Stringy Canonical Forms in Positive Geometry

• Stringy canonical forms are α′-deformations of rational canonical forms that interpolate between polytope geometry and string amplitudes via meromorphic integrals.
• They employ tropical geometry, residue analysis, and saddle-point methods to reveal the combinatorial structure and factorization properties of positive geometries.
• Applications span generalized permutohedra, associahedra, cyclohedra, and Grassmannian integrals, offering a bridge between mathematical physics and modern scattering amplitudes.

Stringy canonical forms are $\alpha'$-deformations of canonical rational forms associated to positive geometries, particularly convex polytopes, realized as meromorphic integrals whose structure simultaneously encodes both the geometry of Newton polytopes and physical principles underlying string amplitudes. These forms interpolate between the worldsheet integrals of string theory and the canonical functions of polytopes connected to scattering amplitudes, and are central to the combinatorial and geometric approach to understanding particle and string scattering from the perspective of positive geometries (Arkani-Hamed et al., 2019, He et al., 2020, 2002.04528).

1. Definition and Construction

Given a set of subtraction-free polynomials $\{p_I(x)\}$ with positive coefficients in real, strictly positive variables $x=(x_1, \dots, x_d)$ and positive parameters $c_I > 0$, the stringy canonical form is defined as: $$\mathcal I_{\{p\}}(X, c) = (\alpha')^d \int_{\mathbb{R}_+^d} \left( \prod_{i=1}^d \frac{dx_i}{x_i} \right) x_i^{\alpha' X_i} \prod_{I=1}^m p_I(x)^{-\alpha' c_I}$$ where $X_i$ are monomial weights and the exponents $(X_i, c_I)$ must lie in the interior of the Minkowski sum $\sum_{I} c_I [p_I]$, with $[p_I]$ the Newton polytope of $p_I$. For a general polytope $\{p_I(x)\}$0, this integral is a meromorphic function in $\{p_I(x)\}$1 which, in the limit $\{p_I(x)\}$2, reproduces the canonical rational form $\{p_I(x)\}$3 associated to $\{p_I(x)\}$4 (Arkani-Hamed et al., 2019). This setup generalizes the Koba–Nielsen integrals of string theory to arbitrary positive geometries.

2. Geometric and Combinatorial Framework

The stringy canonical form provides a geometric bridge:

• $\{p_I(x)\}$5 limit: The integral reduces, via Laplace’s method and tropical techniques, to the canonical form of the Newton-sum polytope $\{p_I(x)\}$6, which can be interpreted as the volume form on the dual polytope $\{p_I(x)\}$7 cut out by constraints $\{p_I(x)\}$8 (Arkani-Hamed et al., 2019, 2002.04528).
• Tropicalization: Tropicalizing the integrand gives piecewise-linear functions whose nonnegativity defines the convergence region. This establishes direct links to the Minkowski sums of polytopes and tropical geometry.
• Residue structure: At any finite value of $\{p_I(x)\}$9, simple poles correspond to the facets of $x=(x_1, \dots, x_d)$0, and residues along these poles yield lower-dimensional stringy forms for the facets, making the family of stringy canonical forms closed under taking residues.

3. Factorization, Binary Geometries, and $x=(x_1, \dots, x_d)$1-Equations

For certain polytope families—most notably generalized permutohedra, associahedra ($x=(x_1, \dots, x_d)$2 type), and cyclohedra ($x=(x_1, \dots, x_d)$3 type)—the stringy canonical forms exhibit remarkable rigidity and factorization properties (He et al., 2020):

• Rigid stringy integrals: The big polyhedron has the combinatorics of a simplex, and facet parameters can be expressed uniquely in terms of $x=(x_1, \dots, x_d)$4.
• Binary geometry: The configuration space, parametrized by $x=(x_1, \dots, x_d)$5, has the property that the only boundaries arise at $x=(x_1, \dots, x_d)$6 or $x=(x_1, \dots, x_d)$7, and the approach to $x=(x_1, \dots, x_d)$8 forces incompatible $x=(x_1, \dots, x_d)$9. In cluster algebraic cases, perfect $c_I > 0$0-equations of the form $c_I > 0$1 hold, with the compatibility degree $c_I > 0$2 controlling their mutual intimate structure.
• Factorization at finite $c_I > 0$3: Poles arise as certain $c_I > 0$4, causing the integral to factorize into products of lower-dimensional stringy canonical forms associated to the subdivided building sets of the polytope. These properties are tightly controlled in cluster and generalized permutohedron settings, extending dualities familiar from string theory (He et al., 2020).

4. Saddle-Point Pushforward, Scattering Equation Map, and Physical Limits

The exponent in the integrand defines a potential function $c_I > 0$5 whose saddle-point equations,

$c_I > 0$6

establish a diffeomorphic map $c_I > 0$7 from the positive orthant to the interior of $c_I > 0$8 (Arkani-Hamed et al., 2019). This provides the pushforward mechanism: $c_I > 0$9 In the string theory context:

• $$\mathcal I_{\{p\}}(X, c) = (\alpha')^d \int_{\mathbb{R}_+^d} \left( \prod_{i=1}^d \frac{dx_i}{x_i} \right) x_i^{\alpha' X_i} \prod_{I=1}^m p_I(x)^{-\alpha' c_I}$$0: The canonical form encodes the bi-adjoint $$\mathcal I_{\{p\}}(X, c) = (\alpha')^d \int_{\mathbb{R}_+^d} \left( \prod_{i=1}^d \frac{dx_i}{x_i} \right) x_i^{\alpha' X_i} \prod_{I=1}^m p_I(x)^{-\alpha' c_I}$$1 amplitude in the kinematic polytope (associahedron).
• $$\mathcal I_{\{p\}}(X, c) = (\alpha')^d \int_{\mathbb{R}_+^d} \left( \prod_{i=1}^d \frac{dx_i}{x_i} \right) x_i^{\alpha' X_i} \prod_{I=1}^m p_I(x)^{-\alpha' c_I}$$2: The map matches the CHY scattering equations/Gross-Mende saddle-point equations. Thus, stringy canonical forms interpolate between field theory (polytope canonical form) and string theory, with the saddle-point structure providing a diffeomorphic correspondence between integration variables and polytope geometry (Arkani-Hamed et al., 2019, 2002.04528).

5. Specialized Classes: Cluster Algebras, Generalized Permutohedra, and Grassmannians

Stringy canonical forms extend uniformly to large classes of polytopes:

• Associahedra ($$\mathcal I_{\{p\}}(X, c) = (\alpha')^d \int_{\mathbb{R}_+^d} \left( \prod_{i=1}^d \frac{dx_i}{x_i} \right) x_i^{\alpha' X_i} \prod_{I=1}^m p_I(x)^{-\alpha' c_I}$$3) and cluster algebras: Stringy integrals on these geometries match open-string worldsheet integrals; the $$\mathcal I_{\{p\}}(X, c) = (\alpha')^d \int_{\mathbb{R}_+^d} \left( \prod_{i=1}^d \frac{dx_i}{x_i} \right) x_i^{\alpha' X_i} \prod_{I=1}^m p_I(x)^{-\alpha' c_I}$$4-variables and perfect $$\mathcal I_{\{p\}}(X, c) = (\alpha')^d \int_{\mathbb{R}_+^d} \left( \prod_{i=1}^d \frac{dx_i}{x_i} \right) x_i^{\alpha' X_i} \prod_{I=1}^m p_I(x)^{-\alpha' c_I}$$5-equations encode the combinatorics of planar scattering.
• Cyclohedra ($$\mathcal I_{\{p\}}(X, c) = (\alpha')^d \int_{\mathbb{R}_+^d} \left( \prod_{i=1}^d \frac{dx_i}{x_i} \right) x_i^{\alpha' X_i} \prod_{I=1}^m p_I(x)^{-\alpha' c_I}$$6) and degenerations: Similar structures hold, sometimes yielding infinitely many rigid and binary degenerations with perfect $$\mathcal I_{\{p\}}(X, c) = (\alpha')^d \int_{\mathbb{R}_+^d} \left( \prod_{i=1}^d \frac{dx_i}{x_i} \right) x_i^{\alpha' X_i} \prod_{I=1}^m p_I(x)^{-\alpha' c_I}$$7-equations.
• Generalized permutohedra: Every such polytope can be associated to a stringy canonical form; while in general the $$\mathcal I_{\{p\}}(X, c) = (\alpha')^d \int_{\mathbb{R}_+^d} \left( \prod_{i=1}^d \frac{dx_i}{x_i} \right) x_i^{\alpha' X_i} \prod_{I=1}^m p_I(x)^{-\alpha' c_I}$$8-equations are not perfect, the configuration space remains binary and the associated integral rigid (He et al., 2020).
• Grassmannian string integrals: Regulating the positive Grassmannian with Plücker minors powered by kinematic variables yields stringy canonical forms whose $$\mathcal I_{\{p\}}(X, c) = (\alpha')^d \int_{\mathbb{R}_+^d} \left( \prod_{i=1}^d \frac{dx_i}{x_i} \right) x_i^{\alpha' X_i} \prod_{I=1}^m p_I(x)^{-\alpha' c_I}$$9 limits are canonical forms of tropical Grassmannian polytopes, with interesting connections to cluster algebras, twisted cohomology, and the geometry of amplitudes (Arkani-Hamed et al., 2019).

6. Algorithmic Computation and Resolution: The Blow-Up Algorithm

Evaluation of leading $X_i$0 asymptotics of stringy canonical forms is algorithmically accessible through sector decomposition and toric blow-ups:

• Blow-up method: The integration region is recursively subdivided into regions where the integrand is a product of monomials (normal-crossing). An adaptation of Hironaka’s polyhedra game ensures every Newton polyhedron can be reduced to an orthant via allowed blow-up and rescaling moves (2002.04528).
• Matrix formulation: At each step, exponent matrices encode the geometry of the polynomials; sector decompositions and row operations enable systematic factorization or row reduction, expediting computation of the leading order coefficient.
• Complexity: The algorithm terminates and outputs rational results but may generate numerous spurious poles/sectors that must be post-processed for true polytope facets.
• Physical and geometric implications: The same machinery applies to Feynman parameter integrals, highlighting deep connections between stringy canonical forms, positive geometry, and standard techniques in particle physics (2002.04528).

7. Physical Interpretation and Open Directions

Stringy canonical forms generalize numerous physical amplitudes:

• Tree-level open-string amplitudes: The Koba–Nielsen integral is the archetype, deformed to arbitrary positive geometries.
• Factorization and corrections: Binary geometry and factorization at finite $X_i$1 guarantee the consistent analytic structure, UV softness, and systematic control over higher $X_i$2 corrections reorganized by polytope facets.
• Connections to cluster algebra, moduli, and worldsheet models: For cluster types $X_i$3, $X_i$4, and their degenerations, perfect $X_i$5-equations, binary geometry, and rigid integrals have been conjectured as universal features; the emergence of infinite degeneration classes and direct product geometries points to a broader landscape of positive geometries.
• Outstanding problems: Classification of binary vs. non-binary geometries, extension to higher genera, explicit $X_i$6-series via $X_i$7-hypergeometric functions, and worldsheet/string worldvolume realization remain prominent open questions (He et al., 2020).

Principal References

Topic	Paper Title	arXiv ID
Foundation of stringy canonical forms	"Stringy Canonical Forms"	(Arkani-Hamed et al., 2019)
Binary geometries and factorization	"Stringy canonical forms and binary geometries from associahedra, cyclohedra and generalized permutohedra"	(He et al., 2020)
Algorithmic blow-ups/resolution	"Blowing up Stringy Canonical Forms: An Algorithm to Win a Simplified Hironaka's Polyhedra Game"	(2002.04528)