A note on freeness

Mordechai Katzman · 2025

Read A note on freeness by Mordechai Katzman (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-16573.txt.

Introduction

Mordechai Katzman settles a tidy algebra question in a note on arXiv in 2025. Take a field extension, one number system inside a larger one. Then build power series rings over each. He asks when the larger ring is a free module over the smaller one. Free means it has a clean basis, like a vector space. The answer: freeness holds precisely when the extension is finite. The note ends with a follow-up question. What is the projective dimension of the larger ring in the other cases?

Catalog record

CREATORMordechai Katzman
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE5 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2512-16573

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A note on freeness
Mordechai Katzman

Abstract
. In this brief note we show that for a field extension K / F K/F , S = K ⁡ [ [ 𝐱 ] ] S=K[\![\mathbf{x}]\!] is a free R = F ⁡ [ [ 𝐱 ] ] R=F[\![\mathbf{x}]\!] -module precisely when K / F K/F is finite. We then raise the question what is the projective dimension of S S ?



2020 Mathematics Subject Classification
Primary 13A35, 13D07, 13J10, 13C10

0.
Introduction

Throughout this paper let
K
be a field extension of
F
, let
\mathbf{x}=x_{1},\dots,x_{n}
be a sequence of variables,
let
R=F[\![\mathbf{x}]\!]
, and let
S=K[\![\mathbf{x}]\!]
. We let
{\mathfrak{m}}
denote the maximal ideal
\mathbf{x}R
.

In this note we prove that
S
is
R
-free if and only if
K
is a finite extension of
F
.

Our motivation originates from the study of regular rings of prime characteristic where a fundamental theorem by Ernst Kunz (
[
2
, Theorem 2.1]
)
states that a local ring
A
of prime characteristic
p
is regular if and only if it is reduced and
A
is a flat
A^{p}
-module. The proof of that theorem
reduced to proving that if, with our notation,
F=K^{p}
, then
S
is a flat
R
-module. When
K
is a finite extension of
F=K^{p}
(the
F
-finite
case) it is straightforward to show
that
S
is
R
-free, (and hence
R
-flat), but in general it was not known whether this is the case in general.
We settle this:
a corollary of our main theorem is that
K[\![\mathbf{x}]\!]
is free over
(K[\![\mathbf{x}]\!])^{p}

precisely in the
F
-finite case.

1.
When is
S
a free
R
-module?

For any
f\in R
and
d\geq 0
define
\delta_{d}(f)
to be the summand of degree
d
in
f
.
Throughout this section
\Lambda
will denote an index set.
We say that a sequence

\left(f_{\lambda}\right)_{\lambda\in\Lambda}
of elements in
R

has
degree-wise finite support
if
for all
d\geq 0
,

\{\lambda\in\Lambda:\delta_{d}(f)\neq 0\}

is finite.

1.1 Lemma
.

Let
T
be a
R
-free module with free basis

\{e_{\lambda}\}_{\lambda\in\Lambda}
.

(a)

The completion
\widehat{T}
of
T
consists of
the elements of
\prod_{\lambda\in\Lambda}Re_{\lambda}

of degree-wise finite support, and

(b)

\widehat{T}=T
if and only if
\Lambda
is finite.

Proof.

The elements of

\widehat{T}=\lim_{\begin{subarray}{c}\longleftarrow\\
d\end{subarray}}\frac{T}{{\mathfrak{m}}^{d}T}\cong\lim_{\begin{subarray}{c}\longleftarrow\\
d\end{subarray}}\bigoplus_{\lambda\in\Lambda}\frac{R}{{\mathfrak{m}}^{d}}e_{\lambda}

are precisely those
(f_{\lambda})_{\lambda\in\Lambda}\in\prod_{\lambda\in\Lambda}Re_{\lambda}
for which. for any
d\geq 0
,
the set
\{\lambda\in\Lambda:f_{\lambda}\notin{\mathfrak{m}}^{d}\}
is finite.

If
\Lambda
is not finite, let
\iota:\mathbb{N}\to\Lambda
be an injection
and let

f_{\lambda}=\left\{\begin{array}[]{ll}x_{1}^{n}e_{\lambda}&\text{if }\lambda=\iota(n)\text{ for some }n\in\mathbb{N}\\
0&\text{otherwise}.\end{array}\right.

Now
f
has degree-wise finite support but
\{\lambda\in\Lambda:f_{\lambda}\neq 0\}
is not finite, so
f\notin T
.
∎

1.2 Theorem
.

(a)

If
S
is
R
-free, it has finite rank.

(b)

S
is
R
-free if and only if
K
is a finite extension of
F
.

Proof.

Since
S
is complete, if it is
R
-free the previous Lemma implies that it has finite rank.

If
S
is
R
-free, say
S\cong R^{k}
, tensoring with
R/{\mathfrak{m}}
yields
K\cong F^{k}
and so
[K:F]=k
.

If
B
is a finite
F
-basis for
K
, it also a free basis for
S
as an
R
-module.
∎

1.3 Remark
.

Note that
S
is always
R
-flat: this follows from properties of general filtered modules, see

[
1
, Chapter III, §5]
.

1.4 Corollary
.

Let
K
be a field of prime characteristic
p
;

K[\![\mathbf{x}]\!]
is a free module over
(K[\![\mathbf{x}]\!])^{p}

if and only if
[K:K^{p}]<\infty
.

Proof.

Since
K^{p}[\![\mathbf{x}]\!]
is a free
(K[\![\mathbf{x}]\!])^{p}
-module
(with free basis
\{x_{1}^{i_{1}}\dots x_{n}^{i_{n}}:0\leq i_{1},\dots,i_{n}<p\}
)
it is enough to show that
K[\![\mathbf{x}]\!]
is a free module over
K^{p}[\![\mathbf{x}]\!]

if and only if
[K:K^{p}]<\infty
, and this follows from Theorem
1.2
.
∎

2.
The projective dimension of
S

Throughout this section we assume that
K/L
is an infinite
extension.
Recall that projective modules over local rings are free
(
[
4
, Theorem 2.5]
), and so our assumption implies
\pd_{R}S>0
.

2.1 Question
.

What is the projective dimension of
S
?

Work by Barbara L. Osofsky and others
on related problems (e.g.,
[
5
]
)
suggests that answering this question in general might
involve
going down deep set-theoretical rabbit holes.
As an example of this we prove the following.

2.2 Theorem
.

Assume that
K
is countable and that

2^{\aleph_{0}}
is the
m
th infinite cardinal
\aleph_{m}
.
Then
\pd_{R}S\leq m+1
.

Proof.

Since
S
is flat over
R
,

[
3
, Théorème 1.2]

implies that
S
is a direct limit
of free
R
-submodules of
finite rank.
The cardinality of the set of free
R
-submodules of
finite rank is bounded by the cardinality of the
set of finite subsets of
S
, which equals the cardinality of

S
, namely
\aleph_{0}^{\aleph_{0}}=2^{\aleph_{0}}=\aleph_{m}
.
Now the result follows from
[
5
, Theorem 2.44]
.
∎

2.3 Example
.

As an example of a set-theoretical rabbit-hole note that
one can now conclude that

\pd_{{\mathbb{Q}}[\![\mathbf{x}]\!]}{\mathbb{Q}}(t)[\![\mathbf{x}]\!]=1,

Read the complete text