Thursday, April 16, 2009

The Great Simplification – Part II

After slashing the original definition of Axiom 6, here comes a similar reduction of Axiom 7, originally introduced in The trapped Arrow of Time – Part III.

The new
Axiom 7 Q-Loops
axiom VII n    
Axiom 7.1 are technical definitions:  completely Q-ordered sets and the closed hull of a set. 

Axiom 7.2-4 introduce a substitute for Jordan-Curves. 7.2 defines the property of being connect for a set in terms of topology. 7.3 defines the relation of being topologically separated for two points.  Please note that we don’t require that the whole space is T1. 7.4 defines J-Curves as sets that contain for each point at least one separated partner and fall apart exactly if a separated pair is removed. This definition requires implicitly the Axiom of Choice, therefore it’s flagged.

For the moment let’s assume that J-Curves exist -later on we will claim the existence of rather specific ones- and see whether the definition meets our expectations. Axiom 7.5 introduces the set of connected subsets into which a separated pair splits the J-Curve. By definition for J-Curves, there must be at least 2 of them.    
Lemma S 2 Segments  
lemma s 2 
The lemma shows that –as intended- a separated pair splits a J-Curve in just 2 segments. And these 2 segments contain at least each a point for a separated pair, that is we’ve got two pairs that mutually separate each other.

Sounds  familiar? Well, be aware that the term J-Curve was introduced and some of it’s properties shown without reference to the initial Q-Order, yet –as intended- on a J-Curve there exists a natural Q-Order, as defined by Axiom 7.6.
Lemma S 3 J-Order 
lemma s 3

Be aware that by no means all J-Curves correspond to q-ordered sets, as shown below. 
The above yellow Rhombus  is a  J-Curve, yet it’s not a Q-ordered set but rather built by of 2 Q-ordered sets, the left and right side.

It appears as if we might start out just with some topology with some nice properties … and [re-]construct the Q-Order. For the moment we ask only –consistent with our whole approach- that every totally q-ordered set shall be consistently embeddable into some J-Curve, where consistency means that original Q-Order and derived Q-order of the curve are the same.

Axiom A 7.7 Consistent Embedding 
axiom VII n 7

Axiom A 7.7 has backward consequences for the Q-topology.
Lemma S 4 Connected Space 
lemma s 4  
As immediate consequence of Axiom 7.6, J-Curves finally do exist. J-Curves connect the whole space (S 4.2), which is hence a connected topological space (S 4.3).

Axiom 7.8 Local Orientation
axiom VII n 8

This Axiom establishes an intrinsic relation between J-Curves –remember they are closed- and the underlying Q-Topology. For each point –respectively its closed hull- there shall exist at least one neighborhood, sufficiently large that it can be split into two subsets –sometimes called local future and local past, or local input and local output- such that  each of these subsets is J-convex – i.e. any two points can be connected by a J-Curves, but sufficiently small that any J-Curve that connects between the sets contains at least one external point.

The picture below illustrates the concept.
grid J-curves
The two sets are
{ {(0,0),(-1,0)}, (-1,0), {(-1,0),(-1,1)}, (-1,1), {(-1,1),(0,0)}, {(-1,1),(0,1)}, (0,1),{(0,1),(0,0)} }
{ {(0,0),(0,-1)}, (0,-1), {(0,-1),(1,-1)}, (1,-1), {(1,-1),(0,0)} {(1,-1),(1,0)}, (1,0),{(1,0),(0,0)} }

Axiom 7.8 requires at least two dimensions (or two J-Curves). As to be shown, it captures the  underlying  essence of the Hawking-construction  for regular curves, which in the original text is scattered between local properties of the manifold –existence of local convex neighborhoods in terms of the Manifold-Topology, global causality-conditions –strong causality-, all needed to effectively define Regular Curves, and finally the properties  defined by the construction as such.

This ends our preliminary presentation of the new Axiom 7 Q-Loops.

Tuesday, April 14, 2009

The Great Simplification - Part I (corrections)

As stated clearly in the presentation, this BLOG documents Work in Progress, not at all final results. Some decades ago, when I started to study seriously Mathematics, I always wondered: how the hell were those powerful initial axioms and definitions found, which then gave origin to such powerful theories? Most of my teachers (and most Text-Books) presented only the final results in the sequence Axiom, Axiom, Definition, Definition, Lemma, Lemma, Lemma .. and out of box jumps a wonderful theorem. Actually it took me some time to discover that the way the final results were presented had little, if any, to do with how they had been constructed. Similar I suspect most traditional papers as published by Journals are –may be due to the harsh space-limits of Journals and time-limits of potential readers- rather an intent to impress than to explain making understand how. (I do doubt however that it truly saves time for the really interested reader, because unless he or she knew already, they have to reconstruct a living body of knowledge by analyzing only its bar bones). 

So here goes a new round in my construction process: While working on a new post about Dedekind-Completeness and its extension to partial orders and Q-orders I got stuck in some little, tiny detail: from Net-Topology I know that there should be never adjacent open or adjacent close points. And Hawking-Topology requires that pieces of world-lines have to be continuous, hence connected, images of [0,1], which boils down to exactly the same requirement as for Net-Topology, yet it turned out to be impossible to deduce this simple property from the Axiom VI in its former form. Reluctant to dump not only already written pages but some 20 Lemmata or so and their proofs, I tried first –what I suspect many do- to patch the initial Axioms –in this case Axiom 6   -the results can be seen in The trapped Arrow of Time  Part II- but finally decided to start over again.

The reasons: Though very common, in my feeling for esthetics these patches damage any beauty of a true Axiom-System. Second the former versions used a concept –Lines- that has its own flaws already on conceptual level: it requires the Axiom of Choice twice for its definition, which makes it very cumbersome to use later on and its in a way unphysical, as a physical process can never correspond to a geometric world-line, again a consequence of the uncertainty-principle, both considerations mentioned already in The trapped Arrow of Time - Part I.

Yet as both Hawking and Petri use Lines very intensively, I gave them at least a try, but finally decided to drop the concept Line as something fundamental and tried to rewrite Axiom 6 and Axiom 7 without using it … and received as gratification a great simplification without –as to be shown- loosing essence in modeling, that is if there are lines, then they behave as before.

Here the new Axiom 6 corrected
Axiom 6 Q-Topology            
axiom VI n    
The Axiom starts with the definition of U-neighborhoods of a point using the Q-Relation itself. Each element of neighborhood U shall by member of some limiting pair (A 6.1.1), for each member of a pair there shall be a representative (A 6.1.2), and finally a point of the same interval closer to the point than some already included, shall be likewise included (A 6.1.3). For the notion closer see Going Backward, Going Forward - Part I and Going in Circles - Part II.

The picture below may be helpful:
Grid Open Intervalls
A blue neighborhood { (0,0), {(0,0),(-1,1)}, {(0,0),(-1,-1)},)}, {(0,0),(1,-1)},)}, {(0,0),(1,1)} }, a red neighborhood { {(0,0),(0,1)}, (0,1), (0,0) }.
The color-families green and purple represent similar as before related pairs w.r.t. the point (0,0).  

Lemma S 1 shows that arbitrary unions and meets of two U –neighborhoods are U –neighborhoods.    corrected 
lemma s 1   
Axiom 6.2-6 Q-Topology    
axiom VI n 2-6   
As before, we introduce closed points as all those which can be distinguished from any other point by using the Q-Relation. An open set of the Q-topology includes for each closed point  a U-neighborhood. Lemma S 1 essentially proves already that O is a topology. We claim consistency of concepts: what is closed as Q-Relation shall be closed as Q-Topology. The complete cone of a point shall be open.

Here ends the presentation of the simplified Axiom 6.

Saturday, April 4, 2009

A Glimpse of the Big Picture

I’ve been asked whether there is a single text, that comprises the most essential of Q-Orders. Well – there is not, or not yet. To get an idea of what is and what not yet, may be the below picture helps.

The Big Picture
The Big Picture

The left side shows in a very simplified manner the tower of mathematics beneath contemporary, classical General Relativity Theory. The right side, as far as I’m aware  less solidified and standardized yet, the tower of mathematics beneath contemporary usage of Petri-Nets in Informatics.

I had this picture already in my office about 30 years ago. Now, I thought, if one would like to relate seriously the truly interesting part on top of the left tower with something may be interesting on top of the right tower –meaning by proofs and not by analogies- then one would need a mathematical bridge between both towers, starting already on some quite low-level of both towers.  This is, where the work on Q-orders started. They should permit both types of domains –Real and Countable- and should produce one single category of a topology to relate both towers.

To complicate the issue, I discarded partial orders as the funding concepts, as both from Physics and Net-Theory we knew that it are the cycles, that produce basic invariants, on the very end even enable measurement: while we can measure our time as cycles, we can’t measure space without using time. [I know Carlo Rovelli will most strongly disagree].

I did know already the red elements towards the center, they were developed while I was still a GMD. 10 years later, Olaf Kummer and  Mark-Oliver Stehr (1) give a quite complete résumé of what has been found out. Yet –though published already in 1976- I was not aware of the proposal of Hawking et. al (1) for a New Topology for Space-Time, i.e. the blue elements towards the center.

May be if I’d known, my life’s history would have been different. Yet I did know already then that the basic invariant of embedding Petri-Nets into (1+n)-Vector Spaces seemed to be the group of conformal transformations. So after leaving GMD in 1985 I accumulated notes and proofs on predecessors for Q-Orders, yet without any serious break-through, still I succeeded giving Axioms 1-5 for Q-orders their current form. (See Going in Circles Part I to III).

After getting back more seriously, about in 2007 or so, and using the resource Internet (+some additional dollars, unfortunately many seminal papers are still sold, while they should be free for humanity), I got across the cited paper from Hawking and a later companion by David Malament (3). Suddenly there was a correspondence already worked-out: a structure on both sides, whose geometrical invariant is the conformal group.

So the only task remaining was to find an axiomatic definition for the Q-Topology, that covered both sides –the Hawking Topology and the Net Topology-. Yet Hawking and Malament use for definitions and proofs many features intrinsically related with lower parts of their tower, starting with the standard definition of paths, which carries automatically the Hausdorff-properties of Reals into the Topology to be defined, over concepts like locally convex, which make sense in a Real-Linear-Vector-Space setting, yet not in the right tower etc. etc. etc. So it took some time to get to the current axioms of Q-Topology (Axiom 6) and Q-Loop-Topology (Axiom 7), which both do only rely on concepts available on both sides of the Big Picture. (The trapped arrow of Time Part I-III).

So what is finished –at least in my electronic scrap-book- is the basic bridge. And I will continue to present its definition and related results during the next weeks. Specifically I will introduce a rich set of models, all by themselves important, for Q-Orders, which will construct something may be close to the fundaments for middle-tower that might be of some use by itself. Done, I’ll proceed to recompile the essays into one single paper, to be published may be through my arXiv account.

Yet I’m fully aware that there is still neither Informatics nor Physics in the picture, which both start on top of their respective towers. The fundamental problem: we have no tools yet to formulate equations or even quantify invariants. Though it’s known that the Hawking-Topology allows to reconstruct the metric and it’s known that the conformal group of transformation corresponds to a single central source of gravity, but as far as I know –and found googleing the Internet- nobody has investigate yet the full way back: i.e. given a Einstein (or Einstein-Cartan) field-equation and posing some reasonable constraints on its right –Energy-Tensor- side, what are the effects on the underlying Casual Structure? (Though Alfonso García-Parrado  and Miguel Sánchez (4) may give some hints).

As these tools are still missing, it doesn’t make much of sense either to speculate about the formal relations between the big tower underlying contemporary Quantum-Mechanics and the modest elements presented so far.

There is however the sketch of a work-program to complete and solidify the middle-tower, once finished the above presentation.  

  1. Going downwards, it appears attractive to introduce the concept of a Q-Manifold. The Q-Manifold will be defined using the Complex or Quaternion (alas 2 and 4 dimensional Minkowski-Spaces) as base-space, using q-continuous functions instead of the usual Euclidian ones. The advantage: it appears as if a Q-Manifold is automatically smooth.
  2. As defining measures on S1 is quite standard and the q-continuous images of S1 correspond to Q-Loops, it should be possible to get some notion of distance for Q-Loop-Spaces.
  3. Similar remembering that the Tangent-Space may be defined as local equivalence classes of paths (alas Q-loops in our model), it should be possible to have a sort of Tangent-Space for Q-Loop-Spaces.

My only hope: my conditions of work will permit to continue … and the readers of this BLOG don’t get too impatient too soon as truly seminal posts are still month away.

The Axioms 
q theory axiom text-001   

(1) Olaf Kummer, Mark-Oliver Stehr: Petri's Axioms of Concurrency - A Selection of Recent Results , Proceedings of the 18th International Conference on Application and Theory of Petri Nets, Toulouse, June 23-27, 1997, Lecture Notes in Computer Science 1248, © Springer-Verlag , 1997
(2) S. W. Hawking A.R. King and P. J. McCarthy, A new topology for curved space-time which incorporates the causal, differential and conformal structures Journal of Mathematical Physics Vol. 17, No 2, February 1976
(3) D. Malament, The class of continuous timelike curves determines the topology of spacetime Journal of Mathematical Physics, July 1977, Volume 18, Issue 7, pp. 1399-1404
(4) Alfonso García-Parrado, Miguel Sánchez,
 Further properties of causal relationship: causal structure stability, new criteria for isocausality and counterexamples, arXiv:math-ph/0507014v2