I tend to agree with your premise, though I wouldn't say the two are mutually exclusive. In fact, I imagine favoring authorship would more often trend to maintainable code than not, depending on what is being optimized for (writing less code, etc.)
I think in the case you described, instead of handling edge cases within that function, it might have been better to create an entirely new function to be called in those cases. You could then go a step further and identify shared logic, extract those and call them separately. At least that's what I tend to do when I find myself having to branch logic, especially established logic. Obviously I'm assuming a lot of the details here, and most likely what y'all ended up doing was the best right thing for your project/team.
I think in the case you described, instead of handling edge cases within that function, it might have been better to create an entirely new function to be called in those cases. You could then go a step further and identify shared logic, extract those and call them separately. At least that's what I tend to do when I find myself having to branch logic, especially established logic. Obviously I'm assuming a lot of the details here, and most likely what y'all ended up doing was the best right thing for your project/team.