Sign up ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Let one piece of literature be one string. Let's define our alphabet to be sufficient to represent all literature (e.g. we may need a page-turn character, etc). So, since the collection of current literature is finite, it is a regular language. However, we usually talk about such languages as context sensitive and natural which are harder for a computer to deal with. So, what am I missing here?


share|cite|improve this question
BTW, there is a way to include 2-D formulae: convert it to LaTeX or treat it as a generalized grammar (two or more concatenation points on each character in the language). – Enjoys Math Mar 22 '12 at 5:13
When people talk about natural language, they're not talking about the existing corpus of natural language, but I suppose the potential corpus of natural language. – Qiaochu Yuan Mar 22 '12 at 5:13
It's also concerned with existing stuff. e.g. Computer, draw a circle. That's easy to interpret if that's all you want the computer to do, but if you add in a bunch of other requirements like it must interpret all ways to say draw a circle or draw an elementary 2D shape using language X, graphics API Y, position R, and only do it in certain contexts, then it's much harder. But I see your guys' points. :D – Enjoys Math Mar 22 '12 at 5:48
Whether natural language is regular or not is just a small part of how problematic it is "for computers to deal with". The classical example: Compare Time flies like an arrow to Fruit flies like a banana. Structurally they are identical, and it's hard to imagine a computer program being able to distinguish them without lots and lots of contextual "knowledge". – mrf Mar 22 '12 at 6:59

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.