In works of this period one frequently meets with expressions of pride and joy in the wonderful results that had been achieved in comparative linguistics in the course of a few decades. Thus Max Müller writes: “All this becomes clear and intelligible by the light of Comparative Grammar; anomalies vanish, exceptions prove the rule, and we perceive more plainly every day how in language, as elsewhere, the conflict between the freedom claimed by each individual and the resistance offered by the community at large establishes in the end a reign of law most wonderful, yet perfectly rational and intelligible”; and again: “There is nothing accidental, nothing irregular, nothing without a purpose and meaning in any part of Greek or Latin grammar. No one who has once discovered this hidden life of language, no one who has once found out that what seemed to be merely anomalous and whimsical in language is but, as it were, a petrification of thought, of deep, curious, poetical, philosophical thought, will ever rest again till he has descended as far as he can descend into the ancient shafts of human speech,” etc. (Ch 41 f.). Whitney says: “The difference between the old haphazard style of etymologizing and the modern scientific method lies in this: that the latter, while allowing everything to be theoretically possible, accepts nothing as actual which is not proved by sufficient evidence; it brings to bear upon each individual case a wide circle of related facts; it imposes upon the student the necessity of extended comparison and cautious deduction; it makes him careful to inform himself as thoroughly as circumstances allow respecting the history of every word he deals with” (L 386). And Benfey, in his Geschichte der Sprachwissenschaft (1869, see pp. 562 f. and 596), arrives at the conclusion that the investigation of Aryan languages has already attained a very great degree of certainty, and that the reconstruction of Primitive Aryan, both in grammar and vocabulary, must be considered as in the main settled in such a way that only some details are still doubtful; thus, it is certain that the first person singular ended in -mi, and that this is a phonetic reduction of the pronoun ma, and that the word for ‘horse’ was akva. This feeling of pride is certainly in a great measure justified if we compare the achievements of linguistic science at that date with the etymologies of the eighteenth century; it must also be acknowledged that 90 per cent. of the etymologies in the best-known Aryan languages which must be recognized as established beyond any reasonable doubt had already been discovered before 1870, while later investigations have only added a small number that may be considered firmly established, together with a great many more or less doubtful collocations. But, on the other hand, in the light of later research, we can now see that much of what was then considered firm as a rock did not deserve the implicit trust then placed in it.

IV.—§ 2. New Discoveries.

This is true in the first place with regard to the phonetic structure ascribed to Proto-Aryan. A series of brilliant discoveries made about the year 1880 profoundly modified the views of scholars about the consonantal and still more about the vocalic system of our family of languages. This is particularly true of the so-called palatal law.[15] So long as it was taken for granted that Sanskrit had in all essential points preserved the ancient sound system, while Greek and the other languages represented younger stages, no one could explain why Sanskrit in some cases had the palatals c and j (sounds approximately like the initial sounds of E. chicken and joy) where the other languages have the velar sounds k and g. It was now recognized that so far from the distribution of the two classes of sounds in Sanskrit being arbitrary, it followed strict rules, though these were not to be seen from Sanskrit itself. Where Sanskrit a following the consonant corresponded to Greek or Latin o, Sanskrit had velar k or g; where, on the other hand, it corresponded to Greek or Latin e, Sanskrit had palatal c or j. Thus we have, for instance, c in Sansk. ca, ‘and’ = Greek te, Lat. que, but k in kakša = Lat. coxa; the difference between the two consonants in a perfect like cakara, ‘have done,’ is dependent on the same vowel alternation as that of Greek léloipa; c in the verb pacati, ‘cooks,’ as against k in the substantive pakas, ‘cooking,’ corresponds to the vowels in Greek légei as against lógos, etc. All this shows that Sanskrit itself must once have had the vowels e and o instead of a; before the front vowel e the consonant has then been fronted or palatalized, as ch in E. chicken is due to the following front vowel, while k has been preserved before o in cock. Sanskrit is thus shown to be in some important respects less conservative than Greek, a truth which was destined profoundly to modify many theories concerning the whole family of languages. As Curtius said, with some resentment of the change in view then taking place, “Sanskrit, once the oracle of the rising science and trusted blindly, is now put on one side; instead of the traditional ex oriente lux the saying is now in oriente tenebræ” (K 97).

The new views held in regard to Aryan vowels also resulted in a thorough revision of the theory of apophony (ablaut). The great mass of Aryan vowel alternations were shown to form a vast and singularly consistent system, the main features of which may be gathered from the following tabulation of a few select Greek examples, arranged into three columns, each representing one ‘grade’:

I II III
(1) pétomaipótēeptómai
(s)ékhō (s)ókhos éskhon
(2) leípōléloipaélipon
(3) peúthomai—eputhómēn
(4) dérkomaidédorkaédrakon
(5) teínō (*tenjo)tónostatós

It is outside our scope to show how this scheme gives us a natural clue to the vowels in such verbs as E. I ride, II rode, III ridden (2), G. I werde, II ward, III geworden (4), or I binde, II band, III gebunden (5). It will be seen from the Greek examples that grade I is throughout characterized by the vowel e and grade II by the vowel o; as for grade III, the vowel of I and II has entirely disappeared in (1), where there is no vowel between the two consonants, and in (2) and (3), where the element found after e and o and forming a diphthong with these has now become a full (syllabic) vowel i and u by itself. In (4) Sanskrit has in grade III a syllabic r (adrçam = Gr. édrakon), while Greek has ra, or in some instances ar, and Gothonic has ur or or according to the vowel of the following syllable. It was this fact that suggested to Brugmann his theory that in (5) Greek a, Lat. in, Goth. un in the third grade originated in syllabic ṇ, and that tatós thus stood for *tṇtós; he similarly explained Gr. déka, Lat. decem, Gothic taihun, E. ten from *dekṃ with syllabic m. I do not believe that his theory is entirely correct; but so much is certain, that in all instances grade III is characterized by a reduction of the vowel that appears in the two other grades as e and o, and there can be no doubt that this reduction is due to want of stress. This being so, it becomes impossible to consider lip the original root-form, which in leip and loip has been extended, and the new theory of apophony thus disposes of the old theory, based on the Indian grammarians’ view that the shortest form was the root-form, which was then raised through ‘guna’ and ‘vrddhi.’ This now is reversed, and the fuller form is shown to be the oldest, which in some cases was shortened according to a process paralleled in many living languages. Bopp was right in his rejection of Grimm’s theory of an inner, significatory reason for apophony, as apophony is now shown to have been due to a mechanical cause, though a different one from that suggested by Bopp (see above, p. [53]); and Grimm was also wrong in another respect, because apophony is found from the first in noun-formations as well as in verbs, where Grimm believed it to have been instituted to indicate tense differences, with which it had originally nothing to do. Apophony even appears in other syllables than the root syllable; the new view thus quite naturally paved the way for skepticism with regard to the old doctrine that Aryan roots were necessarily monosyllabic; and scholars soon began to admit dissyllabic ‘bases’ in place of the old roots; instead of lip, the earliest accessible form thus came to be something like leipo or leipe. In this way the new vowel system had far-reaching consequences and made linguists look upon many problems in a new light. It should be noted, however, that the mechanical explanation of apophony from difference in accent applies only to grade III, in contradistinction to grades I and II; the reason of the alternation between the e of I and the o of II is by no means clear.

The investigations leading to the discovery of the palatal law and the new theory of apophony were only a part of the immense labour of a number of able linguists in the ’seventies and ’eighties, which cleared up many obscure points in Aryan phonology and morphology. One of the most famous discoveries was that of the Dane Karl Verner, that a whole series of consonant alternations in the old Gothonic languages was dependent on accent, and (more remarkable still) on the primeval accent, preserved in its oldest form in Sanskrit only, and differing from that of modern Gothonic languages in resting in some instances on the ending and in others on the root. When it was realized that the fact that German has t in vater, but d in bruder, was due to a different accentuation of the two words three or four thousand years ago, or that the difference between s and r in E. was and were was connected with the fact that perfect singulars in Sanskrit are stressed on the root, but plurals on the ending, this served not only to heighten respect for the linguistic science that was able to demonstrate such truths, but also to increase the feeling that the world of sounds was subject to strict laws comparable to those of natural science.

IV.—§ 3. Phonetic Laws and Analogy.

The ‘blind’ operation of phonetic laws became the chief tenet of a new school of ‘young-grammarians’ or ‘junggrammatiker’ (Brugmann, Delbrück, Osthoff, Paul and others), who somewhat noisily flourished their advance upon earlier linguists and justly roused the anger not only of their own teachers, including Curtius, but also of fellow-students like Johannes Schmidt and Collitz. For some years a fierce discussion took place on the principles of linguistic science, in which young-grammarians tried to prove deductively the truth of their favourite thesis that “Sound-laws admit of no exceptions” (first, it seems, enounced by Leskien). Osthoff wrongly maintained that sound changes belonged to physiology and analogical change to psychology; but though that distribution of the two kinds of change to two different domains was untenable, the distinction in itself was important and proved a valuable, though perhaps sometimes too easy instrument in the hands of the historical grammarian. It was quite natural that those who insisted on undeviating phonetic laws should turn their attention to those cases in which forms appeared that did not conform to these laws, and try to explain them; and thus they inevitably were led to recognize the immense importance of analogical formations in the economy of all languages. Such formations had long been known, but little attention had been paid to them, and they were generally termed ‘false analogies’ and looked upon as corruptions or inorganic formations found only or chiefly in a degenerate age, in which the true meaning and composition of the old forms was no longer understood. Men like Curtius were scandalized at the younger school explaining so many even of the noble forms of ancient Greek as due to this upstart force of analogy. His opponents contended that the name of ‘false analogy’ was wrong and misleading: the analogy in itself was perfect and was handled with unerring instinct in each case. They likewise pointed out that analogical formations, so far from being perversions of a late age, really represented one of the vital principles of language, without which it could never have come into existence.

One of the first to take the new point of view and to explain it clearly was Hermann Paul. I quote from an early article (as translated by Sweet, CP 112) the following passages, which really struck a new note in linguistic theory: