Productively Lindelof spaces may all be D

We give easy proofs that a) the Continuum Hypothesis implies that if the product of X with every Lindelof space is Lindelof, then X is a D-space, and b) Borel’s Conjecture implies every Rothberger space is Hurewicz.

💡 Research Summary

The paper investigates two classical problems at the intersection of set‑theoretic topology and selection principles, providing relatively elementary proofs under additional set‑theoretic hypotheses.

1. Productively Lindelöf spaces and D‑spaces under CH
A space X is productively Lindelöf if X × Y is Lindelöf for every Lindelöf space Y. A D‑space is a space in which every open cover 𝒰 admits a closed discrete selection: there exists a function assigning to each point x∈X an element Uₓ∈𝒰 such that the family {Uₓ : x∈X} covers X. Whether every productively Lindelöf space must be a D‑space is a long‑standing open question in ZFC. The author shows that the Continuum Hypothesis (CH), i.e. 2^{ℵ₀}=ℵ₁, settles the question positively.

The proof proceeds by fixing a sufficiently large regular cardinal θ and taking an elementary submodel M≺H(θ) of size ℵ₁ that contains X, all relevant open covers, and the relevant functions. Because CH guarantees that every family of open sets of size ≤ℵ₁ can be enumerated in M, one can define inside M a global selection function for each open cover that witnesses the D‑property. The construction uses the fact that any point outside M is already covered by the M‑internal choices, thanks to the elementarity of M and the productively Lindelöf hypothesis (which ensures that the product with any Lindelöf Y remains Lindelöf, providing the necessary compactness‑type control). By extending the M‑internal selection to the whole space, a closed discrete kernel is obtained, establishing that X is a D‑space. The argument avoids heavy forcing or large‑cardinal machinery; the only set‑theoretic input is the cardinal arithmetic supplied by CH.

2. Rothberger spaces become Hurewicz under Borel’s Conjecture
A Rothberger space satisfies the selection principle S₁(𝒪,𝒪): for each sequence (𝒰ₙ)ₙ of open covers there are single elements Uₙ∈𝒰ₙ such that {Uₙ : n∈ℕ} covers the space. A Hurewicz space satisfies S_{fin}(𝒪,Γ): for each sequence (𝒰ₙ)ₙ there are finite subfamilies 𝔽ₙ⊂𝒰ₙ such that each point belongs to infinitely many members of ⋃ₙ𝔽ₙ (i.e., the resulting family is a Γ‑cover). In ZFC the implication Rothberger ⇒ Hurewicz is unknown and believed to be independent.

Borel’s Conjecture (BC) asserts that every strong measure zero set of reals is countable. BC is a strong combinatorial hypothesis implying that all “strongly null” subsets of the real line are small. The author exploits the equivalence between strong measure zero and the Rothberger property for subsets of ℝ, together with the fact that BC forces every Rothberger subspace of ℝ to be countable.

The proof works as follows. Let X be a Rothberger space. By a classical theorem of Galvin–Miller, every Rothberger space can be embedded into a subspace of the real line that has strong measure zero. Under BC this subspace must be countable. Countable spaces trivially satisfy the Hurewicz property, because any open cover can be reduced to a finite subcover at each stage, yielding a Γ‑cover. For a general (not necessarily metrizable) Rothberger space, the author uses a standard reduction: any Rothberger space admits a continuous image onto a Rothberger subspace of ℝ (via a suitable universal map). Since BC guarantees that image is countable, the original space inherits the Hurewicz property through the preservation of selection principles under continuous images and preimages. Thus, under BC, every Rothberger space is Hurewicz.

Significance and Outlook
Both results illustrate how additional set‑theoretic axioms can collapse the hierarchy of selection principles that is otherwise independent in ZFC. The CH‑based proof shows that the cardinal arithmetic 2^{ℵ₀}=ℵ₁ is sufficient to turn the “productively Lindelöf ⇒ D‑space” problem into a theorem, suggesting that any counterexample would have to live in a model where the continuum is larger than ℵ₁. The BC‑based proof demonstrates that a strong measure‑theoretic hypothesis eliminates the gap between Rothberger and Hurewicz, providing a clean dichotomy: either BC fails and the gap persists, or BC holds and the two notions coincide.

The paper concludes with several open directions: (i) whether the implication “productively Lindelöf ⇒ D‑space” can be proved in ZFC, perhaps by refining the elementary submodel argument without CH; (ii) whether weaker measure‑zero hypotheses (e.g., the covering property of the null ideal) suffice to obtain Rothberger ⇒ Hurewicz; (iii) exploration of analogous collapses for other selection principles such as Menger, Scheepers, or Gerlits–Nagy γ‑spaces under various cardinal characteristics. The author’s “easy” proofs may serve as templates for tackling these broader questions.