This is a second post from JW, following on from Part I.
The Galois group of as a geometric fundamental group.
In this follow-up post, I’d like to relay something Peter Scholze told me last fall. It concerns the Galois group , and how this is isomorphic to the étale fundamental group of some geometric object
, which is defined over an algebraically closed field. (Of course,
is isomorphic to the étale fundamental group of
, but that’s tautological.) We can even ask that this isomorphism be “natural” in the sense that there is an equivalence of categories between finite étale covers of
and finite étale
-algebras. This is the sense in which the absolute Galois groups of a perfectoid field
and its tilt
are naturally isomorphic, cf. the comments following my first post. Anyway, one afternoon during his visit to Boston, Scholze told me the following theorem:
Theorem 1. Let be a complete algebraically closed valued field containing
. There exists an “object”
defined over
, which has the property that there is an equivalence of categories between finite étale covers of
and finite étale
-algebras.
(I will explain later what sort of thing actually is–in brief, it is the quotient by
of the punctured perfectoid open disc over
.)
Incredulous, I demanded an explanation, which he gave later that evening, at an Indian restaurant in Harvard Square, with Hadi Hedayatzadeh also present. I left this conversation with a giddy feeling that Theorem 1 could bring a lot of clarity to the local Langlands program. Geometry is easier than arithmetic, after all. If you want to classify representations of , now all you have to do is classify local systems on
, which is surely easier. I envisioned the
-adic local Langlands correspondence for
falling away in a tidy puff.
None of this came to pass (yet). Still, Scholze’s theorem and its proof are really elegant stuff. What follows is a motivated exposition of two curry-stained pages of notes from that conversation last fall. In what follows, will always refer to a perfectoid field in characteristic
, and
will always refer to a perfectoid field in characteristic 0, the idea being that
as usual.
In the last post, we considered the tilting process , which inputs a perfectoid field in characteristic 0 and outputs a perfectoid field in characteristic
. Then there is an equivalence of categories between étale
-algebras and étale
-algebras.
The tilting process works in families as well:
Theorem 2. Let be a perfectoid space over a perfectoid field
in characteristic 0. Then
is a perfectoid space over
, whose underlying topological space is homeomorphic to that of
. There is an equivalence of étale sites
.
Thus if we have a perfectoid space in characteristic , any two of its un-tilts have equivalent étale sites (and therefore the same étale fundamental group). This draws our attention to the problem of un-tilting entire perfectoid spaces. Theorem 1 will be proved by un-tilting a certain perfectoid space in two ways: one will involve the mysterious object
, and the other will involve the Fargues-Fontaine curve.
Adic spaces and perfectoid spaces.
We need a little background on adic spaces and perfectoid spaces. Let me just recall the main gadgets: one starts with a pair consisting of a topological ring
and a bounded open subring
, such that the topology on
is induced by an ideal of
(there are other restrictions as well). Then
is the set of equivalence classes of continuous valuations
on
which satisfy
for all
. Under the right hypotheses on
,
forms a topological space equipped two sheaves of rings
and
, whose global sections are
and
, respectively. General adic spaces are formed by gluing together affinoid spaces of the form
.
A basic example is , which has two valuations: the one with
(the special point) and the one with
(the generic point, which we’ll call
). Another is
, which is the adic version of the closed unit disc (note that
for all valuations
). (This is like the Berkovich unit disc but with another class of exotic points added, corresponding to valuations of rank 2.) If that’s the closed disc, what’s the open disc? (It can’t be of the form
, since affinoids are always compact.) One way to construct it is to glue together an ascending sequence of closed discs, but the most direct way is to start with
and to take its fiber over the generic point of
, meaning the set of continuous valuations on
for which
. This is the generic fiber of the formal unit disc
. We will write this as
, where
is the generic point of
.
Now if is a perfectoid field, one has the notion of a perfectoid affinoid
over
: this means approximately that
is a
-algebra,
is an
-algebra, and the Frobenius map is surjective on
. A typical example is
, the perfectoid closed disk. A perfectoid space over
is an adic spaces admitting a covering by perfectoid affinoids over
. For instance, let
I’ll call the perfectoid open disc over
. The tilting process
locally looks like
, where
and
. Then
, the perfectoid disc over
.
Let be a perfectoid field in characteristic
, and let
be any non-unit. Let
be the punctured open disc, so that
is the set of continuous valuations on
for which
and
.
I am now going to write down two really different un-tilts of . One is simply
, for any un-tilt
of
. For the other, we notice that
isn’t just a perfectoid space over
, it’s also a perfectoid space over the field
. That is, the map
induces a morphism
in which maps to the generic point. Thus there is a map
and this presents
as a perfectoid space over
. So, rather like a Necker cube which pops in and out,
is simultaneously a perfectoid space over the bases
and
.
Let be any un-tilt of
. The other un-tilt of
will be a perfectoid space
over
whose tilt is
. Consider the ring
. Then we have
, and
This calculation shows that the tilt of is the set of continuous valuations on
for which
. Let
where
refers to the valuation on
pulled back from the valuation on
(which is an unramified extension of
). That is,
is the set of valuations on
satisfying
and
. Then the tilt of
is the set of valuations on
satisfying
and
, which is exactly
.
As the notation suggests, is the base change to
of an adic space
:
this being the set of continuous valuations on for which
and
. Note that if
is an un-tilt of
, we get a valuation on
by pulling back a valuation on
through the map
. The valuations arising this way all satisfy
. Thus
contains the set of un-tilts of
, in a ready-made geometric object (an adic space over
).
RMB commented on the previous post, asking for an interpretation of the “non-classical” points of . If
is any perfectoid field in characteristic 0, then the
-points of
correspond to injections
. The “classical”
-points correspond to injections where
is finite, but one expects there are plenty of points of
which do not arise this way.
From here it is not difficult to show that:
Proposition 1 The adic space attached to the Fargues-Fontaine curve is isomorphic to the quotient
.
From now on I’ll use to denote the adic space (over
), rather than the projective curve; then
is a perfectoid space, equal to
. We have shown that the tilt of
and is equal to
. Now, the Frobenius map
induces an automorphism of
, which in turn induces an automorphism of
and its tilt
. On the other hand, there is the automorphism
of
, which is
-linear and sends
to
. The composition of
and
induces the absolute Frobenius on
, which induces the identity on
, since
and
are equivalent valuations. This shows that the tilt of
is equal to
. Therefore:
Proposition 2 The étale site of is equivalent to the étale site of
.
I would now like to specialize a bit. Let be an algebraically closed valued field containing
. The roles of
and
will be played by
and
, respectively, and we will specialize
to be the perfectoid field
. The above proposition tells us that the category of finite étale covers of
is equivalent to the category of finite étale covers of
. At this point we apply a theorem of Fargues-Fontaine:
Theorem 3 After base-changing to an algebraically closed field, is simply connected.
This theorem is a consequence of the classification of vector bundles on , which winds up looking a lot like the same classification for the projective line over a field. As a consequence, there is an equivalence of categories
between finite étale covers of
and finite étale
-algebras. Combining this with Prop. 2, we get equivalences between the following:
- Finite etale covers of
,
-
Finite etale covers of
,
- Finite etale
-algebras.
And therefore we get a surprising isomorphism:
All we have to do now to prove Thm. 1 is to descend this picture from down to
. The group on the right has an action of
; this action corresponds to an action of
on
, which for
is given by the familiar formula
These actions of
and
combine to give an action of
on
, in which
acts by
.
In which case, we have an equivalence
-
-equivariant finite etale covers of
,
- Finite etale
-algebras.
Lastly, Theorem 1 promised an object over
, not
. But now we can just use the “easy” un-tilt of
, namely
, so long as we can check that the action of
lifts to
. It does, and you can even give formulas for the action of an element
on
(they involve limits, even for
).
(Pedantic note: the right way to view is that it is the generic fiber of
, the universal cover of the multiplicative
-divisible group
over the base
. It so happens that
is representable by
–I explain this in my Arizona Winter School lectures. Taking generic fibers, we find that
is a
-vector space object in the category of perfectoid spaces over
, and that
acts on
.)
The tilting equivalence (Thm. 2) now shows that
Theorem 4 There is an equivalence of categories between finite -equivariant étale covers of
and finite étale
-algebras.
This is the precise form of Thm. 1 that we wanted. For the object , we can attempt to take the quotient
. This quotient looks horrid–we are quotienting by a group action whose orbits are far from being discrete. But, should a reasonable category be found for
, the étale fundamental group of
can only be
. Brilliant!
So p-adic Gal(Q_pbar/Q_p)-representations are equivalent to Q_p^\times-equivariant vector bundles over D_C^*? This seems almost but not quite entirely (un)like the theory of (phi,Gamma)-modules. What is the actual relationship between these two constructions?
Well, not exactly: You don’t get vector bundles, but Z_p – (or Q_p – ) local systems again. Nonetheless, one can regard this as some version of (phi,Gamma)-modules: Over perfectoid spaces in general, one also has a version of the equivalence between Z_p – local systems and phi-modules over an appropriate sheaf; composing the two functors gives something that can be expressed in terms of the usual (phi,Gamma)-module. Interestingly, the object D_C^* / Q_p^* is already canonically defined over Q_p, which gives rise to an object whose absolute fundamental group is the product of two copies of the absolute Galois group of Q_p. (I think it is really a product, not a semidirect product, but I didn’t check carefully.) Also, the object D_C^* / Q_p^* (or even its version over Q_p) appears naturally in Jared’s construction of the Lubin-Tate tower at infinite level.