The thing is that I think 4e expects you to go in the other direction. You start with the decision to be a Defender. What kind of Defender?
And you've already lost anyone who is playing D&D more for the character performance experience than for a tactical combat experience. You don't start with the decision to use the Kuleshov effect. You start with the needs of the narrative and the emotional content.
Having a role communicates both what you are supposed to be doing in a fight, and lets the designers make sure you have access to the tools to do that.
For a lot of people who play D&D, it's an incredibly narrow and specific and largely irrelevant aspect of playing D&D.
I do think I should reiterate that this doesn't mean it's not important for the people it's important for, or that D&D should ignore this or pretend that people don't want this. But tying it to class like 4e did wasn't good design for D&D, because the design told you that this was VERY IMPORTANT, a key part of your decision to play a particular class, and for a lot of people, it really just wasn't.
And part of why 4e fell flat for a lot of players is that mismatch between what the rules suggested were important and what was actually important to those players. It wasn't just about the setting not meeting expectations. It was also about the rules not meeting expectations.
that isn't all that 4e Fighters can do
Yeah, I mean, we can all have hobbies outside of our job, but 4e was pretty clear that Defender was a Fighter's
job. And if I played my Dex fighter as a bow-using damage dealer who ignored marking, I'd be doing a bad job at doing the Ranger's job, and a bad job at doing my job as the party's defender.
I'm ultimately making the case that combat roles are mostly irrelevant to class for a lot of D&D players. The combat role is part of the medium. It's set dressing. It's paint. It's celluloid. It's story structure. It's meant to fade into play. It's not fun for those players for it to be something you need to pay attention to when selecting a class.
So if as a D&D designer you tie your classes to combat roles, you're going to wind up having a lot of players see that as a problem. As restrictive. As "like an MMO." As purely combat-focused. As artificial. As solving a problem they didn't have. As meaningless to their reasons for picking a class. If 6e has hard-coded class roles, it will have the same problem, because the problem results from the design.
5e's iteration (roles are things you can choose mechanics for if roles are important to you) isn't perfect, but it's better for more people than 4e's iteration (roles are things classes do), which itself was an improvement on 3e's iteration (roles aren't something the rules really need to worry about at all). The lesson that 5e has appeared to learn is that people don't need to care about combat role to choose a class. Combat role can be part of it, but it really doesn't have to be. This seems like a correct assessment from where I'm sitting.