It sounds like we are trying to address several problems at once. The question asks if doweling is advisable for a slab edge that is supported at a beam. A seismic area is noted.
For lateral or horizontal forces (i.e. seismic, sliding, etc.), Sundale makes a good point about a diaphragm structure. Sundale aptly notes vertical forces for further consideration.
Lateral forces will also include plastic shrinkage. With the slab placed over a rigid surface and anchored with dowels, cracking can occur to compensate for the shrinkage. This is why monolithic placements are preferred.
For vertical forces in non-expansive soils, compaction is important. If the soil settles, the slab becomes a suspended structure and must have sufficient strength and/or reinforcement to span from point-of-support to point-of-support.
As the beam width increases, and as the slab settlement increases, the slab rotation at the beam will also increase. Thus, prying can become an issue.
These preceding 2 paragraphs make a case for positive- and/or negative- moment reinforcement.
Any walls setting on the slab must also accomodate the rotation, slope, deflection, etc. of a settling slab as it spans from the edge beam.
Now for expansive soils. Vertical forces on slabs can be divided into two groups, edge lift and center lift.
If the soil shrinks, then the slab becomes a suspended structure. See the preceding paragraphs.
If the soil heaves, then the slab will be lifted. And, as it is doweled to the grade beam, it will lift the foundation. If the foundation is pier supported, the piers will be affected.
As a slab is heaved, positive and negative bending moments must be considered, possibly to include reinforcement. Likewise, shear results at the slab edge as the foundation beam is lifted (and possibly piers).
Conclusion: Following this long winded discussion, the question might arise, what's a girl to do?
Joints provide a means to isolate (small) components of a structure and to control the forces that act on the structure. Floating slabs can be designed to act as a diaphragm to resist lateral forces, yet left free to float as vertical forces act. Sundale hinted at this by forming a lip or edge around the slab to allow vertical expansion. A plastic hinge can also be created from slab to grade beam.
By doweling a slab to the edge beam, continuity is established which increases the area of influence for other forces on a given structure. If the designer can account for those forces and address them, then he has accomplished his due dilengence.
If he can't, then isolate until a good design is achieved.
Maybe this will summarize the considerations. (I am presently reviewing another engineer's work on this subject.)
Michael Mills
Tulsa, Ok.