Quote:
Originally Posted by lastcardcharlie
How much of math reduces to pure manipulation of symbols according to certain rules, like in chess, has been an important question. I think a rough analogy is:
opening position = axioms
rules of chess = rules of logic
legally reachable positions = theorems
There seems to be an unfortunate misunderstanding of what the practice of modern mathematics is as alluded in your post, i.e., many people think that mathematics is simply the "pure manipulation of symbols according to certain rules"; however, what these people fail to understand and appreciate is how those symbols are to be "interpreted" for there are infinitely many ways to "interpret" them. Mathematics is much "richer" than chess and that a more appropriate analogy is how can the "rules of chess" ( board, pieces and movement ) be changed to retain the concepts we would like to have and in a "manageable" way for humans. I'm not merely speaking of something like Chess960, but the board could be enlarged and other "types of pieces" can be included. Then, among "all the variants of chess-like games", what are the right concepts to retain and which "variants" are the most interesting to analyze? Obviously, if we extend the board to 100x100, it isn't "manageable" even though one could have a "chess-like game"; similarly, if we have hundreds of different types of pieces with different characteristics, that would be "unmanageable". Mathematics tries to find those concepts that bear fruit in giving "interesting results", whether they are theorems or more abstractly, other fields of mathematical study. Naturally, for those ideas to be meaningful, they need to be expressed through symbols so that others may be able to comprehend; however, even if they are expressed, they may not necessarily lead to any understanding except for a very small group. Also, a "chess-like game" may not be very interesting practically, if for example, there isn't much "balance" (e.g., if the first player often wins, or one tiny mistake and it's game over!), or little "conflict" ( the opposing "forces" should come into contact without much waiting ), or no "resolution" ( not much point if the game takes more than thousands of moves to finish ) or is extremely difficult/simple to analyze ( e.g., finding candidate moves, and the typical "tree of analysis" isn't significantly larger or smaller than what one typically finds in normal chess). In mathematics, one is sometimes asking for which "reasonable" set of axioms ( which we may or may not know for a long time, if ever! ) can one get these concepts that we would like to have.
I think chess and mathematics are similar in the higher level thinking processes, but they obviously also differ for several reasons. Also, arguably what is fashionable in chess opening theory often has very good reasons; analogously, algebraic geometry is popular because of its applicability to solve problems ( that have been considered important in the last fifty years ) generally.
In chess, at the lowest level, it's something like "if White plays x, then Black can only play one of y1, y2, ... because z1, z2, ... are bad because of ..." with lines in the "tree of analysis"; more abstractly, the logic might be " if White tries to execute the plan of X1, X2, ..., then Black must respond with the plan of Y1, Y2, ...". Opening theory is often like that, too: e.g., some lines are abandoned because either there is a clear plan leading to a known advantage or clear equality or there are "several roads" making it less likely to be "critical". Also, for a chess player OTB/"corres", it's helpful to envision some "target positions" to achieve: e.g., if I can get to one of these "nice positions", there should be some advantage or equality is achieved.
In pure mathematics, at lower levels, it's something like "if these are the axioms, using the rules of logic, these are the theorems". At the higher levels, it's more like "to solve this type of problem, the ideas X1, X2, ... have been shown not to bear fruit; therefore, some new ideas need to be explored, i.e., likely ones that have not been found as of yet" or "the central ideas of theory X turn out to be the concepts Y1, Y2, ...; are there other important ideas?". Usually, these ideas are laid out for us in a "nice way" in textbooks; if not, then that usually means there is some room for discovery. A good question in mathematics is "What are the most important algebraic structures to study and why?". As a simple example: "What are the most important characteristics of binary operations?" [ It turns out that commutativity is
not nearly as important as associativity. ] Thus, to explore symmetries, it is clear to consider groups and therefore group theory is highly relevant, but it is not as important to consider only abelian groups. Today, there are probably not many visionary mathematicians trying to create vast frameworks as in the past ( such as Grothendieck ), but there are still many undiscovered fruitful ideas awaiting us.