Hacker News Reader

Comment by carlsborg | Hacker News Reader

How do wasps learn how to build those honeycomb structures that make up their nests. If a wasp is brought up in isolation and is able to do this, clearly its inherited.

View on HN

That's the inheritance of instinct rather than the inheritance of learning. If one generation of wasp colony "learned" that building their structures under blue canopy was good (researchers protected these structures) and red canopy was bad (researchers destroy these structures each time wasps make them ) and you take the offspring of this colony and raise them in isolation and they still build under blue canopy and avoid the red canopy, then you can perhaps say "learning" was inherited.

It's like how infants inherit the instinct to cry and suckle. Inheritance of learning would be if the mother (since childhood ) was taught to wink and she'd get more attention and food/milk and her infant winked at her instead of crying when the infant wanted attention.

This is a good response.

Building a little (but speculatively... I don't know much about wasps and their nest-building instincts): I often wonder what (if any) fractions of behavior that gets chalked up to "instinct" are actually the result of some biological/mechanical affordance, normal learning, and reasoning.

(I don't mean to deny instinct as such, but I guess I feel like the razor should exclude it until it's the only option left?)

To use the hexagonal example, the term "instinct" implies that the creature could do this thing in a variety of ways, but doesn't. But it wouldn't quite be "instinct" in the sense we mean it if the hexagonal structure is just a coincidental emergent byproduct of some detail of the species' visual system, physiology, etc.

Even in humans, which are relatively short on instincts as animals go, you can see cases where the "instinct" and the "behavior" are not the same thing. Humans have a suckling instinct, which allows us to drink at birth. But it's not an instinct that says "ok, here's how human reproduction works, and here's how milk is produced, and you need to drink from things that look like this and then when you do these things, your hunger will be slaked"; it's an instinct that says when you are tickled here in this manner, turn a bit and start some muscle contractions which result in sucking. They don't cry to say "hey, I need to suckle"... they cry because they feel bad. They don't know "understand" the bad feelings or know how to make them go away. But a combination of a few basic instincts and the environment combine over time to teach them how to eat for the first time.

Many other animals have much more complicated instincts, but even then you can often see the instinct starts as something very simple. A fawn "knows how to walk" at birth, but, look at video of a fawn at birth. I say it's a lot more like "a fawn knows to recognize it is falling and trigger an only-modestly complicated reaction to stab the closest leg out in that direction", a neural circuit so simple with a bit of effort you could almost assemble it by hand in some sense. A little bit of control circuitry on top gives it a small amount of mobility.

I think one of the reasons "nervous systems" are so evolutionarily successful is that they do give that mechanism to go from simple instinct to a much richer learned behavior, and it's obviously a combination of many things to get to that point. It's amazing how some of these things work on not necessarily all that many "instincts".

Furthermore, it has nothing to do with epigenetics. It’s good old genetics.

Reply on news.ycombinator.com