Replication implies that we "know" how it was done. Anything else must fall into the category of demonstration. However, let's not argue semantics; I think your point is: "We think Gordon's method for moving stones is so simple and so efficient that it is very likely to have been known and used by Stone Age/Bronze Age people, but in relation to the erecting of the stones we ought to use a method that is also the most likely one to have been actually used by these people."

Whilst I understand your reasoning from a strictly archaeological viewpoint, I think there is scope to say: "Let's see if Gordon's method of erecting stones is a viable alternative to the other theories." The BBC have already demonstrated how stones could have been hauled up ramps, we don't need to replicate their technique. If Gordon's method turns out to be a good one, then we merely have another contender for the title of "most likely method" and the archaeological establishment can add it to their bag of tricks (or not) as they see fit.

This is why I strongly believe that we should demonstrate the stone rowing well before we attempt the erection. Let's get stone-rowing accepted first. The erection then becomes: "We know that stone-rowing is a great way to transport blocks, let's see if the same levers and logs idea could have been used to erect the stones too." Then it's an experiment and not a proof.

I think we need to know whether we can drop stones into holes reliably. If we can do it then the method is viable and should be demonstrated so that it can be considered on its merits. If we can't then maybe we either stick to stone-rowing and avoid the issue altogether or consider other methods that involve low towers (using Gordons lifting method) and hauling up to vertical by rope. But I see this as a "Plan B", rather than "Plan A", as you seem to be advocating.