Many people at ELU seem to view the relationship between seekers and givers of knowledge as being essentially transactional in nature: the job of the questioner is to state his or her question simply and clearly, and the job of the answerer is to provide an answer that fulfills the questioner's requirements as nearly as possible, provided that the question itself meets certain minimum standards. To this end, they are quick to close questions that they feel are "bad," sometimes resorting to ill-fitting close reasons in order to do so, as though the questioner has failed to fulfill his or her part of the bargain by submitting a question that does not measure up to what they have decided are the minimum standards for questions. I believe that this view fundamentally misreads the dynamic underlying ELU and the rest of the Stack Exchange network, by omitting any consideration of some of the most important parties in the relationship: those who will come later.
When Jeff Atwood and Joel Spolsky founded Stack Overflow, their vision for the site was "Experts Exchange that doesn't suck." Experts Exchange is a programming Q&A site that had gained a reputation for underhanded tactics such as showing up in Google searches with what looked like freely viewable answers that were actually very difficult to access without paying for membership. Stack Overflow, Jeff and Joel decided, would do well by doing good: it would succeed by offering something of value to the programming community without charge, making it freely available not only to regular users but to casual Googlers as well.
This ethos has been engineered into the sites of the Stack Exchange network at a low level. In addition to the well-known "anyone can ask a question" feature, these sites are designed to surface content on search engines so that if someone goes to Google or Bing with a question about programming—or photography, video games, cooking, English, whatever—and a question like it has been answered somewhere on the network, they'll see that question and have easy access to the answer. Get a few of your questions answered that way, the theory goes, and you'll get to know Stack Exchange as a resource, and eventually you'll start participating yourself, and maybe you'll explore the rest of the network and find a site where you're one of the experts who can answer questions rather than just a supplicant who asks them. By any measure, that strategy has worked fabulously well, and is likely to continue doing so.
When we answer a question at ELU, therefore, we don't just provide an answer for the person who originally asked the question. We provide it for the hundreds or even thousands of people who might benefit from our answer in the future. And that obligates us to look for ways to provide the best answer we can, even if the way the question has been asked isn't entirely to our liking.
So suppose someone comes along and asks whether it's better to say "It is cold outside" or "It is frigid outside." We could respond by brusquely replying (in a comment, probably) that they both mean pretty much the same thing, and vote to close as General Reference, because get a dictionary, dummy. That's one way we could handle it. Alternatively, we could say that cold comes from the Old English ceald, and that frigid derives, ultimately, from the Latin frigidus, and is cognate with the French frigide. And we might observe that English seems to have so many more pairs of synonyms for basic concepts than most other European languages do. And we might note that whenever you come across a pair of synonyms for a basic concept, it's likely that one of the words will be of Anglo-Saxon origin and the other will derive from French or Latin, and that when that happens the Latinate word is invariably perceived as "fancier" than the Anglo-Saxon word. And that's because for several centuries after the Norman conquest, French was the language of the nobility in England while English was the language of the commoners, and even today the French-derived words in our language are generally perceived as being appropriate for a more formal register than their English equivalents, and isn't it interesting how the royal ambitions of William the Conqueror continue to affect the language we use everyday a thousand years later. And just like that, we've written a kickass answer that is likely to educate and entertain literally thousands of people over the next several years. And all because instead of just seeing a bad question, we chose to see it as a question that could—logically, reasonably, and in a way that gives the questioner the information she needs—lead to a good answer.
"But phenry, won't a permissive attitude toward subpar questions be bad for the site? Instead of teaching people how to write better questions, you're just enabling them." You know what? I really don't care. I'm not a teacher, and this is not my classroom. My role is not to punish people for writing questions that aren't up to my standards. It's to make the world a better place by increasing the amount of knowledge in it. When you vote to close a question because you just don't like the way it was asked, the person you're really punishing is the guy who comes along a year from now who could have benefitted from a really great answer to that question, but doesn't get it because the first guy didn't recite the proper incantation when asking it, and we told him—in essence—to go away because he sucked at asking questions. That shouldn't be what we do.
When someone asks a question that you just don't like, instead of reflexively voting to close it, make an effort to find the diamond in the rough. Instead of closing as General Reference, consider whether there are nuances or connotations worth exploring that wouldn't necessarily be clear from a terse dictionary definition. Instead of closing as "unclear what you're asking," consider addressing both of the possible interpretations of the question, and explaining the difference between the two. Instead of closing as "primarily opinion-based," consider exploring why different populations might disagree on the appropriate language to use in the given scenario, and how those populations differ from one another. And if all else fails—if you ask for clarification and the original questioner doesn't respond in a reasonable amount of time—remember that if you have enough reputation to vote on closing questions, you have enough reputation to edit them too, and consider taking that bad question and making it into a good one yourself. Yes, it's more work than reflexively voting to close, but no one ever said making the world a better place would be easy.