134 Sifdn and Horsok Vocabularies. [No. 2. 



the characteristics, however pre-eminently, are by no means exclu- 

 sively, Gyarung among the continental tongues, any more than they 

 are exclusively Ta-gala among the insular ones. Among the latter, 

 Humboldt considers that the Ta-gala (a specimen by the way of the 

 inseparable prefix) preserves the primitive type of the whole group ; 

 and that that type is revealed in the Gyarung I am inclined to assert, 

 without however forgetting that my investigation is far from com- 



Tbus, in our Gyarung vocabulary the words, cry) laugh, be silent, run, or four 

 out of twenty-four verbs, instead of a single prefix, have a double and even a treble 

 supply in the simple imperative form there used ; as Da-ka-kru from the root Kru ; 

 Ka-na-re from the root Re ; Na-ka-chum from the root Chum; Da-nara-gyuk 

 from the root Gyuk. Hence, compounding as before, we have from the last cited 

 simple term, Danarasagyuk, cause to run ; Madanarasagyiik, do not cause to run ; 

 Danarasagyungti, I who cause to run ; Manarasagyuti or Madanarasagyuti, he 

 who does not cause to run ; I believe also that the reiterative form Matarmang is 

 quite as usual as the substitutive form Marmang, and Matsazangti, for Matasa- 

 zangti, as Masazangti, time and tense notwithstanding. Repetition and other 

 changes above illustrated in the prefixes, belong much less to the roots, infixes and 

 suffixes whether in verbs or nouns, and when the root is repeated the prefix is com- 

 monly dropt, as has been explained as to substantives. But there are instances in 

 the verbs of root repeated and yet prefix retained, though the vocabulary affords 

 none such as its Kalarlar, round, which is a root repeated yet retaining its prefix ; 

 whilst the adjectives of the vocabulary, unlike the substantives also afford several 

 instances of the doubly and trebly reiterated prefix, as Kamgnar, sweet ; Ka-raa- 

 gnar from the root of gnar, and Kavandro, cold, Ka-va-na-dro from the root dro. 

 The elided forms, however, and particularly Kamagnar show that leaning towards 

 dissyllabism which has been dwelt on, — perhaps too strongly, though it assuredly 

 be a most marked feature of this tongue, and one too which Leyden's mistake as to 

 his own sample verb shows to be pre-eminently proper to Tagala ; for " tolog, to 

 sleep" is not the radical form of the word, as he assumes, but a compound of the 

 root and its customary prefix, ta, with the vowel harmonised to that of the root. 



The prefixes are the great variants, and besides being so much repeated, they can 

 be transposed and interchanged almost at pleasure owing to their synonymous cha- 

 racter, and these variations of the prefixes, with the elisions consequent on much 

 reiteration of them, constitute the greatest part of that enigma which Leyden em- 

 phasizes ; though it be in the actual use of the speech much less excessive (I still 

 speak of Gyarung), than his samples would lead any one to suppose. 



In the above samples of Gyarung I have given the verbs alone, without the added 

 pronouns of Leyden's Tagalan instances — such additional complication being ra- 

 ther suited to create wonderment than to promote sound knowledge. 



