Now, it very often happens that a man wants to say something, but has not yet made up his mind as to what to say; and in this moment of hesitation, while thoughts are in the process of conception, the lungs and vocal chords will often be prematurely set going, and the result is [m] (sometimes preceded by the corresponding voiceless sound), often written hm or h’m, which thus becomes the interjection of an unshaped contradiction. Not infrequently this [m] precedes a real word; thus M’yes (written in this way by Shaw, Misalliance 154, and Merrick, Conrad 179) and Dan. mja, to mark a hesitating consent.
This will make it clear why words beginning with m are so often chosen as adversative conjunctions: people begin with this sound and go on with some word that gives good sense and which happens to begin with m: mais, maar. The Dan. men in the mouth of some early speakers is probably this [m], sliding into the old conjunction en, just as myes is m + yes; while other original users of men may have been thinking of men = meden, and others again of Low German men: these three etymologies are not mutually destructive, for all three origins may have concurrently contributed to the popularity of men. Modern Greek and Serbian ma are generally explained as direct loans from Italian, but may be indigenous, as may also dialectal Rumanian ma in the same sense, for in the hesitating [m] as the initial sound of objections we have one of those touches of nature which make the whole world kin.[78]
XVI.—§ 9. Object of Etymology.
What is the object of etymological science? “To determine the true signification of a word,” answers one of the masters of etymological research (Walde, Lat. et. Wörterb. xi). But surely in most cases that can be achieved without the help of etymology. We know the true sense of hundreds of words about the etymology of which we are in complete ignorance, and we should know exactly what the word grog means, even if the tradition of its origin had been accidentally lost. Many people still believe that an account of the origin of a name throws some light on the essence of the thing it stands for; when they want to define say ‘religion’ or ‘civilization,’ they start by stating the (real or supposed) origin of the name—but surely that is superstition, though the first framers of the name ‘etymology’ (from Gr. etumon ‘true’) must have had the same idea in their heads. Etymology tells us nothing about the things, nor even about the present meaning of a word, but only about the way in which a word has come into existence. At best, it tells us not what is true, but what has been true.
The overestimation of etymology is largely attributable to the “conviction that there can be nothing in language that had not an intelligible purpose, that there is nothing that is now irregular that was not at first regular, nothing irrational that was not originally rational” (Max Müller)—a conviction which is still found to underlie many utterances about linguistic matters, but which readers of the present volume will have seen is erroneous in many ways. On the whole, Max Müller naïvely gives expression to what is unconsciously at the back of much that is said and believed about language; thus, when he says (L 1. 44): “I must ask you at present to take it for granted that everything in language had originally a meaning. As language can have no other object but to express our meaning, it might seem to follow almost by necessity that language should contain neither more nor less than what is required for that purpose.” Yes, so it would if language had been constructed by an omniscient and omnipotent being, but as it was developed by imperfect human beings, there is every possibility of their having failed to achieve their purpose and having done either more or less than was required to express their meaning. It would be wrong to say that language (i.e. speaking man) created first what was strictly necessary, and afterwards what might be considered superfluous; but it would be equally wrong to say that linguistic luxuries were always created before necessaries; yet that view would probably be nearer the truth than the former. Much of what in former ages was felt to be necessary to express thoughts was afterwards felt as pedantic crisscross and gradually eliminated; but at all times many things have been found in language that can never have been anything else but superfluous, exactly as many people use a great many superfluous gestures which are not in the least significant and in no way assist the comprehension of their intentions, but which they somehow feel an impulse to perform. In language, as in life generally, we have too little in some respects, and too much in others.
XVI.—§ 10. Reconstruction.
Kluge somewhere (PBB 37. 479, 1911) says that the establishment of the common Aryan language is the chief task of our modern science of linguistics (to my mind it can never be more than a fragment of that task, which must be to understand the nature of language), and he thinks optimistically that “reconstructions with their reliable methods have taken so firm root that we are convinced that we know the common Aryan grundsprache just as thoroughly as any language that is more or less authenticated through literature.” This is a palpable exaggeration, for no one nowadays has the courage of Schleicher to print even the smallest fable in Proto-Aryan, and if by some miraculous accident we were to find a text written in that language we may be sure it would puzzle us just as much as Tokharian does.
Reconstruction has two sides, an outer and an inner. With regard to sounds, it seems to me that very often the masters of linguistics treat us to reconstructed forms that are little short of impossible. This is not the place to give a detailed criticism of the famous theory of ‘nasalis sonans,’ but I hope elsewhere to be able to state why I think this theory a disfiguring excrescence on linguistic science: no one has ever been able to find in any existing language such forms as mnto with stressed syllabic [n], given as the old form of our word mouth (Falk and Torp even give stmnto in order to connect the word with Gr. stóma), or as dkmtóm (whence Lat. centum, etc.) or bhrghnti̯es or gu̯mskete (Brugmann). Not only are these forms phonetically impossible, but the theory fails to explain the transitions to the forms actually existing in real languages, and everything is much easier if we assume forms like [ʌm, ʌn] with some vowel like that of E. un-. The use in Proto-Aryan reconstructions of non-syllabic i and u also in some respects invites criticism, but it will be better to treat these questions in a special paper.
Semantic reconstruction calls for little comment here. It is evident from the nature of the subject that no such strict rules can be given in this domain as in the domain of sound; but nowadays scholars are more realistic than formerly. Most of them will feel satisfied when moon and month are associated with words having the same two significations in related languages, without indulging in explanations of both from a root me ‘to measure’; and when our daughter has been connected with Gr. thugáter, Skt. duhitár and corresponding words in other languages, no attempt is made to go beyond the meaning common to these words ‘daughter’ and to speculate what had induced our ancestors to bestow that word on that particular relation, as when Lassen derived it from the root duh ‘to milk’ and pictured an idyllic family life, in which it was the business of the young girls to milk the cows, or when Fick derived the same word from the root dheugh ‘to be useful’ (G. taugen: ‘wie die magd, maid von mögen’), as if the daughters were the only, or the most, efficient members of the family. Unfortunately, such speculations are still found lingering in many recent handbooks of high standing: Kluge hesitates whether to assign the word mutter, mother, to the root ma in the sense ‘mete out’ or in the sense found in Sanskrit ‘to form,’ used of the fœtus in the womb. A resigned acquiescence in inevitable ignorance and a sense of reality should certainly be characteristics of future etymologists.