Plethystic algebra (Q557586)

Let \(k\) be a commutative ring. A \(k\)-biring is a \(k\)-ring that represents a functor from the category of \(k\)-rings to itself. A \(k\)-plethory is a biring \(P\) with an associative map of birings \(\circ :P\odot _{k}P\rightarrow P\) and unit \(k\left\langle e\right\rangle \rightarrow P,\) where \(\odot \) is an analogue of tensor product on birings. A \(k\)-plethory is a non-linear generalization of a cocommutative bialgebra: along with the ``product'' and unitary maps above there are the notions of coaddition \(\Delta ^{+}\) and comultiplication \(\Delta ^{\times }\) with counits \(\varepsilon ^{+}\) and \( \varepsilon ^{\times }\) respectively. For any commutative \(k\)-bialgebra \(A\) the symmetric algebra \(S\left( A\right) \) is an example of a \(k\)-plethory. One of the motivations for this article is the following example. Let \( \Lambda \) be the ring of symmetric functions in infinitely many variables \( \left\{ x_{i}\right\} \) -- this naturally has the structure of a \(\mathbb{Z}\) -plethory. Let \(\psi _{n}=x_{1}^{n}+x_{2}^{n}+\cdots .\) Then \(\psi _{n}=\sum_{d\mid n}w_{d}^{n/d}\) for some \(w_{1},w_{2},\dots \in \Lambda .\) These \(w_{i}\)'s are used in the Witt vector operations. In fact, given a plethory \(P\) if we define \(W_{P}\left( R\right) \) to be the set of \(k\)-ring maps \(P\rightarrow R\) for some \(k\)-ring \(R\) then \(W_{\Lambda _{p}}\left( R\right) \) is isomorphic (as rings) to the ring of \(p\)-typical Witt vectors, where \(\Lambda _{p}\) is a certain sub \(\mathbb{Z}\)-plethory of \(\Lambda .\) In this way one can obtain (from the paper): ``a definition [of Witt vectors] given purely in terms of algebraic structure rather than somewhat mysterious formulas''. Given a morphism \(P\rightarrow Q\) of plethories, the author proves the following reconstruction theorem. Let \(\mathcal{C}\) be a category with all of its limits and colimits, and let \(U\) be a functor from \(\mathcal{C}\) to the category of rings. If \(U\) has a left and right adjoint as well as the property that a map \(f\) in \(\mathcal{C}\) is an isomorphism if \(U\left( f\right) \) is, then \(C\) is a category of \(P\)-rings for some plethory \(P\); furthermore \(U\) is the forgetful functor from \(P\)-rings to rings. Here a ``\(P \)-ring'' is a \(k\)-ring with an action by \(P\). Let \(\mathcal{O}\) be a Dedekind domain, \(\mathfrak{m\subset }\mathcal{O}\) an ideal, \(P\) an \(\mathcal{O}\)-plethory, \(Q\) an \(\mathcal{O}/\mathfrak{m}\) -plethory, and suppose we have a surjective map \(P\rightarrow Q\) of plethories. A \(P\)-ring \(R\) is a \(P\)-deformation of a \(Q\)-ring if it is \( \mathfrak{m}\) torsion free and the action of \(P\) on \(R/\mathfrak{m}R\) factors through \(P\rightarrow Q.\) Then there is an \(\mathcal{O}\)-plethory \( P^{\prime }\) which is universal among those which are \(P\)-deformations of \(Q\) -rings. This \(P^{\prime }\) is called the amplification of \(P\) along \(Q.\) Finally, the notion of a linearization of a plethory is described. From this we get \(A_{P}\), the set of elements of \(P\) acting additively on a \(P\)-ring, and \(C_{P},\) the cotangent space to the spectrum of \(P\) at zero. Then, under certain conditions, \(A_{P}\) is a cocommutative twisted \(k\)-bialgebra, there is a coaction of \(A_{P}\) on \(C_{P},\) and \(A_{P}\rightarrow C_{P}\) is \(A_{P}\) -coequivariant.