I just listened to the Robert Hill performance on forte-piano and it was beautiful, but the meter was clearly regular, with use of "tempo ritardando" and "a tempo" in artistic manner at the phase (better: form/section) boundaries (as Rainer mentioned elsewhere). You admit already that you extend the 3rd beat repetitively, as in:
One, two, threeee, one, two, threeee, one, two, threeee ...
My issue is not about using rubato, it's about having REGULAR extra time, such that EVERY 3 beat bar has 3.x beats in it. It is this specific trait that I would like to see identified as inegalite, if it is authentic.
This discussion is getting into a useless preciosity. These things are ofter interrelated. You're so concerned about my 3rd beat (which Rubsam also does! did you listen to his recording?) that you didn't notice I'm also lingering the first beat. You're considering that the way I play those 16ths are the "straight rhtyhm", but it depends on what you have as a reference. if you have the first beat as reference, my 16ths are slightly faster.
Lingering the third beat cannot be identified as inégalité according to most of the treatises I have read, since inégalité asks for lingering the first note of two, or a series of them. But Baroque practices asks for lingering new harmony content where it appears. In this piece, most of the time the 3rd beat changes the harmony. just like the Chacone which I'll soon record: it is a dance whose 2nd beat should be lingered. In the first measures, this is okay, because Bach changes the harmony chords in the second beat. But after some variations, the harmonic changes occur in the 3rd, not the 2nd beat, so it's better to linger the 3rd instead of the 2nd one. There is also another point in Baroque performance, which does add a lot of "extra time" in those bars, as you said: people at that time were consedring musical language in comparison with a verbal one. Each phrase (sometimes they say it's a "paragrapha!") must be played as if it was SAID, in a speech. In slow and expressive pieces, there is plenty of "space" between the end of a phrase and the beginning of the new one. When I first tried to play this way, I left a slight little space. Then I played the Allemande of Bach's 4th French Suite to a Brazilian harpsichordist, he said I should "speak" more. He adds a "space" of a 16th note between each phrase. Considering the very slow tempo of an allemande (yes, Bach's allemandes should be played slow, not brisk as pianists usually do), it's a lot of "extra time".
So don't pay too close attention to the words and the definitions. I recommend not to isolate these things. And Robert Hill does play differently than I do, but as I said baroque practice allows lots of liberty to the performer. I myself struggles a lot in trying to incorporate these things in a natural and "baroque" way. Thanks to Tureck, Gould, Schiff, Perahia, Hewitt, and every other pianist who played Bach, I have NO REFERENCE, because none of them were concerned with an authentic performance. And now we have lots of prejudice against rubato in Bach keyboard music due to this tradition of pianists playing metronomically, with no evidences at all. And even when I talk to early music specialists, it still remains difficult to me, since they play in old instruments, and are usually skeptical of transcribing these baroque intentions into a modern piano. =\
Here it is what Rubsam says about playing Bach on a modern piano.
This is what encouraged me to study this kind of thing.