Carleman linearization¶

A lifting method that represents a finite-dimensional nonlinear dynamical system as an infinite-dimensional linear system over monomials, then truncates it for approximation.

Version: v2 · 2026-09-19 · History
Domain-specific #: 3596
Domain group: Formal Sciences
Origin domain: Mathematics
Subdomains: Dynamical Systems, Linearization Methods → Mathematics

Core Idea¶

Carleman linearization embeds nonlinear polynomial dynamics into a linear hierarchy governing all state monomials. Differentiating each monomial substitutes the original dynamics and couples it linearly to other monomials; retaining finitely many degrees yields a computable linear approximation. The abstraction is therefore identified by a declared carrier, a transformation or constraint over that carrier, and an invariant that tells an analyst whether the named structure is genuinely present.

The load-bearing residual is not the broad topic of dynamical systems. It is infinite linear embedding of nonlinear dynamics through monomial observables. That residual remains recognizable when examples, notation, scale, or implementation change, but it disappears if the carrier is mistyped, the condition that the lifted hierarchy is formally equivalent before truncation and truncation order, closure rule, convergence domain and error assumptions are explicit fails, a neighboring object is substituted, or notation and topical resemblance replace the constitutive test.

How would you explain it like I'm…

No faithful explanation at this level. All three generators judged it N/A: a five-year-old picture collapses into 'pretend the curvy path is a straight line' (ordinary local linearization) or 'break it into easy pieces', whereas Carleman linearization is an exact infinite linear embedding over all monomials that is only approximate once truncated.

Tracking All the Powers

Some things change according to rules with squares in them, like 'this grows by x times x,' which makes them hard to solve. Carleman's trick is to keep track not just of x but also of x times x, x times x times x, and so on, as separate quantities. Then the rule for how each one changes can be written using only adding and multiplying by fixed numbers, which is the easy 'linear' kind of rule. The catch is that each quantity depends on the next bigger power, so the list never ends. If you keep only the first few, you get an approximate answer, and how good it is depends on how many you keep and on where you start.

Monomial Lifting to a Linear System

Carleman linearization turns a nonlinear system with polynomial rules into a linear one by changing what you track. Instead of just the state x, you follow all its monomials, x, x², x³, and so on (and products of components for several variables). Differentiating a monomial and plugging in the original equations gives its rate of change as a linear combination of other monomials, usually including higher-degree ones. The full infinite list of equations is formally equivalent to the original system, but you can only compute with finitely many, so you cut it off at some degree. That truncation gives a linear approximation whose accuracy depends on the cutoff order, how the dropped terms are handled, and where the solution stays.

Carleman linearization embeds a nonlinear dynamical system with polynomial right-hand side into an infinite-dimensional linear system whose variables are all the monomials of the state. Differentiating a monomial with respect to time and substituting the original dynamics produces a linear combination of other monomials, typically of equal or higher degree, so the lifted hierarchy is linear in the monomial observables. Before truncation this hierarchy is formally equivalent to the original dynamics. Truncating at a finite degree yields a finite linear system that approximates the nonlinear dynamics and can be analyzed or solved with linear methods. The validity of any such approximation depends on explicitly stated choices: the truncation order, the closure rule used to handle monomials beyond the cutoff, the domain on which the approximation converges, and the error assumptions. Carleman linearization should not be confused with local Jacobian linearization, which approximates dynamics near an equilibrium rather than lifting them exactly into a linear hierarchy.

Scope of Application¶

Carleman linearization belongs to dynamical systems and is useful where the analyst can specify a nonlinear ordinary differential or discrete system, state variables, monomials of increasing degree, infinite lifted state vector, block linear operator, truncation order, initial condition and approximation error, then evaluate the lifted hierarchy is formally equivalent before truncation and truncation order, closure rule, convergence domain and error assumptions are explicit. The scope is broad within that domain but bounded by the need for the lifted hierarchy is formally equivalent before truncation and truncation order, closure rule, convergence domain and error assumptions are explicit. The entry records a descriptive analytical identity; practical use requires the governing domain's evidence, standards, and safety obligations.

Clarity¶

The abstraction clarifies a crowded vocabulary by making the lifted hierarchy is formally equivalent before truncation and truncation order, closure rule, convergence domain and error assumptions are explicit the center of the account. A claim should name the carrier, the governing operation or relation, the applicable assumptions, and the recognition test. A bare label is insufficient because the name Carleman linearization can be used for a formal identity, an implementation, or a neighboring result unless carrier and convention are stated.

Manages Complexity¶

Without the abstraction, an analyst must reason directly over many local details: the carrier roles, admissibility assumptions, competing conventions, derived invariants, boundary cases, and proof or validation obligations specific to Carleman linearization. Carleman linearization compresses them into the roles in the structural signature. That compression permits comparison across instances without erasing the variables that determine validity. It also exposes which details may be varied safely and which are constitutive.

Abstract Reasoning¶

Identify the carrier. State what the elements, states, objects, or observations are: a nonlinear ordinary differential or discrete system, state variables, monomials of increasing degree, infinite lifted state vector, block linear operator, truncation order, initial condition and approximation error. Reject examples whose alleged carrier belongs to a different problem.
Lock the constitutive rule. Express the lifted hierarchy is formally equivalent before truncation and truncation order, closure rule, convergence domain and error assumptions are explicit independently of one notation or implementation.

Knowledge Transfer¶

Knowledge transfers strongly among subfields of dynamical systems because they reuse a nonlinear ordinary differential or discrete system, state variables, monomials of increasing degree, infinite lifted state vector, block linear operator, truncation order, initial condition and approximation error, Differentiating each monomial substitutes the original dynamics and couples it linearly to other monomials; retaining finitely many degrees yields a computable linear approximation., and type the carrier, state every parameter and convention in the definition, test that the lifted hierarchy is formally equivalent before truncation and truncation order, closure rule, convergence domain and error assumptions are explicit, compare the nearest accepted identity, and report counterexamples, uncertainty, and limiting cases.

Relationships to Other Abstractions¶

Current abstraction Carleman linearization Domain-specific

Parents (1) — outgoing typed relations

Carleman linearization is a kind of Representation Prime

The method represents nonlinear state evolution in a lifted linear coordinate system.

Hierarchy path (1) — routes to 1 parentless root

Carleman linearization → is a kind of Representation → presupposes Abstraction

Neighborhood in Abstraction Space¶

Carleman linearization sits in a crowded region of the domain-specific corpus (31^st percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Dynamical Systems & Chaos (17 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25