72 THE STUDY OF SPEECH CURVES. 



a staid humor of a mild degree. Both these phrases might have been 

 spoken with rising closure without destrojdng the general tenor of the 

 impression, but the humorous turn and the contrast would have been 

 lacking. Throughout the record the melody is one that is appropriate to 

 ceremonial oration with a constant humorous twist to it. The unusually 

 long pauses between the phrases, with the low and monotonous pitch, 

 aid in the ceremonious expression. Some of the vowels are abnormally 

 long for purposes of emphasis. 



The melody of this portion of the Depew speech can be approximately 

 represented in musical notation as in figure 65. There is no division 

 by bars, as the prose accent is not so regular as the musical one. Never- 

 theless there is a decided rhythmical effect, and there is a division into 

 phrases showing similarity and contrast. 



It is often necessary to obtain on the spot the data for a melody plot; 

 a method simpler than that of tracing gramophone curves must be used. 

 A convenient method consists in placing a speaking tube before the mouth 

 or over the larynx and recording the vibrations on a smoked drum by 

 means of a very small tambour. The vibrations in such records are of 

 the same pitch as the glottal tone, but the form of the waves is not that 

 of a speech curve. 



The melody plot of the beginning of a record obtained in this way 

 of "Der Fichtenbaum" from a cultured BerUner (Baron von Hagen, a 

 Prussian major) is given in figure 66. The poem begins 



Ein Fichtenbaum steht einsam 

 Im Norden auf kahler Hoh'. 



It was spoken from memory without any suggestion whatever. 



The first Hne is evidently made up of two thought units, or two men- 

 tal images, namely, the fir tree in its place ("Ein Fichtenbaum steht") 

 and the lonehness ("einsam"). Using the terms proposed in "Elements 

 of Experimental Phonetics," page 553, there are two centers of thought 

 or two centroids. The fundamental law of melody seems to be that each 

 thought unit is characterized by a convex melody; it appears here in the 

 general rise and fall in each unit. The other laws of melody are also 

 exempUfied; thus, the initial vowel of a unit is rising (" ein, " etc.), the final 

 melody falling ("steht," etc.). The sudden fall at the end of [x] in " Fich- 

 tenbaum" is presumably a matter of disturbance of articulation and has 

 nothing to do with the intended melody. The double convexity in " ein- 

 sam" I would explain as arising from a peculiar form of emphasis. The 

 word might be spoken with prepondering stress on the first syllable, as 

 would be the case in the midst of a prose sentence; here the melody would 

 probably be a simple convexity. The word may also be spoken with two 



