My thinking is the flanges were pressed onto the arbor. The simple logic of 'how did they put it together in the first place with those flanges on the arbor' along with the practicalities of hogging a relatively small diameter arbor from a large diameter blank weigh in favor of the flanges being pressed on. At the Stanley Works, I am sure they had a hydraulic press and fixtures to press the flanges onto the arbors and seat them against a shoulder. I've had experience with disassembling press fit machine parts. Rather than rely entirely on brute force, which often is a bit more than was required to press the parts together, I'd suggest 'giving things a little help'. Apply heat to the flanges with a small torch flame. Something like a very small brazing tip on an oxyacetylene torch, or a 'Mapp' torch will work. Rotate the arbor as you apply the heat to the flanges, keeping the torch flame well back from the arbor. This will expand the flange slightly. You can keep a puller setup on the arbor and flange, maintaining a strain on the puller screw. As you add heat, takeup on the puller screw. After you have a good hard strain on the puller screw and have gotten the flange hot, it should break loose. Another old mechanic's trick, when this sort of setup does not break things loose, is to give the head of the puller screw a solid blow with a steel hammer. Not a love tap, but a real blow with a steel hammer having some weight to it will often cause things like this to break loose- often with a good, rewarding 'bang' when it happens.
Jim Christie's post reminds me of the old time saw mill arbors. The collars (or flanges) against which the saw blade was clamped, were heated and shrunk onto the arbors, seated against a shoulder. After the arbor was setup in its bearings, some saw filers or saw mill operators would true and 'cone' the face of the collar or flange in place. This was done using an old lathe compound slide clamped onto anything handy. Some sawmills, running large circular saws ('round saws') would require that the faces of the flanges on the arbor be faced at a slight angle. This put a 'cone' onto the sawblade when it was at rest. At speed, the circular saw would 'flatten out'.
For the OP's grinder, I am sure if care is taken, the flanges will seat squarely against the shoulders on the arbor/motor shaft. Cleaning up the mating surfaces with a fine oilstone to remove burrs once things are apart is a good idea. Heating the flanges (the kitchen oven works well for this if you can get away with it) will expand them so they may be reassembled on the arbor with a light driving rather than a pressing. A piece of steel pipe, ends faced square, will make a good tool for either driving or pressing the flanges back onto the arbor when reassembly time rolls around (sorry about that pun- could not resist it).
BTW: the coefficient of expansion of steel is 0.0000056 inches/inch degree F. (I carry this around in my head as I am an old time engineer, more like a dinosaur).
If the arbor is 5/8" diameter, a 0.625" diameter bore heated a total of 300 degrees F will expand 0.00105". 0.001" may not sound like much expansion, but on a small diameter bore with a press fit, it can make life a whole lot easier.
I take the view that a press fit results in a forced stretching of the metal in the female part. When I design or make up parts to go together with interference fits, I prefer to use shrink fits. On existing work, such as getting a large shaft coupling hub off a shaft 'out in the field', a bit of heat with a torch makes the job a lot easier. Press fits, in my experience, often require more force to break apart than was required to assemble the parts.