Allow me to rephrase the question, just to be sure Iβm understanding it correctly.
Assume that we have a project where some of the artifacts are either incomplete, missing or just plain wrong. How could we approach our testing in order to ensure that we are getting the most value out of our time?
Luckily for us, the artifacts, such as requirements documentation, are a very small part of the testerβs arsenal. The reference materials, can be missed, since we can still use inference, conference, and experience as oracles in order to understand our product.
This is a HUGE subject, so rather than explore it more deeply here, I will refer you to a set of blog posts which explain it much better than I ever could.