Alphabeta Math
LemmaStatement: AI-adaptedProof: AI-generatedSession-authored (Fable 5 assisted)precheck passverified 2026-08-04 (gpt-5.6-sol-codex-subscription)
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

The graph of the piecewise-linear map oscillating between 00 and 11 on the intervals [1/(n+2),1/(n+1)][1/(n+2), 1/(n+1)] is path-connected, its closure adds the segment {0}×[0,1]\{0\} \times [0,1], and that closure is connected, is not path-connected because no path joins the segment to the graph, and is not locally connected

Statement

Write ι\iota for the canonical natural of R\mathbb{R} (The canonical natural ι(n)=n1F\iota(n) = n \cdot 1_F of a field), so that 1/(n+1)1/(n+1) means 1/ι(n+1)1/\iota(n+1), and recall that N\mathbb{N} contains 00. For nNn \in \mathbb{N} put

In  :=  [1n+2, 1n+1](0,1],I_n \;:=\; \Bigl[\tfrac{1}{n+2},\ \tfrac{1}{n+1}\Bigr] \subseteq (0,1],

so that I0=[1/2,1]I_0 = [1/2, 1] and nNIn=(0,1]\bigcup_{n \in \mathbb{N}} I_n = (0,1] (Intervals of R\mathbb{R}: the nine order-convex forms, nondegeneracy, and length). Define εn:=0\varepsilon_n := 0 for nn even and εn:=1\varepsilon_n := 1 for nn odd, and let

f:(0,1][0,1]f : (0,1] \to [0,1]

be the function that is affine on each InI_n with f(1/(n+1))=εnf(1/(n+1)) = \varepsilon_n for every nNn \in \mathbb{N}; explicitly, for xInx \in I_n,

f(x)  =  εn+1  +  (εnεn+1)x1n+21n+11n+2.f(x) \;=\; \varepsilon_{n+1} \;+\; (\varepsilon_n - \varepsilon_{n+1}) \cdot \frac{x - \frac{1}{n+2}}{\frac{1}{n+1} - \frac{1}{n+2}} .

The two clauses agree at each shared endpoint 1/(n+1)1/(n+1), both giving εn\varepsilon_n, so ff is a well-defined function; f(1)=ε0=0f(1) = \varepsilon_0 = 0; and on each InI_n the map ff runs affinely between 00 and 11, so it takes both values 00 and 11 on InI_n, at the two endpoints. Let

G  :=  {(x,f(x)):x(0,1]}    R2,G \;:=\; \{\, (x, f(x)) : x \in (0,1] \,\} \;\subseteq\; \mathbb{R}^2 ,

the graph of ff, with R2\mathbb{R}^2 carrying the product topology, which is the metric topology of dd_\infty (For n1n \ge 1 the product topology on nn copies of the usual topology of R\mathbb{R} is the metric topology of dd_\infty on Rn\mathbb{R}^n, and hence also of d1d_1 and d2d_2, so Rn\mathbb{R}^n as a product and Rn\mathbb{R}^n as a metric space are one space, Rn\mathbb{R}^n as the set of functions nRn \to \mathbb{R}, and d1d_1, d2d_2, dd_\infty are metrics on it, Metrizable space: a topological space whose topology is induced by some metric; metrizability is topological, the metric is not), and let subsets carry the subspace topology (Subspace topology: the traces of the open sets, its closed sets and its bases, the continuity of the inclusion, and the characteristic property of a map into a subspace). Then:

  1. ff is continuous, and GG is homeomorphic to (0,1](0,1]; hence GG is path-connected (Paths, path-connected spaces and path components), connected, and locally connected (Locally connected and locally path-connected spaces: a neighbourhood base of open connected, respectively open path-connected, sets at every point).
  2. The closure is G=G({0}×[0,1])\overline{G} = G \cup (\{0\} \times [0,1]).
  3. G\overline{G} is connected (Separation of a topological space, connected and disconnected spaces, clopen sets, and connected subsets).
  4. G\overline{G} is not path-connected; more precisely, no path in G\overline{G} joins a point of {0}×[0,1]\{0\} \times [0,1] to a point of GG.
  5. G\overline{G} is not locally connected at any point (0,t)(0,t), t[0,1]t \in [0,1], so it is not locally connected.

There is no trigonometric function anywhere in this construction. Every piece of ff is affine, and the oscillation comes from the alternating endpoint values εn\varepsilon_n alone.

Facts & Assumptions

Given: The intervals InI_n, the function ff, the graph GG, and R2\mathbb{R}^2 with the product topology; π0,π1:R2R\pi_0, \pi_1 : \mathbb{R}^2 \to \mathbb{R} denote the two projections.

[A1]

An affine map xc+mxx \mapsto c + mx of R\mathbb{R} into R\mathbb{R} is continuous: (c+ms)(c+mt)=mst|(c+ms)-(c+mt)| = |m||s-t|, so for m0m \ne 0 the ball of radius δ/m\delta/|m| around tt maps into the ball of radius δ\delta, and a constant map is continuous outright (Open ball, closed ball and sphere in a metric space, Metrizable space: a topological space whose topology is induced by some metric; metrizability is topological, the metric is not, The absolute value makes R\mathbb{R} a metric space: d(x,y)=xyd(x,y) = |x-y| is a metric, its open balls are the intervals (xr,x+r)(x-r, x+r), and it is unbounded, The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement, Continuity of a map of topological spaces at a point and globally).

[A8]

A nonempty subset of R\mathbb{R} bounded above has a least upper bound, and for every ε>0\varepsilon > 0 some element of it exceeds supε\sup - \varepsilon (Complete ordered field (least-upper-bound property), Epsilon characterisation of the supremum).

[A9]

XX is locally connected at xx when every open UxU \ni x contains an open connected VV with xVUx \in V \subseteq U; a homeomorphism hh carries such a VV to h[V]h[V], which is connected as a continuous image and open because a homeomorphism is an open map, so local connectedness is a topological property (Locally connected and locally path-connected spaces: a neighbourhood base of open connected, respectively open path-connected, sets at every point, A continuous image of a connected space is connected, and connectedness is a topological property claim 1, Homeomorphism, open map, closed map, embedding, and what it means for a property to be topological).

Proof

technique · direct
1.1

ff is continuous. For nNn \in \mathbb{N} the restriction of ff to the closed set InIn+1=[1/(n+3),1/(n+1)]I_n \cup I_{n+1} = [1/(n+3), 1/(n+1)] is continuous, being affine on each of the two closed pieces by [A1] and agreeing at the shared endpoint, so the finite closed cover clause of [A2] applies with two pieces; likewise ff is affine, hence continuous, on [1/2,1]=I0[1/2,1] = I_0.

A1A2
1.2

(0,1](0,1] is order-convex, hence a connected subset of R\mathbb{R} by [A4], and it is path-connected: for x,y(0,1]x, y \in (0,1] the map tx+t(yx)t \mapsto x + t(y-x) is continuous by [A1] and takes values in (0,1](0,1] by order-convexity.

A1A4
1.3

(0,1](0,1] is locally connected: a basic open subset of it is the trace of an interval of R\mathbb{R}, hence order-convex, hence connected by [A4]; so the open connected subsets form a neighbourhood base at each of its points, which is [A9].

A4A9
1.4

GG({0}×[0,1])\overline{G} \subseteq G \cup (\{0\} \times [0,1]). Let (x,y)G(x,y) \in \overline{G}. Every point of GG lies in [0,1]×[0,1][0,1] \times [0,1], which is closed, being a product of closed sets whose complement is a union of basic open sets; so (x,y)[0,1]×[0,1](x,y) \in [0,1] \times [0,1] by [A5]. If x=0x = 0 the point lies in {0}×[0,1]\{0\} \times [0,1].

A5
1.5

A second consequence, used twice below: for every real ρ>0\rho > 0 there are x0,x1(0,ρ]x_0, x_1 \in (0, \rho] with f(x0)=0f(x_0) = 0 and f(x1)=1f(x_1) = 1. Indeed [A7] gives a natural k1k \ge 1 with 1/k<ρ1/k < \rho; with n:=k1Nn := k-1 \in \mathbb{N} the interval In(0,ρ]I_n \subseteq (0, \rho], and the two endpoints of InI_n carry the values εn\varepsilon_n and εn+1\varepsilon_{n+1}, which are 00 and 11 in one order or the other.

A7
2.1

The sets Wn:=(1/(n+3),1/(n+1))W_n := (1/(n+3), 1/(n+1)) for nNn \in \mathbb{N}, together with W:=(1/2,1]W^{*} := (1/2, 1], form an open cover of (0,1](0,1] in its subspace topology, and ff restricted to each is a restriction of one of the continuous maps of step 1.1; so ff is continuous by the open cover clause of [A2].

step 1.1A2
3.1

The map g:(0,1]R2g : (0,1] \to \mathbb{R}^2, g(x):=(x,f(x))g(x) := (x, f(x)), is continuous by [A3], its components being the inclusion and ff, both continuous by step 2.1 and [A2]; it is injective, since g(x)g(x) determines xx; and its image is GG.

step 2.1A2A3
3.2

{0}×[0,1]G\{0\} \times [0,1] \subseteq \overline{G}. Let t[0,1]t \in [0,1] and let (a,b)×(c,d)(a,b) \times (c,d) be a basic open set containing (0,t)(0,t). By [A7] there is a natural k1k \ge 1 with 1/k<min{b,1}1/k < \min\{b, 1\}, and putting n:=k1Nn := k - 1 \in \mathbb{N} the interval InI_n lies in (0,b)(0, b), since 1/(n+1)=1/k<b1/(n+1) = 1/k < b. As ff runs affinely between 00 and 11 on InI_n, [A4] gives xInx \in I_n with f(x)(c,d)f(x) \in (c,d): the image of InI_n under ff is order-convex and contains 00 and 11, hence contains tt and every point near it inside [0,1][0,1]. Then (x,f(x))G(x, f(x)) \in G lies in the basic set, so (0,t)G(0,t) \in \overline{G} by [A5].

step 2.1A4A5A7
4.1

The corestriction g0:(0,1]Gg_0 : (0,1] \to G is a continuous bijection by step 3.1 and [A2], and its inverse is the restriction of π0\pi_0 to GG, which is continuous by [A3] and [A2]; so g0g_0 is a homeomorphism and G(0,1]G \cong (0,1].

step 3.1A2A3
4.2

Suppose instead x>0x > 0, so x(0,1]x \in (0,1] and f(x)f(x) is defined. Let ε>0\varepsilon > 0. By step 2.1 and [A3] there is δ>0\delta > 0 such that f(s)f(x)<ε|f(s) - f(x)| < \varepsilon for every s(xδ,x+δ)(0,1]s \in (x-\delta, x+\delta) \cap (0,1]. The basic set (xδ,x+δ)×(yε,y+ε)(x-\delta, x+\delta) \times (y-\varepsilon, y+\varepsilon) contains (x,y)(x,y), hence meets GG by [A5] in a point (s,f(s))(s, f(s)); then yf(x)yf(s)+f(s)f(x)<2ε|y - f(x)| \le |y - f(s)| + |f(s) - f(x)| < 2\varepsilon. As ε>0\varepsilon > 0 was arbitrary, y=f(x)y = f(x) and (x,y)G(x,y) \in G. With step 3.2 and step 1.4 this proves claim 2.

step 2.1step 3.2step 1.4A3A5
4.3

Claim 5. Fix t[0,1]t \in [0,1] and let U:=G(R×(t1/4, t+1/4))U := \overline{G} \cap (\mathbb{R} \times (t - 1/4,\ t + 1/4)), an open subset of G\overline{G} containing (0,t)(0,t). Suppose VV is open in G\overline{G}, connected, with (0,t)VU(0,t) \in V \subseteq U. Then VV contains a set G((η,η)×(tη,t+η))\overline{G} \cap ((-\eta,\eta) \times (t-\eta, t+\eta)) for some η>0\eta > 0 by [A3], and that set meets GG by step 3.2, at a point whose first coordinate xx satisfies 0<x<η0 < x < \eta.

step 3.2A3A5
5.1

Hence GG is path-connected, connected and locally connected, these being carried across the homeomorphism of step 4.1 from step 1.2 and step 1.3, using [A6] for connectedness and [A9] for local connectedness. This is claim 1.

step 1.2step 1.3step 4.1A6A9
5.2

A useful consequence of claim 2, used twice below: if pGp \in \overline{G} has π0(p)>0\pi_0(p) > 0 then p=(π0(p),f(π0(p)))p = (\pi_0(p), f(\pi_0(p))), so π1(p)=f(π0(p))\pi_1(p) = f(\pi_0(p)).

step 4.2
5.3

Claim 4. Suppose γ:[0,1]G\gamma : [0,1] \to \overline{G} is a path with γ(0){0}×[0,1]\gamma(0) \in \{0\} \times [0,1] and γ(1)G\gamma(1) \in G, and write k:=π0γk := \pi_0 \circ \gamma and h:=π1γh := \pi_1 \circ \gamma, both continuous by [A3] and [A2]. Then J:={u[0,1]:k(u)=0}J := \{\, u \in [0,1] : k(u) = 0 \,\} is closed in [0,1][0,1], being the preimage of the closed set {0}\{0\}, it contains 00, and 1J1 \notin J since π0(γ(1))>0\pi_0(\gamma(1)) > 0.

step 4.2A2A3A5
6.1

Claim 3: GG is connected by step 5.1 and GGGG \subseteq \overline{G} \subseteq \overline{G}, so G\overline{G} is connected by [A6].

step 5.1A6
6.2

Let c:=supJc := \sup J, which exists by [A8]. Every open set containing cc contains an interval around it, which by [A8] meets JJ; so cJc \in \overline{J}, and JJ is closed in [0,1][0,1] while J[0,1]=[0,1]\overline{J} \subseteq \overline{[0,1]} = [0,1], so cJc \in J by [A5]. Hence k(c)=0k(c) = 0 and c<1c < 1.

step 5.3A5A8
6.3

So π0[V]\pi_0[V] is a connected subset of R\mathbb{R} by [A2, A3, A4], hence order-convex, and it contains 00 and x>0x > 0; therefore [0,x]π0[V][0,x] \subseteq \pi_0[V]. By step 1.5 with ρ:=x\rho := x there are x0,x1(0,x]x_0, x_1 \in (0,x] with f(x0)=0f(x_0) = 0 and f(x1)=1f(x_1) = 1, so VV contains points p0,p1p_0, p_1 with π0(pi)=xi>0\pi_0(p_i) = x_i > 0, and π1(pi)=f(xi)\pi_1(p_i) = f(x_i) by step 5.2.

step 5.2step 1.5step 4.3A2A3A4
7.1

By continuity of hh at cc there is δ>0\delta > 0 with c+δ1c + \delta \le 1 and h(u)h(c)<1/4|h(u) - h(c)| < 1/4 for all u[c,c+δ]u \in [c, c+\delta]; put t:=h(c)t := h(c). Moreover k(c+δ)>0k(c+\delta) > 0, since c+δ>c=supJc + \delta > c = \sup J puts c+δc+\delta outside JJ while k0k \ge 0 everywhere by step 1.4.

step 5.3step 6.2A3
8.1

The restriction of kk to [c,c+δ][c, c+\delta] is continuous on a connected space by [A4] and step 1.2, so its image is order-convex and contains k(c)=0k(c) = 0 and k(c+δ)>0k(c+\delta) > 0; hence [0,k(c+δ)][0, k(c+\delta)] lies in that image. By step 1.5 with ρ:=k(c+δ)\rho := k(c+\delta) there are x0,x1(0,k(c+δ)]x_0, x_1 \in (0, k(c+\delta)] with f(x0)=0f(x_0) = 0 and f(x1)=1f(x_1) = 1, and therefore u0,u1[c,c+δ]u_0, u_1 \in [c, c+\delta] with k(ui)=xi>0k(u_i) = x_i > 0.

step 1.5step 7.1A4
9.1

By step 5.2, h(ui)=f(k(ui))=f(xi)h(u_i) = f(k(u_i)) = f(x_i), so h(u0)=0h(u_0) = 0 and h(u1)=1h(u_1) = 1; but both lie within 1/41/4 of tt by step 7.1, giving 1=h(u1)h(u0)h(u1)t+th(u0)<1/21 = |h(u_1) - h(u_0)| \le |h(u_1) - t| + |t - h(u_0)| < 1/2, which is false. So no such path exists, and since G\overline{G} contains points of both kinds by claim 2, it is not path-connected. This is claim 4.

step 5.2step 7.1step 8.1
10.1

Hence VV contains a point with second coordinate 00 and a point with second coordinate 11, both of which must lie in (t1/4,t+1/4)(t - 1/4, t+1/4) because VUV \subseteq U; that gives 10t+t1<1/21 \le |0 - t| + |t - 1| < 1/2, which is false. So no such VV exists and G\overline{G} is not locally connected at (0,t)(0,t), by [A9]; this is claim 5.

step 4.3step 6.3A9

Remarks

  • Why continuity is checked on an OPEN cover and never on the closed one. The intervals InI_n form a closed cover of (0,1](0,1] with infinitely many members, and the closed pasting lemma is false for infinite covers, the standing witness being R\mathbb{R} covered by its closed singletons: every restriction of the indicator of {0}\{0\} is continuous and the map is not, so the closed pasting lemma needs finiteness. The proof therefore pastes only two closed pieces at a time, producing continuity on a slightly larger closed interval, and then uses the open cover clause, which carries no finiteness restriction.

  • What each claim is for. Claim 3 with claim 4 gives a connected space that is not path-connected; claim 1 with claim 4 gives a path-connected set whose closure is not path-connected; claim 1 with claim 5 gives a locally connected set whose closure is not locally connected. Each of the three is used as a witness later on this page.

  • The failure is exactly at the added segment. By claim 2 the only points of G\overline{G} not in GG are those of {0}×[0,1]\{0\} \times [0,1], and claim 5 locates the failure of local connectedness at each of them. At every point of GG the space G\overline{G} still looks like (0,1](0,1], since GG is open in G\overline{G} — its complement {0}×[0,1]\{0\} \times [0,1] is closed — so no pathology occurs away from the segment.

  • Both endpoint values are attained on every piece, and that is the whole mechanism. The proof never uses any property of ff beyond continuity and the fact recorded in step 5.3: arbitrarily close to 00 the function takes the value 00 and the value 11. Any function with that property and a path-connected graph would serve.

Depends on

Used by

Dependency tree · next 3 levels

Direct dependencies and their dependencies through the next three levels: 148 results over 23 levels. An arrow runs from a result to what uses it, and this result sits at the bottom with a heavier outline. Click the chart to enlarge it.

Sources