@TedShifrin I asked when you started because I've developed a fascination with Reconstruction and its effects recently. And I knew there was no way you started at UGA when it was still segregated, but I thought there may be a chance you overlapped with some people who did
But it was 30 years between that and when you started
So I've got no particular insight with this and am just experimenting, so please don't stab me if I'm being an idiot, but would it be worthwhile to send this through $e^{rz}$ where $R_1 < r < R_2$? The idea being that it'd be periodic on one but not the other might allow us to work with it
So assume $f:A(1,R_1)\to A(1,R_2)$, then take a circle in $A(1,R_2)$ and look at its pre-image, that'll be bounded away from $|z| = 1$ and $|z| = R_1$
So if you take some connected region near the $|z| = 1$ circle in the annulus (that is disjoint from the pre-image of the circle), that'll have to map to a connected set in the image which is disjoint from the circle and it has to all lie on one side
Could we try to consider $f^n$ and see what it does to $A(R_1,R_2)$? Like maybe if we could get that they're equicontinuous and then do some Arzela-Ascoli stuff to it
Yeah our complex class was a bit weak. In hindsight Danny's was better, they got pretty far and apparently have been talking about stuff like elliptic curves and all, but it seems both spent a bit too much time reviewing 203
Yeah. Like I still definitely like Nori, his lectures are quality and he's definitely fun, but I think he was a bit too light. He took way too long on Cauchy's integral formula since he was reviewing a lot of basic facts about convergence, and I think he started off first week literally reviewing stuff like complex numbers and norms
I mean I'm not saying he's like, the model for how all classes should be taught, I could see a case being made that he teaches an honors complex analysis class and then everyone else does something intermediate
iirc i heard from someone that it used to be a thing a lot of other stem people took because of it's relevance for PDEs, which a lot of other stem people take
but i think structural changes in other departments made this stop happening
i wasnt just wildly speculating about curriculum changes
Galois theory with Emerton, complex with Marianna, descriptive set theory with Marianna, combinatorics with Laci, formal languages with Kurtz, harmonic with Schlag
The only case I could see not doing Galois (and this probably a shit idea) is if I don't get into algorithms and then try to do a reading course on the stuff right now
Musical set theory provides concepts for categorizing musical objects and describing their relationships. Many of the notions were first elaborated by Howard Hanson (1960) in connection with tonal music, and then mostly developed in connection with atonal music by theorists such as Allen Forte (1973), drawing on the work in twelve-tone theory of Milton Babbitt. The concepts of set theory are very general and can be applied to tonal and atonal styles in any equally tempered tuning system, and to some extent more generally than that.
One branch of musical set theory deals with collections (sets and...
Suppose
$$
f(x) = \arctan(x)
$$
We can write
$$
f^{(2)}(x) = 2x\left( f^{(1)}(x) \right)^2
$$
Subject to some boundary condition $f(a) = c, f(b) = d$.
I'm not an expert in this boundary value problems, but I'd like to know what kind of iteration I can use to solve such kind of problem.
I've...
Hello, we say $f$ is increasing $\Longleftrightarrow x\leq y\Rightarrow f(x)\leq f(y)$ or $f$ is increasing $\Longleftrightarrow x\leq y\Longleftrightarrow f(x)\leq f(y)$
So in differential geometry, $dx_i(E_j)=\delta_{ij}$ where $dx_i$ is an element of the basis of the dual space of the tangent bundle and $E_j$ is a standard vector field. Is this correct?
So then for $X=a_1E_1+...+a_nE_n$, $dx_i(X)=a_i$, correct?
What do you think, learning abstract algebra from older books like Soviet mathematician Kurosh's "A course in higher (abstract) algebra" book published in 1968 good (or is that book on equal status with the modern ones)? Thanks in advance!
If I have done right then $\chi$ is one of the roots of $\chi^2 (1 + k^2 cos^2(\frac{\beta h}{2}) - \chi (2 - 2k^2sin^{2}\frac{\beta h}{2})) + (1 + k^2 sin^2(\frac{\beta h}{2})) = 0$
With regards to notation of permutations how does on distinguish one-line notation from cyclic permutations? Like $(1, 3, 2, 4)$ could mean $1\to1$ or $1\to 3$ depending on whether it's cyclic or not
We know what a cantor set is. Now consider trying to do a cantor set like construction on the "unit rational interval" $\Bbb{Q} \cap [0,1]$ instead, what will happen
Since the pseudoboundary is countable, it has measure zero
I wonder what happens if we modify the construction of the cantor set such that we remove closed middle thirds intervals instead of open intervals. Perhaps we also end up with a countable set that is dense since the left and rightmost end for each interval in the construction is left untouched
in particular, the points 0 and 1 should be left behind among countably many others
Unrelated question: What are the applications of $\omega_1$ in other fields of mathematics outside of set theory. So far I am only aware of measure theory and the long line?
@Secret if you give $\omega_1$ the order topology, you get a pretty weird topological space, which can serve as a basis for counterexamples to several statements in topology. (e.g. you can construct a sequentially continuous function that is not continuous)
I see. I think I can comprehend such function as based on the fact that no sequences of ordinals can converge to $\omega_1$ without having $\omega_1$ as one of its elements
I am more interested on its appplications to theorems analysis and geometry, besides constructing counterexamples
Hi, just a quick question: Suppose we have two sets of edges I and J, such that |I|<|J|. Furthermore, we know that if we match these sets to a set of nodes called V, the resulting graphs (V,I) and (V,J) both will have at most one circuit. Is there a way to prove that there always exists an element e in J\I, such that (V,I+e) has at most one circuit?
One thing that puzzles me about ZF and ZFC is what motivates us to have a hierarchy of infinite well orderings that has larger cardinality than the previous. I mean, I can understand why we need different infinite cardinalities because of powersets and cantors theorem, but I see no motivation on why we need well orderings beyond countable well orderings
The fact that no (countable) sequences can converge to $\omega_1$ basically means $\omega_1$ is in some sense unreachable from below, thus its existence depends entirely on hartog's number ( or axiom of choice if using ZFC). But what is a rationale to justify defining them?
@AlessandroCodenotti @AkivaWeinberger may be interested
Put it in the context of hartog numbers, why are we interested in having ordinals that don't inject to countable ordinals (sure we can easily prove them, but why we want to define such mapping in the first place)?
I always felt the answer to this might lie in some theorems in real and functional analysis, I might need to dig deeper...
@Secret I don't know about Hartogs number, but from a topological point of view, I interpreted the weird properties of $\omega_1$ as an example that for a topological spaces, sequences alone are insufficient to describe the topology, unlike for metric spaces. Note that in real and functional analysis, a lot of the spaces that people care about are metrizable or pseudometrizable, so this pathologies don't arise.
But there are spaces in functinoal analysis that are not first-countable, which means that their topology is not characterized by sequences, for example dual spaces of infinite dimensional topological vector spaces with the weak topology
This is not directly related to $\omega_1$, though
But it means that there are spaces in functional analysis which behave like $\omega_1$
in the regard that the toplogical properties can't be deduced from the convergence of sequences alone
I see, I might need to read more on that later, but a preliminary check shows these dual spaces are quite important in defining linear operators in some hilbert spaces as well distributions
though at my current knowledge, I am not sure if uncountable well orderings in such space is useful though, I think I can only answer that after a trip to functional analysis
but for many topologies, we can do just fine without thinking much about well orderings in general, I think...
We only need linear order in order to define nets (generalisation of sequences)
I am not terribly worried about the usefulness of axiom of choice (because we know how important it is in many functional analysis theorems) What I am puzzled about is the utility of $\omega_1$ and ordinals that represent higher cardinality well ordering
there's nothing circular about it. Let $f: A \to B$ be a monomorphism. we claim that the unique map $0 \to A$ satisfies the UMP of the $\operatorname{ker}(f)$. Let $g: X \to A$ such that $f \circ g= 0 = f \circ 0$, then as $f$ is a monomorphism, we get $g=0$, so $g$ factors through the zero object
you need to prove that $0 = f \circ 0$ for every morphism $f$