One thing that occurs to me is that if one wants to find the meaning of the word meaning, one first has to decide what meaning means.
Are we trying to find the true meaning of the word "meaning," as it is written down in some platonic universe somewhere? Or are we trying to decide what people usually mean when they saying "meaning?" Or are we trying to decide the most useful meaning we can ascribe to "meaning?"
Besides looking at the way people learn meaning, we can also consider the process of learning foreign languages. It seems that, on the whole, there is a correspondance between the words of two languages, especially if they are both Indo-European languages, say. People can come to an agreement that two words from the two languages mean the same thing. At least in "the same" context. This suggests that there is some concept of "meaning" that goes beyond the word qua word.
[ My god! To think that the day would ever come when I, even on usenet, would be writing the sort of thing that uses the word "qua." If I'm not careful people may start thinking I'm an intellectual! ]
One data point: In the process of learning French, I eventually decided that the only proper thing to do was to use an all French dictionary instead of a French-English dictionary. I don't know whether this helped me learn more effectively or not, but I discovered that what I was doing was looking up the French definition for a word I didn't know and then pondering the definition until I finally figured out what the English equivalent would be. It was only when I had the English equivalent that I had the _feeling_ of really knowing the word. And yet as I read a novel in French, I was very definitely not consciously translating into English.
One big disappointment for me was reading Simenon, who I always liked because his writing was so very "French." But what I discovered was that when I read him in French he seemed much less "French" to me than when I read him in English. Presumably this was because my comprehension in French was much less good and I was settling for only half-understanding many words and not really picking up the nuances of style.
Learning Russian was a different experience and I really liked learning to think in a new way, although even after four years I never came close to my proficiency in French.
One thing about nineteenth century Russian authors, especially Turgenev, is that they are really into trees and, to a lessor extent, flowers. And it would drive me bats reading Turgenev to try and remember all those different words for trees and flowers. But, after all, I'm the worst person in the world when it comes to trees and plants. Visitors to Hawaii ask me "What kind of tree is that?" and I just shrug. So after a while it occurred to me that I wouldn't be able to recognize those trees that Turgenev was talking about if I saw them anyway and what did it matter whether the goddamn word meant "ash" or "birch," as far as I was concerned it was just a tree, and that was as much as I needed to know. But that attitude didn't work. I found that I just wasn't comfortable until I'd looked the word up once again (for about the tenth time!) to see what the English equivalent was. And yet, again, I wasn't consciously translating while I was reading.
Another interesting process is that of reading mathematics, because almost every journal article in mathematics defines a few new terms. An example of such a definition might be "A ring is called _prime_ if whenever I and J are non-zero ideals, the product IJ is also non-zero." That's far from the most complicated example I could have chosen, but it seems like a formidible definition to have to remember. And yet when reading a mathematical paper I can most often remember the definitions the first time I read them and not have to leaf back to find them later. Certainly in listening to mathematical talks this is an essential skill, because a speaker will rarely redefine a term when he re-uses it later on in the talk unless (rarely) someone in the audience asks.
To me, the important question here -- and one which I find myself unable to answer -- is "In what form do I store this mathematical definition?" It seems that it is more than just a string of words, for when I read the theorems and lemmas using the new term I do not stop to consciously translate.
A similar thing happens, for me at least, with acronyms. Acronyms are words in themselves and when I, at any rate, encounter an acronym I do not consciously translate back into what it stands for. And yet when I first encounter a new acronym I find myself uncomfortable until I have been told what the letters stand for even though it may be clear what the "meaning" of the acronym is without knowing that.