This is generally a horrible idea, as anyone who's worked in a localized IDE can attest to. Programmers rely heavily on having one common vocabulary. When the compiler gives me the error "missing type specifier - int assumed", I can share this exact error message with others, for example here on SO, and it will be familiar to those others so they can tell me what it means. If the compiler instead generated error messages in Danish, I'd be limited to getting help from the relatively few programmers who speak Danish.
Suddenly my vocabulary is no longer the same as someone in the same position in Germany, France or Japan. We can no longer exchange code, bugs, bug fixes or ideas.
A developer in Spain wouldn't be able to use my code because it was literally written in another language. And if I had trouble with my code, others would be helpless to debug it, because it wouldn't even compile under their localization settings (and if it did, it'd still be unreadable to them).
Ultimately, a programming language is a language. It may have borrowed some words from English, but it is not English, and you do not need to understand English to program in it, any more than I need to understand latin in order to speak English (English borrows latin words as well).
You might as well ask for a multi-lingual English. What would be the point? Yes, it would in theory allow people who didn't speak English to... speak English. It just wouldn't be the same English as every other English-speaker speaks, so it wouldn't actually enable communication between them.
The keyword if
in a programming language is not the same as if in the English language. They mean different things, even though one was obviously inspired by the other.
The delegate
keyword in C# does not mean the same thing as "delegate" in English. Nor does while
, return
or "constructor". They are not english words, they are keywords or concepts in C++, Java, C#, Python or any other programming language.