Basically, I'll elaborate on what Laurie said.

A polygon is made up of 3 or 4 edges - the face is the surface between those edges, the vertice is the point that connects the edges together to make a triangle, or a quad. 

The face and the edge cannot bend. They can only be manipulated based on the positioning of the vertices that join them together.

All a morph does is translate the vertice(s) in space from one X, Y and/or Z position to a new X, Y and/or Z position. The edges do not bend. They can be lengthened or shortened, but they will always be a straight line between two vertices.

If you want a cube to become a sphere, you have to add more edges and vertices in order to have more surface space (faces) to shape the sphere. Does that make sense?

Following that pattern, you will always always always have more faces in a figure's face than in any other area of the body. Typically, the face will generally make up at least 25% of a models total poly count. Usually more like 30%, maybe even more, depending on how animated the face needs to be.

Joints need more geometry (multiple polys) in order to obtain smoother bends. Mouths and eyes need more geometry in order to open and close and create facial expressions. 

This is basically how polygon models work.

SubD's are somewhat different. It's a different algorythm that creates SubD's (catmull-clark) than polygons, but basically speaking, it's still polygons that make up subDs. The difference is, that the subD mesh is divided dynamically, while polygons are divided manually. If you think of a cube being stretched into a shape, as a subD surface, it will automatically divide itself equally as the cube is stretched and pulled, while a polygon cube has to be divided "by hand". The geometry lines you see on a subD surface are generally considered a control cage. Those are the lines you see and manipulate, but in between them are tons of lines (edges and verts, if you will) that you don't see. They're still there, you just don't see them. (I'm sure a nerd will come along and crit this paragraph - i'm writing in laymens terms)

D|S uses subDs. Poser uses polygons. 

I consider subDs as being a hybrid between NURBS and Polygons

Some argue that subDs are better than polygons. Others say polygons are better than subDs. The benefit of polys over subDs is that polygon models can be fully optimized. If your poly budget is strict, you can put every poly exactly where you need it down to the last vert on a pinky toe and leave them out where you dont. You don't have that option with subDs as their division is always equalized between edges.

Personally I think that its really just a matter of personal preference. Some people love working with NURBS. I can't stand them, i think I'd rather pull my teeth out with pliars and gargle with rubbing alcohol after each one than to have to build a human figure using nurbs. However, nurbs, just as subDs, have their purpose and can be much faster in certain situations, and provide a much smoother surface. 

Most people model using polygons, and when everything is said and done, when you render your scene, everything gets converted to polygons, because that's what render engines understand. Polygon models export as polygon models. SubD's and NURBS get converted to polygons first and then exported as polygon models. 

Feel free to ask anything that's confusing about this. 

 

~Shane