Yes, we humans do manage to deal with the vagaries of natural language, but I don't really see this fact as being a good defense of similar issues cropping up in artificial languages. Why do we have programming languages for developing software at all? It's because specifying the solutions to engineering problems is damned near impossible in English, at least when it comes to the nitty-gritty details. We need programming languages for the precision with which they allow us to specify the operation of systems. Anything that goes against this end ought to be considered a misfeature.I would ask a different question: why do we have multiple programming languages at all? Rather than one, clear, precise way of specifying an algorithm, we have hundreds of them. Many people seem to think that this makes some sort of sense, that different languages do a better job on different problems and/or with different people.
Would there be any point in having multiple programming languages if all of them were essentially the same with only minor syntactical variations between them?
The opinion that computer languages should be based on some concept of mathematical elegance is pretty common... perl and perl alone is pursuing a different path, focusing on "expressiveness" in analogy with natural languages.
My personal opinion is that there isn't really any proof whatsover that one approach is better than the other: demonstrating that "mathematical elegance" is best would require some very difficult social science experiments, and the guys who are proponents of "mathematical elegance" not coincidentally just want to do mathematical proofs. So instead people fall back on introspection, and it seems that some people like one way, and some like another.
Some issues to consider (though they might be side issues):
- The standard of writing (books and documentation) in the perl world is very high... programmers who like thinking "linguistically" also make eloquent writers?
- The pearl at the center of perl culture is the spirit of collaboration: CPAN, perlmonks, comp.lang.perl.*, and so on. Is there a reason that this particular language has inspired this?
- Occasionally the "mathematical" crowd take a stab at designing a language to replace "natural language". The result never seems to catch on. Look up Loglan/lojban. And maybe: "sapir-whorf", "general semantics", and "babel-17".