You can do all of those multi-touch manipulations on a Macintosh trackpad (zoom, pan, rotate, scale, etc). However, that trackpad would still be categorized as a form of a mouse -- correctly, in my opinion.
All of these gestures can be (and are, given that 3D modeling is historically done on desktop) handled with a standard mouse using a combination of the scroll wheel and modifier keys.
Whether it's your fingers or an on-screen pointer, it's the same paradigm in the sense of it being the same model of interaction. You move a pointer around and activate controls on the screen by touching them. I'm not knocking gestural controls, just saying if I had to classify them, I'd say they're an evolution of the mouse or touchpad rather than a whole new model.
And they aren't an evolution in all aspects, either. Multi-touch controls are easier for some things, harder for others. Fine-grain manipulation, for example selecting cells on a spreadsheet, or playing an FPS video game, are harder with touch controls than with a device like a mouse. They've also got a size constraint (the size of your fingertip) that makes many interfaces harder to use.
All of these gestures can be (and are, given that 3D modeling is historically done on desktop) handled with a standard mouse using a combination of the scroll wheel and modifier keys.