jugum, and so on with a large proportion of the native element in our speech. These correspondences are sometimes striking even to the casual observer, as in six and sex, for example, at other times it requires considerable knowledge of the subject to perceive them. Since Latin is in one sense an older language than our own (though from another point of view all Indo-European languages are equally old, as being independently descended from the parent stock), we expect to find the words less decayed in Latin than in English, especially modern English, which has undergone so many changes since the Anglo-Saxon time. But this is not always the case, for different IndoEuropean peoples have had different habits of linguistic conservatism. Thus our word snow shows an Indo-European initials which the Latin (8)nix, nivis, and the Greek (o)víþa, (8)nípha, have lost, and so in many instances. Comparison of cognates has shown that the changes in human speech, arbitrary as they seem to be, proceed in accordance with definite and ascertainable laws. For any united body of people form certain habits of utterance which affect their whole language in a remarkably uniform way, so that when these are once discovered one can predict with reasonable certainty what form the original word will take in a given dialect. Thus we find that it was the unconscious habit of the Spanish provincials to change Latin f to h at the beginning of a word, so that fabulare, to speak,' became hablár, facere became hacér, and so on. Hence we look for the Latin falco, falcon,' under hin the Spanish dictionary, and find it appearing as halcón; formosus appears as hermoso, and so on. Such habits, or laws of sound-change,' are equally noticeable in studying the development of the various Indo-European languages from the parent-speech. They may be followed 6 out in minute detail, and their existence and regularity have made Comparative Grammar almost an exact science. It is in great part our knowledge of such laws as this that enables us to distinguish with precision between cognates and borrowed words. For when a word is simply taken by one language from another, it suffers at first comparatively little change in its form. Thus we know that the Indo-European p was preserved in Latin but became f in Germanic, and this makes it easy to recog nize the Latin pecus and our fee as cognates, that is, as the common descendants of an Indo-European word. Both originally meant cattle,' a sense which the Latin has kept, while in English fee has come to mean 'property' in general, and has then suffered further changes of meaning. Similarly we recognize piscis and fish as cogWhen, however, we find piscatorial in English, and piscatorius in Latin, we have no hesitation in recognizing the former as borrowed from the latter, and so in thousands of instances. nates. We have said that Comparative Grammar is almost an exact science. The qualification is necessary, for appearances are deceitful, in linguistics as in life, and phenomena have to be examined with the greatest care, even when all the facts are in our possession, which is rarely the case. As an example of the difficulties that beset the investigator, take the so-called New England ŏ.' It is well known that the long sound of o represented by oa in road, tends, in the natural speech of New England, to become a short open o, difficult for those who are unfamiliar with it to imitate, but equally hard for a born Yankee to avoid. Thus we have coat, toad, lõad, boat, and so on. We should at once infer that a New Englander would say goat, but in fact, this word is never so pronounced, but is always gōat, without the shortening. The reason is clear enough in this case, for the phenomena are all recent, and the facts are known. The goat is not a familiar animal in this region, so that the name for it is rather a literary than a popular word; it is not properly in the dialect, and hence does not share in its peculiarities. The tendency in question had thus no opportunity to make itself felt as in coat and toad, which every child learns not from books, but from common talk. Suppose, however, we were dealing with a word that became obsolete two or three thousand years ago. We should be at a loss to account for the exception to the law,' and might even be tempted to provide goat with some strange etymology or ascribe a peculiar quality to its vowel in order to explain the facts. Furthermore, there are many opportunities for analogy and for hybrids and cross-breedings in language. Thus our nephew is, in a manner, both a native and a borrowed word. In Anglo-Saxon it was nefa (f pronounced as v), and this is cognate with the Latin nepos, nepotis. The French neveu, however, which is the Latin word in a decayed form, influenced the English word, and the result is our nephew, which is neither Anglo-Saxon nor French, but an amalgam of both. Such processes as this may have been operative at any time in the history of the IndoEuropean family, and their action interferes a good deal with the certainty of prehistoric etymologizing. Still, when all deductions have been made, there remains enough that is regular and undoubted to substantiate the claim of Comparative Grammar to be a true science. It has occupied some of the keenest intellects among scholars during the past century, and the results justify us in speaking with great confidence about the relations of the Indo-European tongues to each other and to the parent speech which has been obsolete for so many thousand years. These results afford the only firm basis for investigating the history of words. In particular, they enable us to reason with assurance about certain very ancient processes in the growth of the inflectional languages, as we shall have to do in the next chapter. CHAPTER XIII THE DEVELOPMENT OF WORDS I. ROOTS, STEMS, INFLECTION1 THE origin of language, as we have already seen, is an unsolved problem; yet the study of single words reveals many facts about the development of their form which make the question of their origin at least a simpler one. We find in our own words, on comparing them with other languages from which they have been borrowed or which have had a common origin with our own, certain obvious joints, as it were, which show clearly that the words have been built up of simpler elements by a process of aggrega 1 This chapter sets forth the main principles that have governed the development of words, as to their form, in our family of languages. For the sake of clearness, distinctions have been somewhat sharply made, and cautionary provisos have been omitted whenever there was danger that they would rather confuse than enlighten. Any orderly statement of these phenomena must be somewhat misleading; for, though the development of language is in general what is indicated, no single process ceased to act at the moment when another process began. The actual operations of speech-making in the Indo-European family must have been almost unimaginably complicated. The most minutely scientific investigation can arrive at only a part of the truth. A general outline must therefore be content to omit qualifications and parentheses. In particular, however, the reader should remember that the growth of stems did not immediately put an end to the root-period, and that the rise of inflection did not instantly put an end to the development of stems. Everything was gradual. The old processes survived alongside of the new, and only died out after long periods of time. A further caution is necessary. The principles here set forth accord with the accepted results of philological science. In other words, they |