Recently, I answered one of the lateral-thinking puzzles here. I still believe that my answer was the best of the lot, although there were quite a few creative answers as well.I don't see how it is too broad.
The Lateral-thinking tag allows for creative answers, but till what extent? Most of the lateral-thinking puzzles are closed as too broad. We need to provide a benchmark on what should constitute a lateral-thinking puzzle or we should remove the tag once and for all.
There is a real problem, but removing the tag is unlikely to help it
Sid observes that most questions tagged lateral-thinking are closed as too broad. Beastly Gerbil agrees: "Most get closed". This is true only if "most" means 22%, but it's certainly true that a lot of lateral-thinking puzzles are really bad.
But will killing the lateral-thinking tag actually make a difference to that? I don't think so. I think most makers of bad lateral-thinking puzzles are (1) newcomers to Puzzling or (2) longer-standing Puzzling users who don't learn easily from their mistakes. See for yourself: take the list of lateral-thinking puzzles, look at the ones with poor scores, and see who's posting them. It's almost all people with very low rep -- new users and users who have never managed to post puzzles that the community likes much.
But those are exactly the people who will not refrain from posting this sort of puzzle merely because there isn't a tag for it, and who are not likely to be posting this sort of puzzle because they think the description of lateral-thinking encourages it. I bet at least 90% of bad lateral-thinking puzzles are from people who have never read the tag description, or at most have seen it in passing when they typed in the tag name having already written the puzzle.
Let's take a look. I'll go through the 50 latest lateral-thinking puzzles, identify ones that (1) have been closed or (2) have negative net score, and see what there is to be seen.
Lateral Creative Question -- poster has 19 rep and no other questions; lateral-thinking tag was added by someone else after it was posted.
Costing calculation -- poster has quite a bit of rep, but has posted many many questions with negative scores; lateral-thinking tag was added by someone else after it was posted.
Figure out the missing card -- poster has posted 7 questions, all of which have negative scores and four of which have actually been deleted; in this case lateral-thinking was put there by the questioner.
Five Apples from Moscow Puzzles -- poster has 13 rep and two questions, both on negative score; lateral-thinking tag was added by someone else after the question was posted.
Reading in the dark, faster than light -- poster has quite a bit of rep (though lots of questions with negative score); lateral-thinking tag was added by someone else after the question was posted.
Tell me the color of the wall behind you -- poster has 140 rep, 5 questions (3 negative, one 0 and deleted, one positive); lateral-thinking tag is original.
The Peculiar and possibly impossible seating arrangement -- poster has 1 rep, two questions both on negative scores; lateral-thinking tag is original; this is a bad puzzle but not for the usual too-broad reason but because it's easy and boring; its badness has nothing to do with its lateral-thinking-ness.
The 100 Point Maze -- poster has 4 rep and no other questions, lateral-thinking tag was added by someone else.
How does the hit-man assassinate his target using only two questions? -- poster has 1 rep, two questions both on negative scores (same one as above).
Story of Fuzzy Nazi -- poster has 240 rep and quite a few questions, lateral-thinking tag is original, question is bad not for the usual lateral-thinking reasons but just because it's too easy.
25 horses, find the 3 fastest ones -- poster has 23 rep and only one other question, lateral-thinking tag was added by someone else.
OK, I'm getting bored and will stop here. But the pattern is very clear, and is just what I claimed above: the people posting silly overbroad lateral-thinking questions are almost all either new to the site or incorrigible, and they generally don't even know the lateral-thinking tag exists. Removing the tag will not make their bad questions go away or improve.
The tag seems reasonable to me
There are decent lateral-thinking puzzles. There are even some on Puzzling. The tag gives a reasonable idea of their content. My own pet criterion ("can I imagine a reasonable person using it in a search?") it passes easily.
There's nothing wrong with the tag. The problem is with the puzzles, and the people making the problematic puzzles mostly don't even know the tag exists. Removing the tag will not help.
This one line from Gareth's answer (which you definitely should go read) is, I think, the heart of the matter, and motivated me to write this followup to it.
Tags are typically used in two ways:
The lateral-thinking tag can, and should, be useful for both of those.
The whole premise of a "lateral-thinking" puzzle is that it's a "puzzle solved through an indirect and creative approach, using reasoning that is not immediately obvious and involving ideas that may not be obtainable only using logic." In other words, think outside the box, because the clues are all there, but it's going to take an unobvious or unexpected way of thinking about the problem to connect them together to get the answer. These kinds of puzzles certainly can be very interesting and clever—and if I was looking for an example of a puzzle that requires creative, outside-the-box thinking, searching for them via the tag is probably the only practical way.
A puzzle setter should also be able to indicate to the solvers that the puzzle requires some lateral thinking to solve. While it's true that tags ideally would be only metadata used to categorize puzzles, they often serve a dual purpose by also being part of the puzzle, at least in that they act as a hint about how the puzzle works. The "lateral-thinking" tag is a very useful signpost to solvers that they need to think outside the box, without the setter having to state it explicitly in the puzzle body. And as I pointed out elsewhere, we often even leverage the lack of this tag to purge answers that try to use (wildly) lateral approaches to "solve" a problem—in other words, by default we assume lateral thinking is not intended, and the tag is then actually necessary to indicate when it is.
So let's not punish the tag just because it attracts bad puzzles.
The tag wiki actually points out the (minimal) requirement for a lateral-thinking puzzle to be at all successful:
When writing a lateral thinking puzzle (this goes for all puzzles on the site, but especially for this genre), one must take care to compose it in such a way that there is a unique correct answer, otherwise the question is likely to be closed as "too broad".
$\hspace{60ex}$(emphasis in original)
... and therein lies the problem. As Gareth pointed out, lateral thinking puzzles (for some reason) seem to attract puzzle setters whose track record here suggests are, well, inexperienced at puzzle setting. The lateral-thinking tag is quite literally—by design, and of necessity—an invitation to the solver to try all kinds of things and see what might stick. The problem is that people can be quite inventive and find all kinds of stuff that, arguably, solves the puzzle. Here far more than anywhere else, it is vital that the puzzle be composed to preclude all other reasonable* answers but the one intended.... and if it seems that the tag tends to attract mainly inexperienced puzzle setters, it's no wonder their puzzles fail to do just that. They end up, as the tag wiki warns up front, being "too broad" and should be closed.
But it's not like good lateral-thinking puzzles don't, let alone can't, exist. The Sheikh dies is a good example of a lateral thinking puzzle, and its accepted answer is a good example of applied lateral thinking. Make all the statements true is also a well constructed, if simple, lateral thinking puzzle—when you see the right answer you know it is the right answer.
Unfortunately, good examples are somewhat hard to find. Some of the highest rated puzzles in the tag are probably not lateral-thinking at all: How can 64 = 65? is about applying the math rigorously, and not at all about lateral thinking; Find the letters that complete these five patterns and Which country is INDIA in? are more pattern recognition/steganography; and Doorknob's "Wrap-up" answer to This is it. This is the one. Find your wife admits that "I didn't know [steganography] existed when I made this."
And the very highest rated puzzle in the tag, A double-agent with a conundrum, has an accepted answer that does not even work, so while it is lateral-thinking, it's not a particularly satisfying puzzle.
Maybe what we need are some good examples of what works, and what totally doesn't—and why—so we have a resource to which we can refer people who want to try their hand at a lateral-thinking puzzle, to help them avoid the common pitfalls, narrow their puzzle's scope, and have a better chance at posting something that works: the "Lateral-Thinking Puzzles: What (Not) To Do" guide.
Any volunteers?
* Reasonableness of answers is actually just a big a problem as the construction of the puzzle, if not more so, and is probably why people look on this tag topic dubiously. "Lateral thinking" should not be free license for the solver to arbitrarily relocate the puzzle to another planet, or decide the characters in the puzzle scenario are dead, or whatever. In a well constructed lateral-thinking puzzle, the right answer is one that uses what the puzzle gave you in a creative and unusual way, and does not—and probably must not—invent facts out of thin air to make a "solution" work. As I expressed it on a puzzle of my own, "lateral-thinking is not an excuse to indulge in wild flights of fancy; I can't close every loophole… and shouldn't have to".
THIS question is about the tag, a part of the puzzle postings themselves.
To read and/or contribute thoughts on the poor quality of answers to puzzles in the lateral-thinking tag, see: 'Lateral Thinking' is a poor excuse for a bad answer.
Removing this tag on the grounds that it is often applied to bad puzzles would be like banning warning labels on the grounds that they are so often attached to dangerous products. The tag is not the problem.
But perhaps there could be a way for individual users to filter out puzzles with this tag, so that they don't appear in the user's stream?
I have several reasons as to why I think we should get rid of the tag:
Its been an issue for a long time.
We have already have 4 questions (1, 2, 3, 4) regarding the tag, most suggesting it should be destroyed. For some reason it hasn't yet been done.
Quite a few get closed.
There have been 500 questions of which 122 have been closed. This is $22.4$% - almost a quater of all questions have been closed.
The tag is the 23rd most closed - and the 3rd most closed tag of tags with at least 50 posts.
Personally I think this is too high. This is perhaps the biggest reason we should get rid of it.
It has an awful average score
Those 500 posts have accumulated a score of 4816 giving an average question score of $9.632$.
However, there have been 111 posts in the last year getting a total score of 854 and an average of $7.693$. That is a big drop in average score.
Taking this further:
Months ago | 0-6 | 6-12 | 12-18 | 18-24 | 24-30 | 30-36 | 36-42 | Score | 86 | 768 | 1483 | 398 | 1092 | 828 | 161 | Number of posts | 31 | 80 | 152 | 47 | 110 | 70 | 10 | Average Score | 2.8 | 9.6 | 9.8 | 8.5 | 9.9 | 11.8 | 16.1 |
Last X Months | 42 | 36 | 30 | 24 | 18 | 12 | 6 | Score | 4816 | 4655 | 3827 | 2735 | 2337 | 854 | 86 | Number of posts | 500 | 490 | 420 | 310 | 263 | 111 | 31 | Average Score | 9.6 | 9.5 | 9.1 | 8.8 | 8.9 | 7.7 | 2.8 |
The score is just going to keep going down
Not only this but we are also getting less questions tagged as such.
A lot have a negative score
102 of the 500 questions have a score of 0 or lower. That is $20.4$%
It doesn't help with searches
The tag itself doesn't define a type of puzzle and almost always is used with other tags.
It has a small score per view ratio
The tag has a score per thousand views of 12.47 which is very small compared to other tags.
Burninate it. While the idea may have been good at the start of the site, as it has progressed and developed people have begun to realise that it is a very broad tag which isn't helpful to the site.
lateral-thinking(Ps., my ^vote here indicates agreement that the tag could use morediscussion, not necessarily banishment, and agreement that Sid's answer was the best.) $\endgroup$