Jean-Philippe Rameau · section 33 of 109

Chapter nineteen

Continuation of the preceding chapter, showing that melody arises from harmony

At first, harmony seems to arise from melody, since the melodies produced by the individual voices become harmony when united. Yet a path first had to be determined for each voice so that they could agree together. Whatever melodic order we observe in each individual part, the parts will scarcely form good harmony together—to say nothing of its being impossible—unless that order is dictated by the rules of harmony. Nevertheless, to make this harmonic whole more intelligible, teaching begins with how to make a melody. Suppose some progress is made: the ideas gained vanish as soon as another part must be joined to it. We are no longer masters of the melody. While searching for the path one part must take in relation to another, we often lose sight of our original intention, or at least must change it. Otherwise, the constraint imposed by this first part does not always allow us to give the others as perfect a melody as we might wish. Harmony, then, guides us, not melody. An accomplished musician can indeed conceive a beautiful melody suited to harmony. But where does this fortunate ability come from? May nature have contributed? Certainly. If nature has denied him this gift, however, how can he succeed? Only through rules. And where do we obtain these rules? That is what we must examine.

Does the first division of the string immediately offer two sounds from which melody can be formed? Certainly not: a man singing from octave to octave would not make a very beautiful melody. The second and third divisions of the string, from which all harmony arises, provide no sounds more suitable for melody, since a melody consisting only of thirds, fourths, fifths, sixths, and octaves would still not be perfect. Harmony is therefore generated first, and the rules of melody must necessarily be drawn from it. This is indeed what we do when we take the harmonic intervals just mentioned separately to form a fundamental progression, which is not yet melody. But when these intervals are placed together above one of their constituent sounds, they naturally follow a diatonic path determined by their own progression as they serve in turn as one another's foundation. From these consonant and diatonic progressions we then derive all the melody required. Thus harmonic intervals had to be known before melodic ones. Everything that can be taught to a beginner as melody consists of these consonant intervals, if that can be called melody. In chapter XXII we shall see that the ancients, too, derived their modulation only from melody, whereas it arises from harmony.

Once we know this consonant progression, adding three sounds above a bass note costs no more effort than adding only one. Here is how the two approaches are explained: you may put a third, fifth, or octave above a bass; or, you must put the third, fifth, and octave above it. To use either approach, these intervals must be known; once known, they are no harder to employ together than separately. Next, the part that made the third must make the fifth if the bass descends a third: this is precisely how we explain it. But if, in the bass's various progressions, we find the third here, the octave there, and the fifth elsewhere, we must always know what should follow each according to the bass's different movement. Thus, without realizing it, we teach four-part composition while presenting it only in two parts. Since I must know the succession of every consonance according to the bass's movement, and each consonance occurs in turn, employing them together is no harder than employing them individually. Moreover, if I cannot distinguish them all at once, I need attend only to each separately. Passing from one to another, I find the means to compose perfect harmony in four parts, from which I then draw all the knowledge needed to reach perfection. The accompanying explanation also leaves no room for error—if we may trust the experience of several people who knew only note values and, on their second reading of our rules, composed harmony as perfect as one could wish. Furthermore, if this composer can enjoy hearing his own productions,* his ear gradually develops. Once he becomes sensitive to the perfect harmony to which these beginnings lead, he can be certain of success, which depends entirely upon these first principles.

There can be no doubt, then, that once these four parts are familiar, we can reduce them to three or two. But what knowledge can two-part composition give us, even if we master it perfectly? Such mastery is almost impossible, since no foundation guides us, and everything taught this way remains barren. Either our memory cannot suffice, or everything can scarcely be encompassed; in the end one must add: Cætera docebit usus—practice will teach the rest. Do we wish to proceed to three and four parts? So little is said about them that we see one must possess genius and taste as fully developed as those great masters' merely to understand their teaching. Concerning four parts, Zarlin* says that they can scarcely be taught on paper and leaves them to composers' discretion: composers can train themselves through his preceding rules for two and three parts. Our views are quite opposed. We have just said that harmony can be taught only in four parts, where everything particular to it occurs in only two chords, as we have pointed out throughout; reducing these to three or two parts is very easy. Zarlin, however, does not fully define either two or three parts, and admits that he cannot define four, after conceding that perfect harmony consists of four parts, which he compares to the four elements. (Chapter 58, folio 281. To conclude, if the rules of harmony given until now have failed to yield perfect knowledge, the principle we have proposed is a sure goal by which to attain that knowledge, allowing nothing to escape.

* This is partly why we provide rules of accompaniment.

* Terza Parte, chapter 65, folio 320.