Impossibility of a normative grammar.
Language has sometimes been regarded as a voluntary or arbitrary act. But at others the impossibility of creating language artificially, by an act of will, has been clearly seen. "Tu, Caesar, civitatem dare potes homini, verbo non potes" was once said to a Roman Emperor. And the æsthetic (and therefore theoretic as opposed to practical) nature of expression supplies the method of discovering the scientific error which lies in the conception of a (normative) Grammar, establishing the rules of correct speech. Good sense has always rebelled against this error. An example of such rebellion is the "So much the worse for grammar" attributed to Monsieur de Voltaire. But the impossibility of a normative grammar is also recognized by those who teach it, when they confess that to write well cannot be learned by rules, that there are no rules without exceptions, and that the study of Grammar should be conducted practically, by reading and examples, which should form the literary taste. The scientific reason of this impossibility lies in the principle that we have demonstrated: that a technique of the theoretical amounts to a contradiction in terms. And what could a (normative) grammar be, but precisely a technique of linguistic expression, that is to say of a theoretic fact?
Didactic organisms.
The case in which Grammar is understood merely as an empirical discipline, that is to say, as a collection of schemes useful for learning languages, without any claim whatever to philosophic truth, is quite different. Even the abstractions of the parts of speech are in this case both admissible and useful. And we must tolerate as merely didascalic many books entitled "Treatises of Linguistic," where we generally find a little of everything, from the description of the vocal apparatus and of the artificial machines (phonographs) which can imitate it, to summaries of the most important I results obtained by Indo-European, Semitic, Coptic, Chinese, or other philologies; from philosophical generalizations as to the origin or nature of language, to advice on format, calligraphy and the arrangement of notes relating to philological work. But this mass of notions, here administered in a fragmentary and incomplete manner about language in its essence, about language as expression, resolves itself into notions of Æsthetic. Nothing exists outside Æsthetic, which gives knowledge of the nature of language, and empirical Grammar, which is a pedagogic expedient, save the History of languages in their living reality, that is to say, the history of concrete literary productions, which is substantially identical with the History of literature.
Elementary linguistic facts or roots.
The same error of taking the physical for the æsthetic, from which the search for the elementary forms of the beautiful originates, is made by those who go in search of elementary linguistic facts, decorating with that name the divisions of the longer series of physical sounds into shorter series. Syllables, vowels and consonants, and the series of syllables called words, all these elements of speech, which give no definite sense when taken alone, must be called not facts of language, but mere sounds, or rather sounds abstracted and classified physically.
Another error of the same sort is that of roots, to which the most distinguished philologists now accord but small value. Having confused physical with linguistic or expressive facts, and considering that the simple precedes the complex in the order of ideas, they necessarily ended by thinking that the smallest physical facts indicated the simplest linguistic facts. Hence the imaginary necessity that the most ancient primitive languages had a monosyllabic character, and that historical research must always lead to the discovery of monosyllabic roots. But (to follow up the imaginary hypothesis) the first expression that the first man conceived may have had not a phonetic but a mimetic physical reflex; may have been externalized not in a sound but in a gesture. And assuming that it was externalized in a sound, there is no reason to suppose that sound to have been monosyllabic rather than polysyllabic. Philologists readily blame their own ignorance and impotence, when they do not always succeed in reducing polysyllabism to monosyllabism, and rely upon the future to accomplish the reduction. But their faith is without foundation, and their blame of themselves is an act of humility arising from an erroneous presumption.
For the rest, the limits of syllables, as those of words, are altogether arbitrary, and distinguished somehow or other by empirical use. Primitive speech, or the speech of uneducated man, is a continuum, unaccompanied by any consciousness of divisions of the discourse into words or syllables, imaginary beings created by schools. No true law of Linguistic can be founded on such divisions. Proof of this is to be found in the confession of linguists, that there are no truly phonetic laws of the hiatus, of cacophony, of diæresis or synæresis, but merely laws of taste and convenience; that is to say, æsthetic laws. And what are laws of words which are not at the same time laws of style?
Æsthetic judgement and the model language.
Finally, the search for a model language, or for a method of reducing linguistic usage to unity, arises from the superstition of a rationalistic measure of the beautiful, from that concept which we have called false æsthetic absoluteness. In Italy we call this the question of the unity of the language.