Knowing part of the password gives you a good start at building a list of possible passwords, but that could be a massive list depending on the max length and charset. Say it's a max of 50 chars and you have 14, that leaves 36 chars worth of entropy in the unknown range.
By knowing part of the hash you can invalidate a lot of those, but without the whole hash you need to search the whole unknown range looking for matches to your part of a hash. The time this takes is considerable, even with GPU help (a tool like oclHashCat) you are talking about decades of work since the part of the hash you know doesn't speed it up; you need to run the hash and then compare it to your fragment. You can't stop as soon as you find one that matches, either, because you have no way of knowing that its the only match and therefore must exhaust the entire space.
If you do exhaust the space with only one match, you have found the password. If you have multiple results you have not found the password. I think the answer to "is it possible still to determine the password" is a solid "maybe" since you will only know if there is 1 possible password until after you check.