That's a nice way of putting it Roy. There's several of us using slightly different language to describe the same phenomenon.
To pitch it from another angle Wiz. I've seen a 1960s US article on a using a RAS which suggests using a saw blade set vertically in the plane of the blade disc, but at 45 deg to the direction of feed along the fence (when viewed from vertically above the plane of the table), and cutting mostly off the side of the teeth to mould a large coving/concave moulding of similar diameter to the sawblade!
I've also seen an article by a respectable contributor to one of the posh mags on how to make a sled clearly not understand that a saw can cut off the side of its teeth like this.
To try and explain the issue. When you cut on a table saw the blade has only a very moderate guiding effect. This because as above the teeth are happy within reason to cut in pretty much any direction. e.g. you can feed (carefully) a workpiece into the blade at quite a sharp angle (e.g. 45 deg on a cross cut fence - but not recommended unless you are feeling brave) to the plane of the disc and it will still cut. The resulting kerf will be a bit weird - basically part of the ellipse the blade appears to form when viewed along the line in which the workpiece is being fed.
Another way of experiencing this. I've been know to (very naughtily - i wouldn't do it now) feed a piece freehand into the blade without using a fence to rip/nibble out a short slot in the end of a piece - working to a marked out line. You would find if you were foolish enough to do this that you have to control the position of the piece sideways as well as the blade is quite as happy to cut sideways as it is in the direction of the slot.
To get back to the mitre slots. They determine the line in which the piece is fed to the blade. So when the piece contacts the blade it will still as above cut regardless of whether or not it's aligned parallel to the slots.
If the blade is out of parallel to the slots, then the problem is that you will get an elliptical section kerf, also as above. This kerf will however (because the blade has no real preference in terms of which direction it cuts) still be straight and parallel to the slots/direction of feed.
The blade does not need to be set up parallel to the line of the slots to get a straight cut. It does if you want to simultaneously get a vertical sided parallel kerf, and to avoid the trailing side of the blade roughing up one cut face. (that zinging noise)
My guess is that the misunderstanding gets overlooked if you are pernickety about setting the blade parallel to the mitre slots or the direction of movement of your sliding table - because if in that case you reference off the plane of the blade disc to square a 90 deg cross cut it will cut square. i.e. it cuts square, but only because in that specific situation by aligning off the blade you end up square to the slots. Right result, wrong reason.
If on the other hand the alignment of your blade is a bit sloppy relative to the slots/direction sliding table travel (as was the case initially with my Robland, i wasn't experienced enough to check it out) you'll be left wondering how it is that the damn thing is not cutting square or cleanly despite your fence being aligned accurately at 90 deg to the blade.
The reason is of course the above.....
This probably just muddies the water even further, it really needs a drawing to be clear. You're probably right about needing to check the alignment of the saw.
ian