This is as absurd as prohibiting hyperlinks because the site being pointed to might have copyright violations

Not to mention completely non-enforceable. You're forbidding people to say "David et al. [3] did a work in this area"

Ok, that makes sense. But is it an absurd proposition to say that people who release non-open articles should not be allowed to directly build on the work of those who did it for free?

So how about if part of the agreement of accessing an open article was that any published work citing it would have to be open too. is that absurd in principle? and is it really non-enforceable?

An offshoot of this would be that anyone who goes on to cite something in a non-open article, could academically be accused of not reading the article their citing, and therefor mis-attributing information. This could lessen there creditability or even be grounds for a specific type of plagiarism.

There's no law which would allow somebody to enforce such a thing. Citing somebody's work doesn't have a copyright effect.

Couldn't it form part of the access agreement to the free article?