Many words are commonly used in multiple senses. For instance, the word “myth” in some contexts means “sacred story” and in other contexts means “popular misconception.” Both usages are solidly established and it is pointless to debate which is the correct one. Far from trying to arbitrate which is “correct,” therefore, a good dictionary will list both meanings, assigning a different number to each.
As regards the overwhelming majority of the words in your composition, you can and should expect that your reader’s sense of each word, in context, will agree with yours well enough so that neither of you need consult a dictionary. If you and your reader do not understand the same word in the same way, all a dictionary can really tell you is whose sense of the word is more commonplace or ordinary. So if you have reason to suspect that you and your reader might have different notions of a particular word’s meaning (which might both be in “the” dictionary, as successive numbered definitions), you should generally just explain in your own terms what you mean by it; and then your reader must accept that that is what the word means as used by you in this piece of writing. In that case, you may want to check with a dictionary just to make sure that your chosen and stipulated meaning is within the ballpark of established usage. If it is not, cast about for another word, one that is more commonly associated with the meaning you wish it to convey.
Your Encarta or World Book or Collier’s Encyclopedia is likewise almost exclusively a repository of common knowledge. What you can find in one, you can generally find in four others readily enough—five being the traditional “rule of thumb” for the “common knowledge” exception to the document-your-sources rule. In that case you need not and really should not cite. (You have to cite if you quote, but don’t quote from such things.) Again, it is usually a very good idea to consult these sources; it is just citing them, and worse yet quoting them, that marks you as an academic tyro or greenhorn.
Wikipedia is a special and somewhat controversial case. The Carleton College library has put up a particularly good page on the subject, with links to others. Wikipedia is simply bigger than any other encyclopedia, by whole orders of magnitude, and so of course it contains information that no other encyclopedia can match, let alone four others. Some material there is excellent, some is utterly abysmal in quality, and all of it is subject to change at any time. The changeability of this material is a good thing in its way, as bad information can be corrected and new information can be added; but it also can frustrate one of the main purposes of scholarly citation, which is to give the reader access to the very same documents that the writer has consulted and employed. (Actually, if you give a precise date in citing a Wikipedia article, your reader can access the exact same source, since Wikipedia archives past versions of its articles.) In any case, the worldwide community of “Wikipedians” who continually make and remake this work generally agree that an essential mark of quality in a Wikipedia article is the citing of sources for any information therein that is not common knowledge; and so the truly scholarly way is to rely only on the Wikipedia articles that do cite sources, and in their case to follow these citations, track these sources down, and finally to use and cite them rather than the Wikipedia article itself.