Actually I think this is a good litmus test for whether an individual is encyclopaedically notable. If there are no sources for basic biographical data other than the individual themselves, in other words if there has never been a reputably published biography or profile, then I don't believe we can have an article.
I would see no problem with an article about a scientist which was entirely about his professional contributions, and didn't so much as mention a date of birth, middle name or anything vaguely personal. The article should be primarily about their work, and that "biographical" information is purely incidental.
I suspect that for very early scientists (eg, 1700s or earlier) we may be in this situation anyway.
Steve