← all systems · see it in The System
Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them)
In theoretical physics and mathematical physics, analytical mechanics, or theoretical mechanics is a collection of closely related formulations of classical mechanics. Analytical mechanics uses scalar properties of motion representing the system as a whole—usually its kinetic energy and potential energy.
Properties published scienceour proposal
Sourced fields, and our own readings.
| aliases | Hamiltonian mechanics |
| born at | 1744 (Euler, Maupertuis) to 1760-1788 (Lagrange); Hamilton 1833-1834 |
| born precision | decades |
| boundary | Inside: the practice of writing a mechanical system as one scalar function (the Lagrangian L = T - V on configuration space, or the Hamiltonian H on phase space), counting its constraints and degrees of freedom, choosing generalized coordinates, deriving equations of motion from the stationary action (Euler-Lagrange) or the Legendre-transformed form (Hamilton's equations), and reading the geometry that follows (Poisson brackets, the symplectic form, Liouville's volume, Noether's conservation laws, integrability, KAM tori and chaos), together with the Hamiltonian systems it describes. Outside: Newtonian vector mechanics (same content, other formulation), the forces and bodies themselves, the mathematics it runs on (calculus of variations, differential equations, vector calculus, linear algebra), the theories it hands over to (quantum mechanics, field theory, relativity, statistical mechanics) and the people and machines that carry it. |
| nested level | A part of classical mechanics (its energy-based formulations, the classical-mechanics walk's L6 'structure' floor); runs on people's senses and thought and on computers; hands its structure on to quantum mechanics (Poisson bracket to commutator), to field theory and to statistics (Hamiltonian Monte Carlo). |
| scale | social |
Measures (19) published science
| Lyapunov time of the inner Solar System | about 5 Myr |
| Lyapunov time of Pluto's motion | 20 Myr |
| Symplectic integration step (Sussman and Wisdom 1992) | 7.2 days |
| Conventional integration step | 40 (Jupiter) / 0.5 (with Mercury) days |
| Speed-up of the averaged equations | about 1000 times |
| Probability of very high Mercury eccentricity over 5 Gyr | on the order of 1 % |
| Equations for the Sun and ten planets (1834) | 30 second-order or 60 first-order, reduced to one function equations |
| Order and number of equations | 2n first-order (Hamilton) vs n second-order (Lagrange) equations |
| Generalized coordinates for N particles with C constraints | 3N - C coordinates |
| Leapfrog stability limit | step approaching 2 (for H = q^2/2 + p^2/2) time units |
| Kinetic focus of the quartic oscillator | 0.646 of the half-period fraction |
| HNN energy error, two-body problem | several orders of magnitude below baseline ratio |
| HNN size and training | 3 layers, 200 hidden units, 2000 gradient steps |
| Leapfrog energy error | order of step squared for H when (q,p) error is order step squared |
| Energy lost to viscous drag (Galley's nonconservative Lagrangian) | dh/dt = -alpha |v|^(n+1); n = 1 Stokes drag, n = 2 nonlinear drag energy per unit time |
| Energy error growth, unbiased (Brouwer's law) | grows as the square root of the number of timesteps (~ t^1/2) relative energy error vs time |
| Port-Hamiltonian power balance | dH/dt = -grad H^T R grad H + y^T u <= y^T u (R positive semi-definite) power |
| Energy error growth, biased | grows linearly with time (~ t) relative energy error vs time |
| Symplectic integrator's own energy error | bounded, independent of time; order epsilon dt^2 without correctors, dt^K + epsilon^2 dt^2 with correctors of order K (up to eleven) |
Open contentions (3) open contention
Where sources or readings disagree. Both stay on the map.
- C-413 Where do thoughts come from?
booked: No theory is proven, so the map lists the main ones side by side and doesn't pick one. It shows where thinking is handled in the brain, not where thoughts start.
the other reading: spontaneous-thought and default-mode accounts; predictive-processing accounts; Jerry's view, labelled as his: thoughts arrive as patterns from outside one's own slice of consciousness - C-1214 Is the 'principle of least action' really about the least?
booked: The map follows the view that the action is stationary, a minimum or a saddle but never a true maximum, and keeps 'least action' as a name.
the other reading: least action, as Maupertuis named it; 'Even Lagrange wrote that the value of the action can be maximum'; the classical principle as a consequence of quantum interference (stationary-action page) against a formalism that 'does not introduce new physics' (analytical mechanics page); particles 'picking out the path' read as a way of speaking, not as purpose - C-1494 Can friction be written into the principle of least action?
booked: The map keeps both readings: friction as the exception, and newer forms that build it in.
the other reading: friction as the named exception to action principles (stationary-action page: 'The only major exceptions are cases involving friction'); Rayleigh's dissipation function 'not sufficiently comprehensive' for history-dependent, nonlocal or nonlinear dissipation; a doubled-variable action handles generic nonconservative systems (Galley 2013); dissipation as a structure of the formalism itself (R(x) in port-Hamiltonian systems)
Cross-scale patterns seen here (21) our proposal
Shapes that recur when the same kind of problem is solved at different scales, and how each shows in Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them). Each links to every other system that shows it.
- same role, different builds (structural)
One job (predicting a mechanical system's motion) done by different builds with the same content: Newton's force vectors per particle, Lagrange's one scalar on configuration space, Hamilton's 2n first-order flow on phase space, Hamilton-Jacobi's one function; 'Both theories ... describe the same phy - symmetry breaking (name the cut) (structural)
Hamilton's principle with fixed end points couples paths only through the time-symmetric half of the Green function, so it cannot give dissipation; Galley names the cut ('"Breaking" the symmetry by introducing two sets of variables') and the broken direction (initial data forward, equality condition - templating (structural)
The action principle as a reused module across physics ('from classical mechanics through quantum mechanics, particle physics, and general relativity'), carried into ecology (Volterra 1937), statistics (HMC) and machine learning (HNN, LNN), each time unchanged in form: write one scalar, vary it. - The Payment to Entropy (structural)
The lossless limit, with the toll bolted on: a Hamiltonian flow hands energy between kinetic and potential with no toll (phase volume kept, reversible in time); the toll enters only as add-ons: a dissipation function ('energy is not conserved when a dissipation function has effect'), friction as the - trade one property for another (structural)
Order for number: Hamilton trades n second-order equations for 2n first-order ones (Lagrange's transformation turns 30 into 60). LNNs buy freedom from canonical coordinates; HNNs pay for exact conservation with needing them. Leapfrog buys exact volume and pays with H kept only approximately. - causal ceiling (partial)
In the chaotic Solar System any finite knowledge of the initial state is used up: 'a small change in initial conditions will significantly change the trajectories after 100 Myr', with a 5 Myr Lyapunov time; the limit comes from the flow's divergence, not from one instrument. Partial: the sources fra - conservation law plus closure (partial)
The action principle leaves one slot open: each application 'requires a specific Lagrangian describing the physics'; once L is supplied, Noether's theorem turns each symmetry into a conservation law. - namespace completeness (the map predicts its gaps) (partial)
Hamilton's optico-mechanical analogy arranged four boxes: geometrical optics and wave optics; classical mechanics and an empty slot. The arrangement predicted the slot's properties ('classical mechanics might be viewed as the geometrical optics ... of an "undulatory motion"'), a wavelength (de Brogl - same shape, opposite sign (partial)
Hamilton's two equations: q-dot = +dH/dp and p-dot = -dH/dq, one mechanism (the gradient of H turned by the symplectic form) with the sign reversed between the conjugate pair; 'coordinates and momenta are independent variables with nearly symmetric roles'. - same toolkit, independent builds (partial)
The Kepler-splitting symplectic integrator was built twice in 1991, by Wisdom and Holman and 'independently' by Kinoshita et al., from the same parts (Kepler motion plus small tugs, operator splitting). Dark twin to watch: both groups read the same celestial-mechanics literature (shared ancestry of - smooth to sharp in finite time (partial)
Two sightings. The action, a minimum for short paths, becomes a saddle at a finite time, the kinetic focus (0.646 of the half-period for the quartic oscillator). And an HMC trajectory meeting high curvature 'almost immediately propels ... towards infinite energies'; numerical, not an exact singulari - The carry (partial)
Two faces. The wait: near KAM tori chaotic diffusion is so slow it 'may be negligible for a very long time, eventually as long as the age of the universe' (Nekhoroshev). The lookahead: Hamilton turned 30 second-order (or 60 first-order) equations for the Sun and ten planets into the search for one f - Umwelt: each system carves its own slice of one shared medium (partial)
The formalism carves its own slice of the shared mechanical world: the forces-constraints channel takes in only generalized forces along the allowed directions, while 'constraint forces do no virtual work and drop out'; each of the eight channels (notation, lecture, writing, thought, data, machine, - who opens the door (partial)
In a port-Hamiltonian model the receiving system's doors are built ahead of time: power enters and leaves only through declared port pairs (u, y), and the stored energy can rise no faster than the supply (dH/dt <= y^T u). A port, by the catalog's terms; booked partial because the doors belong to the - borrowed-then-held (surface)
Hamilton's method was first done in optics (systems of rays, 1828-1831), then held by mechanics as its own (1834), then lent back to optics (electron focusing 1925) and on to wave mechanics. - lightning (surface)
A divergence in HMC: a trajectory flies to infinite energy almost at once, and the sampler is changed after it (the proposal rejected, the divergence flagged as a diagnostic). No stored difference released; surface. - nested ladder (surface)
Symplectic implies volume-preserving: 'By Liouville's theorem, each symplectomorphism preserves the volume form on the phase space'; leapfrog sits on the upper rung (Neal: volume kept exactly), Euler on neither. Two rungs only, surface. - outside-to-inside (surface)
Integrating out the bath: the Q oscillators start outside the open part, and come inside its own action as a nonlocal, history-dependent nonconservative potential K_eff (Galley). The outside phase is shown first; but this is a move in the description, not a gradient the system takes in, so surface. - partner-becomes-part (surface)
Port-Hamiltonian composition: the power-conserving interconnection of port-Hamiltonian systems is again a port-Hamiltonian system, so two partners joined at their ports become one system of the same kind. No inheritance in play; surface. - runaway from an unstable balance (surface)
Betancourt's satellite: a stable orbit needs momentum matched to the pull; too little and it crashes into the mode, too much and it escapes. Runaways seen (divergences; Mercury's eccentricity rising past 0.9 in 13 steps; exponential separation of nearby trajectories), but the sources read do not sho - seed (surface)
One scalar (L or H) unpacks every equation of motion of a system, and the form unpacked across 14 non-classical domain groups. Read as templating more than a Seed; the sources show reuse, not new combinatorics opened by its arrival. Surface (Seed check: maybe, unchanged).
Valence: what it trades, and with whom (17) published science
Each exchange with a neighbour, read from both sides: who gains, who pays, and what carries it.
- Human (Homo sapiens) via sys:sensory-environment · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) gain
- Vector calculus (the calculus of fields: gradient, divergence, curl; the practice and the structures it works on) via sys:sensory-environment · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) gain
- Quantum mechanics via sys:mental-representation · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) gain
- Cotangent bundle via sys:phase-space · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) gain
- Harmonic oscillator via sys:potential-energy · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) gain
- Statistical mechanics via sys:phase-space · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) gain
- Solar System via sys:gravitational-field · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) gain
- Computer via sys:computer · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) gain
- Energy via sys:stress-energy · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) gain
- Statistics: the practice of collecting and reasoning from data, the intuitive statistics it grows from, and the structure (samples, distributions, estimators) it works on via sys:computer · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) even
- Linear algebra (the practice of solving, transforming and taking apart linear systems with vectors and matrices, and the structures it works with) via sys:sensory-environment · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) gain
- Artificial neural networks (trained sequence-to-sequence models) via sys:computer · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) even
- Calculus (the practice of handling change and accumulation by limits, and the structures it works on) via sys:sensory-environment · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) gain
- Differential equations (the practice of writing and solving laws of change, and the equations it works on) via sys:sensory-environment · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) gain
- Classical mechanics (Newtonian mechanics: bodies, forces and motion, and the practice of describing them) via sys:sensory-environment · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) gain
- Optics (light's behaviour: reflection, refraction, interference, lenses and instruments) via sys:mental-representation · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) gain
- Chaos and nonlinear dynamics (the practice, and the chaotic behaviour it studies) via sys:phase-space · Hamiltonian and Lagrangian mechanics (analytical mechanics: classical mechanics built from energy functions and the stationary action, and the practice of using them) gain
Partners and parts published science
Built from / builds (12)
- Symplectic manifold
- Configuration space of a physical system
- Phase space
- Hamilton-Jacobi equation
- Noether's theorem
- D'Alembert's principle of virtual work
- The stationary-action principle
- Euler-Lagrange equations
- Legendre transformation
- Hamilton's equations
- Poisson bracket
- Classical mechanics (Newtonian mechanics: bodies, forces and motion, and the practice of describing them)
Sits inside / holds (3)
- Human (Homo sapiens)
- Physics (the field)
- Classical mechanics (Newtonian mechanics: bodies, forces and motion, and the practice of describing them)
expressed by (12)
- Generation/Synthesis
- Recombination/Composition
- Scaffolding/Enablement (Meta-Tooling)
- Templating / Standardization
- Interoperability/Transduction (Bridging)
- Encapsulation & Modularity (Abstraction)
- Search/Foraging
- Mapping/Modeling
- Uncertainty Reduction/Learning
- Hypothesis Portfolio / Multiple Hypothesis Tracking
- Offline Simulation & Replay-Guided Planning
- Quality Control & Error Correction
floor filled by (24)
- Cotangent bundle
- Human (Homo sapiens)
- Quantum mechanics
- Symplectic manifold
- Configuration space of a physical system
- Harmonic oscillator
- Phase space
- Hamilton-Jacobi equation
- Solar System
- Noether's theorem
- Computer
- Energy
- Calculus (the practice of handling change and accumulation by limits, and the structures it works on)
- Differential equations (the practice of writing and solving laws of change, and the equations it works on)
- KAM theorem
- D'Alembert's principle of virtual work
- Euler-Lagrange equations
- Legendre transformation
- Hamilton's equations
- Poisson bracket
- Holonomic constraints
- Hamiltonian Monte Carlo
- Hamiltonian neural networks
- Classical mechanics (Newtonian mechanics: bodies, forces and motion, and the practice of describing them)
represents (2)
- Habitat and ecosystem
- the human idea of the Hamiltonian formalism, held by science (REPRESENTS the rung-1 structure)
Kinds (7)
- Classical mechanics (Newtonian mechanics: bodies, forces and motion, and the practice of describing them)
- Hamilton-Jacobi equation
- Lagrangian mechanics
- Routhian mechanics
- Relativistic Lagrangian mechanics
- Classical field theory (Lagrangian densities)
- Nambu mechanics
wired (1)
Notes published science
- A computer pays a tiny toll on every arithmetic step, and in the best planetary integrators those tolls add up like a drunkard's walk: the energy error grows only as the square root of the number of steps (Brouwer's law). Let one step be slightly biased and the error grows in a straight line instead (src:rein-tamayo-2015-whfast).
- Hamilton's own principle can't do friction honestly, because it fixes both ends of the path in time. Doubling every variable, one copy running forward and one backward, lets the same principle give the exact energy lost to drag (src:galley-2013-nonconservative).
- The Hamiltonian's 'energy' has wandered far from joules: chemists use Gibbs free energy, and engineers have put the same form on electricity markets, macroeconomics, predator-prey ecology and the training of neural networks, where H has no energy meaning at all (src:munoz-arias-2026-port-hamiltonian).
Layers
The ways it takes in its world, where they meet, and how it acts back, each from L1 (where it meets the world) up.
a Hamiltonian system as nature runs it (the state in phase space, the generator H, the flow, what the flow keeps, regimes)
- L1: The state: a point (q, p) in 2n-dimensional phase space; configuration space is its shadow
- L2: The generator: any smooth function H on a symplectic manifold (usually T + V); fictitious momenta can be added to any continuous state space
- L3: The flow: Hamilton's equations, the Hamiltonian vector field and its flow
- L4: What the flow keeps: H when time-independent, phase volume (Liouville), the symplectic form, reversibility
- L5: Regimes: integrable motion on tori; near-integrable (KAM: most tori survive, resonant ones break); chaos and very slow diffusion
analytical mechanics as physicists describe it (reader's view: mechanical systems, constraints and degrees of freedom, generalized coordinates and the scalar functions, equations of motion, phase space, symmetry and invariants, use and hand-over)
- L1: Mechanical systems: particles with finitely many degrees of freedom (a bead on a wire, a pendulum, the Sun and planets), and by extension fields and fluids
- L2: Constraints and degrees of freedom: holonomic constraints counted, d'Alembert's principle removes the constraint forces (virtual work zero)
- L3: Generalized coordinates and the scalar functions: T, V, the Lagrangian L = T - V; momenta p = dL/dq-dot and the Hamiltonian H by Legendre transform
- L4: Equations of motion from a principle: the stationary action gives the Euler-Lagrange equations (n second-order) or Hamilton's equations (2n first-order)
- L5: Phase space and its geometry: Poisson brackets, the symplectic form, canonical transformations, Liouville's conserved volume
- L6: Symmetry and invariants: Noether's theorem, cyclic coordinates and reduction, Hamilton-Jacobi; integrable, KAM and chaotic regimes
- L7: Use and hand-over: prediction and long integrations, quantization (Poisson bracket to commutator), field theory and relativity, Hamiltonian Monte Carlo, learned Hamiltonians
the boardroom (shared floors)
- L3: Generalized coordinates and the scalar functions T, V, L, p and H
- L4: Equations of motion from the stationary action (Euler-Lagrange) or the Legendre-transformed form (Hamilton's equations)
- L5: Phase-space geometry: Poisson bracket, symplectic form, Liouville volume, canonical transformations
- L6: Symmetry and invariants: Noether, cyclic coordinates, Hamilton-Jacobi, integrability, KAM and chaos
- L7: Use and hand-over: prediction, long integrations, quantization, field theory, sampling, learned models
way in: Data: measured positions and momenta
- L1: Position and momentum readings from a real pendulum (Schmidt and Lipson's data)
- L2: Noisy readings that do not strictly conserve energy (the real pendulum has a little friction)
way in: Forces and constraints on the bodies (nature's side)
- L1: Bodies held by constraints and pushed by applied forces (a bead on a wire held by a reaction force; gravity)
- L2: Constraint forces do no virtual work and drop out
- L3: Generalized forces along the generalized coordinates; non-conservative ones allowed if they satisfy d'Alembert's principle
way in: Hands: writing derivations and correcting them
- L1: Written explanations and found errors in one's own solutions
way in: Hearing: lectures and exercise classes
- L1: Spoken explanation in exercise classes ('fits in a 90-min exercise class')
way in: Learned model: a network fitted to observations
- L1: Observations of mass-spring, pendulum and two-body systems, including pixels of a pendulum
- L2: A network outputs one scalar 'energy-like' value whose gradients give the motion
way in: Machine: integrated trajectories
- L1: Trajectories of the planets integrated with a symplectic integrator
- L2: The integrator's step: leapfrog shear steps keep phase volume exactly
way in: Sight: reading equations and worked examples
- L1: Worked examples: step-by-step solutions of Lagrangian problems on the page
way in: Thought: ideas carried between problems
- L1: Contested origin (C-413). One recorded route: Hamilton carried the idea of his optics of rays into dynamics
way out: Computing: integrations and samples
- L1: Trajectories and samples written out (distant proposals for Metropolis)
- L4: Leapfrog steps run
- L7: A Hamiltonian and an integrator chosen
way out: Dissipation: energy leaving the mechanical system
- L1: Energy not conserved: it leaves the mechanical description
- L4: A dissipation function (Rayleigh) added to the equations
way out: Hand-over to quantum mechanics
- L1: The quantum formulations it suggests: the path integral and the Schrodinger equation
- L7: Poisson bracket replaced by the commutator
way out: Motion: the state carried along its phase path
- L1: The state moves along its phase path
- L3: The Hamiltonian flow
way out: Teaching: worked examples in class
- L1: Explanations given in exercise classes
- L7: The four-step worked-example method chosen
way out: Writing: derivations and papers
- L1: Published derivations, from Hamilton's 1834 paper in the Philosophical Transactions to exercise sheets
- L7: A solution reduced as far as possible ('Reduce the equations if possible')