A remark on the rank conjecture (Q1611687)

The main result of this paper is about the action of \(\lambda\)-operations on the homology of linear groups. More precisely, for any commutative ring \(A\) with identity, the image of \(H_n(GL_r(A),\mathbb Z)\) in \(H_n(GL(A),\mathbb Z)\) is killed by a certain operator built on powers of \(\lambda\). This theorem is then used to study the ``rank conjecture'' and various related conjectures about \(K_n(A)_{\mathbb Q}\). Recall that, viewing \(K_n(A)_{\mathbb Q}\) as the primitive part of \(H_n(GL_n(A),\mathbb Q)\) by setting \(F_r^{\text{rank}}K_n(A)_{\mathbb Q} := \text{Image} (H_n(GL_r(A),\mathbb Q))\cap K_n(A)_{\mathbb Q}\) for \(r\geq 1\). Here the image is that of the natural map \(H_n(GL_r(F),\mathbb Q) \rightarrow H_n(GL(F),\mathbb Q).\) The rank conjecture (attributed to Suslin) asserts that, for any infinite field \(F\), \(K_n(F)_{\mathbb Q}\) is the direct sum of \(F_r^{\text{rank}}K_n(F)_{\mathbb Q}\) and a certain precise part given by the gamma filtration. A related conjecture (labeled 2 here) states that the natural map \[ H_n(GL_r(F),\mathbb Q) \rightarrow H_n(GL_{r+1}(F),\mathbb Q) \] is injective. Both conjectures are known to hold for number fields (Borel-Yang). Here the author formulates a conjecture (labeled 12) about how the rank conjecture, combined with an assumption on the kernel of \(H_n(GL_p(F),\mathbb Q) \rightarrow H_n(GL(F),\mathbb Q)\), could give rise to a different point of view on Bloch's comparison isomorphism between \(K_n^{(p)}(F)\) and \(CH^p(F,n)_{\mathbb Q}\), and he proves part of this new conjecture.