& Fractals Vol. 7, No. 5, pp. 795-819, 1996 Copyright 0 1996 Elsevier Science Ltd Printed in Great Britain. All rights reserved o%w779/96 $15.00 + o.cu
Sets
HEMION of Bielefeld,
33615 Bielefeld,
Germany
Abstract-This paper is concerned with a theory of discrete partially ordered sets which satisfy a number of additional axioms. In particular, the sets will be assumed to be ‘stable’. That is, if a and b are two elements of such a set, with the property that all elements above b are also above a, and all elements below a are below b then-necessarily-we must have (as expected) a c b. It is found that such sets have certain geometrical structures which are analogous to the usual Euclidean geometry. A probability theory for a class of such sets is proposed, and it is found that sets which contain particle-like chains of elements are most likely. Copyright 0 1996 Elsevier Science Ltd.
1. INTRODUCTION
How many axioms are necessary to specify the usual kinds of mathematical analysis which are common in the description of physics? Surely many tens, if not hundreds of axioms are needed! Even a definition of the real number system requires a surprisingly large number of assumptions. Then more axioms are needed to specify Euclidean space. Still more are required to specify the complex numbers and the rules of complex analysis. Many more are needed if we delve into differential manifolds. If algebraic topology is also brought into the picture, then a great number of further axioms are again needed. Has anyone ever gone to the trouble of making a complete list? If so, perhaps it would be worthwhile to check the logical connections between all of these axioms. It may be that some small set of seemingly primitive axioms have unexpected consequences that are in conflict with-or perhaps imply-some of the axioms which appear more towards the beginning of the list. In this paper, we will deal with only a very few, extremely primitive axioms. In particular we will consider the theory of discrete partially ordered sets. Now it is certainly true that discreteness and ordering are things which are not usually considered to be at the beginning of the subject of mathematical analysis. On the other hand, in modern physics these concepts are usually adjoined somewhere onto the main body of the whole theory in a way which confuses the strength of the assumptions which are being made. Therefore, to clarify matters, we will examine these concepts here in an abstract way, free of all peripheral assumptions. 2. BASIC
2.1.
DEFINITIONS
Confluence, discreteness,stability
A partially ordered set (or ‘poset’) is a set X, together ‘c’ C X x X with the properties (1) (4 a) E S, for all a E X, (2) if (a, 6) E G and (b, a) E c then a = b and (3) if (a, b) E G and (b, c) E c then (a, c) E C. 795
with an ordering
relationship
Of course it is more common to write ‘a s h’, rather than ‘(a, h) E 5:‘. The expression n ~1h explicitly excludes the possibility that o = h. In addition to this, we will use the notation ‘17 :?=Ci h’ to denote the idea that a and h are unrelated: that is. a Dd b o u $ h and a 8 h. Most of the posets with which we shall be dealing will be infinite, and so it will often be necessary to find a shorthand way of describing specific examples. For this the following definition will be helpful. Definition. Let (X, 5) be a poset and let 3 LCC XxX. generates c: if s is the smallest partial ordering containing E.
We will say that E
Definiticm. Let CJE X be a typical element of our partially ordered set. Then we will denote by +, the set of elements of X which are less than, but not equal to, a, i.c. a,, - {u E X: ci
A class of partially ordered sets