Timo Flink 90 The two positions listed are like opposite poles of a magnet. Each argu- ment has its counter-argument that seem to cancel it out. Both positions have been reached using the reasoned eclectic method. The differences lie in how both “parties” handle the divided external evidence and the contradictory internal evidence. Transcriptionally, either reading may lie behind the other11. Intrinsic probabilities are used in contradictory ways. The same evidence is used both for and against a given reading, depend- ing on the scholar making the decision. It appears that the external evidence is inconclusive. It is too similar in both cases unless one wants to emphasize certain manuscripts at the expense of others. Both readings find support in early text-types in geo- graphically diverse places and both readings were in existence already in the second century. This leaves the internal evidence. Within the intrinsic probabilities, should the rule of thumb be that any author is consistent with his/her language (favours υ ός) or should variance be allowed (favours κλεκτ ς). How many times any author has to write the same expression before it can be said to be a tendency from which the author does not deviate? This question is still ardently debated12. The problem with intrinsic probabilities is that in this case the chosen variant affects the understanding of the theology, style and language of the Fourth Evangelist i.e. the intrinsic probabilities themselves. It is in a state of flux due the nature of the known variants – every aspect changes with the chosen variant reading. It appears that the decision in this case should be keyed to the transcriptional probabilities13. There is a problem with it too. Which criterion (if any) takes preference: contextual (divided), inter- textual (divided) or doctrinal harmonisations (favours κλεκτ ς)? What that was seen a problematic concept: Jesus as κλεκτ ς is human but not divine (this was the problematic deduction). Hence, there was a desire to remove the term itself by substituting another term less likely to be misconstrued. There is, however, a cave-out to this anti-Adoptionistic argument. Is “the Son of God” any less perceptible to misapplica- tions than “the Chosen One of God”? Hebrew Scriptures use the term “Son of God” to refer to humans (Ps 2,7) as well as to non-human beings (Job 1,6). Thus, substituting υ ς for κλεκτ ς does not necessary solve the problem perceived by some scribes. Perhaps the substitution has more to do with a desire to harmonize the text to the known usage of the Fourth Evangelist rather than with doctrinal bias. 11 B. Aland, “Der textkritische und textgeschichtliche Nutzen”, 34; 12 See, e.g., the contrasting views of B.M. Metzger, who believes any author varies his/ her style occasionally, and J.K. Elliott, who believes that when an author writes the same expression umpteen times, it becomes a tendency and (s)he is unlikely to depart from it. See B.M. Metzger, “Review of Principles and Practice of New Testament Textual Criticism”, TLZ 117 (1992) 32-33; J.K. Elliott, “Thoroughgoing Eclecticism in New Testament Textual Criticism”, in B.D. Ehrman and M.W. Holmes (eds), Status Quaestionis, 321-35. 13 Fee, “Textual Criticism of the New Testament”, 127-55.