488 PROCEEDINGS OF THE AMERICAN ACADEMY. 



der that it is the only interrogative conjunction he knows, as it is the 

 ouly oue which has migrated into the Romance languages. Whether 

 this si is due originally to an influence from the Greek language, I dare 

 not decide.' — The examples for this paragraph are taken by Ussing 

 from Praun (p. 74 f.), but the inferences drawn from them by these two 

 scholars are different. Ussing holds that the almost exclusive use of si 

 in indirect questions instead of other particles is evidence of late author- 

 ship ; Praun, that such was ' die Richtung der Volksprache ' in the clas- 

 sical period. This phenomenon of the almost exclusive use of si with 

 which Ussing concludes his paragraph is really the only point in it that 

 has any force, for the preceding details are unimportant. Thus, there is 

 nothing surprising in the absence of mem from Vitruvius, since it is not 

 found in Catullus, Tibullus, or Pliny the Elder (Schmalz, Lat. Gramm? 

 p. 360). On the other hand, num does not ' entirely disappear ' from late 

 authors, for it is found in an indirect question in Orosius 1, 19, 9. 

 Boethius has numne {Herm. Sec. p. 46, line 12 Meiser), and Arnobius has 

 numquid 46 times (Schmalz, ibid.). The word nonne in indirect questions 

 is exclusively Ciceronian (Schmalz, p. 361). As for -ne, Caesar and 

 Sallust have it only half-a-dozen times each, whereas Tacitus has it nearly 

 thirty times, so that nothing about the date can be argued from its absence 

 from Vitruvius. We should not be surprised at missing an in Vitruvius 

 in the simple indirect question with quaero or other verbs meaning 'ask,' 

 because it is not commonly found in the ante-classical or classical period 

 except in connection with scio and verbs of doubting {Thesaurus, s. v. an, 

 p. 7 ff.). What then is left of Ussing's observation ? Nothing but eight 

 examples in which si is said to be used in indirect questions in Vitruvius 

 (seven quoted by Ussing, to which add 162, 17: quaeratur solum si sit 

 perjietuo solidum). But a closer examination of these examples will 

 show that half of them may be eliminated at once. I mean the two with 

 probari and those with animadvertere and hedicare. In all of these except 

 one (53, 12) we have the indicative in the clause with si, and none are 

 indirect questions but all are conditional protases used instead of indirect 

 questions (see Praun, pp. 70 and 72 on the two examples with proba?-i). 

 This leaves only the four cases with quaero, which certainly cannot be 

 called into evidence for late authorship, since quaero si is found in the 

 Augustan period, for instance in Propertius (2, 3, 5) and Livy (29, 25, 8 ; 

 39, 50, 7). The only truthful observation, therefore, which can be made 

 about Vitruvius's habits in expressing indirect questions is that he seldom 

 employs the ' sentence-question ' 22 and only in the phrase quaero si. 



22 For other kinds of indirect questions in him, see Praun, p. 75 f. 



