Colors Appear To Have Almost-Universal Symbolic Associations — LessWrong