- #36
robwilson
- 60
- 11
Auto-Didact said:This is a strawman argument. When arguing to use dynamics, topology, analysis or geometry as a tool I am specifically not arguing for complicated hypotheses over simple ones. Instead I am arguing for different perspectives which tend to have a completely different, often unconventional, foundation - i.e. unknown to most (not part of a standard curriculum).
Mathematical disciplines that are unknown to many tend also to be slightly intimidating to most, because it might seem to be more complicated than the simple alternative they already know, even if it is conceptually just as simple or even simpler than the more conventional alternative. An example of this is the exterior calculus and differential forms over standard "simpler" vector calculus and multivariable calculus learned in school.
Such "simpler" alternatives are easier in a specific context and purely perceived as generally being simpler due to them already being familiar and spoonfed from a young age, but they are in actuality not really conceptually or mathematically simpler, just different. From the broader mathematical viewpoint, they usually contain assumptions which prevent them from being directly applicable or generalizable to other theories, while the alternative formulations tend to have less or no such problems.
To paraphrase Feynman: "You can't make imperfections on a perfect thing, you need another perfect thing."
This is simply not true in general, and in fact only becomes true once the correct framework has already been identified. But that is precisely the problem we are discussing: how does one identify the correct framework in the first place? What if this framework has not yet been discovered or invented? Algebra itself being non-specific i.e. framework independent is usually of little help in this identification and selection process, especially during the beginning stages.
Discovery of novel frameworks in mathematics is an experimental process of trial and error which requires intuition, not merely computation or deduction; that only comes later once everything has already been worked out. If one is in the beginning stages of creating new mathematical frameworks or generalizing older ones (e.g. the generalization of Euclidean to non-Euclidean geometry) one is typically necessarily unable to use algebra reliably because the correct framework upon which to do algebra simply has yet to be identified.
Let me give you an example. Since this is a thread about symmetry, let's discuss the up/down left/right symmetry of spin in the standard model. Isolated from everything else, one only needs the quaternion group of order 8 to describe these symmetries. To take the classical limit of large numbers of spins, one works in the group algebra. The structure of this algebra is (in physics language) a Euclidean spacetime plus four particles. The spacetime also has a multiplication on it, which describes the macroscopic phenomenon of magnetism. Three of the particles have both left and right spins, one of them only has left spin. There is a finite symmetry that gives the weak doublets of the standard model, if we interpret the particles in the natural way as neutrino, electron, proton and neutron. So one gets all the magneto-weak structure of the standard model and classical physics simultaneously, but without mass or charge, just from the quaternion group. Pure algebra, no geometry required. One needs a bit more to get the whole standard model, but not much.