← all systems · see it in The System
Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data
In mathematics, a vector space is a set whose elements, often called vectors, can be added together and multiplied ("scaled") by numbers called scalars.
Properties published scienceour proposal
Sourced fields, and our own readings.
| aliases | Vector space (of translations) |
| born at | 1888 (Peano); candidates 1750, 1844, after 1920 |
| born precision | year; contested |
| boundary | Inside: the practice of defining, building and using vector spaces (the axioms, linear combinations, bases and coordinates, dimension, linear maps, norms and inner products, Hilbert and Banach spaces) and the structure it works on. Outside: the fields of scalars it is built over (real, complex, rational numbers), matrices as the computing tool (its sibling), the physics, data and minds it is used to describe, and the people and machines who carry it. |
| nested level | A part of linear algebra and of mathematics; built over a field of scalars; carried by mathematicians, teachers, physicists, statisticians and computers; the state-space language of quantum theory and machine learning runs on it. |
| scale | social |
Measures (12) published science
| vector axioms | 8 axioms |
| first axiomatic definition (Peano) | 1888 year |
| characteristic clock: axioms to use (Peano 1888 to development after 1920) | about 32 years |
| Euler's dependence and Cramer's paradox | 1750 year |
| Weyl's dimensional axiom (Raum-Zeit-Materie) | 1918 year |
| state-space dimension of 53 qubits | 2^53 (about 10^16) dimensions |
| Hilbert-space dimension cost of a real-number rewrite of a complex quantum system | 2 factor (doubling) |
| real-number bound beaten (superconducting test) | 43 standard deviations |
| real-quantum-theory constraints violated (photonic test) | over 4.5 standard deviations |
| syntactic analogy questions answered by vector offsets | almost 40 % |
| characteristic clock: word2vec training on 1.6 billion words | less than 1 day |
| Johnson-Lindenstrauss target dimension | O(log n / eps^2) for n points, distances within (1 +/- eps) dimensions |
Open contentions (4) 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: see C-413 in research/CONTENTIONS.md - C-1188 Which parts of the brain do advanced mathematics?
booked: The map records that experts use a network across the frontal, parietal and temporal lobes, not language areas, while noting this is inferred from related fields and not tested for each field by name.
the other reading: language-grounded accounts the abstract names as a rival theory - C-1219 Does quantum theory need complex numbers?
booked: The map keeps both: experiments showing complex numbers are needed in standard quantum theory, and a 2025 preprint on an equivalent theory using real numbers only.
the other reading: complex numbers indispensable (Renou 2021; Chen 2022; Li 2022); a real-number quantum theory consistent with the postulates, experimentally indistinguishable (Hoffreumon & Woods 2025, preprint) - C-2986 When was the vector space born?
booked: The map keeps the competing readings; none is ruled.
the other reading: Euler's 1750 question of dependence (Dorier: the first concepts built in linear equations); Grassmann's 1844 Ausdehnungslehre, which already held independence and dimension (Dorier: 'far ahead of his time'); Peano's 1888 axiomatic definition; the theory's real start after 1920, a frame from the 1930s (Dorier); the structure as a rung-0 role, with the idea at rung 11 (C-011 reading); related C-1187 for linear algebra
Cross-scale patterns seen here (13) our proposal
Shapes that recur when the same kind of problem is solved at different scales, and how each shows in Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data. Each links to every other system that shows it.
- nested ladder (structural)
Vector space ⊃ normed space ⊃ inner product space; Hilbert spaces are Banach spaces whose norm comes from an inner product; topological vector spaces hold both. - templating (structural)
One module, the eight axioms, reused unchanged on arrows, tuples, functions, polynomials, wavefunctions and word vectors; every n-dimensional space over F is the same as F^n. - trade one property for another (structural)
Johnson-Lindenstrauss: fewer dimensions bought with more distortion (k grows as 1/eps^2); PCA: dimension traded against information loss; a real-number quantum theory pays for dropping complex scalars with double the dimension. - borrowed-then-held (partial)
Dimension was borrowed from geometric space (degrees of freedom, three dimensions) and later held inside the theory (Weyl's dimensional axiom, then the invariance theorem). - causal ceiling (partial)
Every space has a basis only by the axiom of choice; no specific basis of R over Q is known: a limit on what can be written down set by the structure, not by effort. - namespace completeness (the map predicts its gaps) (partial)
Dimension classifies every space over a field up to isomorphism: one space per dimension, so the shelf is complete and predicts each member (F^n). - outside-to-inside (partial)
Vectors began as arrows for forces outside (Bellavitis 1833 equipollence; physical quantities) and moved inside an abstract axiom system (1888) that no longer needs space. - same role, different builds (partial)
Holding quantities that add and scale is done by arrows (geometry), n-tuples (arithmetic), functions (analysis), firing-rate vectors (neuroscience) and learned word vectors (language models). - seed (partial)
Eight axioms unpack into all the examples and, after 1920, into new discoveries and renewed proofs across mathematics and the sciences. - symmetry breaking (name the cut) (partial)
A space has no preferred directions; choosing a basis (the cut) fixes coordinates, and everything computed afterwards is in that basis, while dimension stays basis-free (Weyl). - The Payment to Entropy (partial)
Each projection to fewer dimensions (PCA, random projection) hands on the data with some distance or variance lost; PCA is built to minimise the loss, Johnson-Lindenstrauss bounds it at (1 +/- eps). - the whole sets the local reference (partial)
Each vector's coordinates are read against a basis of the whole space; the dimension, set by the whole, is the same whatever basis is chosen. - Umwelt: each system carves its own slice of one shared medium (partial)
One vector, three slices: geometric arrows, arithmetic tuples, structural elements; machines and brains take still other slices (embeddings, firing-rate vectors).
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 · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data gain
- Quantum mechanics via sys:matter · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data gain
- Real number (the practice of measuring the continuum, and the complete ordered field) via sys:sensory-environment · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data gain
- Complex numbers (C) via sys:sensory-environment · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data gain
- Large language model via sys:computer-memory · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data gain
- Human brain via sys:electrochemical-gradient · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data gain
- Shapes (geometry): the practice and the structure via sys:sensory-environment · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data gain
- Mathematics (the field: the practice and the structures it studies) via sys:sensory-environment · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data gain
- Computer via sys:computer-memory · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data 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-memory · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data gain
- 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 · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data gain
- Function (mathematics): the assignment of exactly one output to each input, and the practice of defining, drawing, composing and using it via sys:sensory-environment · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data gain
- Physics (the field) via sys:matter · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data gain
- Matrix (mathematics): the rectangular array of numbers and its algebra, determinants and eigenvalues, matrices as linear maps, their history, and their use in physics, graphics and machine learning; the practice and the structure both booked via sys:written-record · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data gain
- Set theory: the study of sets (collections taken as single objects), membership, Cantor's infinities and the axioms, and the practice of the people and machines who work with them via sys:sensory-environment · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data mixed
- Module (mathematics) via sys:sensory-environment · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data even
- Series (mathematics): sums of infinitely many terms, and the practice of summing, testing and using them via sys:sensory-environment · Vector space: sets whose elements can be added and scaled, as mathematics describes it, with bases and dimension, linear maps, inner products, and their use as the state spaces of physics and data gain
Partners and parts published science
Built from / builds (8)
- Projective space
- Affine space
- Field (algebra)
- Linear maps (maps that keep addition and scaling)
- Abelian group
- Basis (linear algebra)
- Vector bundle
- Linear algebra (the practice of solving, transforming and taking apart linear systems with vectors and matrices, and the structures it works with)
Sits inside / holds (4)
- Mathematics (the field: the practice and the structures it studies)
- Linear algebra (the practice of solving, transforming and taking apart linear systems with vectors and matrices, and the structures it works with)
- Euclidean vector
- Matrix (mathematics): the rectangular array of numbers and its algebra, determinants and eigenvalues, matrices as linear maps, their history, and their use in physics, graphics and machine learning; the practice and the structure both booked
expressed by (6)
- Generation/Synthesis
- Recombination/Composition
- Templating / Standardization
- Interoperability/Transduction (Bridging)
- Encapsulation & Modularity (Abstraction)
- Validation of Effectiveness
floor filled by (22)
- Projective space
- Affine space
- Human (Homo sapiens)
- Quantum mechanics
- Finite Hilbert space H_F (fermions)
- Matter
- Euclidean vector
- Large language model
- Human brain
- Computer
- Statistics: the practice of collecting and reasoning from data, the intuitive statistics it grows from, and the structure (samples, distributions, estimators) it works on
- Field (algebra)
- Linear maps (maps that keep addition and scaling)
- Matrix (mathematics): the rectangular array of numbers and its algebra, determinants and eigenvalues, matrices as linear maps, their history, and their use in physics, graphics and machine learning; the practice and the structure both booked
- Systems of linear equations
- Functional analysis
- Axiom of choice
- Abelian group
- Basis (linear algebra)
- Banach space
- Vector bundle
- Linear algebra (the practice of solving, transforming and taking apart linear systems with vectors and matrices, and the structures it works with)
represents (1)
Kinds (18)
- Minkowski spacetime
- Affine space
- Algebraic structures (a set with operations)
- Tangent space
- Inner product space
- Finite Hilbert space H_F (fermions)
- Four-vector
- Normed vector space (Banach space when complete)
- Cotangent space
- Vector bundle
- Euclidean vector
- Covariant vector (covector)
- Function space
- Real vector space
- Complex vector space
- Finite-dimensional vector space
- Banach space
- Module (mathematics)
wired (2)
Notes published science
- Every vector space has a basis only if you accept the axiom of choice: in standard set theory the two statements are equivalent, so the real numbers have a 'basis' over the fractions that is guaranteed to exist and that nobody has ever written down (src:wp:vector-space, src:fernandes-2026-bases-ac).
- Quantum theory could be rewritten with only real numbers by doubling the size of every state space, but in networks with independent sources the two versions predict differently; 2022 experiments sided with complex numbers by 43 and over 4.5 standard deviations, and a 2025 preprint argues the question is still open (src:renou-2021, src:chen-2022-real-qt, src:li-2022-real-qt, src:hoffreumon-woods-2025).
- Peano wrote down the modern axioms in 1888, in the last chapter of a book, and almost nobody used them for 30 years; Grassmann had most of the ideas in 1844, and an expert reviewer called his treatise 'a failure' (src:dorier-1995).
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 vector space as mathematicians describe it: quantities that add and scale, taken in as arrows, tuples or functions, bound by the vector axioms, built into spans and bases, related by linear maps and classified by dimension, given lengths and angles, and used as the state spaces of physics and data (reader's view)
- L1: Quantities in the world that add and scale: forces and velocities, solutions of homogeneous linear equations, wavefunctions, datasets and word contexts.
- L2: Taking it in: arrows (geometric mode), n-tuples (arithmetic mode), elements of a space (structural mode); for machines, records as real-valued vectors.
- L3: The rule: eight vector axioms over a field (Peano 1888).
- L4: Building inside: linear combinations, span, independence, subspaces, bases, coordinates; Hamel bases by the axiom of choice.
- L5: Relating spaces: linear maps, kernels, eigenvectors, isomorphism; classification by dimension.
- L6: Added structure: norms, inner products, Banach and Hilbert spaces, function spaces, tensor products, high-dimensional geometry.
- L7: Use: a unifying language across mathematics and the sciences; quantum state spaces; PCA, word vectors and attention; neural manifolds.
the boardroom (shared floors)
- L3: The rule: a set over a field with addition and scalar multiplication satisfying eight axioms (an abelian group under addition, scaling compatible with the field). Peano's 1888 definition.
- L4: Building inside a space: linear combinations, span, linear independence, subspaces, bases and coordinates; infinite bases by the axiom of choice.
- L5: Relating spaces: linear maps, kernels and images, eigenvectors, isomorphism; every space classified by its dimension over its field.
- L6: Added structure and the infinite: norms and inner products (lengths and angles), Banach and Hilbert spaces, function spaces, tensor products, high-dimensional geometry (Johnson-Lindenstrauss).
- L7: Use: a unifying language for mathematics and the sciences; the state spaces of quantum theory; data science (PCA, word vectors, attention); neural manifolds.
way in: Data: records, measurements and word contexts arranged as vectors
- L1: Large datasets and corpora (1.6 billion words).
- L2: Each word or record becomes a high-dimensional real vector.
way in: Neural population activity read as a point in state space
- L1: Recordings of thousands of neurons at once.
- L2: Each moment's firing rates form one vector; the trajectory runs on a low-dimensional subspace.
way in: Sight: arrows drawn in a plane or space (forces, velocities, displacements)
- L1: Arrows drawn from one point; their sum the parallelogram's diagonal.
- L2: The geometric mode of thought: a vector is a class of parallel, same-directed, equal-length arrows.
way in: Sight: vectors written as tuples or columns of numbers
- L1: Marks: ordered pairs and n-tuples of numbers.
- L2: The arithmetic mode: a vector is an n-tuple, added and scaled entry by entry.
way in: Systems of linear equations: the problems that first raised dependence
- L1: Equations whose unknowns may not all be fixed: Euler 1750, Cramer's paradox.
- L2: The solution set read as a space; rank as the invariant that sizes it.
way in: The physical world's linear quantities: forces, velocities, waves and quantum states that add and scale
- L1: Quantities with magnitude and direction (forces, velocities).
- L2: Measured into the solutions of linear equations: wavefunctions solving the Schroedinger equation, added and scaled.
way in: Thought: a space held in mind (an inner channel)
- L1: A vector space held and turned over in mind; where such a thought comes from is contested.
- L2: In experts, advanced mathematics engages a bilateral frontal, intraparietal and ventrolateral temporal network, sparing language areas.
way in: Words: definitions and axioms heard or read (lectures, textbooks)
- L1: The definition said or written: a set with addition and scaling satisfying the axioms.
- L2: The structural mode: vectors are elements of vector spaces; three modes of description named in teaching research.
way out: Prediction to experiment: Born-rule probabilities tested on prepared quantum systems
- L1: Qubits or photons measured; the real-number bound beaten.
- L2: A network experiment with independent sources run.
- L7: Probabilities computed from a state vector and projection operators.
way out: Reduction: data projected onto a few directions and written back out
- L1: Low-dimensional coordinates written out, minimising information loss.
- L4: Points projected onto a random or principal subspace.
- L6: The directions of most variance found (eigenvectors).
way out: Speech: explaining a vector in the three modes (teaching)
- L1: Sound to the class.
- L2: Saying it as arrows, as tuples or as elements of a space.
- L7: Choosing a mode of description for the listener.
way out: Writing: coordinates, matrices and proofs on paper or screen
- L1: The written definition: Peano's final chapter of 1888.
- L2: The hand writes tuples, matrices and the axioms.
- L4: Coordinates chosen on a basis; computations moved onto numbers.