Conversation

QC (in Berlin 7/29-???)
@QiaochuYuan
Replying to
@QiaochuYuan
the construction is being done to satisfy a certain *specification* and the specification is the point. the specification is that real numbers are supposed to provide an analytic foundation for euclidean geometry. they do a lot of other things but that's where they come from
2
1
60
QC (in Berlin 7/29-???)
@QiaochuYuan
let us consider the humble circle. in the cartesian way of doing things (which is not at all how euclid did them) a circle is a set of points in the plane R^2 (the set of pairs of real numbers) satisfying an equation like x^2 + y^2 = r^2. why describe circles this way?
Image
1
1
31
QC (in Berlin 7/29-???)
@QiaochuYuan
the reason is that thinking of circles, and geometric shapes generally, in terms of sets of points described by real number coordinates mechanically produces correct answers to classical geometric questions, about the intersections between circles, or whatever else
1
39
QC (in Berlin 7/29-???)
@QiaochuYuan
thinking about points in the plane in terms of coordinates lets us reduce geometric questions to algebra. this is, broadly speaking, one of the most successful mathematical strategies ever, and mathematicians have taken it very far
1
2
63
QC (in Berlin 7/29-???)
@QiaochuYuan
this would not work if, for example, we tried to use the rational numbers instead of the real numbers. circles with only rational coordinates would be "missing" a lot of points (with irrational coordinates) and that would produce the wrong answers to various questions
1
1
40
QC (in Berlin 7/29-???)
@QiaochuYuan
so, okay, that's pretty good, why does anyone have a problem with the real numbers then (and they do)? the problem is that we pay a very bizarre price for filling in the gaps: almost every specific real number is literally indescribable, because there are too many of them!
1
1
46
QC (in Berlin 7/29-???)
@QiaochuYuan
the problem is that no matter how you choose to describe real numbers, there are only countably many possible descriptions (e.g. only countably many programs that spit out strings of digits), but uncountably many real numbers! it's wack
1
3
48
QC (in Berlin 7/29-???)
@QiaochuYuan
this makes some people very uncomfortable (and i think that discomfort is justified). what is "real" about the vast majority of the real number line being inaccessible to any form of description whatsoever???
1
5
42
QC (in Berlin 7/29-???)
@QiaochuYuan
for the purposes of simpler questions in euclidean geometry you can get away with working with a much smaller set of numbers, the algebraic reals, which are all describable but you actually need all of the real numbers to do calculus. and we need calculus for a million things
1
37
QC (in Berlin 7/29-???)
@QiaochuYuan
so the real numbers, as usually constructed, are (this is very much in-my-opinion) this philosophically unsatisfying technical kludge we put up with because it lets us put geometry and calculus and a million other things on a rigorous foundation
3
3
46
QC (in Berlin 7/29-???)
@QiaochuYuan
in the meantime this is a good opportunity for me to highlight another thing i wish people talked about more: ime it's rarely useful (at first pass) to ask "what is X?" in mathematics and usually much more useful to ask "what does X do?" or "what is X for?"
7
9
100
QC (in Berlin 7/29-???)
@QiaochuYuan
actually another meta-level pet peeve of mine: i think a lot of people trying to learn undergrad-to-grad-level math don't really internalize that all this stuff is tools that specific people invented in order to solve specific problems, not like eternal timeless stuff
6
22
191
QC (in Berlin 7/29-???)
@QiaochuYuan
the whole modern axiomatic method, not to mention specific axiomatizations like metric spaces or topological spaces or whatever, all that stuff is contingent. in other timelines it could've gone another way. aliens might not have it. it can be evaluated and discarded if desired
3
2
74