Oh, if you want a non-tautological criticism of deception, I've had one for years.
Learning to play the game.
In order to learn to play the game, you have to be able to correctly understand how your actions were the thing that caused particular good/bad effects. Your choice needs to be the overwhelmingly most important factor in the result, otherwise, your learning is inherently invalid. Now, one of the things you need to learn if playing D&D is that probability matters, and thus that reducing the chance that your actions fail by random chance is extremely important. Even outside of explicit mechanics, developing the ability to make wiser, better choices is very much one of the skills needed for becoming better at playing the "game" part of RPG.
Deception breaks this connection. Illusionism within combat makes it impossible for you to actually learn how your choices produce results in the world, because, by definition, they don't. The GM is the cause of everything. They just elect to do the rule-appropriate thing most of the time--but can and will elect to do whatever else, whenever they think it's warranted, and will make sure you can't know. Illusionism in non-mechanical portions has the same effect, just for different reasons: since whatever you choose, the same result happens, the vast majority of the conclusions you draw from the events you've experienced will be simply wrong. The player's perspective necessarily differs from the GM's, that is after all the point of illusionism, and thus that essential connection between choice and result is disrupted.
When you cannot even in principle validly conclude that your actions resulted in a given effect, you can't learn. Being able to make reasonably informed decisions--which includes sometimes making mistakes, or acting rashly, or failing to account for probability--is essential for there to be a game.
Or, if you like, answer a question: If players feel betrayed when they find out they've been deceived, even if they never actually tested the boundaries of their invisible rails, doesn't that still make it a betrayal even if they don't discover it?
If a group of friends plays poker, and one of them has cheated continuously for years but always carefully avoided getting caught, does that suddenly make it not wrong? If a man cheats on his wife for most of their marriage, but the wife never suspects a thing and dies without knowing, does that make his cheating morally okay? If a teacher intentionally teaches someone wrong, but the student never actually finds out that their "teacher" deceived them, does that mean the teacher didn't do something wrong? Etc.
"It's only wrong if you get caught" has never been a valid argument.
