12 L. A. Sherman. 



determine the sentence average in works of fiction. Here 

 of course the matter is mainly narrative or descriptive, thus 

 reaching the imagination of the reader more directly ; also 

 much of the language is quotation and in dialogue. The 

 sentence average is nevertheless often high, as illustrated by 

 De Foe {6S ; but in this author largely due to the abuse of 

 the semicolon). In the heavier sort of composition, as 

 elaborate criticism, where the thought is almost exclusively 

 symbolic, it is noticeable that the sentence average rises. 

 The approximate aggregate of Matthew Arnold is 37 ; of Mr. 

 Lowell, 38 ; of Higginson, 33 ; and of Walter Pater, 36.5. 



As to the implied fact of a literary sentence-rhythmus which 

 remains constant in standard writers through different periods 

 of composition, a few statistics further will serve both for 

 illustration and evidence. Three hundred sentences will 

 generally reveal the sentence-rhythm of any writer who has 

 achieved a style. Finding that this for De Quincey in the 

 Opium Eater (pub. 1821) is 32, we proceed to test it by 

 averages of his Toryism, Whiggism, and Radicalism (1835), 

 California {1852), and China (1857), and find 31.32 for the 

 first, 30.20 for the second, and 31.35 for the third respec- 

 tively. Again, starting with the 23.65 from Machiavelli 

 (1827) for Macaulay, we find 24.36 in vol. I. of the History 

 (written about 1845), 24.14 in vol. V., and 24 in the Pitt 

 essay (1859). There is little variation between the respec- 

 tive averages taken by the hundreds in the authors examined, 

 De Quincey showing greatest range (37.09-28.97), Channing 

 and Bartol least (27.48-24.13 and 17.46-14.08 respectively). 



It therefore seems clear that mathematics can be shown to 

 sustain a certain relation to rhetoric, and may aid in deter- 

 mining its laws. But what, as a psychological fact, sentence- 

 rhythm really means, how far it is common to authors in 

 English and other literatures, and hence a necessary element 

 of style, are questions yet to be considered. It is in order 

 to hasten their treatment by other hands that these observa- 

 tions are published at the present time. 



130 



