List of Local to Global principles

What are some of the local to global principles in different areas of mathematics?

-
community wiki? –  lhf Apr 20 '11 at 11:55
am I supposed to do that? –  quanta Apr 20 '11 at 11:56
when posting a big-list question, please flag it for moderator attention with a request to make it community wiki. @lhf, J.M: thanks for the ping. –  Willie Wong Apr 20 '11 at 12:56

Differential Geometry

The existence of partitions of unity allows one to transfer local results to global ones.

The Gauss-Bonnet Theorem relates the Gaussian curvature (a local quantity) to the Euler characteristic (a global one).

-

Diophantine Equations

Hasse Condition: If a Diophantine equation is solvable modulo every prime power (locally) as well as in the reals then it is solvable in the integers.

Hasse Principle is that the Hasse Condition holds for all quadratic Diophantine equations.

-
See also the Grunwald–Wang theorem and this question –  lhf Apr 20 '11 at 12:05

Graph Theory

A graph has an Eulerian circuit (global) iff every node has even degree (local).

-
Don't you also need that the graph is connected? –  JSchlather May 14 '11 at 20:33

Number Theory

The "original" (in terms of giving rise to the name) local-global principle for quadratic forms over number fields, due to Hasse, has already been mentioned. Here are two further local-global principles in which Hasse was involved.

1. Two (finite-dimensional) central simple algebras over a number field $K$ are isomorphic if and only if their base extensions to central simple algebras over $K_v$ are isomorphic for every completion $K_v$ of $K$. This is essentially the Albert-Brauer-Hasse-Noether theorem.

2. Class field theory can be formulated both for number fields and for their completions, called respectively global class field theory and local class field theory. (It can formulated also for function fields over finite fields, but let's not worry about that here.) Historically global class field theory came first and the proofs of local class field theory originally depended on global class field theory. Eventually Hasse was able to develop local class field theory in a self-contained way and then use it to prove global class field theory.

-

Complex Analysis

Analytic continuation might be viewed as a local-to-global principle.

-

Real analysis

The extreme value theorem can be read as: if a function on a compact set is locally bounded then it's globally bounded. (This is only part of the theorem.) The key word here of course is compact.

-
I just wanted to add that compactness is a very important property precisely because it is often used (in analysis) to extend local properties to global properties. The extreme value theorem is one example among many. –  svenkatr Apr 20 '11 at 18:25
Another example of this type comes from Differential Geometry. For a smooth vector field on a smooth manifold we can define flows locally. On a compact manifold, we can define a flow globally. –  sjm.majewski Feb 2 at 19:23

Euler Characteristic

If you count the faces vertices and edges (local information) of a polygonal shape then you can compute the Euler Characteristic which tell you how many holes the shape has (global information).

-

Number Theory

If an integer $n \not \equiv 0 \pmod m$ (for any $m$) then $n \not = 0$.

-

Complex Analysis

If a $\mathbb C \rightarrow \mathbb C$ function can be differentiated (local) then it can be integrated! (Inspired by Qiaochus answer here).

-

Number Theory

Zeta L-functions and Birch-Swinnerton Dyer conjecture: These are quantifies which are defined in terms of local things (like multiplying together a function on primes) and global information (like class numbers) is extracted from them.

-

Set Theory

The axiom of choice is local, while Global choice is global.

(Global choice, or Axiom E in the NGB set theory, is equivalent to saying that all proper classes are of the same "cardinality", or there is a well-ordering of the universe)

-

$C^*$-Algebras

Let $\mathfrak{A}$ be a unital $C^*$-algebra, $\mathfrak{C}$ a $C^*$-subalgebra of the center of $\mathfrak{A}$ which contains the unit of $\mathfrak{A}$ and for any maximal ideal $x$ of $\mathfrak{C}$ let $I_x$ be the smallest closed two-sided ideal of $\mathfrak{A}$ containing $x$.

Now the local principle by Allan and Douglas says, that $a\in\mathfrak{A}$ is invertible in $\mathfrak{A}$ if and only if $a+I_x$ is invertible in the quotient algebra $\mathfrak{A}/I_x$ for every maximal ideal $x$ of $\mathfrak{C}$.

One can use this result to tackle questions of invertibility and Fredholmness in operator algebras, e.g. one can characterize Fredholmness properties of Toeplitz operators with piecewise continuous symbols by means of easily checked properties of their respective symbols.

-

Analysis

$$\prod_{p}|x|_p = |x|^{-1}$$ gives a way to piece together all local norms to find the value of a global norm.

-
I don't agree with this. The standard absolute value |x| is also a local norm (it's a norm on a completion of Q), so the standard way to express this equation is to move the right side to the left side and say that the product of all normalized absolute values of any nonzero rational number is 1. –  KCd Sep 10 '11 at 19:25

And i was studying completely about Hasse's local-Global principle,which has many applications like ,there is a separate fantastic group called the Tate-Shafarevich Group that measures the extent of Failure of Hasse's local-global Principle.it is given by $Ш(E/K)=\mathrm{Ker}(H^1(K,E)\mapsto \prod_{v}H^1(K_v,E))$ which means literally ,"the non-trivial elements of the Tate-Shafarevich group can be thought of as the homogeneous spaces of $A$(where $A$ is an Abelian Variety defined over $K$) that have $K_v$-rational points for every place $v$ of $K$, but no $K$-rational point." and the TS group has many important applications in BSD conjecture,Iwasawa theory....

-
If $a$ is a square (or $n$th power) modulo every prime power then, since its $p$-adic valuation is even (or a multiple of $n$) for every $p$, $a$ is a square (or $n$th power).
Not really true for $n>2$. See the Grunwald–Wang theorem mentioned in another comment of mine. –  lhf May 6 '11 at 16:59
16 is not an 8-th power but it is mod $p$ for all primes $p$. See math.stackexchange.com/questions/6758/… –  lhf May 6 '11 at 20:22