Perfect set Encyclopedia: Wikipedia & Grokipedia

Main page

Perfect set

from Wikipedia

In general topology, a subset of a topological space is perfect if it is closed and has no isolated points. Equivalently: the set $S$ is perfect if $S=S'$ , where $S'$ denotes the set of all limit points of $S$ , also known as the derived set of $S$ . (Some authors do not consider the empty set to be perfect.^[1])

In a perfect set, every point can be approximated arbitrarily well by other points from the set: given any point of $S$ and any neighborhood of the point, there is another point of $S$ that lies within the neighborhood. Furthermore, any point of the space that can be so approximated by points of $S$ belongs to $S$ .

Note that the term perfect space is also used, incompatibly, to refer to other properties of a topological space, such as being a G_δ space. As another possible source of confusion, also note that having the perfect set property is not the same as being a perfect set.

Examples

[edit]

Examples of perfect subsets of the real line $\mathbb {R}$ are the empty set, all closed intervals, the real line itself, and the Cantor set. The latter is noteworthy in that it is totally disconnected.

Whether a set is perfect or not (and whether it is closed or not) depends on the surrounding space. For instance, the set $S=[0,1]\cap \mathbb {Q}$ is perfect as a subset of the space $\mathbb {Q}$ but not perfect as a subset of the space $\mathbb {R}$ , since it fails to be closed in the latter.

Connection with other topological properties

[edit]

Every topological space can be written in a unique way as the disjoint union of a perfect set and a scattered set.^[2]^[3]

Cantor proved that every closed subset of the real line can be uniquely written as the disjoint union of a perfect set and a countable set. This is also true more generally for all closed subsets of Polish spaces, in which case the theorem is known as the Cantor–Bendixson theorem.

Cantor also showed that every non-empty perfect subset of the real line has cardinality $2^{\aleph _{0}}$ , the cardinality of the continuum. These results are extended in descriptive set theory as follows:

If X is a complete metric space with no isolated points, then the Cantor space 2^ω can be continuously embedded into X. Thus, X has cardinality at least $2^{\aleph _{0}}$ . If X is a separable, complete metric space with no isolated points, the cardinality of X is exactly $2^{\aleph _{0}}$ .
If X is a locally compact Hausdorff space with no isolated points, there is an injective function (not necessarily continuous) from the Cantor space to X, and so X has cardinality at least $2^{\aleph _{0}}$ .

Notes

[edit]

^ Jech, Thomas (2003). Set Theory (The Third Millenium Edition, revised and expanded). Springer Monographs in Mathematics. Springer. p. 40. doi:10.1007/3-540-44761-X. ISBN 978-3-540-44085-7.
^ Engelking, problem 1.7.10, p. 59
^ "Uniqueness of decomposition into perfect set and scattered set - Mathematics Stack Exchange".

References

[edit]

Engelking, Ryszard (1989). General Topology. Berlin: Heldermann Verlag. ISBN 3-88538-006-4.
Kechris, A. S. (1995), Classical Descriptive Set Theory, Berlin, New York: Springer-Verlag, ISBN 3540943749
Levy, A. (1979), Basic Set Theory, Berlin, New York: Springer-Verlag
Pearl, Elliott, ed. (2007), Open problems in topology. II, Elsevier, ISBN 978-0-444-52208-5, MR 2367385

Revisions and contributors Edit on Wikipedia Read on Wikipedia

Perfect set

View on Grokipedia

from Grokipedia

In topology, a perfect set is a nonempty subset of a topological space that is closed and contains no isolated points, meaning that every point of the set is a limit point of the set.^[1]^[2] Equivalently, a perfect set equals its own derived set, consisting entirely of accumulation points.^[3] Perfect sets exhibit several important properties, particularly in the context of real analysis and metric spaces. In the real line

\mathbb{R}

with the standard topology, every nonempty perfect set is uncountable, as established by Georg Cantor.^[2] They are also compact if bounded, and in complete metric spaces, perfect sets can be uncountable while having Lebesgue measure zero, highlighting their role in illustrating pathological behaviors of sets.^[4] A canonical example of a perfect set is the Cantor set, constructed by iteratively removing middle-third open intervals from the unit interval

[0,1]

, resulting in a closed, nowhere dense set that is uncountable and totally disconnected.^[5] Perfect sets are foundational in descriptive set theory, where they underpin the perfect set property: certain classes of definable sets of reals, such as analytic sets, are either countable or contain a nonempty perfect subset.^[6] This property connects perfect sets to broader questions in set theory, including the continuum hypothesis and the regularity of Borel and projective sets.

Definition and History

Formal Definition

In topology, a point

x

in a topological space

X

is defined as a limit point of a subset

P \subseteq X

if every open neighborhood of

x

contains at least one point of

P

distinct from

x

itself.^[7]^[8] The set of all limit points of

P

, denoted

P'

, is called the derived set of

P

. A subset

P

of a topological space

X

is perfect if

P

is closed in

X

and every point of

P

is a limit point of

P

, meaning

P = P'

.^[9] Equivalently,

P

is perfect if it is closed and equals its own derived set.^[10] This formulation assumes familiarity with the basic concepts of topological spaces, such as open and closed sets.^[9]

Historical Context

The concept of a perfect set was introduced by Georg Cantor in 1883 as part of his foundational work on point-set topology.^[11] In his paper "Über unendliche, lineare Punktmannichfaltigkeiten. V," Cantor defined a perfect set as one that coincides with its derived set, meaning every point in the set is a limit point of the set itself.^[11] This definition emerged from Cantor's innovative use of derived sets, a concept he had earlier developed in 1872 to study accumulation points in the real line.^[12] Cantor's motivation for this development stemmed from his investigations into the structure of closed sets on the real line, particularly in the context of uniqueness theorems for trigonometric series representations.^[12] These studies required a deeper understanding of infinite point sets and their limit points to address questions about convergence and representation in analysis.^[13] In the same 1883 paper, Cantor established a key result: every closed subset of the real line can be expressed as the disjoint union of a perfect set (the "perfect kernel") and at most a countable set of isolated points.^[11] This theorem highlighted the role of perfect sets as the uncountable "core" of closed sets, laying groundwork for later decompositions in descriptive set theory. The term "perfect" was chosen by Cantor to emphasize the set's completeness with respect to limit points, reflecting its dense-in-itself nature without isolated elements.^[14] The concept gained broader significance in general topology through Felix Hausdorff's seminal 1914 monograph Grundzüge der Mengenlehre, where perfect sets were integrated into the axiomatic framework of abstract spaces, extending Cantor's ideas beyond the real line to arbitrary topological structures.^[15]

Basic Properties

Closure and Limit Points

A perfect set

P

in a topological space is characterized by the equality

P = P'

, where

P'

denotes the derived set of

P

, consisting of all limit points of

P

. This condition ensures that no point in

P

is isolated: for every

x \in P

and every open neighborhood

U

x

, the intersection

U \cap (P \setminus \{x\})

is nonempty.^[16] The absence of isolated points means that every point in

P

serves as an accumulation point for other elements of the set, reflecting a dense-in-itself structure intrinsic to the subspace topology on

P

.^[17] From

P = P'

, it follows directly that the closure of

P

, denoted

\mathrm{cl}(P) = P \cup P'

, equals

P

itself, confirming that every perfect set is closed in the ambient space.^[18] This closure property arises because

P

already contains all its limit points, preventing any "gaps" that would require adjoining additional points to achieve closedness. The concept of perfect sets, including their closure and limit point properties, was foundational in Georg Cantor's early development of point-set topology.^[18] The equality

P = P'

also implies stability under iteration of the derived set operator: the first derived set

P' = P

, the second

P'' = (P')' = P

, and inductively, the

n

th derived set

P^{(n)} = P

for all finite ordinals

n

. This persistence extends to transfinite iterations in the Cantor-Bendixson derivative process, where

P

remains invariant. In opposition to scattered sets—those for which iterated derived sets eventually empty—a perfect set maintains its full structure as its own perfect kernel under derivation.^[19]

Density Characteristics

A perfect set

P

in a topological space is dense-in-itself, meaning that every point of

P

is a limit point of

P

, or equivalently,

P

contains no isolated points.^[20] In the subspace topology induced on

P

, this property ensures that every nonempty open subset of

P

is infinite, as a finite nonempty open set would consist of isolated points in the subspace, contradicting the absence of isolated points.^[20] The subspace topology on a perfect set

P

inherits the perfectness from the ambient space:

P

is closed in itself, and every point remains a limit point within the subspace, preserving the defining characteristics.^[20] Nowhere dense perfect sets exist in complete metric spaces; such sets are closed with no isolated points but have closures with empty interior, meaning they contain no nonempty open intervals in the ambient space.^[21] Perfect sets can be meager (of first category), as a nowhere dense perfect set is itself meager yet uncountable, illustrating that uncountability does not preclude first-category status in spaces like the reals.^[21] If a perfect set

P

is dense in the ambient space

X

, then

X

has no isolated points: any isolated point

x \in X

would yield an open neighborhood containing only

x

, but density of

P

requires it to intersect every nonempty open set, leading to a contradiction whether

x \in P

or not.^[20]

Examples and Constructions

Canonical Examples

One of the simplest canonical examples of a perfect set in the real line is a closed interval, such as

[0,1]

. This set is closed in

\mathbb{R}

and contains no isolated points, as every point in

[0,1]

is a limit point of the set; for instance, any neighborhood of an interior point contains other points of the interval, and endpoints like 0 have points arbitrarily close within the set from the right.^[22] The ternary Cantor set, often simply called the Cantor set, provides another fundamental example of a perfect set in

[0,1]

. It is constructed iteratively starting from the closed interval

C_0 = [0,1]

: at the first stage, remove the open middle third

(1/3, 2/3)

, leaving two closed intervals

[0,1/3] \cup [2/3,1]

; at the second stage, remove the open middle third of each remaining interval, yielding four closed intervals of length

1/9

; continue this process indefinitely, with the

n

th stage removing

2^{n-1}

open intervals of length

3^{-n}

each. The resulting Cantor set

C = \bigcap_{n=0}^\infty C_n

is closed as an intersection of closed sets, compact by the Heine-Borel theorem, uncountable, of Lebesgue measure zero, and perfect because it has no isolated points—every point in

C

can be approached by sequences of endpoints from the construction approximating ternary expansions using only digits 0 and 2.^[23] Variants of the Cantor set, such as the middle-fifths Cantor set, also serve as canonical perfect sets on the real line. This set is built similarly by iteratively removing the open middle fifth from each remaining closed interval starting in

[0,1]

, resulting in a compact, totally disconnected, perfect set homeomorphic to the ternary Cantor set but still of measure zero.^[24] The Smith-Volterra-Cantor set, known as a "fat" Cantor set, offers a contrasting canonical example that is perfect yet has positive measure. Constructed by removing the middle quarter from

[0,1]

initially (leaving intervals of total length 3/4), followed by removing middle intervals of length

1/4^n

from each subinterval at stage

n

(with total removed measure summing to

1/2

), the set is closed, contains no isolated points, and has Lebesgue measure

1/2

, while remaining nowhere dense and totally disconnected.^[24]

Methods of Construction

The Cantor-Bendixson construction provides a systematic way to obtain the perfect kernel of any closed set in a second-countable topological space by iteratively removing isolated points via the derivative operator. For a closed set

A

, the first derivative

A'

consists of all limit points of

A

, excluding any isolated points. Successor derivatives are defined as $(A^{(\alpha+1)})' = (A^{(\alpha)})' $, with

A^{(0)} = A

, while at limit ordinals

\lambda

A^{(\lambda)} = \bigcap_{\beta < \lambda} A^{(\beta)}

. This transfinite process continues until reaching the smallest ordinal

\theta

such that

A^{(\theta)} = A^{(\theta+1)}

; the fixed point

P = A^{(\theta)}

is then the largest perfect subset of

A

, and

A \setminus P

is countable.^[25] Homeomorphic images offer another fundamental technique for generating perfect sets, as the perfect property is preserved under homeomorphisms. If

P

is a perfect subset of a space

X

and

f: X \to Y

is a homeomorphism, then

f(P)

is closed in

Y

and inherits the absence of isolated points from

P

, since homeomorphisms preserve limits and openness in the subspace topology. More broadly, if

f: P \to Y

is a continuous bijection onto a closed subset of

Y

, the image

f(P)

remains perfect, enabling the transfer of perfect sets between spaces.^[26] Product constructions extend perfect sets to higher dimensions within compact Hausdorff spaces under the product topology. The Cartesian product

P \times Q

of perfect sets

P \subseteq X

and

Q \subseteq Z

, where

X

and

Z

are compact Hausdorff, is itself closed and compact by Tychonoff's theorem, hence closed in the product space. Moreover,

P \times Q

has no isolated points: for any

(p, q) \in P \times Q

, sequences approaching

p

P

while holding

q

fixed, or vice versa, ensure that every neighborhood of

(p, q)

contains infinitely many other points, leveraging the limit point density in each factor. This yields a perfect set of cardinality

2^{\aleph_0}

if both factors are uncountable.^[27] Embedding perfect sets into

\mathbb{R}

can be achieved using binary expansions tied to tree structures on sequences. Identify subsets of

[0,1]

with their binary representations via the map sending a sequence

\sigma \in 2^\mathbb{N}

\sum_{n=1}^\infty \sigma(n) 2^{-n}

. To construct a perfect subset, start with a tree

T \subseteq 2^{<\mathbb{N}}

that is closed (corresponding to a closed set in the Cantor space) and has no isolated nodes, ensuring infinite splitting; the body

[T]

of such a tree is perfect in the product topology. An injective continuous map from the full Cantor space

2^\mathbb{N}

into

[T]

, built by recursively assigning incompatible extensions to binary branches at splitting nodes, embeds a perfect set homeomorphic to the Cantor set into

\mathbb{R}

. Similar tree-based methods using continued fractions construct perfect sets of irrational numbers by restricting expansions to avoid finite terminations.^[28] The transfinite aspect of the Cantor-Bendixson iteration guarantees completeness for complex closed sets, with the derivative rank bounded by

\omega_1

in second-countable spaces, isolating the perfect kernel without overlooking limit points accumulated over uncountable steps. This ordinal-indexed process formalizes the removal of scattered parts, leaving only the non-scattered perfect core.^[25]

Advanced Properties

Cardinality in Metric Spaces

In complete separable metric spaces, also known as Polish spaces, every nonempty perfect set is uncountable. This result is a cornerstone of descriptive set theory and follows from the fact that such a set contains a homeomorphic copy of the Cantor space

2^\mathbb{N}

, which is uncountable.^[29] The proof proceeds by constructing a binary tree of nonempty basic open neighborhoods within the perfect set: starting from a point in the set, split into two disjoint nonempty relatively open subsets at each level, ensuring the branches yield a continuous embedding of the Cantor space due to completeness and separability.^[29] In particular, within the Euclidean space

\mathbb{[R](/page/R)}

, which is a Polish space, every nonempty perfect set

P

has cardinality at least the continuum, satisfying

|P| \geq 2^{\aleph_0}

. A canonical example is the middle-thirds Cantor set, which is perfect, compact, and has exact cardinality

2^{\aleph_0}

, the same as that of

\mathbb{[R](/page/R)}

, via a surjective continuous map from the Cantor set onto

[0,1]

.^[30] Countable perfect sets cannot exist in Polish spaces; any countable closed set in such a space must have isolated points, contradicting the perfect condition.^[29] In more general metric spaces lacking completeness, however, countable perfect sets are possible—for instance, the space of rational numbers

\mathbb{Q}

with the standard metric is countable, closed in itself, and has no isolated points, making it perfect.^[31] Nonetheless, the uncountability theorem holds specifically in the complete separable case, with emphasis on Euclidean spaces where all nonempty perfect sets exhibit at least continuum cardinality.

Measure and Dimension Aspects

Perfect sets in

\mathbb{R}

exhibit a wide range of behaviors with respect to Lebesgue measure. The classical middle-thirds Cantor set is a compact perfect set with Lebesgue measure zero.^[32] In contrast, the closed interval

[0,1]

is a perfect set with positive Lebesgue measure equal to 1.^[33] Furthermore, there exist perfect sets with positive Lebesgue measure that contain no nonempty open intervals, demonstrating that positive measure does not imply the presence of intervals within the set.^[33] A canonical example of such a positive-measure perfect set without interior is the Smith-Volterra-Cantor set, also known as the fat Cantor set. This set is constructed by starting with the interval

[0,1]

and iteratively removing open intervals from the middle of each remaining closed interval, but with lengths chosen to ensure the total removed measure is

1/2

: at stage

n

2^{n-1}

open intervals each of length

1/4^n

are removed, yielding a limiting set of Lebesgue measure

1/2

.^[34] This construction modifies the standard Cantor set process by removing progressively smaller proportions, preserving positive measure while maintaining the perfect property through closure and absence of isolated points.^[34] Regarding null sets, uncountable perfect sets of Lebesgue measure zero exist, as exemplified by the Cantor set, which contrasts with countable sets of measure zero that cannot be perfect due to the presence of isolated points. Such uncountable null perfect sets highlight the distinction between topological density (every point is a limit point) and measure-theoretic size. In terms of fractal dimensions, the Hausdorff dimension provides a finer measure of "size" for perfect sets beyond Lebesgue measure. For the middle-thirds Cantor set

C

, the Hausdorff dimension is given by

\dim_H C = \frac{\log 2}{\log 3} \approx 0.631,

computed via the self-similarity of the set, where the dimension

s

satisfies

2 \cdot (1/3)^s = 1

. This value lies between 0 and 1, reflecting the set's zero Lebesgue measure yet substantial topological complexity. More generally, perfect sets can achieve a spectrum of Hausdorff dimensions from 0 to 1, with positive values common in standard constructions like the Cantor set.^[35]

Connections to Other Concepts

Relation to Compactness

In metric spaces, compact perfect sets are closed subsets with no isolated points that are also totally bounded, ensuring every open cover has a finite subcover. Such sets are necessarily uncountable, as the absence of isolated points implies infinite density around every point, preventing countability.^[22] A canonical example is the Cantor set, which is compact, perfect, totally disconnected, and uncountable while having Lebesgue measure zero.^[23] Not all perfect sets are compact, however. The real line

\mathbb{R}

, equipped with the standard topology, is perfect because it is closed in itself and has no isolated points—every neighborhood of any point contains infinitely many other points—but it fails to be compact due to lacking boundedness, as demonstrated by the open cover

\{(-n, n) \mid n \in \mathbb{N}\}

having no finite subcover.^[36] In

\mathbb{R}^n

with the Euclidean topology, the Heine-Borel theorem characterizes compactness for closed sets as equivalent to boundedness, so a perfect set in

\mathbb{R}^n

is compact precisely when it is bounded.^[22] In general topological spaces, a perfect set remains defined as a closed set without isolated points, meaning it equals its own derived set.^[37] When additionally compact, it inherits the property that every open cover admits a finite subcover, a condition stricter than perfection alone, as compactness controls the "size" of the space via covering properties rather than merely ensuring density of limit points.^[38] This distinction highlights that perfection addresses local density, while compactness imposes a global finiteness constraint.

Role in Decomposition Theorems

The Cantor–Bendixson theorem states that in a second-countable topological space, every closed set

F

can be uniquely decomposed as the disjoint union

F = P \cup C

, where

P

is a perfect set (possibly empty) and

C

is at most countable.^[39] This decomposition isolates the "perfect kernel"

P

F

, which captures its non-scattered structure. The proof proceeds by transfinite iteration of the derived set operator: starting from

F^{(0)} = F

, define $F^{(\alpha+1)} = (F^{(\alpha)})' $ (the set of limit points of

F^{(\alpha)}

) for successor ordinals

\alpha

, and

F^{(\lambda)} = \bigcap_{\beta < \lambda} F^{(\beta)}

for limit ordinals

\lambda

; this process stabilizes at some countable ordinal

\rho

, yielding the perfect kernel

P = F^{(\rho)}

and the scattered part

C = F \setminus P

.^[39] This theorem has key applications in classifying scattered sets, which are precisely those closed sets with empty perfect kernel (i.e.,

P = \emptyset

), and in assigning a Cantor–Bendixson rank to points or sets, measuring the "height" of the iteration needed to remove them; for well-ordered sets like ordinals, the rank corresponds to the ordinal itself.^[39] The theorem generalizes to Polish spaces (complete separable metric spaces), where every closed subset admits such a decomposition, implying that every uncountable closed set contains a nonempty perfect subset.^[39] The full theorem was established by Georg Cantor in 1884, incorporating an observation by Ivar Bendixson from 1883, building on Cantor's introduction of perfect sets and derived sets in his contemporaneous work.^[40]^[41]

Talk Channels

Knowledge Base

Special Pages

Talk Channels

Knowledge Base

Special Pages

Perfect set

Perfect set

Perfect set

Examples

Connection with other topological properties

See also

Notes

References

Perfect set

Definition and History

Formal Definition

Historical Context

Basic Properties

Closure and Limit Points

Density Characteristics

Examples and Constructions

Canonical Examples

Methods of Construction

Advanced Properties

Cardinality in Metric Spaces

Measure and Dimension Aspects

Connections to Other Concepts

Relation to Compactness

Role in Decomposition Theorems

References