Karigiannis's metric-building formula is not unique to the Cayley form: a mathematical preprint presents additional 4-forms that define non-degenerate Riemannian metrics. One explicit construction is an SU(4)-stabilized one-parameter family, which intersects the Spin(7) orbit at t = 1 but is not metric at t = 0.
That finding goes to the heart of the note: it asks which 4-forms can produce metrics through Karigiannis's formula, whether the Cayley form is unique, and how the paper's definitions of non-degeneracy relate. It also addresses part of the Salamon-Walpuski conjecture, which asks whether a vanishing self-wedge forces degeneracy.
To follow the argument, it helps to separate the construction from the terminology. In the note, metricity means that a form succeeds in producing a metric through the recipe. The paper describes extracting a metric by taking a root and then polarising when the norm associated with a Cayley form is positive definite. Polarisation supplies the bilinear form used as the metric; the research question is which inputs can meet that algebraic requirement.
A family, not a one-off exception
The non-uniqueness is not presented as a single accidental form. The beta_t family is stabilized by SU(4) and supplies additional non-degenerate Riemannian metric forms. At t = 1 it intersects the Spin(7) orbit; at t = 0, the form is not metric. The contrast shows that even within one structured family, metric behavior depends on the parameter.
That keeps the conclusion narrower than a blanket claim about all 4-forms. The note shows that the Cayley form is not the sole metric-defining input in the setting studied, but it does not establish a complete classification of metric 4-forms. Its result is a demonstration of non-uniqueness backed by an explicit family.
The word non-degenerate hides a split
Much of the note then turns to the word non-degenerate, which is doing more than one job. Under Definition 5, a 4-form is non-degenerate when every linearly independent triple u, v, w admits an x for which the form evaluates nonzero on u, v, w, x. The paper shows that this condition implies multisymplectic non-degeneracy.
The implication is not an equivalence. In the note's xi example, the form is multisymplectically non-degenerate but degenerate under Definition 5. The degeneracy witness is u = E1, v = E2 and w = F3: for that triple, the Definition 5 requirement fails. The example shows why the two notions should not be treated as synonyms.
Theorem 2 gives equivalent algebraic characterizations of the note's non-degeneracy condition. One uses symplectic double contractions on a quotient; another uses a nonzero triple-wedge expression for independent vectors. Those formulations let the same condition be recognized in different mathematical languages, while leaving intact the distinction between Definition 5 and multisymplectic non-degeneracy.
A conjecture gets a one-way answer
That distinction also frames the Salamon-Walpuski question. The conjecture asks whether alpha wedge alpha = 0 implies that alpha is degenerate; alpha wedge alpha is the form's self-wedge. The note does not complete that implication, but it establishes a related one-way result: strong non-degeneracy implies ordinary non-degeneracy and also implies that alpha wedge alpha is nonzero.
The gap between the two strength levels is concrete. The beta one-half example is non-degenerate but not strongly non-degenerate, demonstrating that strong and ordinary non-degeneracy are not equivalent. In other words, the stronger test is sufficient in the directions proved, but passing the ordinary test does not automatically put a form in the stronger class.
The metric question remains open
A separate part of the analysis examines an area-metric construction. It decomposes G_alpha into a normalized self-wedge term multiplied by alpha, minus a remainder called H_alpha. The expression makes the self-wedge part of the structure, but it does not by itself supply a general rule for when an arbitrary 4-form is metric.
Lemma 4 adds a signature constraint to that construction. It states that rho-star(G_alpha) is congruent to rho(tilde-star alpha), and concludes that G_alpha is always indefinite. For a general reader, indefinite means that it has mixed signs rather than being entirely positive or entirely negative. This is a conclusion about the area-metric construction defined in the note, not about every metric recovered from a 4-form.
The relationship between metricity and self-wedge is therefore asymmetric. The note states that metricity requires a nonvanishing self-wedge. Yet strong non-degeneracy does not guarantee metricity for the beta_t examples outside the stated interval when t is not 0. The converse question—whether nonvanishing self-wedge is enough for metricity—is left unresolved.
Taken together, the results widen the range of forms that can enter Karigiannis's construction, clarify which non-degeneracy implications do hold, and show where equivalence fails. They stop short of a general metricity condition or a full resolution of the Salamon-Walpuski conjecture, so the paper's contribution is a set of exact algebraic examples and implications rather than a final classification.
The front matter identifies the document as the preprint arXiv:2608.20200v2 in math.DG, dated 21 Aug 2026. Its open questions include the precise metricity condition, the converse between self-wedge and metricity, and whether vanishing self-wedge always forces degeneracy.
Paper data and sources
Original title: A note on 4-forms in 8-dimensions
Authors: Sam Close
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-20
DOI: Not available
Original paper · Full text