1. Introduction
A new classification of topological spaces based on
the finite coarse shape theory was introduced in [
1]. Although this theory is abstract and makes sense for any pair
consisting of the category
C and its full and dense subcategory
D, the most important case is if
(
) (homotopy class of all (pointed) topological spaces) and
(
) (class of (pointed) topological spaces that have the homotopy type of some (pointed) polyhedron) because it yields the (pointed) topological finite coarse shape category
). So far, many interesting invariants of the finite coarse shape theory have been investigated. The most important of them are the (relative) finite coarse shape groups of (bi)pointed topological spaces.
In this paper, we introduce the notion of the finite coarse shape path between two points x and in a topological space X. The space X is said to be finite coarse shape path connected if for every pair of points in X there exists a finite coarse shape path in X between these points, and we show that being finite coarse shape path connected is a topological property (Corollary 1). We also investigate an interesting example of the solenoid , which is known to be coarse shape path connected but not shape path connected and prove that there are infinitely many finite coarse shape path components in .
Furthermore, by Theorem 1, we show that, for every , every finite coarse shape path induces a group isomorphism (for a basepoint preserving bijection) from to with some interesting and useful properties (Proposition 2). Example 2 shows that finite coarse shape groups of the same space, in general, depend on the choice of a base point. Hence, since these groups are finite coarse shape invariant, the pointed finite coarse shape type of , in general, depends on the choice of point x. However, there is a large class of topological spaces for which this is not true. We prove that if X is a finite coarse shape path-connected paracompact locally compact space, the pointed finite coarse shape type of does not depend on the choice of the point x (Theorem 2).
2. Preliminaries
Let us now recall the basics of the finite coarse shape theory. A finite ∗-morphism (i.e., ⊛-morphism) between inverse systems and in a category C is a pair which consists of a function (called an index function) and, for every , of a sequence of morphisms , , in C such that
for every pair
,
, there exist
,
, and
such that, for every
,
Under the composition of ⊛-morphisms
and
we understand a ⊛-morphism
such that
Given a category C, a category -C is defined having all inverse systems in C as objects and, for any pair of objects and , having all ⊛-morphisms between and as morphisms with the composition mentioned above as the categorial composition.
A ⊛-morphism
is said to be
equivalent to a ⊛-morphism
, denoted by
, if every
admits
,
, and
such that, for every
,
The relation ∼ is an equivalence relation on each set of ⊛-morphisms between two inverse systems in C. The equivalence class of will be denoted by .
By
-
C we denote a category whose class of objects consists of all inverse systems in
C and whose morphisms are all equivalence classes
. The categorical composition in
-
C is defined by the representatives, i.e.,
where
.
A faithful functor --C is defined using the joining which associates with each morphism of -C the -C morphism such that , for all , , while keeping inverse systems in C fixed. Thus, -C can be regarded as a subcategory of -C.
Analogously, a faithful functor --C is defined using the joining which associates with each morphism of -C the same morphism as morphism of -C, while keeping inverse systems in C fixed. Thus, -C is a subcategory of -C.
Let
C be any category and let
be full and dense subcategory of
C. Let
and
be two
D-expansions of the object
and let
and
be two
D-expansions of the object
. For a morphism
we say that it is
equivalent to a morphism
in
-
D, denoted by
, if
where
and
are isomorphisms of different expansions of
X and
Y, respectively. Note that ∼ is an equivalence relation such that
and
imply
when compositions
and
make sense. The equivalence class of a morphism
is denoted by
.
Using the relation ∼ of -D with each pair (where D is full and dense in C), we associate a category such that
- −
;
- −
For every pair of objects in , the set consists of classes of all -D-morphisms , where and are arbitrary D-expansions of X and Y, respectively;
- −
The composition of classes
and
is defined by
Category is called the abstract finite coarse shape category of a pair , while morphisms in are called finite coarse shape morphisms and denoted by . A finite coarse shape morphisms can be described using a diagram
It is important to emphasize that set is bijectively correspondent with a set - for any D-expansions and of objects X and Y, respectively.
Isomorphic objects X and Y in category are said to have the same finite coarse shape type. This is denoted by .
Mentioned functors --C and --C induce faithful functors and , respectively, by putting
- −
- −
- −
Hence, abstract shape category can be considered as a subcategory of the abstract finite coarse shape category , and is a subcategory of the abstract coarse shape category .
The composition of functors (shape functor) and is called the abstract finite coarse shape functor, denoted by .
Throughout this paper, observed categories C and D will be (, ) and (, ). In other words, we will deal with the ((bi)pointed) topological finite coarse shape category, briefly denoted by (, ).
Recall that the objects of are all the pointed topological spaces , , and morphisms are all the homotopy classes of mappings of pointed spaces , i.e., homotopy classes of functions satisfying . Analogously, objects of are all the bipointed topological spaces , , and morphisms are all the homotopy classes of mappings of bipointed spaces , i.e., homotopy classes of functions satisfying and . We will usually denote an H-map by omitting the brackets unless we need to especially highlight some mapping f and the corresponding homotopy class . When the object class is reduced to all (bi)pointed topological spaces with the homotopy type of some (bi)pointed polyhedron, we obtain a full subcategory ().
By Theorems 1.4.2 and 1.4.7 of [
2], every topological space
X admits an
-expansion
while every pointed topological space
admits an
-expansion
Moreover, by Lemma 2.3 of [
3], given polyhedral resolution
of
X, for every
morphisms
and
of
-
and
-
, respectively, are also resolutions (of
and
, respectively). Hence, by applying the homotopy functor
H to these resolutions, one obtains corresponding polyhedral expansions
and
, which are both determined by the
-expansion
of
X. This is why each finite coarse shape morphism
(morphism of
) can be regarded as a morphism of
between the corresponding pointed spaces (from
to
or from
to
).
For each
and every pointed topological space
the k-th finite coarse shape group is defined in the following way: for each
,
is a group (if
it is moreover an abelian group) having
as underlying set with a group operation defined by the formula
Here, the finite coarse shape morphisms
and
are represented by
-
morphisms
and
, respectively, where
is an
-expansion of
. Note that
denotes the sum in the corresponding
k-th homotopy group
. In the case when
,
is a pointed set of all finite coarse shape morphisms between
and
, i.e., the set
.
For every
and for every finite coarse shape morphism
, a homomorphism of finite coarse shape groups (for
a base point preserving function)
is defined by the rule
for any finite coarse shape morphism
. For every
, this induces a functor
(for
) associating with every pointed topological space
the
k-th finite coarse shape group
. The functor
is called
the k-th finite coarse shape group functor (see [
4] for details).
Finally, to facilitate readers’ access, let us state some basic facts about solenoids. Marked with
, where
is a sequence in
, solenoid is the inverse limit of the inverse sequence
, where
and
, for every
. In [
5], A. van Heemert proved that solenoids are indecomposable continua; hence, they are partitioned into disjoint composants. Krasinkiewicz and Minc proved in [
6] that solenoids are not weakly joinable between any two points belonging to different composants. Afterwards, Š. Ungar proved in [
7] that joinability and shape path connectedness coincide on the class of all metrizable continua; therefore, solenoids are not shape path connected. Nevertheless, according to [
3], Example 3.4, they are coarse shape path connected.
3. The Finite Coarse Shape Path Connectedness
Definition 1. Let X be a topological space, and let . A finite coarse shape path in X from to is a bipointed finite coarse shape morphism . Space X is said to be finite coarse shape path connected if for every pair there exists a finite coarse shape path in X from to .
Let
be a
-expansion of the space
X and let
such that there exist finite coarse shape paths
and
in
X from
to
and from
to
, respectively. Note that
and
are
-
morphisms such that, for all
and
, components
and
are homotopy classes of ordinary paths in
from
to
and from
to
, respectively. Now, the product
is defined as a finite coarse shape path
, where
denotes the homotopy class of the product of ordinary paths
and
.
Furthermore, a trivial finite coarse shape path is a bipointed finite coarse shape morphism represented by - morphism which consists of the trivial loops in for all and . Finally, an inverse of a finite coarse shape path is a finite coarse shape path represented by an - morphism , where denotes an inverse path of a path , for all and .
Remark 1. The composite of a finite coarse shape path and an appropriate bipointed finite coarse shape morphism is a finite coarse shape path. More precisely, if there exists a finite coarse shape path in X from to , then, for every bipointed finite coarse shape morphism , the composite is a finite coarse shape path in Y from to .
According to the previous remark, it is straightforward to prove the following proposition:
Proposition 1. Let X and Y be topological spaces such that for every pair there exists a pair and a finite coarse shape morphism . If X is finite coarse shape path connected, then so is Y.
An important consequence of the Proposition 1 is that the finite coarse shape path connectedness is a topological property.
Corollary 1. Let be a continuous surjection between topological spaces. If X is finite coarse shape path connected, then so is Y.
According to [
3], Example 3.4, solenoids are coarse shape path connected but not shape path connected. In the next example, we show that they are not finite coarse shape path connected either.
Example 1. As stated in the preliminaries, the common definition of the solenoid , where is a sequence in , is as the inverse limit of the inverse sequence of circles and bonding maps , given by , for every . Nevertheless, it can be obtained as the inverse limit of an inclusive inverse system of solid tori.
Let be a solid torus and, for each , let be a solid torus that is wrapped longitudinally times inside the solid torus without folding. In this manner, we get the sequenceand the intersection is homeomorphic to the solenoid . Moreover, , where , , is the inclusion. Indeed, for , where is the inclusion, and for any morphism in pro-Top we have , for every and every , . Therefore, for every and for every , and, consequently, . Now, defined by for satisfies , and it is unique since is injective for every .
Recall that, by [5], solenoids are indecomposable continua. By [8], indecomposable continua have composants, and in solenoids they coincide with the path components. Let be arbitrary elements in two distinct path components. It is obvious that Let us now assume that there is a finite coarse shape path from x to y,wheresuch that Since is a morphism in -, we have that for every there exists such that for every ,(as homotopy clases). Also, since , there is a stationary subsequence of . Let for some , for every . Then, for every there exists and such that , meaning that homotopy class contains a path between x and y in for every n and, consequently, a path between x and y in , which is a contradiction with the asumption that x and y are not in the same path component. Let us recall that an inverse system
is
n-movable,
, provided every
admits a
such that for every
, every polyhedron
P with
and every map
, there exists a map
such that
Now, a topological space
X is said to be
n-movable provided it admits an
-expansion
such that the inverse system
is
n-movable. Furthermore, for an
n-movable topological space, every
-expansion (i.e., an associated inverse system) is
n-movable.
Also recall that, for every metrizable continuum
X, the following statements are equivalent (Ch. VII. of [
9], ref. [
6] and II. 8. of [
2]):
There exists a point such that the pointed space is 1-movable.
For every point the pointed space is 1-movable.
X is shape path connected.
Therefore, by Corollary 3.9 of [
3], the following holds:
Corollary 2. On the class of all pointed 1-movable metrizable compacta, the shape path connectedness, finite coarse shape path connectedness, coarse shape path connectedness and connectedness are equivalent properties.
4. Isomorphisms Induced by Finite Coarse Shape Paths
Let and be arbitrary points in X and let be a finite coarse shape path in X from to . Then, there exists a - morphism such that , where is an -expansion of a bipointed space and, for all and , is a homotopy class of a path in from to .
Let us, for every
and an arbitrary
, define a joining
by the rule
where
is an induced homomorphism (for
a base point preserving function) induced by
.
To show that
, it sufficies to show that
is an
-
morphism. Since
and
are
-
morphisms, for every pair
there exists
such that
and
, for every
. Thus,
holds for every
. Finally, since for every
inequalities
and
hold, we infer that
. This proves that
is an
-
morphism, i.e.,
is a finite coarse shape morphism.
Theorem 1. Let be a finite coarse shape path in X from to . For every , the induces a group isomorphism (for a base point preserving bijection) given by the rule (2). Moreover, . Proof. We will first show that
is a homomorphism for every
. Let
be finite coarse shape morphisms represented by
-
morphisms
and
, respectively. Since (see [
4])
where
is the sum in the corresponding
k-th homotopy group
, it follows that
It is well known (from homotopy group theory) that for every path
, the induced homomorphism
is an isomorphism (for
a base point preserving bijection), and it holds that
. Thus,
is the inverse of
. This proves that
is a group isomorphism for every
(a base point preserving bijection for
). □
Proposition 2. A group isomorphism has the following properties:
, for every finite coarse shape morphism .
, for all finite coarse shape paths and in X from to and from to , respectively.
, where is the trivial finite coarse shape path at .
Proof. Each finite coarse shape morphism
of
, represented by a
-
morphism
, can be regarded as both
and
of
(we will denote all of these morphisms with the same label
, taking care of the context in which they are used). The finite coarse shape group functor
(see [
4]) associates each of them with corresponding homomorphism (for
with corresponding base point preserving function)
and
, respectively. Now, take an arbitrary
. It holds that
and
is proved.
Let
and
be finite coarse shape paths in
X from
to
and from
to
, respectively, and take an arbitrary
. Now
shows that (ii) holds true.
Finally, from
and Theorem 1, for every trivial finite coarse shape path
at
we infer that
holds. □
In [
2], Example II.3.4, the authors constructed a metric continuum in which the change of base point affects the pointed shape. In [
3], Theorem 2.5 states that, for the same continuum, there exists a pair of base points for which the pointed shape differs and the pointed coarse shape is the same. In the following example, we show that the change of base point also affects the pointed finite coarse shape.
Example 2. Let X be the wedge of the dyadic solenoid and 1-sphere with the identifying point *. The sphere is a retract of , and the inclusion induces a monomorphism on so .
We want to show that the finite coarse shape group of X depends on the choice of base point.
Let with the identifying point *, let , and let , and be paths in P as shown in the following figure.
The fundamental group is isomorphic to the free group on generators and .
As in [2], Example II.3.4, we define mapings , f taking the paths , , to , , , respectively and g taking the paths to , , , respectively. The dyadic solenoid is the inverse limit of the sequencewhere is the map that uniformly wraps the domain circle twice around the codomain circle, and one can see that the inverse limit of the sequenceis homeomorphic to for some . Therefore, is isomorphic to the inverse limit of the sequencewhere . We want to show that this inverse limit is trivial.
Let us assume the opposite, i.e., there is an element Here, for each , the i-th coordinate of x is actually a sequence of words in letters u and v such that Also, for every pair , , there exists such thatwhere is the bonding morphism. Since , there exists such that . Furthermore, there exists a subsequence of such that for every .
Without loss of generality, we can assume that is an odd integer.
Now, for we have that for every and there exists such that , . Consequently, there is such that Let us, for every non-zero word , define the integer as the maximal absolute value of the exponents of v appearing in the word y. It is obvious that . Also, let Now we can analyze for various forms of the word .
- 1.
If then , and - 2.
If we can assume that where . Then
- (a)
and - (b)
and - (c)
and - (d)
and
Therefore, if we take sufficiently large l, we can achievewhich is a contradiction. The previous example shows that finite coarse shape groups of the same space, in general, depend on the choice of a base point. Hence, since those groups are finite coarse shape invariant (see [
4]), the pointed finite coarse shape type of
, in general, depends on the choice of point
x. However, there is a large class of topological spaces for which this is not true—we prove that if
X is a finite coarse shape path connected paracompact locally compact space, then the pointed finite coarse shape type of
does not depend on the choice of point
x.
Theorem 2. Let X be a topological space admitting a metrizable polyhedral resolution, and let . If there exists a finite coarse shape path in X from to , then and have the same pointed finite coarse shape type, i.e., and are isomorphic in .
Proof. Let
,
and
be
-
(
-
) expansions of
,
and
, respectively. Let
be a finite coarse shape path in
X from
to
. In the analogous way as in the proof of Theorem II.8.9 of [
2] an isomorphism
of
-
and its inverse
are constructed. Isomorphisms
and
determine the pointed finite coarse shape (iso)morphisms
and
, respectively. Thus,
. □
Example 3. In the solenoid , for every two points belonging to the same composant, there is a finite coarse shape path from to . Then, by Theorem 2, and are isomorphic in .
Theorem 2 is obviously valid for all compact metric spaces. Moreover, by [
10], all paracompact locally compact spaces admit a metrizable polyhedral resolution. That fact yields the following corollary:
Corollary 3. Let X be a finite coarse shape path connected paracompact locally compact space. Then, the pointed finite coarse shape type of does not depend on the choice of point .
Remark 2. It would be interesting and useful to find, if it exists, an example of a space that is finite coarse shape path connected, but not shape path connected. However, that example must be rather exotic since it does not, by Corollary 2, belong to the class of pointed 1-movable metrizable compacta.