CHORDS IN THE LIGHT OF THEIR ORIGIN 275 



asterisked chord in c) 1 and 2 contains Gi? and Gj^, 

 that in 3 contains G and Gi?, that in 4 contains A 

 and Ati: each contains two distinct chord-fifths of a 

 common chord-root. Likewise each of the asterisked 

 chords in d) contains two distinct chord-sevenths of 

 the same root, in ^) two distinct chord-ninths of the 

 same root. Now it may be objected that all these 

 structures are nothing but passing chords. Yes, but 

 they are chords all the same, each is regnant, each is 

 a combination of harmonic percepts and subject to 

 harmonic analysis. Next it may be objected that the 

 above notation of these chords is arbitrary and incor- 

 rect, that in the asterisked chord in a), for example, 

 we might substitute At? for G^ and then the chord 

 would simply be the small ninth-chord of the domi- 

 nant, marked V^. I reply that Ab in this chord would 

 be absolutely false and misleading for two patent 

 reasons. First, because A]7 is a chromatic down- 

 leader with a downward tend whereas the regnant 

 harmony reports a chromatic upleader with an uptend : 

 hence G^. Second, because the step from AU in this 

 chord to Ajl in the next chord reports a progression 

 whereas the relative and regnant harmony report this 

 specific step to be a rising cadence and resolution : hence 

 again, Gjji. On its logical side no one will gainsay that 

 the symbols of notation to be accurate should be 

 selected in conformity with the harmonic idea to be 

 conveyed, and that this should be insisted on even at 

 the cost of certain old and time-honored traditions 

 and conventions, the preservation of which is the func- 

 tion of history but whose usefulness in practice no 

 longer exists. Certainly our 20th-century notation of 



