Can you name any project X you feel fits your hypothesis?
> The fact is that the vast majority of uses are not capable of contributing to open source code
At least they can pay someone capable to do it. Can you find someone capable of evolving Windows 95? No, you can't. Source-code availability is critical for this.
If your company, for instance, depends on PHP3 (or Perl 3, or Python 1 or something similar) you'll find readily available talent that can either port your application to a modern version of whatever it needs, or add whatever functionality you require to the ancient product you depend on.
> Can you name any project X you feel fits your hypothesis?
That seems an odd question to me. Pick a random organization that is using some FOSS package. I bet you that this organization is not going to be prepared to evolve that system, either due to skills or to other resource issues.
How many companies out there are going to be able to port Python or LibreOffice or what-have-you to a new platform? Most don't have the skillset. But even if they did, they won't have the resources to keep the developer working on that rather than the internal projects. Sure, there are exceptions (and I'm glad there are), but those are just that: exceptions.
Even if the theoretical possibility is there, practical considerations don't allow it.
Most companies do not have the resources to contribute to FOSS, and even fewer companies actually do it.
Nor can most organizations continue to evolve open source project X.
try running your .NET 1.0 installer on an ARM-based Windows 8 machine and tell me if it works.
Same is true for OpenOffice or whatever.
The fact is that the vast majority of uses are not capable of contributing to open source code, let alone a complete, regression-tested port to another platform. For practical purposes, there is precious little difference between FOSS and proprietary software for most users: either what is currently available works for them, or it doesn't.