IT IS possible to calculate the size of any non-porous particle from its surface area, for example, by assuming cubic shape (see eq 1.51 Gregg&Sing):
area=6/(density)/(edge length)
You will need very small particles to give an experimentally measure surface area. For example (same example G&S): if you measure a surface area of 2 m2/g of a solid with density = 3 g/cm3, then the size of the cube is (edge length)=1 micron.
The monolayer(s) or single molecule adsorption isotherm methods (Guggenheim-Anderson-Boer, Langmuir, etc) could give you an idea of the maximum endpoint surface since they allow surface area to be estimated at near the minimum molecular surface,usually the solid's maximum surface area. Volume is probably the easiest to determine, is also a function of density, dry or not; since it may possibly be derived from dry density and dry surface area. (assuming non-porosity in its most basic sense). Typically at that point an approximator to a sphere or some other reasonable representation can be made if only volume and surface area is needed. However size is kind of a "compacted" cumulative geometric function of the total surface (external, internal, voids and internal and external irregularities). Size is then represented by shape, with this factor combining the volume between internal and external particle surface with a representative shape factor. If you go with a seiving method you probably get the average projected to minimum cross-section or size at the seiving energy applied for the particlesize range, which is quite approximate but serviceable, though still may not be sufficiently accurate for what you may be doing.