The Euler and Tits forms of the generalized Kronecker quiver #
This file evaluates the Euler and Tits forms of the generalized Kronecker quiver in the
coordinates of its two vertices. The arithmetic of the Tits form pins down exactly where the
boundary of Gabriel's theorem lies: writing n for the number of arrows,
titsForm d = d src ^ 2 + d tgt ^ 2 - n * (d src * d tgt);- it is positive definite iff
n ≤ 1; - it is positive semidefinite iff
n ≤ 2, and forn = 2it is the perfect square(d src - d tgt) ^ 2, whose radical is the line spanned by the constant vector(1, 1)-- the null root of the affine root systemÃ₁.
Since reflecting a quiver changes its orientation but not its underlying graph, the quiver
reflected at tgt has the same Tits form, and hence the same thresholds.
Main results #
TauCeti.Quiver.Kronecker.eulerForm_applyandTauCeti.Quiver.Kronecker.titsForm_apply: the Euler and Tits forms in coordinates.TauCeti.Quiver.Kronecker.titsForm_posDef_iffandTauCeti.Quiver.Kronecker.titsForm_nonneg_iff: the two thresholdsn ≤ 1andn ≤ 2.TauCeti.Quiver.Kronecker.titsForm_reflect_posDef: the quiver reflected attgthas positive definite Tits form under the same thresholdn ≤ 1.TauCeti.Quiver.Kronecker.titsForm_eq_zero_iff_exists_smulandTauCeti.Quiver.Kronecker.titsForm_eq_one_iff: for the Kronecker quiver the radical of the Tits form is the line spanned by(1, 1), and the vectors of Tits norm one are those whose two coordinates differ by one.
References #
This file supplies the positive semidefinite Tits form (a - b) ^ 2 with radical (1, 1) asked
for by the “Kronecker quiver” worked example of
TauCetiRoadmap/RepresentationTheory/QuiverRepresentations/README.md. See Derksen--Weyman, An
Introduction to Quiver Representations, and Assem--Simson--Skowroński, Elements of the
Representation Theory of Associative Algebras I, Ch. II.
The Tits form on the constant dimension vector (1, 1) is 2 - n. This single value decides
both thresholds below.
The Tits form of the generalized Kronecker quiver is positive semidefinite exactly when there
are at most two arrows. Two arrows is the Kronecker quiver • ⇉ •, of affine type Ã₁, the
boundary case of Gabriel's theorem; three or more arrows makes the form indefinite.
With at most one arrow the Tits form is positive definite: these are the quivers of type
A₁ ⊔ A₁ (no arrow) and A₂ (one arrow).
The Tits form of the generalized Kronecker quiver is positive definite exactly when there is at
most one arrow: no arrow gives the disconnected Dynkin quiver A₁ ⊔ A₁, and one arrow gives A₂.
Every other generalized Kronecker quiver falls outside the Dynkin classification.
With at most one arrow the quiver reflected at tgt has positive definite Tits form:
reflecting changes the orientation of a quiver, not its underlying graph, and
TauCeti.Quiver.Kronecker.titsForm_posDef covers the underlying graph.
The Kronecker quiver itself #
The radical of the Tits form of the Kronecker quiver is the line spanned by the constant vector
(1, 1): the null root of the affine root system Ã₁.
For the Kronecker quiver the integer vectors of Tits norm one -- the real roots of the affine
root system Ã₁, positive and negative alike -- are exactly those whose two coordinates differ by
one, that is (m, m + 1) and (m + 1, m).
The Tits form of the Kronecker quiver is not positive definite: it is isotropic on the nonzero
vector (1, 1). Together with TauCeti.Quiver.Kronecker.titsForm_nonneg this says it is positive
semidefinite but degenerate, which is what places the Kronecker quiver on the boundary of Gabriel's
dichotomy.