The dependence of artistic forms in their manifold developments upon a central governing idea or principle has never been more luminously stated than by the illustrious physicist Helmholtz, in the thirteenth chapter of his Tonempfindungen. I venture to think that in applying that truth to the facts of Greek music he was materially hindered by the accepted theory of the Greek modes. The scales which he analyses under that name were certainly the basis of all music in the Middle Ages, and are much more intelligible as such than in relation to the primitive Greek forms of the art [51].
§ 37. Epilogue—Speech and Song.
Several indications combine to make it probable that singing and speaking were not so widely separated from each other in Greek as in the modern languages with which we are most familiar.
(1) The teaching of the grammarians on the subject of accent points to this conclusion. Our habit of using Latin translations of the terms of Greek grammar has tended to obscure the fact that they belong in almost every case to the ordinary vocabulary of music. The word for 'accent' (tonos) is simply the musical term for 'pitch' or 'key.' The words 'acute' (oxys) and 'grave' (barys) mean nothing more than 'high' and 'low' in pitch. A syllable may have two accents, just as in music a syllable may be sung with more than one note. Similarly the 'quantity' of each syllable answers to the time of a musical note, and the rule that a long syllable is equal to two short ones is no doubt approximately correct. Consequently every Greek word (enclitics being reckoned as parts of a word) is a sort of musical phrase, and every sentence is a more or less definite melody—logôdes ti melos, as it is called by Aristoxenus (p. 18 Meib.). Moreover the accent in the modern sense, the ictus or stress of the voice, appears to be quite independent of the pitch or 'tonic' accent: for in Greek poetry the ictus (arsis) is determined by the metre, with which the tonic accent evidently has nothing to do. In singing, accordingly, the tonic accents disappear; for the melody takes their place, and gives each syllable a new pitch, on which (as we shall presently see) the spoken pitch has no influence. The rise and fall of the voice in ordinary speaking is perceptible enough in English, though it is more marked in other European languages. Helmholtz tells us—with tacit reference to the speech of North Germany—that an affirmative sentence generally ends with a drop in the tone of about a Fourth, while an interrogative is marked by a rise which is often as much as a Fifth [52]. In Italian the interrogative form is regularly given, not by a particle or a change in the order of the words, but by a rise of pitch. The Gregorian church music, according to a series of rules quoted by Helmholtz (l. c.), marked a comma by a rise of a Tone, a colon by a fall of a Semitone; a full stop by a Tone above, followed by a Fourth below, the 'reciting note'; and an interrogation by a phrase of the form d b c d (c being the reciting note).
These examples, however, do little towards enabling modern scholars to form a notion of the Greek system of accentuation. In these and similar cases it is the sentence as a whole which is modified by the tonic accent, whereas in Greek it is the individual word. It is true that the accent of a word may be affected by its place in the sentence: as is seen in the loss of the accent of oxytone words when not followed by a pause, in the anastrophe of prepositions, and in the treatment of the different classes of enclitics. But in all these instances it is the intonation of the word as such, not of the sentence, which is primarily concerned. What they really prove is that the musical accent is not so invariable as the stress accent in English or German, but may depend upon the collocation of the word, or upon the degree of emphasis which it has in a particular use.
(2) The same conclusion may be drawn from the terms in which the ancient writers on music endeavour to distinguish musical and ordinary utterance.