A. Words point to things. Plato, in the Cratylus, holds this sort of view. The problem is that there are lots of words that don't seem to point to things at all -- verbs, adjectives, nouns like "tradition" or "consistency".
B. Some words point to things, other words give structure. This is a common mediaeval view. There were categorematic terms (point to things), and syncategorematic terms (words like "and", "the", and so forth), which enable us to structure the categorematic terms. Problem: this doesn't get around the problem brought up in the first view. It is hard to determine exactly how these syncategorematic terms act, and where exactly the meaning is located for them. Does it imply that the basic unit of meaning is the sentence (because these terms only make sense in sentences)? How do we talk about the meaning of words, then? Most people have thought about words as the basic unit of meaning, and to go below that level (to phonemes, for example) is to go below the level of meaning.
C. Words point to concepts, which point to things. So, maybe linguistic meaning flows through our concepts. The problem is that it is no easier to identify concepts than it is to find referents for certain words. What is the concept "tradition"? How do I know that you have the same sense that I do, even of something straightforward, like "table"? This suggests that meaning is a private affair, and that we can never share any part of it. Maybe its true, but the result is a kind of solipsism.
D. Linguistic meaning can only be deduced from the study of the situation in which speech is used. Behaviourist theory of meaning (Leonard Bloomfield) fits here. A stimulus (S) leads someone to speak (r), and a response (R) results from that speech (s):
S---------r........s----------R
Jill is hungry, sees an apple (S) and asks Jack to get it for her (r); this linguistic stimulus (s) leads to Jack getting the apple (R).
Problem: what if Jack doesn't get the apple? Does that mean that the speech has a different meaning? Actions in the world do not always match up with linguistic structures, unless we have some prior understanding of the meaning of the language.
E. Language has meaning only in relation to itself. This is the structuralist position (de Saussure), although it is not limited to that. Wittgenstein also holds this, as do many modern theorists. We do not find meaning by looking for something outside of language, but only by looking at the interconnections within language. This is the distinction between "sense" and "reference". Semantics is concerned with the sense of the words, rather than what they refer to in the world, to determine meaning.
F. There is no single determinate structure. This is post-structuralism, a position we will encounter later in the term. Basically, it suggests that words do not point at things, nor at structures, nor at other words, at least not reliably. The referents break down, reconfigure in ways beyond our power. Furthermore, it can be argued that the history of logic and word usage has served to marginalize certain groups; in effect, define them out of existence. The problem has been with our idea that meaning is connected to words, and that it stays put over time.