Yesterday I ran across this fascinating link about synesthesia, onomatopoeia, and new hints about the origins of language. For those who don't know, synesthesia is when the senses intermix, and in this case, scientists have been looking at places where words and other sensations cross. It turns out that if you show experimental subjects the words "kiki" and "bouba" and ask them to assign one of two meanings to them - pointy or round - the distribution is not at all random. Most people will assign kiki to pointy, and bouba to round. The article draws some fascinating conclusions about language, and whether or not it can be considered entirely arbitrary.
I think these conclusions are very interesting for writers who make up words. Think about it - your gut feeling about whether the sounds of the words "feel" right for the meaning is not only legitimate, but likely to be shared across much of the world. There's something satisfying to me about the idea that on some basic level, the way I might assign random words to meanings by feel will be accepted and found natural by readers of other cultures. I suspect there are also some story idea opportunities available there - either about the historical origins of language or about synesthesia in unexpected contexts!