Cominuscule subvarieties of flag varieties

Benjamin McKay (University College Cork)

arxiv: 2002.08193 · v8 · submitted 2020-02-19 · 🧮 math.AG · math.DG

Cominuscule subvarieties of flag varieties

Benjamin McKay (University College Cork) This is my paper

Pith reviewed 2026-05-24 14:17 UTC · model grok-4.3

classification 🧮 math.AG math.DG

keywords flag varietiescominuscule subvarietiesDynkin diagramshomogeneous varietiesalgebraic geometryLie theory

0 comments

The pith

Every flag variety contains a naturally defined homogeneous cominuscule subvariety whose Dynkin diagram is read from the original diagram.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper proves that any flag variety admits a canonical subvariety inside it that is both homogeneous under a group action and cominuscule. The construction requires no arbitrary choices and works uniformly for every flag variety. The authors give an explicit rule that produces the Dynkin diagram of this subvariety directly from the Dynkin diagram of the ambient flag variety. A reader would care because the result supplies one uniform object that sits inside every flag variety and inherits a simple combinatorial description.

Core claim

We show that every flag variety contains a naturally defined homogeneous cominuscule subvariety. From the Dynkin diagram of the flag variety, we compute the Dynkin diagram of that subvariety.

What carries the argument

The naturally defined homogeneous cominuscule subvariety obtained by a choice-free construction from the Dynkin diagram of the flag variety.

If this is right

The subvariety is homogeneous under the natural group action on the flag variety.
Its Dynkin diagram is obtained by a deterministic rule applied to the Dynkin diagram of the ambient variety.
The construction applies without exception to every flag variety.
The subvariety is cominuscule by definition of the construction.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The diagram rule may allow direct computation of numerical invariants such as dimension or degree for the subvariety in any given case.
The same diagram operation could be tested on other homogeneous spaces to see whether an analogous subvariety appears.

Load-bearing premise

There exists a single natural construction that produces, for every flag variety, a subvariety that is simultaneously homogeneous and cominuscule.

What would settle it

An explicit flag variety whose Dynkin diagram yields no homogeneous cominuscule subvariety under the stated construction rule.

read the original abstract

We show that every flag variety contains a naturally defined homogeneous cominuscule subvariety. From the Dynkin diagram of the flag variety, we compute the Dynkin diagram of that subvariety.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper claims a uniform, choice-free construction of a homogeneous cominuscule subvariety inside every flag variety, readable directly from the Dynkin diagram.

read the letter

The central claim is that every flag variety G/P carries a naturally defined homogeneous cominuscule subvariety whose Dynkin diagram is determined by a rule from the diagram of G/P. This is presented as a single construction that works for all parabolic types without extra choices. If the full paper delivers an explicit definition and verifies the cominuscule property, that would be the main new piece: a uniform diagram-based recipe instead of case-by-case lists. The approach looks useful for people who need consistent examples across types for classification or computation inside flag varieties. The paper does well in aiming for canonicity, which often simplifies later work in representation theory settings. The main soft spot is that the abstract supplies no definition of the subvariety, no check that it is homogeneous or cominuscule, and no sample computation on a low-rank case. Without those steps visible, it is impossible to confirm the construction is truly natural or that it avoids hidden choices for some Dynkin types. The soundness score from the abstract alone is therefore low, but that is expected when only the statement is given. The math itself does not show internal contradictions or circular steps in what is written. This paper is aimed at specialists in algebraic geometry and Lie theory who already work with flag varieties and want a uniform tool for subvarieties. A reader who needs concrete examples or diagram rules for further calculations would get direct value. It deserves a serious referee because the claim, if substantiated in the body, would be a practical addition to the toolkit even if it requires revisions for clarity and examples. I would send it to peer review.

Referee Report

1 major / 0 minor

Summary. The paper claims that every flag variety contains a naturally defined homogeneous cominuscule subvariety and that the Dynkin diagram of this subvariety can be computed directly from the Dynkin diagram of the flag variety.

Significance. If the asserted canonical construction exists and produces a homogeneous cominuscule subvariety for every parabolic type, the result would supply a uniform way to associate cominuscule varieties to arbitrary flag varieties and to read off their root data, which could streamline classification arguments and geometric constructions in the theory of homogeneous spaces.

major comments (1)

[Abstract] Abstract: the central claim asserts the existence of a single natural, choice-free construction producing a homogeneous cominuscule subvariety for every flag variety G/P, yet no definition of the subvariety, no explicit construction, and no verification that the resulting variety is cominuscule or homogeneous appear in the text; without these the claim cannot be checked.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their report. The major comment correctly identifies that the manuscript would be strengthened by an explicit early definition of the subvariety and a self-contained verification of its properties. We will revise the paper to address this.

read point-by-point responses

Referee: [Abstract] Abstract: the central claim asserts the existence of a single natural, choice-free construction producing a homogeneous cominuscule subvariety for every flag variety G/P, yet no definition of the subvariety, no explicit construction, and no verification that the resulting variety is cominuscule or homogeneous appear in the text; without these the claim cannot be checked.

Authors: We agree that the current presentation does not make the construction immediately verifiable from the abstract and early sections. In the revised manuscript we will add an explicit definition of the cominuscule subvariety (as the unique homogeneous subvariety whose tangent space at the base point is the sum of the negative root spaces corresponding to the cominuscule roots read off the Dynkin diagram of G/P), include a direct construction from the parabolic P, and verify homogeneity and cominusculeness by checking the root-system criterion. The Dynkin-diagram rule will be stated as a theorem with proof. These additions will appear in a new subsection of the introduction and in Section 2. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper's central claim is the existence of a canonical, choice-free construction that produces a homogeneous cominuscule subvariety inside any flag variety G/P, with the subvariety's Dynkin diagram read off from that of G/P. No equations, fitted parameters, self-citations, or ansatzes are exhibited in the abstract or description that would reduce the result to a definition or prior self-referential step. The derivation is presented as a direct geometric/algebraic construction in the language of Dynkin diagrams and parabolic subgroups, which is self-contained against external benchmarks in Lie theory and does not rely on load-bearing self-citation or renaming of known results by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract supplies no information on free parameters, background axioms, or new entities; ledger is therefore empty.

pith-pipeline@v0.9.0 · 5535 in / 1038 out tokens · 25674 ms · 2026-05-24T14:17:31.123296+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages

[1]

Flag varieties, 4 3.4

Introduction, 3 3.3. Flag varieties, 4 3.4. Cominuscule varieties, 4

work page
[2]

Statement of the theorem, 5

work page
[3]

Gradings, 7 5.4

Reducing to root systems, 7 5.3. Gradings, 7 5.4. Associated cominuscules in rank 2, 7

work page
[4]

The associated cominuscule subvariety is a cominuscule subvariety,8

work page
[5]

The Hasse diagram of a root system,/seven.oldstyle 7.4

Hasse diagrams, /seven.oldstyle 7.3. The Hasse diagram of a root system,/seven.oldstyle 7.4. The Hasse diagram of a ﬂag variety,; 7.5. The Hasse diagram of a cominuscule variety,32

work page
[6]

Proof of the theorem, 35 :

Finding the Hasse diagram of the associated cominuscule,32 /seven.oldstyle. Proof of the theorem, 35 :. Conclusion, 55 References, 6

work page
[7]

Our eyes immediately spot in that Hasse diagram its uppermost component, which is always the same as that of a unique cominuscule variety

I n t ro d u c t i o n While we cannot draw a ﬂag variety, or even its associated root system, except in low dimensions, we can draw its Hasse diagram. Our eyes immediately spot in that Hasse diagram its uppermost component, which is always the same as that of a unique cominuscule variety. We then predict (correctly, as we will see) that each ﬂag variety ...

work page
[8]

6/seven.oldstyle:., so thatG is a semidirect product of these

p. 6/seven.oldstyle:., so thatG is a semidirect product of these. The unipotent radical of COMINUSCULE SUBV ARIETIES OF FLAG V ARIETIES 5 P is denotedG` Ă P, and a maximal reductive Levi factor is denotedG0 Ă P, so P “ G0˙G`. A ﬂag variety is cominuscule just whenG` is abelian [6] p. 4;8 §5.4.5. There is an involutive elementwPG in the Weyl group ofG so t...

work page
[9]

With denoting a node which could be either aor , the associated cominuscule subvarieties are: G{P ˘G{ ˘P Ar p q p q ℓ ℓ Br ℓ ℓ Cr ℓ ℓ Dr ℓ ℓ´1 continued

S tat e m e n t o f t h e t h e o r e m Theorem /one.oldstyle. With denoting a node which could be either aor , the associated cominuscule subvarieties are: G{P ˘G{ ˘P Ar p q p q ℓ ℓ Br ℓ ℓ Cr ℓ ℓ Dr ℓ ℓ´1 continued ... 6 BENJAMIN M cKAY ...continued G{P ˘G{ ˘P ` E6 E7 continued ... COMINUSCULE SUBV ARIETIES OF FLAG V ARIETIES 7 ...continued G{P ˘G{ ˘P E8 F4 G2

work page
[10]

Gradings

R e d u c i n g t o ro o t s y s t e m s 5.3. Gradings. A root system with a basis of simple rootsα1,...,α r is graded: each root řniαi has grade ř ini. For a ﬂag varietyX“G{P, the root system is also P-graded by the same sum, but only over the noncompact simple roots. The box is the set ofP-maximal roots, terminology which roughly follows [5], [;] p. 7/s...

work page
[11]

T h e a s s o c i at e d c o m i n u s c u l e s u b va r i e t y i s a c o m i n u s c u l e s u b va r i e t y We prove lemma3 on page 5. Proof. Take notation as above for a ﬂag varietyX“G{P. To prove thatp ˘X, ˘Gq is eﬀective and cominuscule, it is suﬃcient to prove this forX an irreducible ﬂag variety, i.e.G simple, as otherwiseX is a product of irred...

work page
[12]

The Hasse diagram of a root system.Recall theHasse diagramof a root system

H a s s e d i ag r a m s 7.3. The Hasse diagram of a root system.Recall theHasse diagramof a root system. Given an irreducible reduced root system with a choice of basis of simple roots, and some ordering of the simple roots, asuccessor of a positive rootα is a /eight.oldstyle BENJAMIN M cKAY positive root of the formα`β for a positive simple rootβ. The H...

work page
[13]

Take the ﬂag variety ; order the roots 1 2 3 4

F i n d i n g t h e H a s s e d i ag r a m o f t h e a s s o c i at e d c o m i n u s c u l e Example 8. Take the ﬂag variety ; order the roots 1 2 3 4. The Hasse diagram ofF4 is COMINUSCULE SUBV ARIETIES OF FLAG V ARIETIES 33 2 3 2 4 1 3 3 2 1 3 41 2 3 41 3 3 41 3 2 4 1 4 4 42 1 3 4 2 4 3 1 2 3 4 while that of the ﬂag variety emerges by removing all edge...

work page
[14]

/five.oldstyle/five.oldstyle/seven.oldstyle: 1 2 3 4 COMINUSCULE SUBV ARIETIES OF FLAG V ARIETIES 53 The Dynkin diagram ofpX,F 4q is this diagram with various nodes crossed

p. /five.oldstyle/five.oldstyle/seven.oldstyle: 1 2 3 4 COMINUSCULE SUBV ARIETIES OF FLAG V ARIETIES 53 The Dynkin diagram ofpX,F 4q is this diagram with various nodes crossed. The ﬂag varietypX,F 4q is not cominuscule. The associated cominuscule variety is: “p P1, PSL2q “p P2, PSL3q “p P3, PSL4q “pQ7, SO9q Proof. Look at the Hasse diagram: 2 3 2 4 1 3 3 ...

work page

[1] [1]

Flag varieties, 4 3.4

Introduction, 3 3.3. Flag varieties, 4 3.4. Cominuscule varieties, 4

work page

[2] [2]

Statement of the theorem, 5

work page

[3] [3]

Gradings, 7 5.4

Reducing to root systems, 7 5.3. Gradings, 7 5.4. Associated cominuscules in rank 2, 7

work page

[4] [4]

The associated cominuscule subvariety is a cominuscule subvariety,8

work page

[5] [5]

The Hasse diagram of a root system,/seven.oldstyle 7.4

Hasse diagrams, /seven.oldstyle 7.3. The Hasse diagram of a root system,/seven.oldstyle 7.4. The Hasse diagram of a ﬂag variety,; 7.5. The Hasse diagram of a cominuscule variety,32

work page

[6] [6]

Proof of the theorem, 35 :

Finding the Hasse diagram of the associated cominuscule,32 /seven.oldstyle. Proof of the theorem, 35 :. Conclusion, 55 References, 6

work page

[7] [7]

Our eyes immediately spot in that Hasse diagram its uppermost component, which is always the same as that of a unique cominuscule variety

I n t ro d u c t i o n While we cannot draw a ﬂag variety, or even its associated root system, except in low dimensions, we can draw its Hasse diagram. Our eyes immediately spot in that Hasse diagram its uppermost component, which is always the same as that of a unique cominuscule variety. We then predict (correctly, as we will see) that each ﬂag variety ...

work page

[8] [8]

6/seven.oldstyle:., so thatG is a semidirect product of these

p. 6/seven.oldstyle:., so thatG is a semidirect product of these. The unipotent radical of COMINUSCULE SUBV ARIETIES OF FLAG V ARIETIES 5 P is denotedG` Ă P, and a maximal reductive Levi factor is denotedG0 Ă P, so P “ G0˙G`. A ﬂag variety is cominuscule just whenG` is abelian [6] p. 4;8 §5.4.5. There is an involutive elementwPG in the Weyl group ofG so t...

work page

[9] [9]

With denoting a node which could be either aor , the associated cominuscule subvarieties are: G{P ˘G{ ˘P Ar p q p q ℓ ℓ Br ℓ ℓ Cr ℓ ℓ Dr ℓ ℓ´1 continued

S tat e m e n t o f t h e t h e o r e m Theorem /one.oldstyle. With denoting a node which could be either aor , the associated cominuscule subvarieties are: G{P ˘G{ ˘P Ar p q p q ℓ ℓ Br ℓ ℓ Cr ℓ ℓ Dr ℓ ℓ´1 continued ... 6 BENJAMIN M cKAY ...continued G{P ˘G{ ˘P ` E6 E7 continued ... COMINUSCULE SUBV ARIETIES OF FLAG V ARIETIES 7 ...continued G{P ˘G{ ˘P E8 F4 G2

work page

[10] [10]

Gradings

R e d u c i n g t o ro o t s y s t e m s 5.3. Gradings. A root system with a basis of simple rootsα1,...,α r is graded: each root řniαi has grade ř ini. For a ﬂag varietyX“G{P, the root system is also P-graded by the same sum, but only over the noncompact simple roots. The box is the set ofP-maximal roots, terminology which roughly follows [5], [;] p. 7/s...

work page

[11] [11]

T h e a s s o c i at e d c o m i n u s c u l e s u b va r i e t y i s a c o m i n u s c u l e s u b va r i e t y We prove lemma3 on page 5. Proof. Take notation as above for a ﬂag varietyX“G{P. To prove thatp ˘X, ˘Gq is eﬀective and cominuscule, it is suﬃcient to prove this forX an irreducible ﬂag variety, i.e.G simple, as otherwiseX is a product of irred...

work page

[12] [12]

The Hasse diagram of a root system.Recall theHasse diagramof a root system

H a s s e d i ag r a m s 7.3. The Hasse diagram of a root system.Recall theHasse diagramof a root system. Given an irreducible reduced root system with a choice of basis of simple roots, and some ordering of the simple roots, asuccessor of a positive rootα is a /eight.oldstyle BENJAMIN M cKAY positive root of the formα`β for a positive simple rootβ. The H...

work page

[13] [13]

Take the ﬂag variety ; order the roots 1 2 3 4

F i n d i n g t h e H a s s e d i ag r a m o f t h e a s s o c i at e d c o m i n u s c u l e Example 8. Take the ﬂag variety ; order the roots 1 2 3 4. The Hasse diagram ofF4 is COMINUSCULE SUBV ARIETIES OF FLAG V ARIETIES 33 2 3 2 4 1 3 3 2 1 3 41 2 3 41 3 3 41 3 2 4 1 4 4 42 1 3 4 2 4 3 1 2 3 4 while that of the ﬂag variety emerges by removing all edge...

work page

[14] [14]

/five.oldstyle/five.oldstyle/seven.oldstyle: 1 2 3 4 COMINUSCULE SUBV ARIETIES OF FLAG V ARIETIES 53 The Dynkin diagram ofpX,F 4q is this diagram with various nodes crossed

p. /five.oldstyle/five.oldstyle/seven.oldstyle: 1 2 3 4 COMINUSCULE SUBV ARIETIES OF FLAG V ARIETIES 53 The Dynkin diagram ofpX,F 4q is this diagram with various nodes crossed. The ﬂag varietypX,F 4q is not cominuscule. The associated cominuscule variety is: “p P1, PSL2q “p P2, PSL3q “p P3, PSL4q “pQ7, SO9q Proof. Look at the Hasse diagram: 2 3 2 4 1 3 3 ...

work page