confines of a related cluster of forms (a “paradigm”) but may extend its influence far beyond. Of a number of functionally equivalent elements, for instance, only one may survive, the rest yielding to its constantly widening influence. This is what happened with the English -s plural. Originally confined to a particular class of masculines, though an important class, the -s plural was gradually generalized for all nouns but a mere handful that still illustrate plural types now all but extinct (foot: feet, goose: geese, tooth: teeth, mouse: mice, louse: lice; ox: oxen; child: children; sheep: sheep, deer: deer). Thus analogy not only regularizes irregularities that have come in the wake of phonetic processes but introduces disturbances, generally in favor of greater simplicity or regularity, in a long established system of forms. These analogical adjustments are practically always symptoms of the general morphological drift of the language.

A morphological feature that appears as the incidental consequence of a phonetic process, like the English plural with modified vowel, may spread by analogy no less readily than old features that owe their origin to other than phonetic causes. Once the e-vowel of Middle English fet had become confined to the plural, there was no theoretical reason why alternations of the type fot: fet and mus: mis might not have become established as a productive type of number distinction in the noun. As a matter of fact, it did not so become established. The fot: fet type of plural secured but a momentary foothold. It was swept into being by one of the surface drifts of the language, to be swept aside in the Middle English period by the more powerful drift toward the use of simple distinctive forms. It was too late in the day for our language to be seriously interested in such pretty symbolisms as foot: feet. What examples of the type arose legitimately, in other words via purely phonetic processes, were tolerated for a time, but the type as such never had a serious future.

It was different in German. The whole series of phonetic changes comprised under the term “umlaut,” of which u: ü and au: oi (written äu) are but specific examples, struck the German language at a time when the general drift to morphological simplification was not so strong but that the resulting formal types (e.g., Fuss: Füsse; fallen “to fall”: fällen “to fell”; Horn “horn”: Gehörne “group of horns”; Haus “house”: Häuslein “little house”) could keep themselves intact and even extend to forms that did not legitimately come within their sphere of influence. “Umlaut” is still a very live symbolic process in German, possibly more alive to-day than in medieval times. Such analogical plurals as Baum “tree”: Bäume (contrast Middle High German boum: boume) and derivatives as lachen “to laugh”: Gelächter “laughter” (contrast Middle High German gelach) show that vocalic mutation has won through to the status of a productive morphologic process. Some of the dialects have even gone further than standard German, at least in certain respects. In Yiddish,[162] for instance, “umlaut” plurals have been formed where there are no Middle High German prototypes or modern literary parallels, e.g., tog “day”: teg “days” (but German Tag: Tage) on the analogy of gast “guest”: gest “guests” (German Gast: Gäste), shuch[163] “shoe”: shich “shoes” (but German Schuh: Schuhe) on the analogy of fus “foot”: fis “feet.” It is possible that “umlaut” will run its course and cease to operate as a live functional process in German, but that time is still distant. Meanwhile all consciousness of the merely phonetic nature of “umlaut” vanished centuries ago. It is now a strictly morphological process, not in the least a mechanical phonetic adjustment. We have in it a splendid example of how a simple phonetic law, meaningless in itself, may eventually color or transform large reaches of the morphology of a language.

IX

How Languages Influence Each Other

Languages, like cultures, are rarely sufficient unto themselves. The necessities of intercourse bring the speakers of one language into direct or indirect contact with those of neighboring or culturally dominant languages. The intercourse may be friendly or hostile. It may move on the humdrum plane of business and trade relations or it may consist of a borrowing or interchange of spiritual goods—art, science, religion. It would be difficult to point to a completely isolated language or dialect, least of all among the primitive peoples. The tribe is often so small that intermarriages with alien tribes that speak other dialects or even totally unrelated languages are not uncommon. It may even be doubted whether intermarriage, intertribal trade, and general cultural interchanges are not of greater relative significance on primitive levels than on our own. Whatever the degree or nature of contact between neighboring peoples, it is generally sufficient to lead to some kind of linguistic interinfluencing. Frequently the influence runs heavily in one direction. The language of a people that is looked upon as a center of culture is naturally far more likely to exert an appreciable influence on other languages spoken in its vicinity than to be influenced by them. Chinese has flooded the vocabularies of Corean, Japanese, and Annamite for centuries, but has received nothing in return. In the western Europe of medieval and modern times French has exercised a similar, though probably a less overwhelming, influence. English borrowed an immense number of words from the French of the Norman invaders, later also from the court French of Isle de France, appropriated a certain number of affixed elements of derivational value (e.g., -ess of princess, -ard of drunkard, -ty of royalty), may have been somewhat stimulated in its general analytic drift by contact with French,[164] and even allowed French to modify its phonetic pattern slightly (e.g., initial v and j in words like veal and judge; in words of Anglo-Saxon origin v and j can only occur after vowels, e.g., over, hedge). But English has exerted practically no influence on French.

The simplest kind of influence that one language may exert on another is the “borrowing” of words. When there is cultural borrowing there is always the likelihood that the associated words may be borrowed too. When the early Germanic peoples of northern Europe first learned of wine-culture and of paved streets from their commercial or warlike contact with the Romans, it was only natural that they should adopt the Latin words for the strange beverage (vinum, English wine, German Wein) and the unfamiliar type of road (strata [via], English street, German Strasse). Later, when Christianity was introduced into England, a number of associated words, such as bishop and angel, found their way into English. And so the process has continued uninterruptedly down to the present day, each cultural wave bringing to the language a new deposit of loan-words. The careful study of such loan-words constitutes an interesting commentary on the history of culture. One can almost estimate the rôle which various peoples have played in the development and spread of cultural ideas by taking note of the extent to which their vocabularies have filtered into those of other peoples. When we realize that an educated Japanese can hardly frame a single literary sentence without the use of Chinese resources, that to this day Siamese and Burmese and Cambodgian bear the unmistakable imprint of the Sanskrit and Pali that came in with Hindu Buddhism centuries ago, or that whether we argue for or against the teaching of Latin and Greek our argument is sure to be studded with words that have come to us from Rome and Athens, we get some inkling of what early Chinese culture and Buddhism and classical Mediterranean civilization have meant in the world’s history. There are just five languages that have had an overwhelming significance as carriers of culture. They are classical Chinese, Sanskrit, Arabic, Greek, and Latin. In comparison with these even such culturally important languages as Hebrew and French sink into a secondary position. It is a little disappointing to learn that the general cultural influence of English has so far been all but negligible. The English language itself is spreading because the English have colonized immense territories. But there is nothing to show that it is anywhere entering into the lexical heart of other languages as French has colored the English complexion or as Arabic has permeated Persian and Turkish. This fact alone is significant of the power of nationalism, cultural as well as political, during the last century. There are now psychological resistances to borrowing, or rather to new sources of borrowing,[165] that were not greatly alive in the Middle Ages or during the Renaissance.

Are there resistances of a more intimate nature to the borrowing of words? It is generally assumed that the nature and extent of borrowing depend entirely on the historical facts of culture relation; that if German, for instance, has borrowed less copiously than English from Latin and French it is only because Germany has had less intimate relations than England with the culture spheres of classical Rome and France. This is true to a considerable extent, but it is not the whole truth. We must not exaggerate the physical importance of the Norman invasion nor underrate the significance of the fact that Germany’s central geographical position made it peculiarly sensitive to French influences all through the Middle Ages, to humanistic influences in the latter fifteenth and early sixteenth centuries, and again to the powerful French influences of the seventeenth and eighteenth centuries. It seems very probable that the psychological attitude of the borrowing language itself towards linguistic material has much to do with its receptivity to foreign words. English has long been striving for the completely unified, unanalyzed word, regardless of whether it is monosyllabic or polysyllabic. Such words as credible, certitude, intangible are entirely welcome in English because each represents a unitary, well-nuanced idea and because their formal analysis (cred-ible, cert-itude, in-tang-ible) is not a necessary act of the unconscious mind (cred-, cert-, and tang- have no real existence in English comparable to that of good- in goodness). A word like intangible, once it is acclimated, is nearly as simple a psychological entity as any radical monosyllable (say vague, thin, grasp). In German, however, polysyllabic words strive to analyze themselves into significant elements. Hence vast numbers of French and Latin words, borrowed at the height of certain cultural influences, could not maintain themselves in the language. Latin-German words like kredibel “credible” and French-German words like reussieren “to succeed” offered nothing that the unconscious mind could assimilate to its customary method of feeling and handling words. It is as though this unconscious mind said: “I am perfectly willing to accept kredibel if you will just tell me what you mean by kred-.” Hence German has generally found it easier to create new words out of its own resources, as the necessity for them arose.

The psychological contrast between English and German as regards the treatment of foreign material is a contrast that may be studied in all parts of the world. The Athabaskan languages of America are spoken by peoples that have had astonishingly varied cultural contacts, yet nowhere do we find that an Athabaskan dialect has borrowed at all freely[166] from a neighboring language. These languages have always found it easier to create new words by compounding afresh elements ready to hand. They have for this reason been highly resistant to receiving the linguistic impress of the external cultural experiences of their speakers. Cambodgian and Tibetan offer a highly instructive contrast in their reaction to Sanskrit influence. Both are analytic languages, each totally different from the highly-wrought, inflective language of India. Cambodgian is isolating, but, unlike Chinese, it contains many polysyllabic words whose etymological analysis does not matter. Like English, therefore, in its relation to French and Latin, it welcomed immense numbers of Sanskrit loan-words, many of which are in common use to-day. There was no psychological resistance to them. Classical Tibetan literature was a slavish adaptation of Hindu Buddhist literature and nowhere has Buddhism implanted itself more firmly than in Tibet, yet it is strange how few Sanskrit words have found their way into the language. Tibetan was highly resistant to the polysyllabic words of Sanskrit because they could not automatically fall into significant syllables, as they should have in order to satisfy the Tibetan feeling for form. Tibetan was therefore driven to translating the great majority of these Sanskrit words into native equivalents. The Tibetan craving for form was satisfied, though the literally translated foreign terms must often have done violence to genuine Tibetan idiom. Even the proper names of the Sanskrit originals were carefully translated, element for element, into Tibetan; e.g., Suryagarbha “Sun-bosomed” was carefully Tibetanized into Nyi-mai snying-po “Sun-of heart-the, the heart (or essence) of the sun.” The study of how a language reacts to the presence of foreign words—rejecting them, translating them, or freely accepting them—may throw much valuable light on its innate formal tendencies.

The borrowing of foreign words always entails their phonetic modification. There are sure to be foreign sounds or accentual peculiarities that do not fit the native phonetic habits. They are then so changed as to do as little violence as possible to these habits. Frequently we have phonetic compromises. Such an English word as the recently introduced camouflage, as now ordinarily pronounced, corresponds to the typical phonetic usage of neither English nor French. The aspirated k, the obscure vowel of the second syllable, the precise quality of the l and of the last a, and, above all, the strong accent on the first syllable, are all the results of unconscious assimilation to our English habits of pronunciation. They differentiate our camouflage clearly from the same word as pronounced by the French. On the other hand, the long, heavy vowel in the third syllable and the final position of the “zh” sound (like z in azure) are distinctly un-English, just as, in Middle English, the initial j and v[167] must have been felt at first as not strictly in accord with English usage, though the strangeness has worn off by now. In all four of these cases—initial j, initial v, final “zh,” and unaccented a of father—English has not taken on a new sound but has merely extended the use of an old one.

Occasionally a new sound is introduced, but it is likely to melt away before long. In Chaucer’s day the old Anglo-Saxon ü (written y) had long become unrounded to i, but a new set of ü-vowels had come in from the French (in such words as due, value, nature). The new ü did not long hold its own; it became diphthongized to iu and was amalgamated with the native iw of words like new and slew. Eventually this diphthong appears as yu, with change of stress—dew (from Anglo-Saxon deaw) like due (Chaucerian dü). Facts like these show how stubbornly a language resists radical tampering with its phonetic pattern.

Nevertheless, we know that languages do influence each other in phonetic respects, and that quite aside from the taking over of foreign sounds with borrowed words. One of the most curious facts that linguistics has to note is the occurrence of striking phonetic parallels in totally unrelated or very remotely related languages of a restricted geographical area. These parallels become especially impressive when they are seen contrastively from a wide phonetic perspective. Here are a few examples. The Germanic languages as a whole have not developed nasalized vowels. Certain Upper German (Suabian) dialects, however, have now nasalized vowels in lieu of the older vowel + nasal consonant (n). Is it only accidental that these dialects are spoken in proximity to French, which makes abundant use of nasalized vowels? Again, there are certain general phonetic features that mark off Dutch and Flemish in contrast, say, to North German and Scandinavian dialects. One of these is the presence of unaspirated voiceless stops (p, t, k), which have a precise, metallic quality reminiscent of the corresponding French sounds, but which contrast with the stronger, aspirated stops of English, North German, and Danish. Even if we assume that the unaspirated stops are more archaic, that they are the unmodified descendants of the old Germanic consonants, is it not perhaps a significant historical fact that the Dutch dialects, neighbors of French, were inhibited from modifying these consonants in accordance with what seems to have been a general Germanic phonetic drift? Even more striking than these instances is the peculiar resemblance, in certain special phonetic respects, of Russian and other Slavic languages to the unrelated Ural-Altaic languages[168] of the Volga region. The peculiar, dull vowel, for instance, known in Russian as “yeri”[169] has Ural-Altaic analogues, but is entirely wanting in Germanic, Greek, Armenian, and Indo-Iranian, the nearest Indo-European congeners of Slavic. We may at least suspect that the Slavic vowel is not historically unconnected with its Ural-Altaic parallels. One of the most puzzling cases of phonetic parallelism is afforded by a large number of American Indian languages spoken west of the Rockies. Even at the most radical estimate there are at least four totally unrelated linguistic stocks represented in the region from southern Alaska to central California. Nevertheless all, or practically all, the languages of this immense area have some important phonetic features in common. Chief of these is the presence of a “glottalized” series of stopped consonants of very distinctive formation and of quite unusual acoustic effect.[170] In the northern part of the area all the languages, whether related or not, also possess various voiceless l-sounds and a series of “velar” (back-guttural) stopped consonants which are etymologically distinct from the ordinary k-series. It is difficult to believe that three such peculiar phonetic features as I have mentioned could have evolved independently in neighboring groups of languages.

How are we to explain these and hundreds of similar phonetic convergences? In particular cases we may really be dealing with archaic similarities due to a genetic relationship that it is beyond our present power to demonstrate. But this interpretation will not get us far. It must be ruled entirely out of court, for instance, in two of the three European examples I have instanced; both nasalized vowels and the Slavic “yeri” are demonstrably of secondary origin in Indo-European. However we envisage the process in detail, we cannot avoid the inference that there is a tendency for speech sounds or certain distinctive manners of articulation to spread over a continuous area in somewhat the same way that elements of culture ray out from a geographical center. We may suppose that individual variations arising at linguistic borderlands—whether by the unconscious suggestive influence of foreign speech habits or by the actual transfer of foreign sounds into the speech of bilingual individuals—have gradually been incorporated into the phonetic drift of a language. So long as its main phonetic concern is the preservation of its sound patterning, not of its sounds as such, there is really no reason why a language may not unconsciously assimilate foreign sounds that have succeeded in worming their way into its gamut of individual variations, provided always that these new variations (or reinforced old variations) are in the direction of the native drift.

A simple illustration will throw light on this conception. Let us suppose that two neighboring and unrelated languages, A and B, each possess voiceless l-sounds (compare Welsh ll). We surmise that this is not an accident. Perhaps comparative study reveals the fact that in language A the voiceless l-sounds correspond to a sibilant series in other related languages, that an old alternation s: sh has been shifted to the new alternation l (voiceless): s.[171] Does it follow that the voiceless l of language B has had the same history? Not in the least. Perhaps B has a strong tendency toward audible breath release at the end of a word, so that the final l, like a final vowel, was originally followed by a marked aspiration. Individuals perhaps tended to anticipate a little the voiceless release and to “unvoice” the latter part of the final l-sound (very much as the l of English words like felt tends to be partly voiceless in anticipation of the voicelessness of the t). Yet this final l with its latent tendency to unvoicing might never have actually developed into a fully voiceless l had not the presence of voiceless l-sounds in A acted as an unconscious stimulus or suggestive push toward a more radical change in the line of B’s own drift. Once the final voiceless l emerged, its alternation in related words with medial voiced l is very likely to have led to its analogical spread. The result would be that both A and B have an important phonetic trait in common. Eventually their phonetic systems, judged as mere assemblages of sounds, might even become completely assimilated to each other, though this is an extreme case hardly ever realized in practice. The highly significant thing about such phonetic interinfluencings is the strong tendency of each language to keep its phonetic pattern intact. So long as the respective alignments of the similar sounds is different, so long as they have differing “values” and “weights” in the unrelated languages, these languages cannot be said to have diverged materially from the line of their inherent drift. In phonetics, as in vocabulary, we must be careful not to exaggerate the importance of interlinguistic influences.

I have already pointed out in passing that English has taken over a certain number of morphological elements from French. English also uses a number of affixes that are derived from Latin and Greek. Some of these foreign elements, like the -ize of materialize or the -able of breakable, are even productive to-day. Such examples as these are hardly true evidences of a morphological influence exerted by one language on another. Setting aside the fact that they belong to the sphere of derivational concepts and do not touch the central morphological problem of the expression of relational ideas, they have added nothing to the structural peculiarities of our language. English was already prepared for the relation of pity to piteous by such a native pair as luck and lucky; material and materialize merely swelled the ranks of a form pattern familiar from such instances as wide and widen. In other words, the morphological influence exerted by foreign languages on English, if it is to be gauged by such examples as I have cited, is hardly different in kind from the mere borrowing of words. The introduction of the suffix -ize made hardly more difference to the essential build of the language than did the mere fact that it incorporated a given number of words. Had English evolved a new future on the model of the synthetic future in French or had it borrowed from Latin and Greek their employment of reduplication as a functional device (Latin tango: tetigi; Greek leipo: leloipa), we should have the right to speak of true morphological influence. But such far-reaching influences are not demonstrable. Within the whole course of the history of the English language we can hardly point to one important morphological change that was not determined by the native drift, though here and there we may surmise that this drift was hastened a little by the suggestive influence of French forms.[172]

It is important to realize the continuous, self-contained morphological development of English and the very modest extent to which its fundamental build has been affected by influences from without. The history of the English language has sometimes been represented as though it relapsed into a kind of chaos on the arrival of the Normans, who proceeded to play nine-pins with the Anglo-Saxon tradition. Students are more conservative today. That a far-reaching analytic development may take place without such external foreign influence as English was subjected to is clear from the history of Danish, which has gone even further than English in certain leveling tendencies. English may be conveniently used as an a fortiori test. It was flooded with French loan-words during the later Middle Ages, at a time when its drift toward the analytic type was especially strong. It was therefore changing rapidly both within and on the surface. The wonder, then, is not that it took on a number of external morphological features, mere accretions on its concrete inventory, but that, exposed as it was to remolding influences, it remained so true to its own type and historic drift. The experience gained from the study of the English language is strengthened by all that we know of documented linguistic history. Nowhere do we find any but superficial morphological interinfluencings. We may infer one of several things from this:—That a really serious morphological influence is not, perhaps, impossible, but that its operation is so slow that it has hardly ever had the chance to incorporate itself in the relatively small portion of linguistic history that lies open to inspection; or that there are certain favorable conditions that make for profound morphological disturbances from without, say a peculiar instability of linguistic type or an unusual degree of cultural contact, conditions that do not happen to be realized in our documentary material; or, finally, that we have not the right to assume that a language may easily exert a remolding morphological influence on another.

Meanwhile we are confronted by the baffling fact that important traits of morphology are frequently found distributed among widely differing languages within a large area, so widely differing, indeed, that it is customary to consider them genetically unrelated. Sometimes we may suspect that the resemblance is due to a mere convergence, that a similar morphological feature has grown up independently in unrelated languages. Yet certain morphological distributions are too specific in character to be so lightly dismissed. There must be some historical factor to account for them. Now it should be remembered that the concept of a “linguistic stock” is never definitive[173] in an exclusive sense. We can only say, with reasonable certainty, that such and such languages are descended from a common source, but we cannot say that such and such other languages are not genetically related. All we can do is to say that the evidence for relationship is not cumulative enough to make the inference of common origin absolutely necessary. May it not be, then, that many instances of morphological similarity between divergent languages of a restricted area are merely the last vestiges of a community of type and phonetic substance that the destructive work of diverging drifts has now made unrecognizable? There is probably still enough lexical and morphological resemblance between modern English and Irish to enable us to make out a fairly conclusive case for their genetic relationship on the basis of the present-day descriptive evidence alone. It is true that the case would seem weak in comparison to the case that we can actually make with the help of the historical and the comparative data that we possess. It would not be a bad case nevertheless. In another two or three millennia, however, the points of resemblance are likely to have become so obliterated that English and Irish, in the absence of all but their own descriptive evidence, will have to be set down as “unrelated” languages. They will still have in common certain fundamental morphological features, but it will be difficult to know how to evaluate them. Only in the light of the contrastive perspective afforded by still more divergent languages, such as Basque and Finnish, will these vestigial resemblances receive their true historic value.

I cannot but suspect that many of the more significant distributions of morphological similarities are to be explained as just such vestiges. The theory of “borrowing” seems totally inadequate to explain those fundamental features of structure, hidden away in the very core of the linguistic complex, that have been pointed out as common, say, to Semitic and Hamitic, to the various Soudanese languages, to Malayo-Polynesian and Mon-Khmer[174] and Munda,[175] to Athabaskan and Tlingit and Haida. We must not allow ourselves to be frightened away by the timidity of the specialists, who are often notably lacking in the sense of what I have called “contrastive perspective.”

Attempts have sometimes been made to explain the distribution of these fundamental structural features by the theory of diffusion. We know that myths, religious ideas, types of social organization, industrial devices, and other features of culture may spread from point to point, gradually making themselves at home in cultures to which they were at one time alien. We also know that words may be diffused no less freely than cultural elements, that sounds also may be “borrowed,” and that even morphological elements may be taken over. We may go further and recognize that certain languages have, in all probability, taken on structural features owing to the suggestive influence of neighboring languages. An examination of such cases,[176] however, almost invariably reveals the significant fact that they are but superficial additions on the morphological kernel of the language. So long as such direct historical testimony as we have gives us no really convincing examples of profound morphological influence by diffusion, we shall do well not to put too much reliance in diffusion theories. On the whole, therefore, we shall ascribe the major concordances and divergences in linguistic form—phonetic pattern and morphology—to the autonomous drift of language, not to the complicating effect of single, diffused features that cluster now this way, now that. Language is probably the most self-contained, the most massively resistant of all social phenomena. It is easier to kill it off than to disintegrate its individual form.