I see learning as increasing the brain-based skill set of the individual, but not the DNA skill set. Learned skills therefore cannot be passed through DNA. I see learning as having the potential to improve the evolutionary fitness of an individual, i.e. his chances at surviving and passing on his DNA.
I see learning as requiring special innate skills such as imitating, experimenting, and the capability of adding to (editing) existing skills. Imitating and experimenting are physical actions, like the actions expressing most other innate skills. Editing skills is an internal process like applying skills.
I see imitation as a skill that can produce action that copies a visually-perceived novel action so that the individual performs an action very similar to the perceived action. it can be learned without the risk of experimenting.
I see editing as a capability that can abstract and memorize the copied action such that it can be performed again and again in the future. I see the memory of the imitated action as similar to an innate skill, except that it is not encoded in the DNA and this cannot be passed through procreation to the next generation.
I see experimenting as a skill that can generate a novel action. Experimenting takes a skill-based action and changes some element of that action. Most experiments fail and could even be deadly, such as trying a new food. Some are successful and should be kept, i.e. converted from a once-only action into a repeatable skill.
Again, I see editing as the capability that can abstract and memorize the novel action into a novel skill and add this skill to the information content of the brain. The novel skill can then lead to repeating the novel action without taking the risk of repeated experimentation.
In essence, experimenting plays the role of meiosis, recombination, and mutation (for DNA), by constantly generating new actions that, if beneficial, generate new skills.