Anyway, it would not be fair to automatically invalidate papers citing retracted ones. Including:
- reproduction attempts
- some minor citation in related works
- it's usually not obvious that a paper is manipulated or even wrong without any bad faith involved
Now, I wish we could update papers with disclaimers and notes, but again, we are dealing with good old PDFs that are never going to be updated…
Good luck even noticing that a paper was retracted.
I wish we had better formats and publication processes.
None of those require the paper you're citing to be free from fraud: seminal works with fraudulent or inaccurate results may still have preparation methods that are applicable to novel, non-fraudulent, experimental techniques and equations are rarely at issue with data manipulation or other fraud.
There's still space to be aware of retracted papers in citations as building directly on those results and managing to show an improvement on doctored data would be suspicious but I'm pretty sure that is the minority of citations in many fields.
Nothing would ever get done if this was the bar for citation. Real advancements take place on the back of fraudulent results in widely cited papers. Actually, even within papers that have fraudulent results, advancements happen.
The motivation for such fraud is because work generally wont get published without positive results, even if the work was excellent. The fraud doesn't take the place of hard/great work. I'm not defending the practice.