What justification is there for preferring R to Matlab, other than cost?
My impression is that the primary problem domains for these two are pretty different.
Matlab is pretty heavily targeted at folks doing numerical analysis [1], e.g. solving differential equations related to physics/engineering/economics, and has its origins as a wrapper around existing fortran implementations for numerical linear algebra. Engineers, especially, use Matlab all over the place.
The R project explicitly is interested with statistics [2], and as someone earlier pointed out, make is pretty easy to do analysis on a dataset. Though, I have pretty limited experience with R, so I could be mistaken.
[1] Cleve Moler. Numerical Computing with MATLAB. http://www.mathworks.com/moler/chapters.html This is the guy who wrote the first Matlab implementations, and founded Mathworks.
[2] The R Project for Statistical Computing. http://www.r-project.org/