Puzzlement for Me

I'm very puzzled now. Recall Devereux's argument in redaction criticism, which I posted here:

Scholars have argued...that chapters 7-9 and chapter 12 are 'later insertions' in Z since (i) they are intrusive in their context and (ii) there is no mention of them in the summary at the beginning of H. Further, in H 3 Aristotle says that it was shown elsewhere--not in Z or H--that forms are neither generated nor produced (1043b16-18). But Z 8 contains an argument for this thesis, so it seems that this chapter was not a part of Z when the reference in H 3 was written. These details suggest that there was an early, 'first' version of Z which is summarized at the beginning of H, and a later expanded version containing the 'inserted' chapters, Z 7-9 and Z 12.

This argument largely rests on the presupposition that in H 3, when Aristotle wrote the phrase, "in other discussions", e)n a)/lloij, then (if he really did write this) he could not have meant some other book in the Metaphysics: according to Devereux, he could not, from within H, have referred back to something already existing in Z by using that phrase. I pointed out that this argument was almost worthless. But I didn't do the obvious thing then, namely, check for other uses of e)n a)/lloij in the corpus. I argued on the basis of antecedent probabilities and analogies.

Well, I checked just now, and it turns out there are at least two other examples where Aristotle uses that phrase in one book of the Metaphysics to refer to something in another book. At 1017b9, in 5.8, that phrase is used to refer to 9.7 (o&te de\ dunato_n kai\ po&te ou1pw, e0n a1lloij dioriste/on))) and at 1046a5, in 9.1, that phrase is used to refer to 5.12 (o3ti me\n ou}n le/getaipollaxw~j h( du&namij kai\ to_ du&nasqai, diw&ristai h(mi=n e0n a1lloij)

So Devereux's argument is entirely worthless, isn't it?

What's bizarre is that Myles Burnyeat endorses the argument also in his Map of Metaphysics Z. In a section entitled, "Various reasons for thinking Z 7-9 a later insertion" he writes:

[iii] H3.1043b16, refers to Z8 as 'in another discussion' (e0n a1lloij). This proves the one-time independence of Z7-9 provided H is continuous with Z, which there is anyway good reason to suppose.

'Proves' is a strong word. I wonder if the appearance of proof comes from Burnyeat's argument having the character of tautology: if H is continuous with Z, then of course a reference to part of Z as if it were not continuous would show that that part was not within Z. That's true 'by definition', given an appropriately strong meaning of 'continuous'. On the other hand, one might just as well say that the H3 reference shows that Aristotle never regarded that discussion (or indeed any discussions in the Metaphysics), as 'continuous' in that very strong sense.