I'm taking your use of the word attribute to mean the 'thing' assigned a percentage in your first example, the state of the boundary in your second, and the status of the rash in your third.
The term used in fuzzy control systems to describe such attributes is crisp, as opposed to fuzzy. It allows, but does not require, the possibility that the attribute could theoretically take on non-binary (or more generally, non-discrete) values. It also requires, however, that the only values actually assigned are discrete.
For example, in your first example, it is possible to contemplate a value being not 0% and not 100%, even if only for the purpose of excluding them from being legal values of the attribute, but the only legal values of the attribute are 0% and 100%. The attribute is said to be crisp, with 0% and 100% the permitted crisp values. The other two examples may be considered in a similar fashion.