13 Narrative Structure and Verbal Aspect Choice in Luke of backgrounding verbs. Thus, the mode of discourse in this passage ap- pears to be the diegetic. This apparently straightforward conclusion, however, could be chal- lenged when the following verse and its aorist verb φαγεν are considered. Logically, the action of ο κ φαγεν also ought to be a backgrounded action, since, in the same way as γετο and πειραζ μενος, it happens off the timeline. It also likewise occurs open-ended over the same 40 days, continuing on after its immediate reference time; i.e., after the 40 days, Jesus continues to be led by the spirit, be tested by the devil, and eat nothing (the whole point of his victory over the devil’s first temptation)31. Thus, none of these actions could be said to be a completed act and suited to an aorist verb32. However, these “real world” considerations of both the action’s time and completeness are not tenable grounds for grammatical analysis33. It is not the event’s “real”, referential nature that dictates its aspect, but the author’s posture as to the conception of the event34. The action of ο κ φαγεν is recorded by an aorist verb, not as being puncticular, but rather as being viewed as a whole. Luke, in this case, selected the perfective aspect, as grammaticalized by the aorist form, to make a summarizing statement about Jesus’ fasting, conceived in its entirety. This is augmented by the prepositional phrase ν τα ς μ ραις κε ναις, which aids in effecting a change from an internal perspective to an external. Such a use of the aorist is commonly treated in grammars as complexive, comprehensive, or constantive35. While these categories may 31 This undermines D.L. Bock’s statement (following Plummer) that, on the basis of the imperfective aspect of πειραζ μενος, “the tempting is contemporaneous with the leading”, since the act of “not eating” is also contemporaneous and it is given a perfective aspect (Luke [Grand Rapids 1994] I, 369-70). See also n. 51. 32 This problem has gone largely unnoticed in the scholarship. M. Goulder, Luke: A New Paradigm (Sheffield 1989) I, 102-3, 297, has, at least, noticed the apparent incongruity, however he resorts to the troublesome notion of “Lukan muddle”, where the tense choices were not the result of authorial discretion, but rather a confused result of a mixing of the language of the sources. 33 Bakker’s words on the “referential bias” in hermeneutics (see above) are apropos here. 34 In the same way, the aorist πε νασεν cannot ipsa verba mean “began to be hungry”, which is how H. Schürmann, Das Lukasevangelium (Freiburg 1969) I, 208-10 and Bock, Luke, 371, take it. The aorist form of this verb is not constrained by lexis to be inceptive. In fact, when the beginning of an action or the entrance into a state is specifically indicated, Luke (and other NT writers) employ a form of the aorist of ρχομαι + inf. (BAGD, 140, s.v. ρχω, 2). Cf. especially Matt 12,1, where the contrast between the aorist πε νασεν, which in this instance denotes a state, and the ρχομαι + inf. construction is evident. 35 E.g., D.B. Wallace, Greek Grammar Beyond the Basics: An Exegetical Syntax of the New Testament (Grand Rapids 1996) 557-58, I; Fanning, Verbal Aspect, 255-61; N. Turner, A Grammar of New Testament Greek: Syntax (ed. J. H. Moulton) (Edinburgh 1963) III, 72; H.W. Smyth, Greek Grammar (Cambridge, MA 1956) §1927.