Practical tensor calculus on embedded submanifolds of arbitrary codimension

This paper introduces a fully extrinsic, parametrization-free, and component-free tensor calculus framework for embedded submanifolds of arbitrary codimension, featuring an algorithmic recursive notation that facilitates both theoretical analysis and practical applications in fluid dynamics, continuum mechanics, and evolving geometry.

The Gist

Imagine you are trying to describe the shape and movement of a piece of paper floating in a 3D room, or a soap bubble, or even a complex, high-dimensional shape that we can't easily visualize. In mathematics, these shapes are called submanifolds.

For a long time, mathematicians have had a very specific, rigid way of doing calculus (the math of change and motion) on these shapes. It's like trying to describe the paper's movement by first gluing a grid of graph paper onto it, writing down coordinates for every single point, and then doing complex calculations based on that grid. This works, but it's messy, hard to compute, and breaks down if the paper twists, turns, or changes shape over time.

The Paper's Big Idea: "The Tree Method"
Vladimir Yushutin proposes a new, cleaner way to do this math. Instead of gluing a grid onto the shape, he suggests looking at the shape from the "outside" (the room it's floating in) and using a special, recursive structure he calls a "row representation."

Think of a tensor (a complex mathematical object that holds information about direction and magnitude) not as a giant spreadsheet of numbers, but as a complete tree.

• The top of the tree is the main object.
• The branches split into smaller pieces (rows).
• The leaves are the actual numbers.

This "tree" structure allows the math to be algorithmic. It means you can write a computer program that handles these shapes by simply following the branches of the tree, no matter how complex the shape is or how many dimensions it has. You don't need to worry about the specific coordinates of the shape; you just follow the rules of the tree.

The Three Main Discoveries
The author uses this new "tree" method to solve three specific problems that were previously difficult or misunderstood:

1. The "Zero Net Push" Rule (Euler Flows):
   Imagine a fluid (like water) flowing perfectly smoothly over a curved surface, like a ball or a saddle. Old math suggested that if the surface had no symmetries (no perfect left-right or up-down balance), the fluid might push the surface in weird ways.

   • The Finding: Using this new method, the author proves that if the fluid is incompressible (it doesn't squish), the total push (momentum) on the entire surface is always zero. Even if the fluid is swirling wildly, the forces cancel each other out perfectly across the whole shape. It's like a group of people pushing a boat from all sides; even if they push randomly, if they are all on the boat, the boat doesn't move forward or backward as a whole.

2. The "Cut" Misunderstanding (Cauchy Stress):
   In engineering, we talk about "stress" inside materials. Usually, we assume that if you cut a piece of material, the force acts only along the cut surface. For flat sheets, this is easy. But for curved, 3D shapes (like a twisted rope or a curved shell), mathematicians have debated whether the force must always stay "flat" against the surface or if it can point "up" or "down."

   • The Finding: The paper argues that previous models were too restrictive. They assumed you could only cut the material in a specific, flat way. The author shows that if you allow for any cut (even a weird, angled one), the math proves that the force doesn't have to stay flat against the surface. It can point in any direction, and the laws of physics (Newton's laws) still hold true. This changes how we model stress in complex, curved materials.

3. Tracking Changing Shapes (Evolving Submanifolds):
   Imagine a soap bubble that is expanding, shrinking, and wobbling. How do you calculate the energy of a pattern drawn on that bubble as it changes?

   • The Finding: The author creates a formula to calculate exactly how the "energy" of a pattern changes as the shape itself moves and morphs. This is done using a "material derivative," which is like a camera that moves with the shape, tracking the changes from the inside while accounting for the shape's movement in the outside world. This provides a precise tool for modeling things like growing biological tissues or deforming membranes.

Why This Matters
The paper doesn't just offer a new theory; it offers a practical toolkit. By treating these complex shapes as "trees" of data, the math becomes:

• Coordinate-free: You don't need to pick a specific grid system.
• Recursive: You can solve big problems by breaking them down into smaller, identical steps (like following a tree branch down to a leaf).
• Universal: It works for shapes of any dimension and any "thickness" (codimension).

In short, the paper provides a new, more flexible, and computer-friendly language for describing how things move, push, and change on curved surfaces, removing the need for messy, old-school coordinate grids.

Technical Summary

Technical Summary: Practical Tensor Calculus on Embedded Submanifolds of Arbitrary Codimension

Problem Statement
Many scientific problems involve tensor fields defined on submanifolds embedded in $\mathbb{R}^n$. Traditional mathematical tools for these problems, such as differential operators and integration rules, are typically introduced intrinsically using local coordinate parametrizations or Ricci index notation. While powerful, this formalism can hinder computations, particularly when dealing with geometric evolution or submanifolds of arbitrary dimension and codimension. Existing extrinsic approaches often restrict themselves to scalar cases, rank-two tensors, or codimension-one scenarios, lacking a unified framework for general-rank tensors in arbitrary codimensions.

Methodology
The paper proposes a fully extrinsic, parametrization-free, and component-free variant of tensor calculus. The methodology relies on the ambient Cartesian derivative in $\mathbb{R}^n$ and the extrinsic level-set description of the submanifold's geometry.

The core of the framework is a recursive row representation of tensors. Instead of using matrix indices or coordinate components, a tensor of rank $q$ is represented as a list of $n$ row-components, each being a tensor of rank $q-1$. This structure is analogous to a complete $n$-ary tree, where leaf nodes store values and internal nodes guide evaluation via recursive dot products. This representation allows for:

• Algorithmic Recursivity: Operations such as projection, gradient, divergence, and curl are defined recursively on the row components.
• Extrinsic Derivatives: The submanifold gradient $\nabla_M$ is defined via the ambient gradient projected onto the tangent space, avoiding the need for intrinsic connections.
• Tangential Projection: A recursive algorithm projects general tensor fields onto the tangent subspace of the submanifold.
• Integration: A component-wise integration definition extends the standard Stokes' theorem to arbitrary rank tensors, yielding an extrinsic Stokes' formula that includes a term involving the mean curvature vector $\kappa$.

Key Contributions and Results
The paper demonstrates the utility of this framework through three distinct applications, deriving novel results for arbitrary dimension and codimension:

1. Euler Equations and Extrinsic Momentum:
   The authors derive a new global conservation principle for incompressible Euler flows on Riemannian manifolds. They show that the extrinsic momentum (the integral of the velocity covector over the manifold) vanishes for any tangential, incompressible flow, regardless of the manifold's symmetries. This leads to an ambient force balance equation where the Young-Laplace force, boundary reaction, and centripetal force sum to zero. This result generalizes the known property of zero net pressure force on fixed domains in $\mathbb{R}^2$ or $\mathbb{R}^3$.

2. Cauchy Stress on Submanifolds:
   The paper revisits the concept of Cauchy stress on embedded submanifolds with positive codimension. By relaxing the assumption that stress tensors must act only on tangential cuts, the authors derive equilibrium equations for generally oriented cuts. They prove that while conservation of angular momentum implies the tangentiality of the stress vector for tangential orientations, it does not imply that the stress tensor itself must be symmetric or purely tangential. The stress tensor can be non-symmetric and non-tangential without violating Newton's laws, provided the normal-at-tangential component of the stress vector cancels out.

3. Dirichlet Energy on Evolving Submanifolds:
   For evolving submanifolds, the authors introduce an extrinsic material derivative for tensor fields of general rank. They derive an expression for the temporal rate of change of the tensorial Dirichlet energy. This derivation generalizes previous results for scalar and vector fields to arbitrary tensor ranks, dimensions, and codimensions, utilizing commutators between the material derivative and the submanifold gradient.

Significance and Claims
The paper claims that the proposed framework offers a practical notation and set of tools immediately usable in mathematical modeling, the analysis of geometry-aware partial differential equations (PDEs), and numerical methods on embedded submanifolds.

The distinctive features highlighted are:

• Generality: It applies to submanifolds of any dimension and codimension and handles tensors of arbitrary rank.
• Computational Suitability: The recursive row representation mirrors the structure of complete trees, making the calculus amenable to algorithmic implementation and theoretical analysis.
• Extrinsic Clarity: By avoiding local coordinates and relying on ambient derivatives, the framework simplifies the handling of geometric evolution and boundary conditions.

The authors conclude that this approach facilitates the derivation of novel conservation laws and equilibrium conditions that were previously difficult to formulate in a general setting, and they suggest future directions in numerical analysis, variational calculus, and shape classification using the developed tools.