Co-rank of weakly parafree 3-manifold groups (Q2180189)

The co-rank of a group is the maximum rank of its free epimorphic image; for a 3-manifold \(M\), the co-rank of \(\pi_1(M)\) is the same as the cut number \(c(M)\). The authors call a group very large if its co-rank is greater than 1, that is, if it maps onto a non-commutative free group: \(c(M)>1\). They call a group large if it has a very large finite-index subgroup. The motivation of this work is as follows. Most compact 3-manifold groups are large. Are they very large? Generally, the answer is negative: in her previous work, the first author constructed a manifold of a special kind, with any given first Betti number, such that its group is not very large. On the other hand, the answer is positive for handlebodies \(H_g\) of genus \(g\ge2\), since \(c(H_g)=g\). Can this be extended to homology handlebodies? It looks plausible because their groups are weakly parafree, i.e., very similar to free groups. The question can be also understood as whether weakly parafree 3-manifold groups are very large. While the authors do not arrive at an answer, they show that, for each genus \(g\), the group of some homology handlebody \(Y_g\) is ``half as large'' as that of the corresponding handlebody: \(c(Y_g)=\lceil\frac12g\rceil\). In particular, the group of the homology handlebody \(Y_2\), which is Thurston's tripus manifold, is not very large. The question of whether examples can be found with even smaller cut number \(c(Y)\), or maybe even with \(c(Y)=1\), i.e., with not very large group, is left open.