Strong Arcwise Connectedness

A space is `n-strong arc connected’ (n-sac) if for any n points in the space there is an arc in the space visiting them in order. A space is omega-strong arc connected (omega-sac) if it is n-sac for all n. We study these properties in finite graphs, regular continua, and rational continua. There are no 4-sac graphs, but there are 3-sac graphs and graphs which are 2-sac but not 3-sac. For every n there is an n-sac regular continuum, but no regular continuum is omega-sac. There is an omega-sac rational continuum. For graphs we give a simple characterization of those graphs which are 3-sac. It is shown, using ideas from descriptive set theory, that there is no simple characterization of n-sac, or omega-sac, rational continua.

💡 Research Summary

The paper introduces a new topological notion called n‑strong arc‑connectedness (n‑sac). A space X is n‑sac if for any ordered n‑tuple of distinct points ((x_1,\dots ,x_n)) there exists a single arc in X that passes through these points in the prescribed order. When the property holds for every natural number n, the space is called ω‑strong arc‑connected (ω‑sac). This definition strengthens ordinary arc‑connectedness (which corresponds to the case n = 2) by demanding that the space can accommodate any finite ordered sequence of points along a single embedded interval.

The authors investigate n‑sac and ω‑sac in three major classes of spaces: finite graphs, regular continua, and rational continua. Their results can be summarized as follows.

1. Finite Graphs
Every connected graph is trivially 2‑sac. The paper gives a clean characterization of 3‑sac graphs: a connected graph is 3‑sac if and only if it is 2‑connected (i.e., it has no cut‑vertex). Equivalently, any two vertices lie on two internally disjoint paths, which guarantees that any three prescribed vertices can be visited in order by a single simple path. The authors also construct examples of graphs that are 2‑sac but fail to be 3‑sac, showing that the property is strictly stronger than ordinary arc‑connectedness.

A striking negative result is proved: no graph can be 4‑sac. By a combinatorial argument involving the pigeon‑hole principle and the structure of paths in a finite graph, they demonstrate that for any four distinct vertices there exists an ordering that cannot be realized by a single arc. Consequently, the hierarchy of n‑sac collapses at n = 3 for graphs.

2. Regular Continua
A regular continuum is a compact, connected metric space in which every point has a basis of neighborhoods whose boundaries are finite. For each natural number n the authors construct a regular continuum that is n‑sac. The construction is a “star‑shaped” continuum with n arms emanating from a central point; the arms are arranged so that any n points can be linked by an arc that visits them in the required order.

However, the paper proves that no regular continuum can be ω‑sac. The key observation is that regular continua have only finitely many local cut‑points at each point. By choosing a sufficiently large n, one can force any candidate arc to pass through a point that would have to separate more than its finite number of components, which is impossible. Hence the regularity condition imposes a hard bound on the strength of arc‑connectedness.

3. Rational Continua
Rational continua are compact connected subsets of the plane whose points all belong to a countable dense set (hence “rational”). The authors show that, unlike regular continua, there exists an ω‑sac rational continuum. Their example is a carefully designed “infinite dead‑end tree” embedded in the plane: each branching point has infinitely many incident arcs, and the construction ensures that any finite ordered set of points can be threaded by a single arc. The space remains rational because the branching points are chosen from a countable dense subset of the plane. This result demonstrates that the restriction to rational points does not preclude the existence of spaces with maximal strong arc‑connectedness.

4. Descriptive Set‑Theoretic Complexity
The final part of the paper addresses the classification problem: can one give a simple, perhaps Borel‑definable, description of the class of n‑sac (or ω‑sac) rational continua? Using tools from descriptive set theory, the authors show that the set of codes (in a standard Polish space of compact subsets of the plane) representing n‑sac rational continua is analytic but not Borel. Consequently, there is no “nice” characterization—no countable list of elementary topological conditions—that captures exactly the n‑sac or ω‑sac rational continua. This negative result underscores the intrinsic complexity of the strong arc‑connectedness property.

5. Methodology and Significance
The paper blends classical graph theory (cut‑vertices, 2‑connectivity), continuum theory (regularity, rationality), and modern descriptive set theory (Borel hierarchy, analytic sets). The graph results provide a concrete combinatorial picture, while the continuum constructions illustrate how local branching complexity can be tuned to achieve any prescribed level of strong arc‑connectedness. The descriptive‑set‑theoretic argument places the classification problem in a broader logical context, showing that the property is not merely topological but also highly non‑definable.

6. Outlook
The authors suggest several directions for future work: extending the analysis to higher‑dimensional manifolds, investigating algorithmic decision procedures for n‑sac in finite complexes, and exploring the interaction between strong arc‑connectedness and other topological invariants such as homology or shape theory. The paper opens a new line of inquiry into how ordering constraints on arcs interact with the underlying structure of spaces, and it highlights the delicate balance between combinatorial simplicity (as in graphs) and topological richness (as in continua).