Proposition 16.6.4. The lax symmetric monoidal functors \[ \Sp \xrightarrow {\Omega ^{\infty }} \CGrp (\An ) \hookrightarrow \CMon (\An ) \xrightarrow {\fgt } \An _* \xrightarrow {\fgt } \An \] admit strongly monoidal left adjoints: \[ (\An ,\times ) \xrightarrow {(-)_+} (\An _*,\wedge ) \xrightarrow {F^{\CMon }} (\CMon (\An ),\otimes ) \xrightarrow {(-)^{\grp }} (\CGrp (\An ),\otimes ) \xhookrightarrow {\bB ^{\infty }} (\Sp ,\otimes ). \] Here, \((-)_+\) adjoins a disjoint basepoint, \(F^{\CMon }\) is the free commutative monoid functor, \((-)^{\grp }\) is group completion, and \(\bB ^{\infty }\) is infinite delooping.
Proof. The functor \[ (\An ,\times )\longrightarrow \bigl (\Ar (\An ),\square \bigr ), \qquad X\longmapsto (\emptyset \to X), \] is symmetric monoidal: the pushout-product formula gives \((\emptyset \to X)\,\square \,(\emptyset \to Y)\simeq (\emptyset \to X\times Y)\), and the unit is \(\emptyset \to *\). Its composite with the symmetric monoidal cofiber functor of Lemma 16.3.4 is \((-)_+\), so adjoining a basepoint is strongly monoidal.
The same conclusion for \(F^{\CMon }\) follows from the bilinear universal property of Lemma 16.4.3. Indeed, for pointed animae \(X,Y\) and a commutative monoid \(M\), there are natural equivalences \[ \begin {aligned} \Hom _{\CMon (\An )}(F^{\CMon }(X)\otimes F^{\CMon }(Y),M) &\simeq \Nat _{\Span (\Fin )^2} \bigl (F^{\CMon }(X)(-)\times F^{\CMon }(Y)(-),M(-\times -)\bigr ) \\ &\simeq \Hom _{\An _*}(X\wedge Y,\fgt (M)) \\ &\simeq \Hom _{\CMon (\An )}(F^{\CMon }(X\wedge Y),M). \end {aligned} \] The middle equivalence is obtained by applying the free commutative monoid adjunction in each variable; the functors \(S\mapsto M(S\times T)\) remain commutative monoids because \(-\times T\) preserves the finite products of \(\Span (\Fin )\). The analogous nullary argument identifies the monoidal units. The group completion functor is strongly monoidal because it is the symmetric monoidal Bousfield localization of Proposition 16.4.1.
It remains to consider infinite delooping. Since \(\Sp \) is additive, evaluation \(U\colon \CGrp (\Sp )\to \Sp \) is an equivalence by Proposition 5.3.17. Its left adjoint, the free commutative group functor \(F_{\Sp }^{\CGrp }\), is symmetric monoidal by Proposition 16.4.1. Hence its inverse \(U\) is symmetric monoidal as well.
Applying Proposition 16.4.2 to the symmetric monoidal adjunction \(\S [-]\dashv \Omega ^{\infty }\) of Proposition 16.6.2 gives a symmetric monoidal left adjoint \(\widetilde L\) to postcomposition with \(\Omega ^{\infty }\). Thus the composite \[ \CGrp (\An )\xrightarrow {\widetilde L}\CGrp (\Sp )\xrightarrow {U}\Sp \] is symmetric monoidal. Its right adjoint \((\Omega ^{\infty })_*F_{\Sp }^{\CGrp }\) is a finite-product-preserving lift of \(\Omega ^{\infty }\) to \(\CGrp (\An )\): evaluation commutes with postcomposition, and \(UF_{\Sp }^{\CGrp }\simeq \id _{\Sp }\). By the uniqueness assertion in Theorem 5.4.6, this right adjoint is \(\bOmega ^{\infty }\). Uniqueness of left adjoints therefore identifies the displayed composite with \(\bB ^{\infty }\), giving its symmetric monoidal refinement. Finally, the right adjoints acquire the displayed lax symmetric monoidal structures by Proposition 14.5.6. β‘