But if you can’t find a proxy task after genuine effort, this is a bad sign about the project.
I would go one level meta of this and say it is ok if you can’t find a proxy task iff you can explain why it should be difficult to find a proxy task, IE the relevant north star is surrounded by epistemological difficulties, AND you can explain why you expect further investigation to be worthwhile despite a lack of proxy tasks. But this is a much higher bar and unfortunately involves more deductive reasoning which is harder to do correctly and more costly for other researchers to understand and validate.
I would go one level meta of this and say it is ok if you can’t find a proxy task iff you can explain why it should be difficult to find a proxy task, IE the relevant north star is surrounded by epistemological difficulties, AND you can explain why you expect further investigation to be worthwhile despite a lack of proxy tasks. But this is a much higher bar and unfortunately involves more deductive reasoning which is harder to do correctly and more costly for other researchers to understand and validate.
Seems reasonable. But yeah, it’s a high bar