Let's say you have created a programming language, maybe because there was a problem which was unfeasible to try to solve using existing languages, what would you do to market it (and so on) in order to make it gain attention in the software engineering (or academic etc. depending on the purpose of the language) industry?
Use it as your primary development tool. When you run into its failings, address them, because you can afford to change basically whatever you want about it before it becomes popular. The language will improve and solidify over time. As you produce more and more great applications written in your language, people will take notice of it and begin to take it seriously as a tool, and they might start using it. That's step one.
Step two is to keep using it, and make it as easy as possible for the people who might want to use it. That means top-notch documentation and examples, actively promoting the language to everyone you can, and getting your users to actively promote it by making it good enough to use as a primary tool.
But it all starts with using it yourself for real tasks. If you never get to that point, neither will anyone else.
Unfortunately there's not a whole lot you can do to MAKE something popular. Take for example the popular game for the apple platform Angry Birds. The company that made that game had created many many games prior that were only marginally used in comparison. Then there was just something about Angry Birds that just struck with people... then they told their friends who told their friends to the point that it seems like almost everyone owns/has played/knows about it. The same thing is true of programming languages. You make it... then its either a success in time or its not. The only real things you can do is try to promote it with friends, colleagues, etc. Certain things will influence its success such as speed, ease of use, functionality, versatility, etc. Its these areas that are most worth your time and effort.
EDIT: Well I actually just thought of something... volume might be with in your control. If you're extremely creative you could just start pumping out hundreds of different programming languages... eventually the odds would get very good that one of them would become popular...
Flip this around and ask yourself why someone would want to use your new language?
What does it provide that existing languages don't provide? Are the advantages big enough to warrant someone taking the plunge ...
Is the language specified? Is the specification complete? Is it stable?
Are the runtime libraries available?
Is there a decent tool chain? Compilers, debuggers, IDE support.
Are there real people using it to do real projects?
Are there examples? Case studies? Tutorials?
If the answer to many of these questions is "err ... no", then it is too soon to market your language.
Why would you create another programming language? Unless you want to learn some fundamental CS concepts to improve your day to day programming writing a programming language in order to make it popular seems like a waste of time. Then again if you're really set on creating a popular programming language then targeting the JVM or the CLR or LLVM is probably the way to go.
That and assuming the this problem that people have affect enough people. I know Ruby got really popular cause of the MVC framework rails. They also had a huge marketing team. Microsoft C# is very popular because they would try to push it on everything. XNA framework uses C# and to do games on their platform (PC, ZUNE, 360, what have you), you got this neat XNA framework that uses C#.
So yeah, it seems like you have to be proactive and push for it or have something so significant that you don't have to do marketing (scala for example). Google Go language for example isn't taking off. I think only time will tell.