merge on non-identical names

Greetings,

I need to conduct a merge on two databases containing information on
organizations, but the organization names are often non-identical and there
is no common unique identifier.  Does anyone know a good way to calculate a
similarity measure on two names, or even better is there a natural language
matching function in an R package?  I did some searches on this but must not
know the right keywords to search.

As an example, here are some possible non-identical names:

Oxfam,  Oxfam USA
American Services, Americam Services   -  (just mis-spelled)
Global Alliance for Action, Global Alliance for the Environment   -  (a
non-match)

Any suggestions are welcome!