# Beside transcendental or uncomputable numbers what other types of numbers are there?

What other types/categories of numbers are there that we know of today (i.e. some one has done some work on them, like Chaitin's uncomputable $\Omega$ number)? Of course there are uncountably many types/categories of numbers (besides the trivial integer, rational, natural numbers, ...) due to the uncountability of the reals.

Are there any generalisations of irrationality/transcendentality for reals by analogy to complex numbers (2-tuples) or $n$-tuples? What I am trying to ask is: are the reals the only playground for number categorisation (such as irrationality, trancendentality, ...) ? or does anyone know of some other structure similar to $\mathbb R$ (by some type of analogy!) with analogous irrationality/algebraicness/transcendentality properties?

-
Actually, there are only countably many types of numbers, because we only have a finite alphabet with which describe each type! –  Zhen Lin Jul 22 '11 at 9:21
You can trivially extend the definition of irrational, algebraic, and transcendental by considering fractions and polynomials involving Gaussian integers. Nothing really comes to mind otherwise. –  anon Jul 22 '11 at 9:25
You may be interested in reading about periods. Interestingly, it seems still open whether $e$ is a period or not. –  t.b. Jul 22 '11 at 12:19
This question is being interpreted as "Tell me some mathematical structures which have 'number' in the name," which in my opinion is not specific enough a question for such a site. It seems that the OP is looking for something more specific, but I don't really understand the question. Could s/he clarify it? –  Pete L. Clark Jul 22 '11 at 21:41
@Arjang: With regard to the second paragraph: you can talk about irrationality and transcendence with respect to any field extension $K/F$. Is this the sort of thing you have in mind? –  Pete L. Clark Jul 22 '11 at 21:42

The Euclidean Constructible Numbers are of great historical importance. Roughly speaking, these are the numbers that we can produce, starting from $0$ and $1$, by using the ordinary arithmetical operations (addition, subtraction, multiplication, and division) as well as taking the square roots of non-negative numbers.

Thus, for example, $$\frac{\sqrt{10-3\sqrt{2}}}{17}$$ is Euclidean constructible.

The reason they are interesting is that they are closely connected to the problem of what are the possible geometric constructions by straightedge and compass. For example, it was proved (possibly by Gauss, certainly by Wantzel) that the number $\cos(20^\circ)$ is not Euclidean constructible. This was the key step in showing that the famous ancient problem of trisecting the general angle cannot be done in general by straightedge and compass.

At the same time, Wantzel proved that the number $\sqrt[3]{2}$ is not Euclidean constructible. This shows that the old problem of Duplicating the Cube is not solvable by straightedge and compass. (You are given the side of a cube, and you want to construct the side of a cube which has twice the volume of the given cube.)

Some thirty years earlier, Gauss had proved that the regular $17$-gon is straightedge and compass constructible, basically by showing that the number $\cos(360^\circ/17)$ belongs to the Euclidean constructible class of numbers. This was the first new result about Euclidean constructibility of regular polygons since ancient Greek times. Supposedly, it was finding this result that made Gauss decide to concentrate on mathematics, and not philology.

Another historically very important class of numbers are the (complex) numbers that are obtainable by using the ordinary operations of arithmetic, together with $n$-th roots for arbitrary integers $n$. Understanding this class of numbers was a key idea in the proof by Galois that there is no general "formula" for solving equations of degree $5$. (About $300$ years earlier, it had been shown by Cardano, and others, that there is a general formula for solving equations of degrees $3$ and $4$). Of course people had known how to solve quadratic equations for far longer than that.

And then there are the endlessly many extensions of the notion of number. You might find the Surreal Numbers of Conway interesting.

-

You might be interested in learning about the $p$-adic numbers.

-

Besides the classes mentioned so far, "there exist" hyperreal numbers which can involve infinite and infinitesimal quantities, interval numbers, and fuzzy numbers which come in a countable infinity of types themselves.

For the last question, when asking about structure, what operations or relations do you pair up with R?

-

Related: I would recommend a good read, Numbers by Ebbinghaus et. al. It is a very interesting book on the history and development regarding numbers.

-

Physical constants, such as the fine structure constant, or ratios between restmasses of elementary particles. Noone knows if it is rational, irrational, transcendental. Maybe it could be surreal or nonstandard. And its definetively not computable (atleast not currently, and not unlikely forever). These could also be completely new types of numbers surpassing all our current classifications of number.

-
I fail to see how a number coming from work in fields of real numbers can be nonstandard or surreal (and not real as well). –  Asaf Karagila Jul 22 '11 at 11:05
@Asaf These numbers come from physical measurements, of the type we can reasonably bound these numbers by two rationals, not by "work in field of reals". If all we know about a number x is a<x<b, for rationals a and b, then x can be surreal –  vittu_lain Jul 22 '11 at 11:16
As far as I know, most physics is done in real/complex spaces (and lately in p-adic spaces as well). Furthermore, physical measurements cannot be accurate simply because we have cutoff at some $2^{-10000000}$ of a picometer where we can no longer measure due to particle sizes. –  Asaf Karagila Jul 22 '11 at 11:35