第119章
- First Principles
- 佚名
- 763字
- 2016-03-02 16:29:02
Anglo-Saxon steorra has been in course of time consolidated into Englishstar, mona into moon, and nama into name. The transition through semi-Saxonis clearly traceable. Sunu became in semi-Saxon sune, and in English son;the final e of sune being an evanescent form of the original u. The changefrom the Anglo-Saxon plural, formed by the distinct syllable as, to our pluralformed by the appended consonant s, shows the same thing: smithas in becomingsmiths, and endas in becoming ends, illustrate progressive coalescence. So,too does the disappearance of the terminal an in the infinitive mood of verbs;as shown in the transition from the Anglo-Saxon cuman to the semi-Saxon cumme,and to the English come. Moreover the process has been slowly going on, evensince what we distinguish as English was formed. In Elizabeth's time, verbswere still frequently pluralized by the addition of en -- we tell was wetellen; and in some places this form of speech may even now be heard. Inlike manner the terminal ed of the past tense, has united with the word itmodifies. Burn-ed has in pronunciation become burnt; and even in writing,the terminal t has in some cases taken the place of the ed. Only where antiqueforms in general are adhered to, as in the church-service, is the distinctnessof this inflection still maintained. Further, we see that the compound vowelshave been in many cases fused into single vowels. That in bread the e anda were originally both sounded, is proved by the fact that they are stillso sounded in parts where old habits linger. We, however have contractedthe pronunciation into bred; and we have made like changes in many othercommon words. Lastly, let it be noted that where the repetition is greatest,the process is carried furthest; as instance the contraction of lord (originallyhlaford) into lud in the mouths of barristers; and, still better, the coalescenceof God be with you into Good bye.
Besides thus exhibiting the integrative process, Language equally exhibitsit throughout all grammatical development. The lowest kinds of human speech,having merely nouns and verbs without inflections, permit no such close unionof the elements of a proposition as results when their relations are markedeither by inflections or by connective words. Such speech is what we significantlycall "incoherent." To a considerable extent, incoherence is seenin the Chinese language. "If, instead of saying I go to London, figscome from Turkey, the sun shines through the air, we said, I go end London,figs come origin Turkey, the sun shines passage air, we should discourseafter the manner of the Chinese." From this "aptotic" form,there is a transition, by coalescence, to a form in which the connexionsof words are expressed by joining with them certain inflectional words. "InLanguages like the Chinese," remarks Dr. Latham, "the separatewords most in use to express relation may become adjuncts or annexes."To this he adds the fact that "the numerous inflexional languages fallinto two classes. In one, the inflexions have no appearance of having beenseparate words. In the other, their origin as separate words is demonstrable."From which the inference drawn is, that the "aptotic" languages,by the more and more constant use of adjuncts, gave rise to the "agglutinate"languages, or those in which the original separateness of the inflexionalparts can be traced; and that out of these, by further use, arose the "amalgamate"languages, or those in which the original separateness of the inflexionalparts can no longer be traced. Strongly corroborative of this inference isthe fact that, by such a process, there have grown out of the amalgamatelanguages, the "anaptotic" languages, of which our own is the bestexample -- languages in which, by further consolidation, inflexions havealmost disappeared, while, to express the verbal relations, new kinds ofwords have been developed. When we see the Anglo-Saxon inflexions graduallylost by contraction during the development of English, and, though to a lessdegree, the Latin inflexions dwindling away during the development of French,we cannot deny that grammatical structure is modified by integration; andseeing how clearly the earlier stages of grammatical structure are explainedby it, we must conclude that it has been going on from the first.
In proportion to the degree of this integration, is the extent to whichintegration of another order is carried. Aptotic languages are, as alreadypointed out, necessarily incoherent -- the elements of a proposition cannotbe completely tied into a whole. But as fast as coalescence produces inflectedwords, it becomes possible to unite them into sentences of which the partsare so mutually dependent that no considerable change can be made withoutdestroying the meaning. Yet a further stage in this process may be noted.