The ideas here are based on a small amount of data and are undoubtedly well-known in some form to many people (perhaps even anyone who has played a two-part invention!).  However they don’t show up in the usual discussions of Bach, and they certainly don’t seem to have influenced performances on recordings.

I’m been playing some pretty basic Bach—three-part inventions (in D and B flat) and a Well-Tempered Klavier I fugue (in B flat).  What is striking is that all of these pieces are in a very specific form you could call Bach’s sonata form.  I’ll say a little about the specifics in a minute, but the important thing is that it is obvious that Bach wants those pieces to be played in a way that reflects this clear dramatic structure.  No one would ever play Mozart or Beethoven without respecting the dramatic structure those composers have set up, frequently based on their own versions of sonata form.  However for some reason pretty much no one does that with Bach.  Is there a concern about “authentic performance”?  How can it be inauthentic if it is clearly intended?

Bach wrote for a harpsichord, so he didn’t have dynamic variation to use in performance. But that doesn’t mean he couldn’t build drama.  Tempo variation and ornaments were parts of his language, and he undoubtedly used them to the same ends.

Bach’s sonata form is in five sections:

1. Entry of the voices
2. Limited development based on the fugal subject as presented and transitioning to a strong cadence in a new key—as an end to the exposition.
3. Development section proper, with a clear departure point and free use of any pieces or rhythms of the fugal subject or other features of the exposition
4. Recapitulation in the subdominant as a clear contrast with development.  In all three pieces Bach goes to some pains to emphasize the recapitulation event.
5. Transition to the tonic as if anything a bigger event, followed by a short kind of victory lap

All three cited pieces do this exactly.

I think you can make a case more generally that people pay too much attention to the “horizontal” structure of Bach’s music—the intricacies of multi-voice writing—and not enough attention to the “vertical” structure—the musical events created by all the voices acting together.  That overemphasis on the “horizontal” leads to performances where the sole objective seems to be making sure the fugal subject is heard clearly regardless of whatever else is going on.  In the extreme such performances can amount to little more than the same thing played over and over again in different keys—because that’s just about all you can hear. It’s strange that this persists in an era fixated on historically correct performance, since it is patently inauthentic for keyboard!

Another side of the same thing is in the comment you sometimes hear about how amazing it is that that real emotional music emerges out of all that complicated polyphony.  I think that point of view is wrong.  Bach wanted to produce music, and the complicated polyphony was his language to produce it.  He manipulated that language the same way any other artist manipulates his medium of creation.  Bach’s ability to master polyphony is jaw-dropping, but that doesn’t mean he regarded it as the primary objective.

There’s a relevant quote from Beethoven: “I believe that Bach’s works for solo violin are perhaps the greatest example in any art form of a master’s ability to move with freedom and assurance, even in chains.” We spend too much time talking about chains.

P.S. My own ability to make a case for this is limited, but I’ve attached a recording of a familiar fugue from book 2 of the Well-Tempered Clavier. This is fugue #2 in C-minor. There is plenty of drama here, but the recapitulation is rather different than what was described above–it’s actually purely dramatic. For the record there is a difference of opinion about whether the piece should end in C major or minor. My edition used C major, and I agree with that because of the surprise C major chord that occurs earlier in the recapitulation.