← all systems · see it in The System
Number and counting (the number sense, the practice of counting, and the structure it works on)
A number is a mathematical object used to count, measure, order and label. ... Counting is the process of determining the number of elements of a finite set of objects; that is, determining the size of a set. ... In psychology, number sense is the term used for the hypothesis that some animals, particularly humans, have a biologically determined ability that allows them to represent and manipulate large numerical quantities.
Properties published scienceour proposal
Sourced fields, and our own readings.
| born at | not dated in the sources read (rung 7; counting attested at least 50,000 years ago; notched bones 44-42 ka) |
| born precision | undated for the number sense |
| boundary | Inside: the representation of how many (numerosity) in animal brains (object files for small sets, the approximate number system, number neurons), the practice of counting (count lists, the cardinal principle, finger counting, tallies, numerals), and the structure that practice names (natural numbers with successor and their extensions). Outside: the sets and sequences counted, the languages and scripts that carry the words and marks, the people and animals who count, and the money, measurement and science that run on number. |
| nested level | Held in nervous systems from insects to humans (rung 7 up), carried by human languages and scripts at rung 11, and naming a mathematical structure booked at rung 0 (roles). |
| scale | organism |
Measures (29) published science
| characteristic clock: 1-knower to Full-Counter (US children) | 1-2 years |
| characteristic clock: 1-knower to Full-Counter (Tsimane' children) | 4-8 years |
| characteristic clock: lexical stability of low number words | 10,000 to over 100,000 years |
| characteristic clock: zebrafish number-neuron tuning window | 3-7 days post-fertilisation |
| start of formal verbal counting | about 2.5 years of age |
| object tracking system capacity | ≤3-4 items |
| subitizing range | up to about 4 items |
| estimation regime (reaction times rise) | up to about 7 items |
| crow NCL neurons' range | up to 30 items |
| newborn cross-modal matching range | 4-18 objects |
| zebrafish neurons tuned >1 at 7 dpf | 14 % of number-selective neurons |
| heritability of non-symbolic numerosity discrimination (adolescents) | 0.32 |
| developmental dyscalculia prevalence | 3-6 % of school-age children |
| number-word stability relative to Swadesh average | 3.5-20 times longer |
| oldest notched bone with notches added at different times (Border Cave) | 44-42 ka |
| Les Pradelles incisions (possible numerical record) | 72-60 ka |
| counting attested archaeologically | at least 50,000 years |
| Awetý number words | 20 words |
| adaptation rates used (slow / fast) | 2 / 8 flashes per second |
| chick discrimination at 3 days | 5 vs 10; 10 vs 20 items |
| mealworm males track females | up to 4 females |
| numerosity-selective units in an object-recognition network (final layer) | 9.6 (3601 of 37,632) % of units |
| age at which number-sense precision peaks | about 30 years |
| lifespan sample for number-sense precision | more than 10,000 participants, 11-85 years people |
| written numerical notations known | close to 200 notations |
| body-based numeration systems recorded (BodyBase) | over 800 systems |
| languages compared for numeral-system efficiency | 30 languages |
| carry settling time, 16-bit addition (ripple vs lookahead) | 31 vs 8 gate delays |
| base of the first positional system (Babylonian) | 60 |
Open contentions (9) 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 research/CONTENTIONS.md C-413 - C-463 Are the Ishango and Lebombo bones the oldest evidence of arithmetic?
booked: The map lists them as possible arithmetic artifacts and notes that this reading is disputed.
the other reading: oldest arithmetic artifacts (some historians); tally or other marks - C-465 Do we need number words and schooling to grasp exact numbers?
booked: The map follows the 2022 finding that counting words and schooling matter, and keeps the view that exact number develops with little teaching alongside.
the other reading: exact number develops with little instruction, domain-general (Butterworth et al. 2008 via da Silva Sinha 2026); Pirahã limits read from test structure and context (Frank et al. 2008 via da Silva Sinha 2026) - C-485 What are numbers, really?
booked: The map doesn't rule on what numbers are and lists the philosophical views.
the other reading: cardinal numbers as sizes of sets (Giaquinto 2017 argues for it); the other rival views Giaquinto lists as facing difficulties (not named in the abstract) - C-486 Did animals inherit their sense of number from a common ancestor?
booked: The map leaves it open, reading number-sensing neurons as separate inventions for now and keeping shared ancestry alongside.
the other reading: deep homology: circuits inherited from a common vertebrate ancestor (Zanon et al. 2026 lists both; Agrillo & Bisazza 2017 ask) - C-487 Is the sense of number inborn?
booked: The map leaves it open, keeping inborn circuits and learning from general abilities side by side.
the other reading: innate circuitry (number neurons in visually naive chicks and 3-dpf zebrafish); emergent from architectural constraints plus domain-general learning (Zorzi & Testolin 2017); arises from the statistics of random feedforward wiring before any learning (Kim et al. 2021); or from training on object recognition alone (Nasr et al. 2019) - C-488 Do we sense number directly, or work it out from size and density?
booked: The map follows the view of a dedicated number sense, and keeps the magnitude-based views alongside.
the other reading: number read from continuous magnitudes or density; a broader magnitude system (Zanon et al. 2026 open question) - C-489 Do we use one system for small numbers and another for large ones?
booked: The map shows two systems, one for about 3-4 items and one for all amounts, and keeps the single-system view alongside.
the other reading: a single scalar-variability system; the small-large divide from task and attention (Zanon et al. 2026) - C-1877 What counts as a number base?
booked: The map keeps the definitions open and rules none.
the other reading: linguists' looser readings: values used in forming numerals (Pelland et al. 2025 report the definitions vary substantially); European finger counting read as base 10 with sub-base 5 (earlier accounts, as Dudojć et al. 2025 report them)
Cross-scale patterns seen here (14) our proposal
Shapes that recur when the same kind of problem is solved at different scales, and how each shows in Number and counting (the number sense, the practice of counting, and the structure it works on). Each links to every other system that shows it.
- same role, different builds (structural)
Number neurons tuned to a preferred numerosity on a log scale in the macaque fronto-parietal cortex and the crow's nidopallium caudolaterale, endbrains that evolved separately; zebrafish pallium and chick NCL show the same code without a layered cortex. - same shape, opposite sign (structural)
The borrow is the carry run backwards: one unit of the next place moved between adjacent columns, leftward as a carry in addition, taken from the left as a borrow in subtraction ('The opposite is a borrow'). - seed (structural)
Ten digits in a place-value system 'can display any non-negative integer'; a small fixed packet unpacks an unbounded set of numerals. - The carry (structural)
Overflow passed on: in place-value notation 'a carry is a digit that is transferred from one column of digits to another column of more significant digits' (each place holds up to base minus one; Babylonian base 60 was the first positional system). The wait: 'no digit position can have an absolutely - trade one property for another (structural)
Numeral systems trade informativeness against simplicity ('a tradeoff between two competing forces: informativeness and simplicity'), and the world's systems sort along the trade: approximate, exact over a small range, and fully recursive, each near-optimal for its point on the curve (30 languages). - Umwelt: each system carves its own slice of one shared medium (structural)
Each animal carves number from its own slice of the shared world: bees count landmarks passed, male frogs count a neighbour's notes and the intervals between them, male mealworms count females by odour, chicks and people count seen and heard sets, and people add words and numerals. One shared Umgebu - borrowed-then-held (partial)
The hand and body borrowed as the counting device (handful, one hand, two hands, Oksapmin body tally), later held in words and marks that keep the base: the hand 'used as external representations of number' shapes 'numerical structure and organization, as well as aspects of numerical language'. - namespace completeness (the map predicts its gaps) (partial)
The naturals' own operations point to the missing numbers: they are 'not closed under subtraction', which the integers fill; 'If every solution of a polynomial equation is considered to be a number, the result is the complex numbers'. The slots were seen before they were accepted (negatives resisted - outside-to-inside (partial)
The count list arrives from outside first: children 'begin by learning to recite the list of count words ... without knowledge of their meaning', and only later do number words recruit numerosity-related parietal cortex (Kersey et al. 2026); the hand used as an external representation of number shap - same toolkit, independent builds (partial)
Tuned, log-scaled number detectors appear in monkey cortex, crow NCL and, from local filtering and pooling units alone, in deep networks trained on objects (9.6% of the final layer) or never trained; 'virtually identical' to real number neurons. Dark twin noted: the networks were biologically inspir - templating (partial)
The Hindu-Arabic system displaced Roman numerals in Europe and 'remains the most common system for representing numbers in the world today': one module, swapped in across languages, still works. - The Payment to Entropy (partial)
Handing quantity on loses some of it. Sense to code: number units 'exhibited approximate tuning that decreased in precision with increasing numbers' (in networks as in real neurons), so the larger the set, the more of it arrives blurred. Speaker to listener: 'Communication is not perfectly informati - who opens the door (partial)
The receiver is built before the traffic: tuned number units exist in untrained, randomly wired networks 'in the complete absence of learning' and 'arise before visual training', as number neurons do in visually naive chicks and 3-day zebrafish; children learn the count list as words 'without knowle - the whole sets the local reference (surface)
Adaptation: what counts as 'many' is read against recent input; after fast tapping or flashes the same set reads as fewer.
Valence: what it trades, and with whom (28) published science
Each exchange with a neighbour, read from both sides: who gains, who pays, and what carries it.
- Real number (the practice of measuring the continuum, and the complete ordered field) via sys:sensory-environment · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Vertebrate (Vertebrata) via sys:electrochemical-gradient · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Human (Homo sapiens) via sys:sensory-environment · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Language (human language) via sys:acoustic-wave · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Insect (class Insecta) via sys:electromagnetic-field · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Writing (the practice of marking language on a surface, and the script plus rules it works with) via sys:written-record · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Money of account (the unit of account: number applied to exchange) via sys:written-record · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Human brain via sys:electrochemical-gradient · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- The world through the senses (sensory environment) via sys:sensory-environment · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Writing system (a script plus the rules binding it to a language) via sys:written-record · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Money via sys:written-record · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Deep neural network (generative model) via sys:electromagnetic-field · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Arithmetic (the practice of calculating, and the operations and laws it works with) via sys:sensory-environment · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Computer via sys:sensory-environment · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Shapes (geometry): the practice and the structure via sys:sensory-environment · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Mathematics (the field: the practice and the structures it studies) via sys:sensory-environment · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Measurement (the practice of comparing a quantity with a unit, and the units and standards it uses) via sys:mental-representation · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Algorithm (the practice of step-by-step procedures and the structure they compute) via sys:written-record · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Fraction (parts of a whole): the ratio sense, the practice of fractions, and the rational numbers they name via sys:sensory-environment · Number and counting (the number sense, the practice of counting, and the structure it works on) mixed
- Ratio and proportion: comparing two quantities by how many times one contains the other, the practice of keeping proportions, and the structure it works on via sys:sensory-environment · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Turing machine: the abstract machine that reads and writes one symbol at a time on an unbounded tape by a finite table of rules, the model of computation of logic and computer science via sys:written-record · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Accounting (the field and practice) via sys:written-record · Number and counting (the number sense, the practice of counting, and the structure it works on) 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 · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Complex analysis (the practice of working with functions of a complex variable, and the holomorphic structure it works on) via sys:sensory-environment · Number and counting (the number sense, the practice of counting, and the structure it works on) 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 · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Numeracy via sys:sensory-environment · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Infinity: the boundless or endless as mathematics and philosophy describe it (potential and actual infinity, Cantor's sizes of infinity, infinity in calculus, geometry, physics and computing) and the practice of the people who argue over it via sys:sensory-environment · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
- Prime numbers (the practice of number theory that finds, proves and uses them, and the structure of the primes it works on) via sys:sensory-environment · Number and counting (the number sense, the practice of counting, and the structure it works on) gain
Partners and parts published science
Built from / builds (11)
- Numeral system
- Approximate number system
- Object tracking system
- Number neuron
- Intraparietal sulcus
- Count list (number words in order)
- Natural number (the structure)
- Measurement (the practice of comparing a quantity with a unit, and the units and standards it uses)
- Ratio and proportion: comparing two quantities by how many times one contains the other, the practice of keeping proportions, and the structure it works on
- Whole-number strategy against fraction magnitude for the answer in a fraction comparison
- Absolute count against proportion for the choice of a population to draw from
Sits inside / holds (6)
- Vertebrate brain
- Language (human language)
- Insect (class Insecta)
- Human brain
- Fraction (parts of a whole): the ratio sense, the practice of fractions, and the rational numbers they name
- Prime numbers (the practice of number theory that finds, proves and uses them, and the structure of the primes it works on)
expressed by (14)
- Generation/Synthesis
- Recombination/Composition
- Templating / Standardization
- Interoperability/Transduction (Bridging)
- Encapsulation & Modularity (Abstraction)
- Sensing/Scanning
- Search/Foraging
- Navigation/Pathfinding
- Uncertainty Reduction/Learning
- Multi-Scale Data Fusion / Triangulation
- Trail-Marking / Externalized Memory (Stigmergy)
- Surprise / Change-Point Detection
- Social Learning / Public Information Use
- Quick Harm Avoidance
floor filled by (24)
- Electromagnetic field
- Acoustic wave (sound: pressure waves in matter)
- Human (Homo sapiens)
- Language (human language)
- Prefrontal cortex
- Writing (the practice of marking language on a surface, and the script plus rules it works with)
- Money of account (the unit of account: number applied to exchange)
- The world through the senses (sensory environment)
- Object tracking system
- Intraparietal sulcus
- Deep neural network (generative model)
- Nidopallium caudolaterale (NCL)
- Mathematics (the field: the practice and the structures it studies)
- Visual cortex
- Hand
- Carry (and borrow)
- Numeral base
- Shapes (geometry): the practice and the structure
- Measurement (the practice of comparing a quantity with a unit, and the units and standards it uses)
- Algorithm (the practice of step-by-step procedures and the structure they compute)
- Ratio and proportion: comparing two quantities by how many times one contains the other, the practice of keeping proportions, and the structure it works on
- Function (mathematics): the assignment of exactly one output to each input, and the practice of defining, drawing, composing and using it
- Complex analysis (the practice of working with functions of a complex variable, and the holomorphic structure it works on)
- Infinity: the boundless or endless as mathematics and philosophy describe it (potential and actual infinity, Cantor's sizes of infinity, infinity in calculus, geometry, physics and computing) and the practice of the people who argue over it
represents (1)
Kinds (8)
- Real number (the practice of measuring the continuum, and the complete ordered field)
- Complex numbers (C)
- Natural number (the structure)
- Integer
- Rational number
- Irrational number
- Hyperreal number
- Surreal number
Notes published science
- The first written list was mostly numbers. At Uruk V, 39 of the 47 signs (83%) were numerals; 150-350 years later the Uruk III list had 838 signs and 87% stood for things and ideas. Writing started as bookkeeping and only then learned to say much else (src:bentz-dutkiewicz-2026-signs).
- A computer vision network trained only to tell cats from cars grew its own 'number neurons': about one unit in ten in its top layer became tuned to how many dots it saw, with the same log-scaled blur real neurons have. Another study found such units in networks that had never been trained at all (src:nasr-2019, src:kim-2021).
- Charles Babbage called the mechanical method of carrying the tens the most important part of his Analytical Engine. The carry is still the slow part: a 16-bit ripple adder needs 31 gate delays to settle, a lookahead adder 8, and people adding in their heads quietly reorder numbers to land on round tens (src:neth-payne-2025, src:wp:carry-lookahead-adder).
- Your number sense keeps sharpening long after school: in a sample of over 10,000 people, its precision peaked at about age 30 (src:halberda-2012).
- Snowshoe hare numbers swing on a 10-year cycle that travels like a wave from Saskatchewan across western Canada to Alaska, and the hares' browsing leaves dark marks in spruce tree rings that track the cycle back to 1751 (src:krebs-2018-hare, src:sinclair-1993-solar-hare).
- A 4,000-year-old scribal habit is still an open maths problem: can every fraction 4/n be written as three unit fractions, Egyptian-style? Nobody has proved it (the Erdős-Straus conjecture), though a typical prime has only a small number of ways to do it (src:elsholtz-tao-2011, src:zelator-2009).
- Play people a random rhythm, have them tap it back, and play them their own tapping: in all 39 groups tested across 15 countries, a few rounds pull the rhythm onto simple whole-number ratios like 1:1:2, though which ratios win depends on the local music (src:jacoby-2024).
- A trained eye parses a written formula with nested brackets in under a fifth of a second, in the visual brain, and for professional mathematicians doing algebra the language areas stay quiet: the work runs in circuits for number and space (src:maruyama-2012, src:amalric-dehaene-2016).
- In 1925 Cecilia Payne worked out that stars are mostly hydrogen and helium; her advisor Russell discouraged the result, and by 1929 he had accepted it himself (src:lodders-2020).
- Every continuous function of many variables, however tangled, can be built from functions of a single variable and plain addition, and the inner one-variable pieces are the same for every function; only the outer pieces change. Kolmogorov and Arnold proved it in 1956-57, against what Hilbert's 13th problem expected. Source: src:toscano-2025.
- Your first number line may be logarithmic: Mundurucu adults and children, with little formal schooling, place numbers on a line so the gaps shrink as numbers grow, and Western adults do the same when they can't count. A straight, even number line may be something school installs. Other researchers say the same data fit an even line. Source: src:dehaene-2008, src:cantlon-2009.
- Ask people for the next number after 9, 15, 21 and they say 27, not 149.5, though a rule can be found that gives either; the Bayesian arithmetic says 27 is millions of times more probable. Source: src:wilson-2015.
- Language models were found doing addition with trigonometry nobody taught them: three models hold each number on a helix (circles of period 2, 5, 10 and 100) and add by rotating the helices like a clock; when they get sums wrong, the authors test whether they failed to 'carry' (src:kantamneni-tegmark-2025).
- To find an Earth twin by its starlight you must see the star wobble at 9 cm per second, a 100 kHz shift over a whole year; a change of one millibar in air pressure inside the instrument would drift the spectrum by about 100 m/s, so the best spectrographs live in vacuum tanks held to a hundredth of a degree (src:fischer-2016).
- Maxwell's demon has been built: by watching a jiggling bead and flipping a potential at the right moment, physicists turned information into free energy with 28% efficiency, paid for by the cost of handling the information (src:toyabe-2010-information-heat-engine).
- 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).
- Does nature need imaginary numbers? A 2021 theory paper and two 2022 experiments said yes (one beat the real-number limit by 43 standard deviations); a 2025 preprint argues a real-number quantum theory fits the same postulates and no experiment can tell them apart. The question is open again (src:chen-2022, src:hoffreumon-woods-2025).
- A cochlear implant is electronics wired straight into a person's hearing at the nerve, skipping the ear's hair cells; researchers call it the most successful neural prosthesis developed to date (src:pisoni-2017-ci).
- A starling in a murmuration keeps track of about 21 neighbours, by count rather than distance, whatever the flock's density; with that number fitted, a magnet-style model predicts how order spreads through the whole flock with no free parameters (src:bialek-2012-flocks).
- In a modern chip, thinking is cheap and moving is expensive: fetching one byte from off-chip memory costs about a thousand times the energy of adding two numbers, and 30-60% of a chip's energy is burned just in its wires (src:leugering-2024-neuromorphic, src:markov-2014-limits).
- A firewall is mostly as good as the people who write its rules: when 37 real corporate rule sets were audited in 2000-2001, only one had just a single mistake, and the number of errors grew with the logarithm of the rule set's size, about ln(RC) + 1.5; a repeat study of 84 rule sets found more than 60% still had rules of the form 'from somewhere to any, allow any service' (src:wool-2004, src:wool-2010).
- Your eye is a three-number spectrometer: any colour you see is just three cone responses, so two different spectra can look identical (src:funt-roshan-2021). Mantis shrimp carry 12 colour receptors, yet tests show they are poor at telling close colours apart (src:streets-2022).
- Children of 4 to 7 mostly agree you can always add one more to numbers, space and time (about four in five), but fewer say these never end: the rule comes before the endlessness. Source: src:sullivan-2023.
- Alex learned 'seven' and 'eight' only as the numerals after six on the list, and then, untaught, applied them to piles of seven and eight things: the order of the list told him the size (src:pepperberg-carey-2012-abs, src:pepperberg-carey-2012). He had learned his first number words out of order, three and four first and one last.
- Number words are some of the slowest-changing words people have: 'one' to 'five' can stay recognisably the same for 10,000 to over 100,000 years, 3.5-20 times longer than ordinary basic vocabulary (src:pagel-meade-2017).
- Tapping your finger fast for a few seconds makes the next cloud of dots look like fewer dots, and the effect follows where your hand tapped, not which hand: the brain's number sense is shared by eyes, ears and actions (src:burr-2017).
- Three-day-old chicks can keep a running total of objects disappearing behind two screens and go to the screen hiding more; newborn babies match 4-18 dots to sequences of sounds by number (src:zanon-2026, src:izard-2009).
- Fetching one number from main memory costs about 800 times the energy of adding two numbers, and on phones about 63% of all energy goes to moving data around, which is why the choice of data structure (how far data must travel) matters so much (src:mutlu-2020-pim-primer). And a food-caching bird's hippocampus seems to stamp each hidden seed with its own sparse 'barcode', an index kept apart from the memory's contents, much like a key in a hash table (src:fang-2026-barcodes).
- In 2024 an online group of mostly amateur mathematicians settled a 60-year-old question: the longest any 5-state machine can run before stopping is exactly 47,176,870 steps. They had to check 181,385,789 machines, and the whole proof was checked by computer. With one more state, a machine called Antihydra stops only if a Collatz-like puzzle about odd and even numbers is false, so nobody knows. Sources: src:bbchallenge-2025, src:wp:busy-beaver.
- Across 21 billion US Facebook friendships, one number stood out: how many better-off friends poorer people have. If low-income children grew up where that connectedness matched what high-income children get, their adult incomes would be about 20% higher on average (src:chetty-2022).
- The Antikythera Mechanism's maker hit a wall that every gear designer still meets: the most accurate known Venus cycle needed a gear with 1151 teeth (a prime number, so it can't be split into smaller gears), so he used a slightly less accurate cycle that could actually be cut, with gears under 100 teeth (src:freeth-2021-antikythera).
- At Facebook one CPU core answered 0 every time it was asked for the whole-number part of 1.1 to the power 53, while giving the right answer for 1.1 to the 52; nothing in the machine's logs showed a fault, and the bad answers silently deleted files (src:dixit-2021-sdc). Google sees 'a few mercurial cores per several thousand machines' (src:hochschild-2021-mercurial).
- Richard Dedekind got the idea that founds the real numbers while teaching beginners calculus in Zurich: leaning on pictures of lines left him so dissatisfied that he resolved to keep meditating until he found an arithmetic answer. He found it on 24 November 1858, sat on it for 14 years, and published only when Heine's and Cantor's papers arrived. He also wrote that space, if it is real at all, need not be continuous (src:dedekind-1872).
- Your computer cannot hold almost any real number, and the trouble is a carry that may never arrive: to round log(x) correctly to the 53 bits a double stores, some inputs need about 117 bits of working precision, and adding a list of numbers in a different order gives a different answer (src:boldo-2023).
- Physicist Nicolas Gisin argues the "real" numbers are badly named: almost all of them carry infinite information, which no finite piece of space can hold, so he proposes calling them "random numbers" and rewriting classical physics with finite-information numbers, which makes it indeterministic (src:gisin-2019).
- Adults stumble on the infinity sign because it is small: when people compare ∞ with digits they treat it like a big ordinary number, answering faster the bigger the number it is paired with, as if ∞ sat just past it on the number line. Source: src:feder-2024.
- Children tend to say 'you can always add one more' before they say 'numbers never end', and the second belief goes with knowing how the count list rolls over to make new number words: the carry comes before the infinite. Sources: src:sullivan-2023, src:chu-2020.
- The pigeonhole principle is named after Dirichlet, but it was in print two centuries earlier as a Jesuit teaching line: two men must have the same number of hairs. A 1624 book made a wager of it, and the 'two people with the same number of hairs' example is still how it is taught. Source: src:rittaud-heeffer-2014.
- The numbers 1 to 7824 can be split into two groups with no a² + b² = c² inside either group, but 1 to 7825 cannot. Proving the 7825 case took 800 computer cores about two days and produced a proof of almost 200 terabytes, which its own authors say looks like 'a giant heap of random information' to a person. Source: src:heule-2016.
- Day-old chicks can pick out a prime: chicks shown only even-numbered sets for an hour then spent longer next to 7 or 11 objects than next to 9, apparently because a prime number of things can never be split into equal groups (src:loconsole-regolin-2022).
- Periodical cicadas live on prime numbers: their broods come up every 13 or 17 years, so the two kinds surface together only once in 221 years; why primes is still argued between dodging predators, avoiding hybrids, and an ice-age origin (src:loconsole-regolin-2022, src:toivonen-fromhage-2020, src:ito-2015).
- The largest known prime, 41,024,320 digits long, was found in 2024 on thousands of rented cloud GPUs in 17 countries, after nearly a year and about $2 million; it was the first record flagged by a fast 'probably prime' test, which started an argument about which day counts as the discovery (src:gimps-2024, src:wp:largest-known-prime-number).
- Truth has a ceiling built in: no consistent rule system for arithmetic can prove every true statement about whole numbers (Goedel, 1931), and no language can define its own truth from inside (Tarski); the only way out is to hand the job up to a bigger language, which then cannot define its own. (src:wp:goedels-incompleteness-theorems, src:wp:semantic-theory-of-truth)
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.
quantity in the world: sets and sequences (reader's view)
- L1: Quantity in the world: discrete items and events, sets of objects, sequences of sounds, flashes and actions, groups of conspecifics, prey and predators
- L2: The senses taking a set in: early visual cortex responds monotonically to how many items there are (summation code); hearing takes in sequences of sounds and pulses; touch and the moving hand take in taps
- L3: Individuation: salience maps mark each object as a local peak (local excitation, lateral inhibition); small sets are held as separate object files (Object Tracking System, about 3-4); dense sets are read as texture instead
- L4: The number sense: an approximate number system (Weber's law, ratio-dependent) carried by number neurons tuned to a preferred numerosity on a logarithmic scale; intraparietal sulcus and prefrontal cortex in primates, nidopallium caudolaterale in crows, pallium in fish
- L5: Counting: the learned count list ("one, two, three" recited first without meaning), its words learned one by one (knower-levels), then the cardinal principle: Full-Counters use counting to determine exact cardinality; number words are heard on the IPS
- L6: Number systems: numerals and number words built on bases taken from the hand and body (five, ten, twenty), tallies and notations, the Hindu-Arabic place-value decimal system; and the structure they name (natural numbers with successor, extended to zero, negatives, rationals, reals and complex numbers); mathematical thought recycles the IPS
- L7: Number put to use on the world: choices by numerical advantage (lions, hyenas, wolves), joining the larger shoal, foraging; counting group members, prey, property and debts; measuring, ordering and labelling
the boardroom (shared floors)
- L4: Object Tracking System: small sets (about 3-4) held as separate object files in visual working memory; works across modalities in infants
- L4: The number sense (approximate number system): number neurons tuned to a preferred numerosity, logarithmic scale, Weber's law; abstract across space, time, modality and action
- L5: Counting: the count list recited, its words learned one by one, then the cardinal principle (Full-Counters); learned from instruction
- L6: Number systems and mathematics: bases from the hand and body, tallies and notations, place-value numerals; the structure of natural numbers and its extensions; mathematical reflection recycles the number-sense regions
- L7: Number put to use: decisions by numerical advantage, group choice, foraging, keeping track of group members, prey, property and debts
way in: Hearing: sequences of sounds counted in time
- L1: Sequences of sound events: tones, roars and calls of conspecifics, frogs' pulsed notes
- L2: Ear and auditory pathway take in timed pulses; the temporal interval between pulses carries species identity and call type
- L3: Interval-counting neurons in the anuran midbrain: spike only after a threshold number of correctly timed pulses; one long gap resets the count
- L4: Numerosity of sequences shares the visual number code: adapting to tones shifts seen flashes and vice versa (probably IPS)
way in: Landmarks passed and steps taken (navigation)
- L1: Landmarks passed along a flight path (bees); steps taken (ants)
- L2: Bees take in spatial detail by scanning items one after another, not at a glance
- L3: Landmark count weighed against the learned distance in choosing the feeder
- L4: Counting-like ability in a very small circuit (the insect brain)
way in: Number words and numerals (the cultural channel)
- L1: Number words spoken by others (count list sequences); numerals on surfaces
- L2: Auditory (and visual) cortex: number words and alphabet sequences activate them alike
- L3: Ordinal sequence representations for number words (and alphabet) in bilateral inferior frontal gyri and anterior temporal lobes
- L4: Bilateral IPS takes number word sequences, overlapping visual numerosity (the port); in trained monkeys prefrontal neurons bind symbols to values
way in: Sight: sets of objects seen (visual numerosity)
- L1: Arrays of objects reflecting light (dot arrays, shoals, groups, food items)
- L2: Early visual cortex: firing that rises monotonically with numerosity (summation code)
- L3: Salience maps individuate objects as local activity peaks; crowding past a density limit turns the array into texture; 1-3 items subitized
- L4: Tuned numerosity representations in the intraparietal sulcus (the port)
way in: Smell: how many senders in an odour
- L1: Olfactory bouquets from up to four females (mealworm beetles)
- L4: Males track the number of females with a ratio-dependence greater than 1:2
way in: Thought: numbers held and worked in the mind (inner)
- L1: A number or a mathematical idea arriving in thought (origin contested: C-413)
- L4: Mathematical reflection runs on bilateral intraparietal and ventral temporal regions of the number sense, not the language areas
way in: Touch and action: one's own taps and finger movements
- L1: Taps and finger movements made by the counter (and touched items)
- L2: Skin and proprioceptive receptors of the tapping hand (see Human's touch and proprioception channels)
- L3: The count of actions is kept in spatial, not hand, coordinates: adaptation follows where the hand taps, not which hand
- L4: Shared number code: adapting to fast or slow tapping shifts the apparent numerosity of flashes and dot arrays
way out: Calling a number of notes (anurans)
- L1: Calls of growing complexity during aggressive interactions
- L7: Matching the number of calls or notes a neighbour produces
way out: Choosing and approaching (animals)
- L1: Approach or join: larger shoals, food-associated numerosities
- L7: Respond by numerical advantage; pick the screen hiding the larger running total
way out: Marks: notches, tallies, numerals
- L1: Notches cut into bone at different times; incisions recording homologous units
- L6: A notation: one mark per unit, or numerals in a place-value system
way out: Pointing, finger counting, moving tokens
- L1: Fingers raised, items touched or moved
- L5: Each item tagged and marked as counted, until none is left unmarked
- L6: Finger-counting conventions (handful, one hand, two hands, hands and feet) set the base
way out: Saying the count (number words aloud)
- L1: Number words in the air, heard by others
- L5: The count list recited in order (count-list elicitation, Move-and-Say)
- L7: Deciding to count aloud or report a number