I define instinct as the greatest possible being.
Just me and Bobby McGee, that’s all. 
I’m with fessie on this one.
We call it flocking instinct when sheep do it, but socialization when teenage girls do it. Same behavior with a little more metacognition.
I’m not sure you can dismiss the mating instinct as chemically depedant and therefore not an instinct. Aren’t our entire brains chemically dependant to some degree or other?
I’d like to take something vknowles and run with it.
“And the “without variation” part is actually wrong on its face. It will have some variation, as does any genetic trait. The more important (survival-wise), the less variation you will see in the population. But there will always be variation”
What if the concept “instinct” has much more variability than suggested in the OP. That is if a behavior that we consider instinct is merely softwired early on. For instance, is it possible that robins could learn to use different materials for their nests? Might they vary the construction slightly if their usual types of trees are unavailable? Don’t they recognize that the nests they want can be built in these new places somehow? These variations would not need to be inheritable. But wouldn’t such variations call into question the “hard wired” nature of the instinct?
I’m sugesting that instincts are closely related to reflexes. I’m not sure a good dividing line between the two is available. “Relatively complex” doesn’t seem to be sufficient. Also, I’d like to propose that humans may in fact have instincts, but that they are not as “hard wired” as those found in other species. That is we have strong desires to perform certain behaviors, and we might even “default” to these behaviors without guidance or serious thought. But we can, in the end, override them if we wish to.
Why yes they, are, and curiously, if you cut off someone’s head, they exhibit no behaviors of any kind. Epimetheus’ distinction between chemical and instinctual behavior is a crock.