I consider "n < 2" to be an expression evaluating to either true (1) or false (0). You can do basic if-else branching by considering true/false statements as conditions to if/elseif/else branching, but true/false cannot be used to evaluate only parts of a function by itself. That's what I intended to express, sorry if it's not evident.
What constitutes a "part of a function" in the article is a little confusing to me, for example: "I consider it as 'cheating' as we now have evaluated only parts of the function." Once you pass functions to a function, are those "part of the function"? If we say no, they're just arguments, it seems to sidestep the issue in an unsatisfying way; I could rewrite all my functions to take all the functions they call as arguments (dependency injection taken to the extreme) and suddenly all my functions have... no parts? That can't be right! On top of that, looking up an entry in a list (even one of functions), given an integer index, in a functional language, is arguably a series of conditional tests (is index 0? is index-1 0? is index-2 0? ...) Overall it's a fair article, but I think there have been a number of esolangs over the years that have tried to turn the idea of the conditional on its head, as well; more references might make it stronger.