here's a version of the semiotic triangle from seb's last post:
it was made famous by ogden & richards in a 1923 book known as 'the meaning of meaning.' it's important to note that the line along the bottom is dashed and labeled 'an imputed relation,' since symbols can only refer to objects (referents, that which is referred to) through a concept or thought.
here's how ferdinand de saussure, writing slightly earlier, saw the relation between thought and language:
where A is the medium of thought (nebulous, undifferentiated), and B is the medium of language (undifferentiated sound material). the vertical lines are the arbitrary connections between the two media.
roland barthes, an interesting theorist who bridges the structuralist/post-structuralist gap in the mid-20th century, used this model: i personally like the more specific labels of the relations between the three vertices of the triangle. later, he turned to this image to explain the structure of the 'myth': here, the sign is a more complicated amalgam and the whole process of sign formation/understanding is recursive. (derrida, in particular, is known for his claims about the infinite play of signs and meanings).
another more recent semiotic triangle is extended to include the term 'definition'. this one is from 1997 and is credited to Suonnuuti, who i've never heard of before. i guess the definition is added based on other conceptual systems or something like that.
here's what i guess could be called a post-structuralist semiotic schema. it's greimas' semiotic square. it's not actually about the structure of a symbol, but the structure of a particular opposition within a text. remember, post-structuralism isn't a philosophy per se, it's just a way of handling texts and their meanings. in this example, the 'deconstructed' binary is 'beautiful/ugly', and the corners represent (more or less) different positions that characters in a story can occupy or represent.
and here's an epistemological model of the semiotic triangle that i think tries to explain what an artificial intelligence would have to have in order to understand signs: pretty, no? i don't really understand it, but it has the most lines, squiggles, and circles so far!
i guess the point is that these kinds of models are very vague. what lines and circles mean is pretty debatable and not even very empirically useful. also, the value of using a little picture to talk about symbols seems dubious at best to me. these kinds of diagrams only make sense when they're responding to one another, or when one philosopher says 'but what if it did work like this...' and uses an illustration to demonstrate their difference from some other tradition. someone should probably launch a study of these kinds of diagrams as a form in the semiotic/linguistic/cognitive fields. maybe someone already has...