I'm not convinced this is an English problem, or even a human problem, though. The problem is the low resolution being used; no matter how fine your stated opinion, if it is being viewed through a low resolution sensing device, it'll get aliased to one of the things the sensing device can actually perceive. (This is also the correct definition of "bias", though that would take a bit longer to describe.) It takes cognitive energy to increase resolution, so a lot of times we'll use "love/hate" as the only two options pretty casually, because it's easy and we don't have an infinite supply of cognitive energy.
(A number of terms in that paragraph are sort of ill-defined but it would take a lot of time to really nail them down.)
For counter examples: if I were to say "I don't drive fast", you would likely not interpret that as "I drive slowly"; if I say "I'm not rich", you wouldn't hear "I'm poor".
Yet in all these classes, we must be using some resolution. So your explanation only begs the question of "why, then, do some kinds of statements get treated to higher resolution than others?".
This seems to be embedded in English, at least. If I say "I don't like chocolate", that's universally interpreted as "I dislike chocolate". But I only meant what I said: I do not possess an active enjoyment of chocolate. The listener just assumes "if yer not with us, yer agin us".
It seems that taking sides is baked pretty deeply into our thought process.