Operations in Milnor \(K\)-theory (Q1013109): Difference between revisions

From the abstract: ``We show that operations in Milnor \(K\)-theory mod \(p\) of a field are spanned by divided power operations... Moreover, the integral case is discussed and we determine the operations \(K_i^M/p\to K_j^M/p\) for schemes over a field.'' From the introduction: ``Voevodsky constructs Steenrod operations on the motivic cohomology \(H^{*,*}(X,\mathbb Z/p)\) of a general scheme over a field \(k_0\), where the prime \(p\neq\text{char }k_0\). However, when \(p\) is odd or when \(p=2\) and \(\sqrt{-1}\in k\), such operations vanish on the motivic cohomology groups \(H^{i,i}(\text{Spec }k,\mathbb Z/p)\) for \(i>0\) of the spectrum of a field extension \(k\) of \(k_0\). Here, we study operations on \(H^{i,i}(\text{Spec }k,\mathbb Z/p)\) for, as Nesterenko-Suslin and Totaro have shown, there is an isomorphism \(H^{n,n}(\text{Spec }k,\mathbb Z/p)\simeq K^M_n(k)\) with the Milnor \(K\)-groups. The same phenomenon happens in étale cohomology, where Steenrod operations, as defined by Epstein, vanish on the étale cohomology \(H^i_{\text{ét}}(\text{Spec }k,\mathbb Z/p)\) of a field if \(p\) is odd or if \(p=2\) and \(\sqrt{-1}\in k\). Under the assumption of the Bloch-Kato conjecture, our operations give secondary operations relatively to Steenrod operations on the étale cohomology of fields.'' Milnor \(K\)-theory: ``Let \(k\) be a field. The \(n\)th Milnor \(K\)-group \(K_n^M(k)\) is the quotient of the \(n\)-fold tensor power \((k^\times)^{\otimes n}\) of the multiplicative group \(k^\times\) of the field \(k\) by the relations \(a_1\otimes\dots\otimes a_n=0\) as soon as \(a_i+a_j=1\) for some \(1\leq i<j\leq n\). \dots Thus, elements of \(K_n^M\) are sums of symbols'', that is, of images in \(K_n^M\) of elements of the type \(a_1\otimes\dots\otimes a_n\). Power operations: ``Given a base field \(k_0\) and a prime number \(p\), an operation on \(K_i^M/p\) is a function \(K_i^M(k)/p\to K_*^M(k)/p\) defined for all fields \(k/k_0\), compatible with extension of fields... It is important for our purpose that our operations should be functions and not only additive functions, the reason being that additive functions will appear to be trivial in some sense''. Divided powers: ``The axioms of divided powers formalize the properties of \(\frac{x^n}{n!}\) in a \(\mathbb Q\)-algebra''. ``Let \(x=\sum_{r=1}^l s_r\) be a sum of \(l\) symbols in \(K_i^M(k)/p\), the mod \(p\) Milnor \(K\)-group of \(k\) of degree \(i\). We define the \(n\)th divided power of \(x\), given as a sum of symbols, by \(\gamma_n(x)=\sum_{1\leq l_1<\dots<l_n\leq l}s_{l_1}\dots s_{l_n}\in K^M_{ni}(k)/p\). Such a divided power may depend on the way \(x\) has been written as a sum of symbols. However, Kahn shows that the above formula gives well-defined divided powers \(\gamma_n\:K^M_{2i}(k)/p\to K_{2ni}^M(k)/p\) for \(p\) odd and \(\gamma_n\:K^M_i(k)/2\to K_{ni}^M(k)/2\) for \(k\) containing a \(\sqrt-1\).'' Results of the paper: ``In this paper, we show that in these cases, divided powers define power operations in the above sense and form a basis for all possible operations on mod \(p\) Milnor \(K\)-theory. In the remaining case, when \(-1\) is not a square in the base field \(k_0\), divided powers as defined above are not well-defined on mod \(2\) Milnor \(K\)-theory. However, we will define some new, weaker operations, and show that these new operations are all the possible operations on mod \(2\) Milnor \(K\)-theory.'' ``We also describe some new operations in integral Milnor \(K\)-theory over any base field \(k_0\). Under some reasonable hypothesis on an operation \(\varphi\:K_i^M\to K_*^M\) defined over \(k_0\), we are able to show that \(\varphi\) is in the \(K_*^M(k_0)\)-span of our weak divided power operations. In Section 4, we extend our results to the case of the Milnor \(K\)-theory of smooth schemes over a field''. ``Our results in Section 5 applied to Quillen's \(K\)-theory of fields say that operations mod \(p\) from Milnor \(K\)-theory to Quillen \(K\)-theory are spanned by divided power operations.'' Subsection 2.3 discusses the relation between the divided powers and Milnor's Stiefel-Whitney classes of a quadratic form.