Are you asking for a terminology for boolean questions, or more general ones? For boolean questions (those admitting only two possible answers, "true" and "false"), all that needs to be done is to add a single symbol, typically a question mark, to our formal notation. You can place it after a statement to turn it into the question "is the following statement true?". You can place it in other places (immediately after a quantifier, on top of an equals sign or inequality, etc.) but these are just simple syntactic transformations that add no expressive power. But the more interesting questions are the more open-ended ones. How fast does this function grow as x gets large? (answer is a function of x, along with a piece of terminology like O() or o()) Can we say anything interesting about ___? (answer is a mathematical statement) How can we precisely define what it means for a transformation to be natural? (answer is the invention of the field of category theory) I'm not sure how you would define a formal syntax for this sort of question, much less a formal semantics. And I'm skeptical about how useful it would be even if you could define it. Andy On Sun, Dec 3, 2017 at 2:06 AM, Scott Kim <scottekim1@gmail.com> wrote:
I'm thinking about teaching the problem solving process in mathematics, and have run into a curious question: can one ask a mathematical question purely in mathematical notation? I believe the answer is no — mathematical questions always require human language in addition to mathematical notation. Problem statements are therefore always extramathematical in nature.
In practice, school kids frequently see problems stated in forms like: 13+78 = ___, which means "what is the sum of thirteen and seventy eight?" But that involves a nonmathematical symbol (the blank), AND ends up misleading kids to think that the equals sign means "the answer is". Very sloppy. We can do better.
This is an extension of something that has always bothered me: if mathematics is so rigorous, then how come it is conducted in a mish mosh of English and formal notation? The practical reason for mixing in human language is clear…like a computer program, a formal mathematical proof is more readable if it is annotated with comments. But too much reliance on human language means that formal proofs are not checkable by computer — a weird situation at best. I suppose that makes me a formalist, or a computer scientist…I've certainly got the latter bias because I don't trust anything I can't program.
— Scott _______________________________________________ math-fun mailing list math-fun@mailman.xmission.com https://mailman.xmission.com/cgi-bin/mailman/listinfo/math-fun
-- Andy.Latto@pobox.com