Derek, with all respect, I have to say that you make a mistake here. ( Of course, you might well succeed in persuading me to change my mind, as you have done in the case of the Carrick mat / bend ! )
The bowline does not have a hitch in it, like the Sheet bend. We can tie a bowline that holds satisfactory well, even if the collar is very loose. We can not do the same with a Sheet bend.
The bowline is a close relative of the Gleipnir, not of the Sheet bend. In the bowline, we simply replace the 50 % mechanical advantage that is offered by the second line that goes through the nipping loop, with the 50 % capstain advantage that is offered by the collar.
The Sheet bend does incorporate a hitch, just as the ABoK#1406, that is simply a doubled and symmetric Sheet bend (as such, it has two hitches.)
I know that we can spend the rest of our lives arguing about this, but why? Is nt it much better, and interesting, and useful, just to make another walk in the Knotland, with the hope we meet some new knots ?
I can see that in order to make any headway here, I will have to demonstrate to your satisfaction that the basic Bowline has the SBCore-i.e. that it incorporates a simple hitch...
I will start by conceding that under certain loading conditions the element in question is configured as a turned nip - as per the gripping element of the Gleipner, and in the Eskimo bwl the clamping function of the simple hitch is removed by the SP connection to the 'tail' of the hitch. Having said that, it is when load is applied to the 'tail' of the hitch, that the very worse characteristic of the Bowline shows itself - i.e. loading on this leg of the loop risks an easy transformation from bwl to 'noose'.
In the spectrum of loadings that a Bowline will meet, the load will vary from 100% on the bight leg, right through to 100% on the 'hitch tail'.
When the load is full on the bight leg, the knot is working as a pure sheetbend and I hope you can agree, in that loading configuration, it IS a simple hitch. At all other ratios of leg loadings, the hitch is functioning at some fraction as a simple hitch and some fraction as a 'turn-nip' - BUT - its function as the 'turn-nip' is the configuration we do NOT want to expose the bowline to.
As for spending my life arguing about this - NO - not interested, but I can spare some time to share perspectives.