But without the source code, you can do neither of these things easily. The likelihood of the a) case occurring is lower by virtue of accessibility.
With the source code and the development history of the product in the open, we can at least see whether a misbehavior was intentional, or exploited logic questionable, and that there might have been a mistake in the original thinking. How easily would one have come to those conclusions if the source was not private? The response to the issue would have been very different, and I fail to see how if the source code wasn't open would it have somehow made it less likely a bug from the 80s wouldn't still be there.
I think it's even more important now than ever to keep source open to make sure bugs can be easily found; even if the benefits are marginal to discover, they are certainly better handled in the aftermath.