Stephen W. Frary 10 familiar words to grammatically smooth readings to a much greater degree than Robinson identifies. Elliott, on the other hand, believes that scribes, influenced by a revival of Attic Greek literary style in the second century A.D., improved the style and expression of the NT writers, and rid the text of “unGreek” expressions to a far greater degree than Robinson or Metzger will attest.24 Wherever such alterations by the scribe can be shown to have occurred, the resulting reading is unlikely to be original. Finally, it is generally agreed that “Readings which conform to the known style, vocabulary, and syntax of the original author are to be preferred.”25 This is not to say, though, that there is a consensus on what particular habits are to be attributed to a given author. Robinson raises the interesting point, for example, that one cannot necessarily extrapo- late from the fact that οὖν is typical of John’s writings that every time a variant reading presents a choice between οὖν and, say, δὲ (actually used more often by John) that we can assume οὖν to be original.26 There are two corollaries derived from these rules that bear mention- ing, one agreed upon and one in dispute. Given that scribes did, on oc- casion, make deliberate changes to the text, there is a consensus that “The reading which is more difficult as a scribal creation is to be preferred.”27 This is based on the assumption that a scribe wishing to “improve” the text would do so in a way so as to remove, not create, difficult-to-under- stand passages. Robinson goes on to hold that this rule should apply only when the difficult reading is not a solitary or poorly attested one. This contradicts the tenets of Elliott’s brand of eclecticism, but nonetheless all text critics will apply the rule in some instances. The disputed corollary is succinctly put: “prefer the shorter reading.” This seems to be a reasonable derivation from the facts, admitted by all, that line skip and other related unintentional variations result in shorter readings while pious expansions and conflations tend to lengthen the text. The fact that these rules covering these eventualities will have already been applied, though, render this additional principle unnecessary and mischievous. Robinson protests that applying this principle where there is no identified expansion or omission due to other factors automatically tends to favor the Alexandrian text type, which tends to be shorter where it varies from either the Western or Byzantine.28 He would eliminate the Elliott, The Greek Text of the Epistles, 8. 24 Robinson, “NT Textual Criticism”, 31. 25 Robinson, “NT Textual Criticism”, 32. 26 Robinson, “NT Textual Criticism”, 30. 27 Robinson, “NT Textual Criticism”, 43. 28