Towards a ”General Theory” of Concretization

Philipp Harland
June 2025

Email: [email protected]
Abstract
There are many different objects in mathematics that do not have
innate geometric interpretations that are ”direct” in some way – that is,
realizations that can be seen, intuitively, as different ”forms” of that object.
This paper, building upon previous work and integrating some new ideas
seeks to provide a general framework for doing so.

Contents
1 Introduction 1

2 Results of Research 2
2.1 Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.1.1 Regular Graphs . . . . . . . . . . . . . . . . . . . . . . . . 2
2.1.2 Hypergraphs . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.2 Manifolds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.3 Algebraic Structures . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.3.1 Finite Groups . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.3.2 Other Algebraic Structures . . . . . . . . . . . . . . . . . 5
2.4 Tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.5 Some commentary . . . . . . . . . . . . . . . . . . . . . . . . . . 5

3 Works Referenced/Utilized 5

1 Introduction
Def. 1.1. Given a category, C, the concretization of C, which we denote C Conc ,
is given by a bijective functor, F: C → K ⊂ SetKn , where K is the smallest
subset of Set up to a ”natural” equivalence relation for which a bijection exists.
Here, K is assumed to either be an infinite ring or an infinite field equipped with
a metric, M. The functor F has to satisfy a key property, namely, isomorphisms
in C Conc have to directly correspond to isomorphisms in C Conc , and so do em-
bedding maps, homomorphisms, etc.

1
Additionally, there has to exist a functor, F −1 , which gives what we call ”re-
construction”, that satisfies the diagram:

F ◦F −1 F −1 ◦F

F
C Conc C
−1
F

We will develop many forms of this functor that apply to various objects through-
out this paper, for example for finite groups in 2.3.1.
We generally assume K = R in this paper for familiarity, but the theory can
also be imported to other fields/rings.

2 Results of Research
2.1 Graphs
2.1.1 Regular Graphs
As expounded on in [Har25a] and [Har25c], we can map a graph, G, to R2 by
constructing a set of points equipped w/ a set of functions, F, which we denote
GConc that follows the axioms:

i : |GConc | = #V (G) (1)

ii : fa,b (aConc
G , bConc
G ) = 1 ⇐⇒ aG adj bG (2)
A different but related concept is ”graph reconstruction”, which provides an iso-
morphism between Graph and ConcGraph.
Def. 2.1.1.1. The reconstruction of a set of points equipped with functions com-
patible with 4 (in the sense that they take on binary values and have the mutual
domain of the set’s points), is the graph isomorphic to it when concretized.

2.1.2 Hypergraphs
As well as the procedure for regular graphs, [Har25a] defined a mode of con-
cretizing hypergraphs.
Def. 2.1.2.1. Given a hypergraph, J, its concretization is a set of points
equipped with a set of functions, F, that follows the axioms:

i : |J Conc | = #V (J) (3)

ii : fa1J ,...,ak J (a1 Conc

J , ak Conc
J ) = 1 ⇐⇒ [adj(a1J , ..., akJ )] = 1 (4)
where [P ] is the Iverson function of P. This is only convenient if the hyperedges
have a Sfixed rank. However, this is an easy fix if we consider J Conc as the
m Conc
union: i=1 J⟨i⟩ , where m is the maximal hyperedge rank of J, and J⟨d⟩ is the
subhypergraph of J s.t. all of its hyperedges are of fixed rank d..

2
2.2 Manifolds
[Har25c] detailed a method of concretizing manifolds which is analogous to the
way we concretize graphs, utilizing the concept of a triangulation. The original
definition was for 2-manifolds, but it can also be applied to 1- or higher dimen-
sional manifolds.
Def. 2.2.1. Given a triangulation, TM of a k-manifold, M, the fineness of TM
w.r.t. a k-ball, O, is the limit:
n
X #vert (O : TM )
F (TM ) = lim (5)
n→∞
i=1
n

where we take the sum over random samplings of positions of O. Here, #vert (O :
TM ) counts the number of intersections. See [Har25c].
We will spare the details of constructing a concretization of a manifold which
were given in the aforementioned paper, except to say that we can generalize the
construction to n dimensions (or, equivalently, n-1 topological dimensions).
Now, we can build some theory:
Def. 2.2.2. Given two points, a, b, on M, the distance of a and b on a (k − 1)-
manifold, M, is defined as:
v  m
Z 1 u k 2X −1     
uX
t (γa,b )′x (t)2  dt ≡ lim 2 i i+1
L γ a,b , γa,b (6)
0 i=1
i
m→∞
i=0
2m 2m

where γa,b : [0, 1] → M Conc is the shortest curve on M connecting a, b, in that it

is the shortest path satisfying γa,b (0) = a, γa,b (1) = b. We may also generalize
the above to other metrics, such as the Manhattan metric or Lq metrics.
Def. 2.2.3. Given a hollow manifold, M, a parametrization of M is a contin-
uous function, f, defined over the interval [0, 1] → M Conc s.t. t∈[0,1] f (t) =
S

M Conc precisely.
Def. 2.2.4. Given 2 (solid) concrete manifolds of topological dimension, k,
S∂M ,T∂N
M, N, their concrete connected sum M + Conc N at two projections of open
k-balls onto their boundaries, S∂M , T∂N , is, up to homeomorphism, defined as:
[
{M, N } ∪ ft (S, T ) (7)
t∈[0,1]

where ft (a, b) = (1−t)a+tb. We can also consider this the union of the actions of
all homotopy shields w.r.t. all parameters t and all ordered pairs in S∂M × T∂N .
Def. 2.2.5. Two concrete manifolds, M, N, are homeomorphic if Rec(M ) ∼ =
Rec(N ).

2.3 Algebraic Structures

2.3.1 Finite Groups
The concretization theory of finite groups is quite simple to develop. As a basis,
we start with representation theory. Say, a group G has a representation at a

3
vector space V, which is n-dimensional, and we have a representation of G as a
subgroup of GL(V ).
Def. 2.3.1.1. Given a matrix, M, the row major vectorization (RMV) of M is
the mapping:
   
m11 . . m1n m11
 . . 
 −→  .  = MRM V
 
M =  . (8)
.   . 
mn1 . . mnn mnn

It is an immediate consequence thst M ∈ M atn×n (F ) =⇒ MRM V ∈ V ecn2 (F ).

Def. Ex. 1.  
  a
a b b
→  (9)
c d c
d
Similarly to how we can explicitly define a relation, R, on a set by defining a
union of tuples of elements, we can do the same for a group:
Def. 2.3.1.2. The semi-concretization of a group, G, is the set of 3-tuples
{(a, b, g)|g = a ◦G b}, which can essentially be thought of as a specific coproduct
of direct products over SetG which satisfies a specific set of properties. This can
be concisely summarized:

a
GSemiConc = V (a, b, c)
a,b∈G,π1 (V )◦π2 (V )=c=a◦G b
a G
rightb c (10)
π3 ◦f
π1 ◦f where, in the diagram to the left, G lefta
G
lefta V f k describes the left group multiplication
of a in G, and G righta describes the
π2 ◦f right group multiplication of a in G.
b

.
Def. 2.3.1.3. Given a group, G, its full concretization at V is, given a vec-
tor space, V, which represents G, the semi-concretization of K, where K =
(GL(V )RM V , ∗), with the product defined as a∗b = RM V (RM V −1 (a)·RM V −1 (b)).
In the case where a finite group does not have a representation theory (it is non-
linear), we may use other methods such as constructing the semi-concretization
of a group which it is isomorphic to by virtue of its presentation.
Even though representation theory is not generally well defined for InfGrp,
there are a few special cases for which it works, e.g. (Z, ·) identified with the
representation over GL2 (R). We will briefly summarize the construction, in-
voking the FTA due to Gauss.
Thm. 2.3.1.1.Qm (Gauss) Every integer, n, greater than 2, has a unique factor-
ization, n = i=1 pekii , ki , ei ∈ Z+ ∀i.

4
  √ 
p (p + p)
We can assign every prime, p, a matrix Ap = √ . Then,
(p − p) p
this representation of Z is generated by {A2 , A3 , A5 , ..., Api , ...}.

2.3.2 Other Algebraic Structures

We can also construct an algebraic structure as a set of 3-tuples, which will
work for any universal algebraic variety (in the sense of universal algebra, see
[BS81]), F. We denote the category of all such algebraic structures as Cat(F)
with morphisms defined by homomorphisms in the variety.
However, this only works for 2-ary algebraic structures, i.e. structures with
operators who have 2 inputs. (for a treatment of such structures, see [BS81])
Though, it is easy to create an analogous construction. We construct the set:

(M, •)Conc = {(a1 , ..., ak , b)|b = •(a1 , ..., ak )} (11)

where (M, •) is an algebraic structure, not necessarily of any specific universal

algebraic variety, equipped with the k-ary operation •.

2.4 Tensors
In N
a mathematical
Ns context, the most familiar definition of a tensor is an element
r
of i=1 V ⊗ r=1 V ∗ for a vector space, V. However, we can realize a different
way of constructing tensors which fits in line with the rest of our ”concretization”
theory:
k
Lk 2.4.1. A concrete tensor+ is a ”sublattice” of Z Q
Def. , which is equivalent to
k
(F +
i=1 di M ink {1}), ∀i di ∈ Z , equipped with a set of i=1 di functions which
+ k
are elements of the function space K ((Z ) )
for a field K, and which are denoted
Km1 ,...,mk with indices 1 ≤ mi ≤ di ∀i ∈ {1...k}.

2.5 Some commentary

Conj. 2.5.1. Are there any classes of objects for which no concretizations
exist?
The evidence seems to point to all objects seemingly being concretizable (under
the axioms of Def. 1.1., that is). However, there is still room for concern that
these convenient constructions may not be applicable to every type of object.

3 Works Referenced/Utilized
[Har25a] - Philipp Harland, A Method for Mapping Graphs to Real Metric Space
[Har25b] - Philipp Harland, Homotopy shields in Rn
[Har25c] - The Fundamental Group of a Graph
[BS81] - S. Burris, H. P. Sankappanavar, A Course In Universal Algebra