Charles Travis's attack on compositional, truth conditional semantics is built around a bunch of lively thought experiments, including cats dipped in puce dye ("Meaning's Role in Truth"), a guy named Sid who grunts when punched in the solar plexus (Unshadowed Thought), and a question about whether wearing a tie made of freshly cooked linguine would count as part of business attire (Ibid.). But one of Travis's examples has received more attention in the literature than any other. It involves one of his recurring characters, Pia, and the leaves of a Japanese maple tree. I'll quote part of the frequently cited passage:
A story. Pia’s Japanese maple is full of russet leaves. Believing that green is the color of leaves, she paints them. Returning, she reports, “That’s better. The leaves are green now”. She speaks truth. A botanist friend then phones, seeking green leaves for a study of green-leaf chemistry. “The leaves (on my tree) are green”, Pia says. “You can have those”. But now Pia speaks falsehood. ("Pragmatics")
There's a lot to say about what happens in that short paragraph, and a lot has been said about it. One thing to say about the example is that Pia's motivation for painting the leaves is odd. Who would want to paint leaves to make the world conform with the belief that leaves are green? In an unpublished paper that takes up the question of the painted leaves (which he has since modified in very interesting ways), Jason Bridges says of Pia's action and utterance, "When I imagine someone doing and saying this, I can’t help but envision her with a fixed, desperate smile".
Jason may be right about the oddity of Pia's actions as described in Travis's example. But leaves get painted for all sorts of reasons, not all of them strange. Stuck to the side of houses, they get painted inadvertently (more here and here); they get painted intentionally as a way of indicating that they are to be removed; and simply because it looks interesting.
The philosopher of language Stefano Predelli, who has a provocative paper that responds on behalf of compositional, truth conditional semantics to the example of the painted leaves, managed to find and get his picture taken next to some actual, vividly painted leaves. (His other pictures are worth seeing as well.)
Searches on Flickr also yielded pictures illustrating another one of Travis's examples, which involves ink that looks black in the bottle but which writes blue (Unshadowed Thought). It turns out that it is hard to tell what color ink will write simply by seeing it in the bottle. Almost all ink in the bottle looks black if the bottle is completely full.
Illustrations of more classic thought experiments surfaced as well, including a barn facade, a mule painted to look like a zebra, and a possible robot cat.