Are we moomin, or are we dancer?
I mainly take issue with the word "error," as I believe it automatically inspires contempt and incuriosity.
Just because we can't discern the adaptive purpose behind something offhand doesn't make it wrong, bad, or pointless.
My friend also pointed out that during periods of stability, systems will lean towards homogeneity, but during periods of upheaval or change, we'll begin to see a turn towards heterogeneity, as if nature is throwing things at the wall to see what "sticks."
A trait that doesn't seem immediately adaptive for an individual in isolation may actually prove useful or adaptive for the larger group, and/or may provide that individual with a unique edge during a period of great change, stress, or scarcity.
In this vein, even fetishes and kinks may serve an adaptive purpose that may not be immediately apparent, and we should remain open to that possibility.
I'd additionally argue that fetishes aren't purely or solely about sex, as I've chatted with many people who've insisted that sex is only a secondary or tertiary motivation for indulging in a given kink or participating in a kink community. With a lot of kinks, sexual arousal seems to become inexorably intertwined with other emotions like love, fear, embarrassment, and empowerment.
Well, it's language from the 1990's, so it grates on modern ears.
I'm not sure whether finding a synonym for it that would be better received by non-scientists would be helpful- because primary literature isn't intended to be read by non-scientists. You could change error to 'abnormality', or 'deviation', but these terms would also be likely to cause offence, in spite of being more neutral in their literal meaning.
For researchers 'error' works well enough- because it outlines the problem 'How do we explain the origin of sexual interests for things that are seemingly irrelevant to or even present an obstacle to reproduction?' and 'Can this be explained as part of the mechanism that constructs 'normal' human libido?'
I also think that we should try to avoid seeking direct adaptive explanations for sexual fetishes or gender identities. I sense that some people perceive an adaptive explanation of those traits as a form of validation, as if they are necessary in order to underpin the worth of individuals who have those traits- and I think we should actually attack that idea because it is irrelevant to morality, rather than seek out adaptive explanations that we may end up having emotional difficulty abandoning if evidence emerges to show they are not true.
There are obvious adaptive explanations for male sexual interest in girls who are underage but fertile, for example, while there isn't an overall scientific consensus on an adaptive explanation for male homosexuality...and many of the existing adaptive explanations for why some of us are gay are perceived as offensive (for example the 'sneaky male' hypothesis).
I am not sure I agree with your friend's idea that 'instability creates heterogeneity in animals' is generally true, more clarity and specificity about what this means might be necessary. The way you phrased it sounds teleological to me- I am cautious of explanations of natural phenomena that require us to anthropomorphise nature, because appealing to an anthropomorphic analogy sometimes indicates there isn't actually a convincing mechanistic explanation.