Romance languages, comprising Italian, Spanish, French, &c.; inasmuch as not only does the classic Latin remain substantially the representative of their common original, but the very stages of their development from it are pre served in documents of successive ages. Thus, in compar ing the vocabularies of Italian and French, it is, in the first place, seen that they to a great extent correspond, this correspondence extending to words which one lan guage is least likely to borrow from another, viz., pronouns, the lower numerals, and names of the most universal and familiar objects. It is only, however, by etymological analysis that their depth of correspondence comes fully into view, it being seen that the ultimate elements or roots are largely common to the two languages, as are also the grammatical affixes by which words are formed from these roots, while general similarity of linguistic structure per vades both tongues. Such intimate correspondence could only result from derivation from a common parent lan guage, which in this case exists in Latin. In other groups of languages the existence of the common parent may be inferred from correspondence of this highest order. Thus there must have existed, at some period, what may be called the parent Slavonic, whence descend the Russian, Polish, Bohemian, <fec. ; and the parent Keltic, whence descend Welsh, Gaelic, Breton, &c., while behind the various branches of the whole Aryan family are dimly to be dis cerned the outlines of a primitive Aryan speech. In like manner, a comparison of the Arabic, Hebrew, Syriac, &c., shows that these must be all derived from a primitive Semitic speech, containing many of the simple root forms, which still exist in its modern descendants, and being already characterised by the principle of internal inflexion. Beyond the limits of these two, the most important lin guistic families, various others have been satisfactorily made out, though hardly with the same completeness of proof. In the Turanian or Tatar family are included the Turkish, Mongol, Hungarian, Finnish, Ostyak, &c.; the Dravidian family takes in the Tamil, Telugu, and various other South Indian dialects ; the Polynesian family comprises the lan guages of the higher race of the South Sea Islands ; the Negro-Kafir family consists of the prefixing languages spoken by most African tribes from the equatorial regions southward ; the Guarani family in South America, the Algonquin and Athapascan families in North America, and the Australian family, each includes a number of tribes ranging over a vast extent of territory, and so on. As to smaller divisions, it is common for languages to occur in groups of several connected dialects, though not forming part of one of the wider linguistic families ; thus the Aztec and Nicaraguan are closely related dialects, as are the Quichua and Aymara, while what philologists describe as isolated languages, as the Basque appears to be, are rather isolated groups of dialects, with no known analogues
beyond a limited district.
If the present state of the philological classification of mankind be compared with that of half a century ago, it will be seen that much progress has been made in referring groups of languages each to a common ancestral tongue. At the same time, greater cogency of proof is now de manded in such classification. The method of comparing a short vocabulary of twenty words or so in two languages is now abandoned, for where an extensive connection really exists, this is much better proved by a systematic com parison, while a few imperfect resemblances in the two lists might be due to accident, or the adoption of words. Nothing short of a similarity in the roots or elements of two languages, as well as in their grammatical structure, too strong to be explained by any independent causes, is now admitted as valid proof of common descent. This limitation, however, by no means amounts to a denial of the possibility of such descent. Thus it is often argued, on the strength of some similarities between Hebrew and Indo-European roots, that the two so distinct Semitic and Aryan families of language are themselves sprung from some yet more remotely ancient tongue. Thus also it has been attempted to connect the Malay and Tatar groups of languages. Either or both of these opinions may be true ; but the general verdict of philologists is, that they are not satisfactorily made out, and therefore cannot be recognised. Under the present standard of evidence in comparing languages and tracing allied groups to a common origin, the crude speculations as to a single primeval language of mankind, which formerly occupied so much attention, are acknowledged to be worthless. Increased knowledge and accuracy of method have as yet only left the way open to the most widely divergent suppositions. For all that known dialects prove to the contrary, on the other hand, there may have been one primitive language, from which the descendant languages have varied so widely, that neither their words nor their formation now indicate their unity in long past ages, while, on the other hand, the primitive tongues of mankind may have been numerous, and the extreme uulikeness of such languages as Basque, Chinese, Peruvian, Hottentot, and Sanskrit, may arise from absolute independence of origin.
itself absolute evidence as to its race-affinities. This is clearly shown in extreme cases. Thus the Jews in Europe have almost lost the use of Hebrew, but speak as their vernacular the language of their adopted nation, whatever it may be ; even the Jewish-German dialect, though consisting so largely of Hebrew words, is philologically German, as any sentence shows : " Ich hab noch hojom lo geachelt" " I have not yet eaten to-day." The mixture of the Israelites in Europe by marriage with other nations is probably much greater than is acknowledged by them ; yet, on the whole, the race has been preserved with extraordinary strictness, as its physical characteristics sufficiently show. Language thus here fails conspicuously as a test of race, and even of national history. Not much less conclusive is the case of the predominantly Negro populations of the West India Islands, who, nevertheless, speak as their native tongues dialects of English or French, in which the number of intermingled native African words is very scanty : "Dem hitti netti na ini watra bikasi dem de fisiman" " They cast a net into the water, because they were fishermen." (Surinam Negro-Eng.) " Bef pas ca j amain lasse poter cunes li;" " Le bceuf n est jamais las de porter ses cornes." (Haytian Negro-Fr.) If it be objected that the linguistic conditions of these two races are more artificial than has been usual in the history of the world, less extreme cases may be seen in countries where the ordinary results of conquest-colonisation have taken place. The Mestizos, who form so large a fraction of the population of modern Mexico, numbering several millions, afford a convenient test in this respect, inasmuch as their intermediate com plexion separates them from both their ancestral races, the Spaniard, and the chocolate-brown indigenous Aztec, or other Mexican. The mother-tongue of this mixed race is Spanish, with an infusion of Mexican words ; and a large proportion cannot speak any native dialect. In most or all nations of mankind, crossing or intermarriage of races has thus taken place between the conquering invader and the conquered native, so that the language spoken by the nation may represent the residts of conquest as much or more than of ancestry. The supersession of the Keltic Cornish by English, and of the Slavonic Old-Prussian by German, are but examples of a process which has for un told ages been supplanting native dialects, whose very names have mostly disappeared. On the other hand, the
