In The Guardian, psychologist Tom "Mind Hacks" Stafford outlines five classic scientific studies that underpin much of today's thinking about how we learn things. One of Stafford's favorites is BF Skinner's 1930s claims that "with the right practice conditions – meaning that correct behaviour is appropriately rewarded – any task can be learned using simple associations." In 1995, Keio University researchers took Skinner's efforts further by training pigeons to discriminate between paintings by Monet and Picasso.
Like (Skinner), they believed that we underestimate the power of practice and reward in shaping behaviour. After just a few weeks' training, their pigeons could not only tell a Picasso from a Monet – indicated by pecks on a designated button – but could generalise their learning to discriminate cubist from impressionist works in general.
For a behaviourist, the moral is that even complex learning is supported by fundamental principles of association, practice and reward. It also shows that you can train a pigeon to tell a Renoir from a Matisse, but that doesn't mean it knows a lot about art.
"The science of learning: five classic studies" (The Guardian)
And here's a PDF of the 1995 paper: "Pigeons' Discrimination of Paintings by Monet and Picasso"
