The Sapir-Whorf hypothesis has been thoroughly shot down for human languages ([citation needed], I know), but I guess it's accurate for programming languages, given that learning a programming language is a prerequisite for learning programming in the first place.

Orestis,

I would be interested in reading more about that if you can find a link.

Truth be told it was the first time I had heard about it. I can't speak much for its validity with respect to human language but, I agree, it makes a lot of sense in the realm of computer programming languages.

Thanks,

-Matt

My reading of it was that rather than disproving Sapir-Whorf, the counter-research has mostly just emphasized the universality of most human languages. (The human languages all are "Thought-complete" in a way that most programming languages are "Turing-complete".) I wouldn't describe that as "thoroughly shot down", particularly because that's not what the Sapir-Whorf hypothesis is ultimately about depending upon your interpretation of the hypothesis. On the other hand, there is interesting evidence in the relative differences between languages and their effect on typical discussion patterns. As with nearly any aspect of psychology I think that it is very hard to conclusively say that a hypothesis is proven or not.

The section "Cognitive Linguistics" from the referenced Wikipedia article has a paragraph summarizing some of George Lakoff's points that many people arguing for or against the hypothesis often don't even agree on the same interpretation of the hypothesis.

I am not a linguist, but I do find it fascinating. There is a lot of great reading on the Sapir-Whorf hypothesis on both sides of the court. I personally feel that the hypothesis is "true", but have not found enough evidence to know exactly "how true", if you know what I mean. I absolutely believe it is accurate for programming languages.

I beg to differ. Different programming languages have different features. Python is the most featured of languages I can think of. It doesn't mean that it should beat all other languages, because, sometimes, the feature "compiled" (C++, ObjC) or "template language" (Php) are more important.

I know the same languages as you do and I don't feel at all that I program in a different way in different languages. However, when more feature are available (as in python) I tend, over the years, to use those features more and more intensively (generators, meta classes and the like).

I think that most of the work may be done with one script language (say python). Then, for specific tasks one might need one compiled language for speed (say C or C++), possibly one "template" language (php/LaTeX). I don't think that you need much more than that. Most of all, it's the different features of the languages that make them useful. If python was both interpreted, compiled (see cython...) and template, I would only use python for the rest of my life.

The comparison with human languages is completely irrelevant. All human languages have more or less the same features and you can express whatever you want in any human language. This is why languages exist in the first place! I speak 5 languages and I really can't imagine which would be better for what.

Olivier,

I'm confused as to what exactly you disagree with. You're making the same points as the post... ie use the language with the features/performance/etc. required for the job.

Am I missing something?

-Matt

I'm quoting your post: "Use whatever aligns itself with the way your mind works".

I utterly disagree with that, and I explained my point. It's only a matter of *features* of the languages, not a matter of how your mind works. Python has most of those required features, so I end up programming in python 99% of the time. On some rare occasion something else is needed. This is not because of how my mind would work differently with another language, but simply because python doesn't have that feature that is needed to solve the problem.

Also, you make a doubtful comparison with human languages, which is not relevant because all human languages have the same features, so you need only one human language to express yourself.

Thanks for your feedback. I agree with your points with regards to the line you quoted. In hindsight it doesn't jive with the rest of the points I'm making in that paragraph.

With respect to Sapir-whorf. My understanding is that what it suggests is fundamentally different than the point you're making. It supposes that a language affects how you think.

This does make sense to me. Take Python - you think "pythonic". A functional programming language would encourage you to think functionally. C encourages you to think procedurally. Etc.

Great post, really help me alot. Thanks.

Cheers,
<a href="http://sain-web.com rel="nofollow">Blog Review