>My problem here is google's attempt to correlate incognito users to their non-incognito history.

To a web server incognito mode isn't a thing. It's a client only thing. You don't know if a user is using incognito mode, or if they just cleared their cookies / cache. There's no way to know the user's intent.

And this is by design (even though it's actually detectable -- try watching Netflix or Amazon, or any similar DRMed content, in incognito mode), because telling the server "hey, I'm in incognito mode" is antithetical to the goal of seeming to blend in.

But I still see a problem with Google's control of both sides of the connection, and with fingerprinting in general.

In your view should Google not allow people to log in to Gmail while in incognito mode? How can someone remain untracked by Google while in incognito mode but also interact with personalized Google services, like email?

By logging into one's account. Surely you see the distinction between deliberately availing oneself of a service and bring tracked on entirely separate websites without being informed, much less consenting.

You're making an argument against tracking in general. That's fine, but it's not what we're discussing. What we're discussing is if sites should treat traffic from browsers in incognito mode differently than traffic from browsers not in incognito mode. Do you think they should? I would argue that sites shouldn't even know whether or not their users are in incognito mode.