LANMAN. — PALI BOOK-TITLES. 685 



.) 



oldest strata of redactional precipitates, and no question at all that 

 the Abhidhamma represents the latest. 28 



Buddhaghosa, in explaining at D.cm., 1.22, how the Tipitaka, as an 

 aggregation of collections (nikayas), may be regarded as five-fold, says 

 that it consists of the Dlgha, Majjhima, Sarjyutta, Anguttara, and 

 Khuddaka, and proceeds : " Apart from the four Nikayas, all the rest, 

 namely, the entire Vinaya and Abhidhamma, and the fifteen aforesaid 

 works, Khuddaka-patha etc., are the word of Buddha." Then, continu- 

 ing with a verse of "the ancients," he says : "And apart from these 

 four Nikayas, Dlgha and so forth, the words of Buddha other than 

 those, are held to be the Khuddaka-nikaya." 



Thapetva caturo p' ete nikaye Digha-adike 

 Tad-anfiarj Buddha- vacanarj Nikayo Khuddako mato ti. 29 

 The Gandha-varjsa expressly says 30 that the Khuddaka-nikaya consists 

 of the usual 15 texts plus the Vinaya and the Abhidhamma. Accor- 

 dingly, if we take Sutta and Vinaya and Abhidhamma as the sequence 

 of the texts in our scheme, doubtless no one will make serious 

 objection. 



Quadriliterals for uncanonical texts. — If the scheme thus far has 

 been rightly settled, we need have no hesitation in designating the titles 

 of the post-canonical books by quadriliterals. Herewith are not included 

 the commentaries (especially those of Buddhaghosa and Dhammapala), 

 which are discussed below. 



Canon 6. — A digraph must be counted as two letters, never as one. 

 This rule, as applied to Canon 5, is so absolutely essential and has 

 been so wholly ignored, that it demands special and separate mention. 

 If, on looking at an abbreviation, we must stop and go through the 

 mental process of considering whether two separately printed characters 

 are to be counted as one or as two, it is obvious that the advantage of 

 a scheme of abbreviations in which the number of letters employed is 

 highly significant, is wholly lost. This will be clear to any one upon 

 examining List 4 as it appears in the original typography, JPTS.1896, 

 pages 103-106. Here D. and Dh. alike are to be understood as unilit- 

 eral ; Vbh. as biliteral ; and Thig. as triliteral, — all being Pitaka texts 

 and intended to be designated with one letter. — Digraphs must on no 

 account be split, as in List 1, where Abhidhana-ppadlpika is designated 

 byAb. 



28 Cp. Pischel's Buddha, p. 6. 



29 D.cm., I. 23; repeated for substance, Dhs.cm., p. 26; also Sdhs., p. 30, 

 in JPTS.1890. Cp. also Childers, Dic'y, P- 282. 



30 Ed. JPTS.1886, p. 57, top; or Minayeff, Recherches, p. 237. 



