VOYAGE NINETEEN · WHAT ORDER LIVES INSIDE APPARENT RANDOMNESS

Primes and Hidden Order — From the Sieve to Bounded Gaps

Prove in Alexandria that primes never end, see repeating remainders in Toulouse, and translate products into series in Saint Petersburg. Continue through zeta zeros in Berlin, average density in Paris and Brussels, and modern sieves and gaps in Oslo, Beijing, Cambridge, and Durham—measuring the distance between ‘a pattern appears’ and ‘it is true infinitely often.’

QUESTION FOR THE ROUTE

Why do primes obey average laws and residue structure even when the next prime is hard to predict, and where must finite computation stop short of proof?

WHAT THE LINE DOES NOT CLAIM

This route does not equate primes with a random sequence, a universal key to cryptography, or one person’s discovery. It separates finite samples from infinite theorems, necessary residue conditions from sufficient primality tests, average density from individual locations, and bounded gaps from the twin-prime conjecture. The route line is an edited comparison, not one document’s proven chain of transmission.

The camera rests on each city while you read, then eases through the runway between scenes. Select any marker or scene link to travel in either direction.

The same route, four questions

A lens never hides a scene or proves a cause. It changes which places you compare first, and the URL preserves your choice.

Whole route Translation networks Commercial demand Observation and institutions Patronage and power

Translation networks · READING QUESTION

What was lost, preserved, or newly created when an idea met another language and audience?

Primes in geometric propositions, congruences in letters, infinite products and series, complex zeros, and combinatorial pseudorandomness did not descend through one notation or text. The question ‘where do multiplicative atoms occur?’ was translated into proofs, tables, series, functions, and computation.

Start with:Alexandria Toulouse Saint Petersburg Berlin Cambridge

Translation, copying, or commentary does not prove that one document traveled directly through the whole route.

Live map · 지도를 불러오는 중…

01 / 15 · c. 300 BCE

Alexandria

01 · c. 300 BCE
Alexandria · Composition
Proving That the Primes Never End — Book IX of the Elements
Translation networks · lens spotlight
Proposition 20 of Euclid’s Book IX shows that however many primes are given, those alone cannot account for every number. It is often summarized as ‘multiply them all and add one,’ but the original argument uses a prime divisor of a number one greater than the product. Infinitely many primes is neither a formula for the next prime nor the whole modern theorem of unique factorization.
PAUSE AND ASK
How can one prove that every finite list of primes must miss another prime?
How the idea changed
Move from counting many examples to one general argument that defeats any proposed finite list.
What this place made possible
The Alexandrian editorial and teaching tradition associated with the Elements preserved arithmetic propositions as a chain depending on prior definitions and results.
How it moved
Greek arithmetic traditions → propositions ordered in Book IX → Greek, Arabic, and Latin editions → repeated teaching of prime infinitude → modern generalizations
Do not overclaim
The exact writing room and date are not known. Proposition 20 does not say that one plus a product of given primes is always prime, and it is not identical to the whole modern unique-factorization theorem.
Evidence sources
- Clay Mathematics Institute — Euclid, Elements IX.20 commentary
  Supports: The structure of Elements IX.20 and its proof of infinitely many primes
Stable link to this scene
AlexandriaAlexandria
02 · c. 240 BCE
Alexandria · Main activity
Finding by Crossing Out — The Sieve of Eratosthenes
Cross out numbers forced to be composite—multiples of 2, then 3, and so on—and the survivors are prime. The procedure turned a definition into repeatable computation and remains a foundation for finding primes in intervals. Surviving descriptions are later, however, so the exact ancient implementation and date are not reconstructed as modern code.
PAUSE AND ASK
Instead of testing numbers one by one, can we first remove those forced to be composite?
How the idea changed
Turn a static definition into a generative procedure of repeatedly removing multiples—and into a question of computational cost.
What this place made possible
Alexandria’s tables, astronomy, measurement, and collecting institutions supplied audiences and media for organizing long sequences and transmitting procedures.
How it moved
Marking multiples → the sieve later attributed to Eratosthenes → manuscript arithmetic teaching → mechanical tables → segmented sieves and modern prime searches
Do not overclaim
This is not modern pseudocode surviving in Eratosthenes’ own hand. Attribution and approximate date rely on later testimony, and the basic sieve is not the optimal method for every modern primality task.
Evidence sources
- MacTutor — Prime numbers
  Supports: The traditional attribution of the sieve and an overview of prime-number history
Stable link to this scene
AlexandriaToulouse
03 · 1640 CE
Toulouse · Letter sent
Seeing Repetition in Remainders Modulo a Prime — Fermat’s Little Theorem
Translation networks · lens spotlight
In a letter to Frénicle dated 18 October 1640, Fermat stated that if p is prime and p does not divide a, then a^(p−1) leaves remainder 1 modulo p. No complete proof survives in that letter. The theorem helps test primality, but its converse fails: some composite numbers pass related tests.
PAUSE AND ASK
Can a repeating remainder pattern modulo a prime be used to test primality?
How the idea changed
Focus on recurring remainders rather than quotients, comparing infinitely many integers through finite congruence classes.
What this place made possible
Fermat’s research alongside legal work in Toulouse and a European correspondence network moved unpublished statements and challenges quickly among peers.
How it moved
Problems on integer powers → Fermat–Frénicle letter of 1640 → Euler’s proof and generalization → congruence notation → probabilistic primality tests
Do not overclaim
The surviving letter contains no complete proof, and passing a^(p−1)≡1 mod p does not guarantee primality. ‘Little theorem’ does not mean mathematically unimportant.
Evidence sources
- MacTutor — Perfect numbers
  Supports: Fermat’s little theorem as stated in the 18 October 1640 letter to Frénicle
Stable link to this scene
ToulouseSaint Petersburg
04 · 1737 CE
Saint Petersburg · Main activity
Connecting a Product over Primes to an Infinite Series — Euler
Translation networks · lens spotlight
Euler developed the divergence of the sum of prime reciprocals and a factorization of a series over all integers into a product indexed by primes. His 1737 work and the Euler product in the 1748 Introductio carried multiplicative structure into analysis. The identity does not directly predict each individual prime.
PAUSE AND ASK
How can the multiplicative structure of primes appear inside an additive series over all integers?
How the idea changed
Leave individual prime tables for a factorization of a series over all integers into an infinite product over primes, making distribution an analytic object.
What this place made possible
The St Petersburg Academy’s salary, publication system, and calculating community let Euler move for years among series, products, and integer questions.
How it moved
Prime factorization → harmonic and infinite series → divergence of prime reciprocals in 1737 → Euler product in 1748 → Dirichlet series and zeta functions
Do not overclaim
The 1737 work and the expression in the 1748 book are not collapsed into one date. Euler products require a convergence-domain distinction and do not list individual primes directly.
Evidence sources
- MacTutor — Leonhard Euler
  Supports: Euler’s St Petersburg activity and his 1737 work connecting primes and series
Stable link to this scene
Saint PetersburgBerlin
05 · 1837 CE
Berlin · Publication
Finding Infinitely Many Primes in Each Eligible Progression — Dirichlet
Dirichlet proved that an arithmetic progression whose first term is coprime to its common difference contains infinitely many primes. The condition matters: not every progression qualifies. Primes were no longer merely scattered points; every admissible residue lane became an infinite distribution problem.
PAUSE AND ASK
Do primes continue forever inside numbers with a specified remainder?
How the idea changed
Refine infinitude of all primes into infinitude within every coprime arithmetic progression, combining characters with analytic tools.
What this place made possible
Berlin’s university, academy proceedings, and research environment provided a venue to publish and inspect a new combination of number theory and analysis.
How it moved
Fermat and Euler congruence patterns → Gauss’s congruence notation → Dirichlet characters and L-series → 1837 theorem → distribution in arithmetic progressions
Do not overclaim
Coprimality of first term and common difference is essential. The theorem does not locate the next prime in each progression or guarantee short-interval uniformity.
Evidence sources
- MacTutor — Peter Gustav Lejeune Dirichlet
  Supports: Dirichlet’s Berlin activity and his 1837 theorem on primes in arithmetic progressions
Stable link to this scene
BerlinSaint Petersburg
06 · 1848 CE
Saint Petersburg · Publication
Bounding Density without an Exact Formula — Chebyshev
Chebyshev rigorously trapped the prime-counting function π(x) on the scale of x/log x and showed that if the limiting ratio existed, it had to equal 1. He did not finish the prime number theorem, but built a strong fence between conjecture and proof.
PAUSE AND ASK
Without proving an exact asymptotic formula, can prime density still be trapped rigorously on the correct scale?
How the idea changed
Replace one plausible approximation with upper and lower bounds and the conditional conclusion ‘if the limit exists,’ raising the level of proof.
What this place made possible
St Petersburg’s university and academy traditions in probability, number theory, and mechanical calculation supported controlling arithmetic functions through inequalities.
How it moved
Prime tables and Gauss’s approximation → Chebyshev functions and inequalities → 1848 memoir → proof of Bertrand’s postulate → 1896 asymptotic theorem
Do not overclaim
Chebyshev did not prove the full statement π(x)~x/log x. His bounds and conditional result are not retroactively replaced by the later theorem.
Evidence sources
- Clay Mathematics Institute — Bombieri on the Riemann Hypothesis
  Supports: Chebyshev’s 1848 bounds on prime distribution and their place on the road to the prime number theorem
Stable link to this scene
Saint PetersburgGöttingen
07 · 1849 CE
Göttingen · Letter sent
Recording an Early Observation in a Later Letter — Gauss and the Logarithmic Integral
In an 1849 letter to Encke, Gauss recalled that while studying prime tables in his early years he had come to expect average density 1/log x and regarded the logarithmic integral Li(x) as a better approximation. The familiar claim that he discovered x/log x at exactly age fifteen overstates the precision of a retrospective account.
PAUSE AND ASK
How did a law like ‘average gap near x is about log x’ emerge from irregular prime tables?
How the idea changed
Shift from predicting the next prime to comparing cumulative counts and local average density.
What this place made possible
Göttingen’s observational, computational, and correspondence setting, together with Gauss’s long-kept tables, supported his 1849 account to Encke of an early empirical observation.
How it moved
Hand prime and logarithm tables → young Gauss’s empirical law → 1849 letter to Encke → comparison of Li(x) and x/log x → formal asymptotic distribution
Do not overclaim
‘Discovered at fifteen’ turns an 1849 recollection into an exact birthday event. The empirical approximation is also distinct from the 1896 proof.
Evidence sources
- Clay Mathematics Institute — Bombieri on the Riemann Hypothesis
  Supports: Gauss’s 1849 letter to Encke and retrospective account of his early Li(x) approximation
Stable link to this scene
GöttingenBerlin
08 · 1859 CE
Berlin · Publication
Hearing Prime Fluctuations through Zeta Zeros — Riemann
Translation networks · lens spotlight
Riemann’s Berlin Academy paper extended the zeta function into the complex plane and related its zeros to fluctuations in prime counting. The Riemann hypothesis says that the nontrivial zeros have real part one half. The prime number theorem was proved without the hypothesis in 1896, and even RH would not become a one-line next-prime formula.
PAUSE AND ASK
How can zeros of a complex function tune the irregular error in prime counting?
How the idea changed
Move beyond one average density to analytic continuation, a functional equation, and many oscillations generated by zeta zeros.
What this place made possible
The Berlin Academy’s monthly reports and election process provided the institutional stage for Riemann’s short, compressed number-theory program.
How it moved
Euler product → Dirichlet L-series → Riemann’s 1859 complex extension and explicit formula → nonvanishing boundary in 1896 → modern zero computation and the hypothesis
Do not overclaim
The Riemann hypothesis is not required for the prime number theorem and remains open. Checking finitely many zeros does not prove an infinite claim, and RH would not be a simple formula for individual primes.
Evidence sources
- Clay Mathematics Institute — Riemann’s 1859 manuscript
  Supports: The 1859 Berlin Academy paper and manuscript connecting prime distribution with zeta zeros
Stable link to this scene
BerlinParis
09 · 1896 CE
Paris · Publication
Proving the Law of Average Density — Hadamard
Hadamard used complex analysis, including nonvanishing of zeta on the line of real part one, to prove π(x)~x/log x. De la Vallée Poussin found an independent proof in the same year. The symbol ~ says their ratio tends to one; it does not say their finite difference stays small or on one side.
PAUSE AND ASK
Why does proving that zeta has no zero on a boundary force the average density of primes?
How the idea changed
Translate a staircase function on integers into analysis of singularities and zero-free regions in the complex plane.
What this place made possible
Parisian universities, academies, and journals gathered readers able to publish and compare a long proof linking complex function theory and number theory.
How it moved
Riemann’s program → Hadamard’s theory of entire functions → nonvanishing on real part one → π(x)~x/log x in 1896 → analytic number theory
Do not overclaim
Asymptotic equivalence ‘~’ does not mean equality at finite x or a monotonically shrinking difference. De la Vallée Poussin’s independent proof in the same year remains visible.
Evidence sources
- MacTutor — Prime numbers
  Supports: The independent 1896 proofs of the prime number theorem by Hadamard and de la Vallée Poussin
Stable link to this scene
ParisBrussels
10 · 1896 CE
Brussels · Publication
An Independent Proof Opens the Study of Error — de la Vallée Poussin
In work published through a Brussels learned journal, de la Vallée Poussin independently proved the prime number theorem and used a zero-free region to obtain finer error information. Placing the Paris and Brussels routes side by side reveals the journals, universities, and correspondence that made proofs inspectable—not only the theorem’s name.
PAUSE AND ASK
Can an independent proof of the same average law also ask a sharper question about its error?
How the idea changed
Ask not only whether the theorem is true, but how a zero-free region controls the approximation’s error.
What this place made possible
Research at Leuven and the Brussels scientific society’s publication network circulated an independent Belgian route distinct from the Paris center.
How it moved
Riemann zeta function → Belgian university research → Brussels society publication → independent 1896 proof and zero-free region → sharper error terms
Do not overclaim
The pin marks the Brussels publication network and is distinct from the author’s main post at Leuven. The two proofs are not merged into one joint collaboration.
Evidence sources
- Clay Mathematics Institute — Bombieri on the Riemann Hypothesis
  Supports: De la Vallée Poussin’s independent 1896 proof and zero-free region
Stable link to this scene
BrusselsOslo
11 · 1919 CE
Oslo · Publication
Showing That the Reciprocal Sum of Twin Primes Converges — Brun’s Sieve
Brun developed a new sieve to prove that the sum of reciprocals of twin-prime pairs converges. Since the reciprocals of all primes diverge, twin primes are dramatically sparser. A finite Brun constant, however, does not decide whether there are finitely or infinitely many twin-prime pairs.
PAUSE AND ASK
Can one prove how sparse twin primes are without knowing whether infinitely many exist?
How the idea changed
Move from a sieve that identifies primes exactly to weighted sieves allowing almost-primes in exchange for upper and lower bounds.
What this place made possible
The university and learned society in Kristiania, together with contacts made in Göttingen, supported Brun’s combinatorial sieve research and Norwegian publication.
How it moved
Sieve of Eratosthenes → Legendre-style inclusion–exclusion → Brun’s weighted sieve → convergence of twin-prime reciprocals in 1919 → modern sieve theory
Do not overclaim
Finiteness of Brun’s constant proves neither infinitely nor finitely many twin primes. Convergence and number of terms are different questions.
Evidence sources
- MacTutor — Viggo Brun
  Supports: Brun’s sieve and his 1919 convergence result for reciprocal twin primes
Stable link to this scene
OsloPrinceton
12 · 1949 CE
Princeton · Main activity
Proving the Prime Number Theorem Again without Complex Zeros — Selberg and Erdős
Selberg’s fundamental formula and Erdős’s combinatorial arguments produced an ‘elementary’ proof of the prime number theorem without complex function theory. Elementary describes the permitted tools, not an easy proof. Disputes over priority and communication also make this a shared result rather than a simple lone-victor story.
PAUSE AND ASK
Does proving the same prime number theorem without complex analysis reveal a different reason it is true?
How the idea changed
Replace zeta zeros with symmetric identities for logarithmic prime sums and combinatorial estimates, redrawing the tool boundary called ‘elementary.’
What this place made possible
Visiting research, seminars, and rapid manuscript exchange at Princeton’s Institute for Advanced Study combined Selberg’s and Erdős’s ideas quickly while intensifying priority conflict.
How it moved
Selberg’s fundamental formula → Erdős’s combinatorial amplification → 1948–1949 announcements and dispute → PNT without complex analysis → elementary analytic number theory
Do not overclaim
‘Elementary’ means neither easy nor historically earlier. The episode is not reduced to a winner’s story in which one person finished everything and the other merely polished it.
Evidence sources
- AMS Bulletin — The elementary proof of the prime number theorem
  Supports: The mathematics and 1948–1949 history of the Selberg–Erdős elementary proof
Stable link to this scene
PrincetonBeijing
13 · 1973 CE
Beijing · Main activity
Leaving Only One Extra Prime Factor in Goldbach’s Problem — Chen Jingrun
Chen proved in full detail that every sufficiently large even integer is the sum of a prime and a number with at most two prime factors, extending his 1966 announcement. This is a summit of sieve methods, not the strong Goldbach conjecture that every even integer greater than two is a sum of two primes. ‘Sufficiently large’ and finite exceptions also remain distinct.
PAUSE AND ASK
If requiring two primes is too rigid, how close can one get by allowing one extra prime factor?
How the idea changed
Distinguish primes from almost-primes with at most two factors, replacing an exact conjecture with a quantified near theorem.
What this place made possible
The Chinese Academy institute, seminars, and publishing environment in Beijing provided an institutional base for preserving and checking long sieve calculations amid political upheaval.
How it moved
Goldbach problem → Hardy–Littlewood circle method → Brun and Selberg sieves → Chen’s 1966 announcement and 1973 full paper → modern additive number theory
Do not overclaim
Chen’s theorem says every sufficiently large even N equals p+P₂, where P₂ has at most two prime factors. It is not a proof that every even number is a sum of two primes.
Evidence sources
- AMS Graduate Studies in Mathematics — Sieve Methods excerpt
  Supports: Chen’s theorem expressing every sufficiently large even integer as a prime plus a P₂
Stable link to this scene
BeijingCambridge
14 · 2004 CE
Cambridge · Main activity
Finding Arbitrarily Long Arithmetic Progressions inside the Primes — Green and Tao
Translation networks · lens spotlight
Ben Green and Terence Tao proved that the primes contain arithmetic progressions of every finite length. The primes in such a progression need not be consecutive primes. Collaboration across Cambridge and UCLA combined additive-combinatorial structure with pseudorandomness, revealing long order within a sparse set.
PAUSE AND ASK
Must an increasingly sparse set of primes still contain arithmetic progressions of every desired length?
How the idea changed
Do not call primes random; transfer a structure theorem for dense sets into a sparse pseudorandomly weighted setting.
What this place made possible
Email, preprints, and international seminars connecting Green in Cambridge and Tao at UCLA combined distinct specialties in one proof.
How it moved
Szemerédi’s dense-set theorem → harmonic analysis and additive combinatorics → von Mangoldt weights for primes → 2004 Green–Tao preprint → more quantitative pattern research
Do not overclaim
Primes in the progression need not be consecutive, and the common difference varies by progression. One map pin does not erase Tao at UCLA or the distributed collaboration.
Evidence sources
- Annals of Mathematics — The primes contain arbitrarily long arithmetic progressions
  Supports: The theorem on arbitrarily long arithmetic progressions in the primes and the authors’ affiliations
Stable link to this scene
CambridgeDurham, NH
15 · 2013 CE
Durham, NH · Main activity
Proving That Primes Come Boundedly Close Infinitely Often — Yitang Zhang
Zhang proved that infinitely many consecutive-prime gaps are smaller than seventy million. Later Maynard–Tao ideas and Polymath collaboration greatly reduced the bound, while the exact gap-two twin-prime conjecture remains open. Finite computation, recurrence below some fixed bound, and infinitely many exact gap-two pairs are three different claims.
PAUSE AND ASK
Without proving exact gap two, can one show that some fixed bounded prime gap recurs infinitely often?
How the idea changed
Reframe the exact gap-two conjecture into a staged breakthrough: first prove that some finite upper bound exists infinitely often.
What this place made possible
A teaching post, library access, and independent research time at the University of New Hampshire, followed by online Polymath verification, connected a quiet manuscript to rapid collective improvement.
How it moved
Goldston–Pintz–Yıldırım method → Zhang’s strengthened distribution estimate → 70-million bound in 2013 → Maynard–Tao multidimensional sieve → Polymath reductions
Do not overclaim
Seventy million was an initial upper bound for infinitely many cases, not every consecutive-prime gap. Later smaller bounds still do not prove the exact gap-two twin-prime conjecture.
Evidence sources
- Annals of Mathematics — Bounded gaps between primes
  Supports: Zhang’s 2013 submitted paper, the 70-million bounded-gap theorem, and his UNH affiliation
Stable link to this scene

FOUR SCALES OF HIDDEN ORDER

Sieve, sort, count, then ask what the evidence cannot settle

Every choice is finite, computed in the browser, and preserved in the URL. Patterns guide conjecture; they do not quietly turn into proofs.

DEFINITION BECOMES A PROCEDURE

Do not hunt primes—remove what cannot be prime

Keep 2, cross out its later multiples; keep 3, then cross out its later multiples. Every composite eventually reveals a smallest prime factor.

primes left

last prime shown

23456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100

The grid computes the result; its color is not a proof that this ancient attribution is exact.

Average density, congruence restrictions, and bounded gaps describe different kinds of order. None is a formula listing the primes one by one.

TOUCH THE MATHEMATICS