There is some confusion possible with the ambiguous word 'unknown'

True, Codd's rule 3 says "Null must always be interpreted as an unknown value"  but 'unknown' is plainly used in the sense of an unitialized value i.e. no data was entered.

'Unknown' in the above example can be a very real value in the sense that it explicitly tells us that the information is not available or cannot be determined; not the same thing at all. This kind of differentiation is very important in many real world data sets.

[oops rule 3 not 4]