I heard back from Dr. Robinson, and he mentioned that there is a better, fuller presentation of his position on line, available here. And he said he does read my blog! Made my day.
Here is the argument Dr. Robinson presented, almost verbatim:
It is peculiar that the na27/ubs4 editors would reason in such a manner, particularly since their own evidence suggests a very simple transcriptional reason for the omission of qelwn in the two (and only two) Greek mss out of approximately 600 that contain this passage, reflecting all known text types.
If the particular archetype of B and D* (also the respective now-lost Greek archetypes of it-d, it-61 vg sy-p cop-sa arm) had read the transposition QELWN PIASAI ME (as is found in F G 1739), the omission could easily have occurred due to homoioteleuton of WN^WN, skipping from the prior ending of DAMASKHNWN to the ending of QELWN. That such a transposition with a consequent omission by h..t. could occur in whatever archetypes may independently have underlay the various versional witnesses is likewise no surprise.
Further, the transposed form would reflect a far more “natural” manner of syntactical expression than the Byzantine PASAI ME QELWN, which peculiarly places the infinitive before the participle. One need only compare the remaining nt instances of qelwn with an infinitive present to see this point amply illustrated: cf. Mt 1:19; 14:5; Lk 10:29; 14:28; 23:8, 20; Ac 24:27; 25;9; Ro 9:22; Heb 12:17; 1Pet 3:10. Apart from the peculiar Byzantine placement here of the infinitive before qelwn, there are no other instances of such in either nt or lxx, which clearly makes the Byzantine construction the “more difficult” reading.
If the preceding scenario is valid, and the reading of B D* arose by h. t. skipping from an archetypal DAMASKHNWN QELWN PIASAI ME, the issue then becomes a choice between the slimly supported QELWN PIASAI ME (F G 1739) or the strongly supported PIASAI ME QELWN — the latter being not only syntactically “more difficult,”, but also supported by the widest variety of mss and versional testimony.
As usual, I suggest the Byzantine reading to be more likely correct than vice versa.
There’s the argument. I’ll respond from my perspective soon.