Proving that the first Galois cohomology group is direct limit of finite quotients

91 Views Asked by At

This question comes from Silverman's Arithmetic of Elliptic Curves, specifically the appendix on Galois cohomology.

I am a cohomology beginner, interested (for now) in understanding just enough to get through a solid proof of Hilbert's Theorem 90: $$H^1(G_{\overline{K} / K}, \overline{K}^*) = 0.$$

I follow the proof in the case where $L / K$ is finite given in Serre's Local Fields, so the result would follow from an exercise in Silverman showing that if $M$ is a module on which $G_{\overline{K} / K}$ acts continuously (given the discrete topology on $M$ and profinite topology on $G_{\overline{K} / K}$, then $$H^1(G_{\overline{K} / K}, \overline{K}^*) \cong \varinjlim H^1(G_{L/K }, M^{G_{\overline{K}/ L}}).$$ Here, $M^{G_{\overline{K}/ L}}$ is the set of elements of $M$ fixed by $G_{\overline{K} / L}$. My background with direct limits is mostly from exercises in Atiyah Macdonald, so I would like to do this by showing that $H^1(G_{\overline{K} / K}, \overline{K}^*)$ has the universal property of direct limits. I can construct the direct system using the inflation maps, but I am stuck on how to define $\varphi : H^1(G_{\overline{K} / K}, \overline{K}^*) \to N$, such that for all finite Galois extensions $L$ of $K$, $\varphi_L = \varphi \circ \iota_L$, where $\iota_L$ is the inflation map from $H^1(G_{L / K}, M^{G_{\overline{K}/L}})$ to $H^1(G_{\overline{K} / K}, M)$, and $\varphi_L : H^1(G_{L / K}, M^{G_{\overline{K}/L}}) \to N$ is given (such that all the $\varphi_L$ are compatible with the direct system). How can we define such a map?

P.S. I know that this result appears in a lot of standard sources, but usually with much more power/background in homological algebra than I am looking for. Weibel has this result for profinite groups, but the proof uses $\delta$-functors, for example. The closest I've seen to what I'm looking for is in David Harari's Galois Cohomology and Class Field Theory, but even there I had trouble translating the general situation he presents into the concrete function that I'm looking for.