{
  "id": "accumulation-point",
  "title": "Accumulation Point",
  "lang": "en",
  "source_lang": "en",
  "summary": "An accumulation point of a set is a point whose every neighborhood contains a point of the set other than itself.",
  "category": "mathematics",
  "subcategory": "analysis",
  "type": "concept",
  "aliases": [
    "Limit point",
    "Cluster point"
  ],
  "generated_at": "2026-11-05",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/accumulation-point/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/accumulation-point/",
      "json": "https://aiwiki.page/zh/accumulation-point.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "An accumulation point of a set is a point that can be approached arbitrarily closely by points of that set other than the point itself. Formally, a point \\(x\\) of a topological space \\(X\\) is an accumulation point of \\(A\\subseteq X\\) if every neighborhood of \\(x\\) intersects \\(A\\setminus\\{x\\}\\). The point need not belong to \\(A\\). Accumulation points are fundamental in mathematical analysis and topology, where they describe the local distribution of sets and help characterize closure, convergence, and compactness."
    },
    {
      "heading": "Definition",
      "text": "Let \\(A\\) be a subset of a topological space \\(X\\). A point \\(x\\in X\\) is an accumulation point of \\(A\\) if\n\n\\[U\\cap\\bigl(A\\setminus\\{x\\}\\bigr)\\ne\\varnothing\\]\n\nfor every neighborhood \\(U\\) of \\(x\\). Equivalently, the condition may be checked only for open neighborhoods.\n\nThe removal of \\(x\\) is essential. Without it, every member of \\(A\\) would automatically satisfy the condition, regardless of whether any other points of \\(A\\) occur nearby.\n\nIn a metric space \\((X,d)\\), the definition becomes\n\n\\[\\forall\\varepsilon>0,\\quad \\exists a\\in A \\quad\\text{such that}\\quad 0<d(x,a)<\\varepsilon.\\]\n\nThus every open ball centered at \\(x\\) must contain a point of \\(A\\) distinct from \\(x\\). On the real line with its usual metric, this says that every interval \\((x-\\varepsilon,x+\\varepsilon)\\) contains such a point.\n\n“Limit point” is a common synonym. “Cluster point” is also used, but terminology varies, particularly when discussing sequences rather than sets."
    },
    {
      "heading": "Examples",
      "text": "A set with one accumulation point\n\nConsider\n\n\\[A=\\left\\{\\frac1n:n=1,2,3,\\ldots\\right\\}\\subseteq\\mathbb R.\\]\n\nIts only accumulation point is \\(0\\). Every interval around \\(0\\) contains \\(1/n\\) for sufficiently large \\(n\\), even though \\(0\\notin A\\).\n\nEach point \\(1/n\\), by contrast, has a sufficiently small neighborhood containing no other member of \\(A\\). Adding \\(0\\) to the set does not change its accumulation points.\n\nAn interval\n\nFor \\(A=(0,1)\\), the set of accumulation points is \\([0,1]\\). Both endpoints are accumulation points despite lying outside \\(A\\), while every interior point is approached by other points of the interval.\n\nRational numbers\n\nEvery real number is an accumulation point of the set of rational numbers \\(\\mathbb Q\\). Indeed, every nonempty open interval contains infinitely many rational numbers. The same statement holds for the irrational numbers.\n\nIntegers and finite sets\n\nThe integers have no accumulation points in \\(\\mathbb R\\): sufficiently small intervals around any real number contain at most one integer, and an interval avoiding that integer can be chosen when necessary.\n\nLikewise, a finite subset of a metric space has no accumulation points. This statement requires care in more general topological spaces, where finite sets can have accumulation points."
    },
    {
      "heading": "Closure, isolated points, and boundary",
      "text": "A point is an adherent point of \\(A\\) if every neighborhood intersects \\(A\\), without excluding the point itself. The set of all adherent points is the closure \\(\\overline A\\).\n\nWriting \\(A'\\) for the set of accumulation points of \\(A\\), one has, in every topological space,\n\n\\[\\overline A=A\\cup A'.\\]\n\nConsequently, \\(A\\) is a closed set if and only if it contains all its accumulation points.\n\nA point \\(a\\in A\\) is an isolated point of \\(A\\) if some neighborhood \\(U\\) satisfies\n\n\\[U\\cap A=\\{a\\}.\\]\n\nThus each point of \\(A\\) is either isolated in \\(A\\) or an accumulation point of \\(A\\), but not both. An accumulation point outside \\(A\\) belongs to the closure without being a member of the set.\n\nAccumulation points should also be distinguished from boundary points. A boundary point has every neighborhood intersecting both \\(A\\) and \\(X\\setminus A\\). An isolated point of \\(A\\) can be a boundary point without being an accumulation point; an interior point of an interval is an accumulation point without being a boundary point.\n\nThe ambient space matters. The set \\(\\{1/n:n\\geq1\\}\\) has \\(0\\) as an accumulation point in \\(\\mathbb R\\), but has no accumulation points when regarded as a space in its own right with the subspace topology."
    },
    {
      "heading": "Characterization by sequences",
      "text": "In metric spaces, accumulation points can be characterized through limits of sequences:\n\n\\[x\\in A' \\quad\\Longleftrightarrow\\quad \\text{there exists a sequence }(a_n) \\text{ of pairwise distinct points of }A\\setminus\\{x\\} \\text{ with }a_n\\to x.\\]\n\nFor the forward direction, progressively smaller balls around \\(x\\) provide the sequence. The points can be chosen distinct because every neighborhood of an accumulation point in a metric space contains infinitely many points of \\(A\\). Conversely, a sequence of points different from \\(x\\) converging to \\(x\\) meets every neighborhood of \\(x\\).\n\nThe same characterization holds in any first-countable \\(T_1\\) space. First countability supplies a countable local neighborhood base; the \\(T_1\\) condition, meaning that every singleton is closed, allows previously selected points to be excluded.\n\nAccumulation of a set versus clustering of a sequence\n\nA cluster point of a sequence is usually defined as a point whose every neighborhood contains terms at infinitely many indices. In a metric space, this is equivalent to being the limit of a subsequence.\n\nThis differs from an accumulation point of the sequence’s range. A constant sequence \\(a_n=a\\) converges to \\(a\\) and has \\(a\\) as a cluster point, but its range is the singleton \\(\\{a\\}\\), which has no accumulation points in a metric space. Sequence clustering counts repeated occurrences; accumulation of a set concerns distinct locations.\n\nBeyond first-countable spaces\n\nIn arbitrary topological spaces, sequences may not detect every accumulation point. The general replacement is a net:\n\n\\[x\\in A' \\quad\\Longleftrightarrow\\quad \\text{some net with values in }A\\setminus\\{x\\} \\text{ converges to }x.\\]\n\nA convergent net of this kind can be constructed by directing the neighborhoods of \\(x\\) by reverse inclusion and choosing a point of \\(A\\setminus\\{x\\}\\) in each neighborhood."
    },
    {
      "heading": "Separation assumptions and infinite neighborhoods",
      "text": "In metric spaces, “every neighborhood contains another point of \\(A\\)” is equivalent to “every neighborhood contains infinitely many points of \\(A\\).” More generally, this equivalence holds in \\(T_1\\) spaces, including every Hausdorff space.\n\nTo see why, suppose a neighborhood contains only finitely many points of \\(A\\setminus\\{x\\}\\). Since finite sets are closed in a \\(T_1\\) space, those points can be removed to obtain a smaller neighborhood containing no point of \\(A\\setminus\\{x\\}\\), contradicting accumulation.\n\nWithout the \\(T_1\\) assumption, the two conditions can differ. For example, let \\(X=\\{p,q\\}\\) have only \\(\\varnothing\\) and \\(X\\) as open sets. Then \\(p\\) is an accumulation point of \\(\\{q\\}\\), because its only open neighborhood contains \\(q\\). No neighborhood can contain infinitely many points of this finite set.\n\nSome authors therefore distinguish ordinary accumulation points from \\(\\omega\\)-accumulation points, whose every neighborhood contains infinitely many points of the set."
    },
    {
      "heading": "Derived sets and perfect sets",
      "text": "The set \\(A'\\) of all accumulation points of \\(A\\) is called the derived set. In a \\(T_1\\) space it is closed, though this need not hold in an arbitrary topological space.\n\nDerived sets can be iterated:\n\n\\[A^{(0)}=A,\\qquad A^{(n+1)}=(A^{(n)})'.\\]\n\nFor a closed set in a \\(T_1\\) space, each step removes the isolated points of the preceding set. The construction extends to ordinal numbers by taking intersections at limit stages and forms the basis of Cantor–Bendixson analysis.\n\nA perfect set is a closed set with no isolated points; equivalently,\n\n\\[P=P'.\\]\n\nThe Cantor set is a standard example. Every one of its points is approached by other points of the set, even though the set contains no nonempty open interval. In the real line, every nonempty perfect set is uncountable."
    },
    {
      "heading": "Compactness and the Bolzano–Weierstrass theorem",
      "text": "Accumulation points are closely related to compactness. Every infinite subset of a compact space has an accumulation point in that space.\n\nA direct proof uses the open-cover definition. If an infinite set \\(A\\) had no accumulation point, each point of the space would have an open neighborhood meeting \\(A\\) in at most that point. A finite subcover would then force \\(A\\) to be finite.\n\nFor metric spaces, the following are equivalent:\n\n- The space is compact.\n- Every infinite subset has an accumulation point in the space.\n- Every sequence has a convergent subsequence whose limit belongs to the space.\n\nOutside metric spaces, these conditions require separate treatment and are not all equivalent.\n\nThe Bolzano–Weierstrass theorem states that every bounded infinite subset of Euclidean space \\(\\mathbb R^n\\) has an accumulation point in \\(\\mathbb R^n\\). Equivalently, every bounded sequence in \\(\\mathbb R^n\\) has a convergent subsequence. This follows from the compactness of closed bounded subsets expressed by the Heine–Borel theorem."
    },
    {
      "heading": "Roles in analysis",
      "text": "Accumulation points specify where nontrivial limits can be studied. In the usual definition of a function limit as \\(x\\to a\\), the point \\(a\\) is required to be an accumulation point of the function’s domain. Otherwise, there may be no domain points satisfying \\(0<d(x,a)<\\delta\\), making the limit condition vacuous rather than determining a unique value.\n\nIn complex analysis, accumulation points appear in the identity theorem. If two analytic functions on a connected open domain agree on a set with an accumulation point inside that domain, they agree throughout the domain. The location of the accumulation point is essential: accumulation only at a boundary point does not generally force this conclusion."
    }
  ],
  "links": [
    "topological-space",
    "mathematical-analysis",
    "topology",
    "neighborhood-topology",
    "metric-space",
    "open-ball",
    "real-number",
    "rational-number",
    "integer",
    "closure-topology",
    "closed-set",
    "subspace-topology",
    "limit",
    "first-countable-space",
    "net-mathematics",
    "hausdorff-space",
    "ordinal-number",
    "cantor-set",
    "compact-space",
    "bolzano-weierstrass-theorem",
    "euclidean-space",
    "heine-borel-theorem",
    "mathematical-function",
    "complex-analysis",
    "analytic-function"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}