[js8]

Ah, OK. But I am not sure what cause are we justifying here - the tenure? I guess the idea of tenure is predicated on beating somebody enough to have him go through graduate and postgraduate studies, so then only people who are likely to really want to work in the field will remain and they will continue working on their own.

It seems to me that there is tradeoff. We can look today at Peter Higgs and say, whoah, what a failure he was after 1964. But could that have been said in 1974? I am not sure. What if in 1975 he would come up with another breakthrough?

So the trade-off is in the timescale on which we judge the scientist's output. If you shorten the timescale, you decrease your accepted risk, and you can miss some rare wins (and I think that's where the Higgs example shines, because it is an example of such rare event). If you make the time scale longer, you accept greater risk of people turning badly. Idea of tenure advocates maximal practical timescale of such trust - one human lifespan, because with tenures arguably the wins are worth more than the accrued losses.