R help - Jun 2007 - polr: attempt to find suitable starting values failed

Hi,

I got a problem with polr function a bit. I have this error:

attempt to find suitable starting values failed

I wonder what would be an appropriate way to solve this problem? I can try to
specify some random starting value but I am afraid that it will cause the MLE to
stuck at some local maxima. Is there anyway to cope with this in a more elegant
way? Any help would be really appreciated. Thank you.

- adschai