Personally I go with Ogre, but I know c++ and Blender and am on Linux. For you, I'd recommend to have a look at Unity and UDK. And hey, without lots of time....forget about it. This should also go into a FAQ section....

I'm not sure unity is that well suited for a minecraft clone, you'd most likely benefit quite a bit by writing your own voxel renderer for such a project.

As for C# it will work on Linux and OS X aswell as long as you don't use XNA(or one of the DX wrappers), (there is an XNA port called monogame in development aswell but i don't think its quite there yet). C#+OpenTK work really well for cross platform development (atleast on desktop pcs).

If you are going to target a wide variety of platforms you really want to make sure you keep as much of your game logic separate as possible since you might have to replace/rewrite other parts of the game. (You cannot target the xbox360 as an indie without using XNA and you cannot target the ipad using XNA, you can however use C# for both platforms and just swap out the rendering/sound/input libraries.)

I have never seen a project failing because of the wrong language/api/framework was chosen. Software projects fail because the people working on it fail.
So stop worrying about languages, just pick one (possibly the one you feel more comfortable with) and start working... you are the bottleneck, always.

Ever heard of the Blender Game Engine?

I have a series of video tutorials that cover the basics: http://www.gamedev.net/topic/625003-video-tutorials-blender-game-engine/

That said: a Minecraft clone is not exactly a trivial project, especially if you're not an experienced game developer - You might want to try a more realistic project.

I realize that people will read this in a beginner forum and think I'm a moron or something, talking about a grandiose game while not having experience making games but, I'm just trying to learn guys

the problem is that you are not learning by asking these things, you're just wasting time. Because people can come here and start listing anything from Java to some exotic language and, maybe, present you with a list of reasons why that would be a good decision. But you wont be able to understand any of those reasons because you're probably at step 1 and they are talking about step 1346.
There is nothing that makes a 3d voxel engine a-la minecraft any different from whatever other game style from a language/tech point of view, so there is nothing different from your question and the other million "what language should I use?" questions on the forum. They all go down the same path, a random sequence of : "you should use C#, Java, Phyton, C++, no C++ is not a good beginner language, yes it is, no it isn't, look at my links, bla bla bla bla bla"... over and over again.

You have a success story right there: Minecraft is coded in Java. What do you need to know more? You want a list? Here's the list for you, you could do it in:

- C / OpenGL
- C++ / OpenGL
- C++/DX9 , 10 , 11
- C#/Xna
- C#/SlimDX
- C#/SharpDX
- C#/OpenTK
- Unity
- Java
- Python / PyPy
- Blender Game Engine
- UDK
- CryEngine
- D/OpenGL
- Google Go/OpenGL

I could go on forever.. you can do Minecraft with all of the above, these are some of your options (there are literally hundreds more) what's the REAL decisive factor? You are! All these tools are capable of accomplish the job... are you?
Just pick a random one and start, brake down your problem into small sections, google for info, code, test, iterate. If anytime during the development you'll feel like the tool/language is holding you down, think again, it isn't.. it's all about you.

Fair enough, but I guess what I'm asking is, what are my options? Like could someone name 10 ways people make games? Presumably there's a pretty small number of *sensible* ways to make a 3D voxel game with building and multiplayer, like methods/frameworks that people who know what they are doing would use, as opposed to silly ways of trying to do it.

There are lots of ways to do it but none of the big engines is really helpful (you can't just hand them your voxel data and expect them to deal with it as you can for games using meshes or sprites) you will have to write mesh generation and physics code yourself anyway (For voxels its best if you do your collision tests against the voxel data rather than creating collider objects to use with a normal physics engine).
This means that the only features you need from a game engine is the ability to create custom meshes at runtime and render 3D graphics, Any 3D engine should let you do this and as you are ignoring most of the engines other features the things you should be looking at are:

1) Platform support (Try to pick a language that is available on all your target platforms and libraries/frameworks/engines that support as many of them as possible, keep your own code as isolated as you can so that moving it to a different engine or swapping out the libraries it uses stays reasonably easy)
2) License costs and restrictions (You won't use most of the functionality provided so cheap/free and as few restrictions as possible are a good idea)

You need to learn the language you want to code in first.
Jumping into 3D, with no knowledge of Java ( as you mentioned ) is doomed to failure. Even blindly doing 3D with a simplistic language such as Python ( Panda 3D http://www.panda3d.org ) isn't going to cut it.

Learn the language you choose to use first, THAN try out game creation.

I. If anytime during the development you'll feel like the tool/language is holding you down, think again, it isn't.. it's all about you.

I disagree .... there are quite a few languages that are notorious for not being able to develop with