Obviously, I'm an armchair/hindsight 20/20 observer here. I have no clue whether it would have been an been an infeasible engineering feat to have more than 4 wheels. Also, I don't really have a clue what a reaction wheel is.
EDIT: Wikipedia to the rescue http://en.wikipedia.org/wiki/Reaction_wheel
In short, add more reaction wheels and end with a much more heavy machine, more complex, expensive, and more expensive to launch.
> Since the reaction wheel is a small fraction of the spacecraft's total mass...
The gist from the Wikipedia article suggests momentum wheels (used in Hubble) are very heavy, but reaction wheels, not so much, due to their use for precision movement.
Obviously I defer to someone that has more than a cursory Wikipedia articles breadth of knowledge on the subject.
Am I misinterpreting Wikipedia, or are the reaction wheels on Keppler more than a "small fraction of the mass"?
Hubble also masses eleven times that of Kepler. (11,000kg vs 1,000kg) I don't know the actual numbers, but attitude control systems possibly take up a larger percentage of spacecraft mass on Kepler than it does on Hubble.
I guess the tech might not have been available for use on Kepler when they were designing it (they seem to have started design work in the late '90s sometime), or it had some kind of limitation or drawback that they couldn't accept.