Plectis

Long paper

The Binary Totient Series

Erdős #249 By July 2026; revised 1 October 2026 Authorship, AI use and citation

Summary. The dyadic totient sections through level e≥1 have an explicit rational basis of size 2e+1. Truncating a scaled tail difference produces an integer whose residue certifies nonintegrality when its distances from both endpoints of the residue interval exceed an explicit error bound. The record develops the resulting finite denominator exclusions, equivalent residue criteria, and sufficient exponential-sum estimates. Rational comparison sequences identify the limitations of particular size, parity, rank, and finite-prefix arguments. The short paper supplies the all-base basis theorem; here each conditional result carries the arithmetic hypothesis still needed to apply it. Those hypotheses are not established in the form required to prove irrationality of S=∑ φ(n)2−n.

This paper covers a claim-bounded problem-specific reasoning surface for Erdős #249: selected checked premises, finite certificates, closed routes, surviving obligations, and their exact evidence bands.

It does not establish the validity of claims tagged Lean, which belongs to the cited kernel-checked source, or a solution to Erdős #249, which remains open.

In this paper

Parameter ranges and quantifiers

The distinction between a finite calculation and an unbounded family is in the quantifiers. A statement such as ∀a≥8 is uniform in a, whereas a list of computed scales is finite and a condition ∀N0∃N≥N0 asks for witnesses arbitrarily far out. The table uses these distinctions even when a uniform statement is conditional on an unproved hypothesis. A bound on a secondary variable, such as J+(a+6)<2⋅2a in the positivity row, does not bound a.

Some identities have no distinguished scale parameter. They are listed with the general identities. An equivalence can hold for all parameters without establishing either of the equivalent arithmetic assertions.

The LCM formulation of Problem 249 requires arbitrarily large scales: irrational_totientSeries_iff_actualLcmOrbitNonintegralitySupply

Irrational(∑n≥1φ(n)/2n)⟺∀a0 ∃a≥a0, Ωa∉Z,

a cofinal statement in the exponent a indexing H=H(2a). A finite calculation alone does not establish it; a further theorem could propagate finite information to arbitrarily large scales.

The table groups statements by their parameter ranges, not by proof status. In particular, some assertions with cofinal quantifiers are already proved, such as the prime constructions and rational comparison sequences. Others are implications whose quantified hypotheses remain open. The horizontal rule separates the cofinal group for ease of reference; it does not mark a boundary between proved and unproved mathematics. The descriptions identify this distinction in that group. The prefixes summarise the quantifiers; complete hypotheses are given in the cited statements. A definition records an object or condition; it does not by itself prove the associated soundness or existence assertion.

In the periodic-weight rows, write w:N→Z and cw(n)=∑d∣nw(d) for n≥1. Positivity and nonvanishing refer to these divisor coefficients, not to the weight w itself. Here “nonzero arbitrarily far out” means that for every N there is n>N with cw(n)≠0; it does not mean nonvanishing on every arithmetic progression. The nonprincipal character χ4 modulo four is 0 on even integers and (−1)(n−1)/2 on odd integers.

Results and unproved conditions, grouped by their quantifiers. The dividing rule separates the cofinal group, not proof status. Complete hypotheses are given in the cited statements.

Result

Source Quantifier prefix Representation

Result

Source Quantifier prefix Representation

scaled totient tail (definition) totientTail ∀N binary-weighted series
integer prefix and real tail two_pow_mul_totient_series_eq ∀N binary-weighted series
D(h,N,L) (definition) windowDiscrepancy ∀h∀N∀L finite binary sums
C(h,N,L) (definition) certifiedKill ∀h∀N∀L finite binary sums
necessary depth inequality certifiedKill_depth_floor ∀h∀N∀L, given ChNL finite binary sums
residue test soundness tail_diff_notMem_int_of_certifiedKill ∀h∀N∀L, given C finite binary sums
completeness of the residue test exists_certifiedKill_iff_tail_diff_notMem_int ∀h∀N finite binary sums
integrality from denominator divisibility tail_diff_int_of_den_dvd ∀N∀h∀r:Q, given S=r, r.den∣2N(2h−1) binary-weighted series
tail-period law eventual_period_of_not_irrational ¬IrrationalS→∃h>0∃N0∀N≥N0 binary-weighted series
least common multiple H(t) periodLcm ∀t LCM periods
a nondivisor below 2t is a prime power eq_prime_pow_of_not_dvd_periodLcm ∀t∀j(0<j<2t), given j∤Ht LCM offsets
totient factorisation at divisor offsets totient_periodLcm_ray_split j∣H(t), with the prime-divisor hypotheses of Proposition 6.51 totient factorisation
integrality on the LCM grid under rationality rational_totient_series_forces_lcm_cone_flatness ¬IrrationalS→∃t1∀t≥t1∀q>0∀m LCM grids
second-difference residue test second_diff_notMem_int_of_certifiedRank2Kill ∀h∀N∀L, given second-difference residue condition finite binary sums
a nonintegral pair on a finite grid exists_nonintegral_pair_of_coneNonflatCert finite Q≠∅; all size and residue hypotheses of Theorem 6.56 finite LCM grids
finite exclusion of integer carries tail_diff_notMem_int_of_survivorKill ∀h∀N∀K, given finite carry exclusion integer recurrences
Farey neighbour denominator law farey_gap a,b,c,d,r,s∈Z; b,d>0, bc−ad=1, as<br, dr<cs rational intervals
rational-approximation criterion irrational_of_den_mul_abs_sub_tendsto_zero uj∈Q, x∈R; eventually uj≠x, and den⁡(uj)|uj−x|→0 general criterion
near-integer irrationality criterion irrational_of_int_mul_near_int ∀ξ:R, ∀q>0∃mz0<|mξ−z|<1/q general criterion
independence from a nonsingular evaluation matrix linearIndependent_of_separatedMinorCertificate a finite family with a square evaluation matrix of nonzero determinant general criterion
dyadic totient-kernel rank =2e+1 linearIndependent_canonicalTotientKernelFamily ∀e≥1 dyadic sections
rationality forces unbounded carry-kernel rank not_irrational_totientSeries_implies_unbounded_carryRank_unconditional ¬IrrationalS→∃v>0∃u∀e carry sections
no fixed D clears all A(n)/n for A=φ∗μ no_fixed_integral_primitive_euler_index ∀D>0 Möbius sums
Lambert-series value identities tsum_totient_div_pow_two_eq_pnat_half_pow value identities, L(μ)=12, L(φ)=2, L(1)=Erdős–Borwein Möbius sums
prime-power unit-gap criterion reducedDenominatorUnitGapCert_prime_pow_iff ∀p prime, e>0, N,K≥0 rational intervals
positive rational-difference lower bound positive_rational_difference_lower_bound ∀ wholepfx:Q, pfx<whole general criterion
L2(μ)=S−12 tsum_totient_half_pow_eq_half_add_moebius_sq unconditional Möbius sums
weighted squared-denominator identity tsum_lambert_linear_weight_sq_pure ∀w:N→R(|w(d)|≤d)∀r∈[0,1) Möbius sums
L2(μ)=S−12 Proposition 5.14 unconditional Möbius sums
L2(1)=ζq(2)−ζq(1) identity tsum_one_div_mersenne_sq_eq_sigma_sub_tau_series unconditional Möbius sums
L2(φ) = gcd moment tsum_totient_div_mersenne_sq_eq_gcd_moment_series unconditional gcd moments
comparison of the weights Propositions 5.14–5.18 unconditional Möbius sums
gcd-divisibility factorizes tsum_pos_pair_both_dvd_half_eq_inv_mersenne_sq ∀d:N,d>0 geometric weights
sum over positive coprime pairs equals 1 tsum_pos_coprime_inv_mersenne_eq_one unconditional (sum over all coprime pairs) geometric weights
Stern–Brocot recursion with stopping cylinderMass_split positive integers a,b Stern–Brocot tree
positive numerator coefficients mobiusNumeratorPolynomial_eq_gcdWord ∀r(Squarefreer)∀k<r divisor and cyclotomic sums
evaluation at 2 gives the numerator mobiusNumeratorPolynomial_eval_two ∀r(Squarefreer) divisor and cyclotomic sums
reduction to the radical of H scaledMobiusShadow_eq_radicalBase ∀H>0∀r>0 Möbius sums
numerator congruence modulo Φm cyclotomic_dvd_mobiusNumeratorPolynomial_sub ∀r(Squarefreer)∀m(m∣r) divisor and cyclotomic sums
gcd of the numerator and Φr(2) mobiusNumerator_gcd_cyclotomicValue ∀r(Squarefreer) divisor and cyclotomic sums
Mersenne factors retained in the denominator upperHalfChannel_product_dvd_den_of_scale_primeFactors_le t≥5, squarefree r; finite P of primes dividing r in (t/2,t]; prime factors of r and the multiplier are at most t divisor and cyclotomic sums
denominator lower bound / exact value lcmHeight_scaledMobiusShadow_den_exact ∀t (exact-value); t≥5 (lower bound) Möbius sums
numerator congruence after adjoining a prime cyclotomic_dvd_primeJump_new_fibre r≥1, p prime, p∤r, m∣r divisor and cyclotomic sums
squared-Mersenne series: partial sum plus tail mobius_square_series_eq_partial_add_tail ∀D:N Möbius sums
cancellation of the affine divisor terms joint35_oldChannel_zero ∀d>0(d∣H); general: finite affine annihilator LCM diagonal
new-divisor counting identity supportCoeff_mul_eq_add_defect ∀a∈A(a>0)∀x>0 divisor counts
bound on zero blocks in divisor counts supportCoeffZeroWindow_length_le_eps_logb_add ∀ε>0∃B, given rationality of the base-2 support series divisor counts
equivalent shift condition for the target 1/2 (Problem 257) shiftWindowZero_iff_half_mem_mersenneAchievementSet iff, no extra parameter Mersenne subsums
totient multiplication with common prime factors totient_mul_eq_overlapFactor_mul ∀x>0 (j fixed) totient identities
totient relative-Euler-product split totient_mul_eq_totient_mul_relativeEulerProduct ∀jx>0 totient identities
formula for φ(2H(t)+j)−φ(H(t)+j) lcmRayArithmeticLetter_eq_deltaTotient ∀t∀j Möbius sums
the finite diagonal sum equals D(H(t),H(t),L) lcmDiagonalArithmeticWord_eq_windowDiscrepancy ∀t∀L Möbius sums
#249 ⟺ cofinal actual-orbit nonintegrality irrational_totientSeries_iff_actualLcmOrbitNonintegralitySupply iff Möbius sums
the LCM tail difference is an integer translate of a multiple of S actualLcmTailOrbit_eq_scaled_totientSeries_sub_prefix ∀a Möbius sums
explicit finite-block approximation abs_actualLcmTailOrbit_sub_rawApprox_lt ∀a∀q Möbius sums
an equivalent residue-interval condition exists_certifiedKill_iff_guardCylinderWitness ∀hN, iff integer recurrences
sign of the three-term totient difference under the stated middle-value bound fixedRankSecondDifference ∀Hj, given MiddleRankTotientExtremal totient differences
exact value on a specified dyadic family fixedRankSecondDifference_dyadic_fixture_exact ∀n (closed-form) totient differences
the rational parity comparison (definition) parityCoboundaryWeight ∀n (definition) binary-weighted series
RN as a sum over positive divisor indices totientTail_eq_tsum_positiveForeignResidueKernel ∀N divisor and cyclotomic sums
RN as a shifted Möbius sum totientTail_eq_tsum_shiftedMobiusPulse ∀N divisor and cyclotomic sums
absolute convergence and regrouping of Lambert sums tsum_lambert_pair_regroup_if ∀wv:N→R(|w(d)|≤d,|v(m)|≤m) divisor and cyclotomic sums
carry divisibility and integral tail differences carryShift_dvd_iff_tailDiff_mem_int ∀Nk, given the carry recurrence and limit integer recurrences
rationality ⇒ periodicity & unbounded rank not_irrational_totientSeries_implies_mod_period_and_unbounded_rank ¬IrrationalS→∃vu∀e integer recurrences
the stated reconstruction cost is at least 3 three_le_totientAdjugateTailCost ∀ finite w:ι→Qx:ι→N, given ∑wiφ(xi)=1 weighted tail bounds
directed residue interval (definition) directedCertifiedKill h,N,L≥0; soundness and the existential-depth equivalence are proved in Proposition 6.69 integer recurrences
rationality and an integer carry with vanishing boundary term binaryCoeffSeries_rational_iff_exists_temperedBinaryOrbit ∀c:N→N(c(n)≤n), iff binary-weighted series
doubling-orbit rigidity doublingOrbit_eq_zero_of_tempered ∀d:N→R, doubling and d(n)/2n→0 imply d≡0 integer recurrences
number of labels needed for exact tail recovery balancedPulse_no_autonomous_decoder ∀m∀State, collapsing-state → no decoder binary-weighted series
endpoint residues do not determine the initial carry affineBinaryOrbit_mod_twoPow_eq ∀a:N→Z∀u0v0∀L binary-weighted series
integer-coefficient block-certificate dichotomy int_coeff_series_irrational_or_bpow_mul_eq_intCast_of_full_block_certificates ∀b≥2∀c∀q, given full-block certificate at q binary-weighted series
divisibility under iterated forward differences dvd_iteratedForwardDifference d∣f(n) for every n; every finite list of shifts integer sequences
totient series as a coprime-pair sum tsum_coprime_pair_pow_eq_tsum_totient_mul_pow ∀n; ∀r∈[0,1) geometric weights
factorisation through a rational-valued linear functional linearDescender_eq_smul_eval rational vector spaces V,W; surjective linear ev:V→Q, linear L:V→W, ker⁡(ev)⊆ker⁡L divisibility and linear algebra
divisibility of a reduced denominator divisor_dvd_divInt_den ∀D>0∀m∣D∀a divisibility and linear algebra
the complementary denominator divides the integer multiplier scalarLocalization_complement_dvd ∀x:Q∀H(H∣x.den)∀c((cx).den∣H) divisibility and linear algebra
largest-exponent nonvanishing criterion scaled_dyadic_sum_ne_zero ∃m∈s (strictly maximal, odd coeff), ∀i≠me(i)<e(m) divisibility and linear algebra
column rescaling and determinant nonvanishing det_residualMonomialMatrix ∀d∀e:Find→N∀z:Find→C (z nonzero) complex matrices
determinant 1 for the divisor-incidence matrix incidenceMobius_det_eq_one ∀N:N Möbius sums
error bound from convergence controlledForeignProjection_of_tail_limit 2H≤D; convergence is supplied for H>0 in Section 6.6.8 Möbius sums
carries realising fixed-precision unit residues fixedPrecisionTropicalNoGo ∀u>0 (fixed precision), ∀ finite carry word divisibility and linear algebra
vanishing of iterated prime dilation differences iteratedDilationDifference_eq_zero_of_cardFactors_le distinct primes p1,…,pr, r>K, gcd(n,∏pi)=1; Ω(a)≤K on A divisor counts
φ(n)>34n for the stated rough integers three_quarters_mul_lt_totient_of_rough_lt_two_pow ∀a≥8∀n>0 (n is t-rough, t=2a, n<22⋅2a) Möbius sums
positivity of short-window totient differences lcmRayArithmeticLetter_pos_of_lt_two_mul ∀a≥8∀j(0<j<2⋅2a) Möbius sums
the negative representative of the true endpoint actualLcm_trueEndpointSurvivor_neg ∀a≥8∀JK (size bound) Möbius sums
integrality forces upper-endpoint residue actualLcm_integral_forces_topEdgeResidue ∀a≥8∀JK (size condition, 2H+J+K+2<2K) Möbius sums
terminal dyadic staircase impossible under the size condition not_actualLcmTerminalDyadicStaircase_of_room ∀a≥8∀JKm (size condition, modulus-wide) Möbius sums
penultimate difference in a partial divisibility pattern puncturedDyadicStaircase_penultimate_eq_half ∀a≥8∀JKm (hypotheses of Proposition 6.86) Möbius sums
upper-endpoint residue-gap ⇒ nonintegral actualLcmTailDiff_notMem_int_of_topEdgeResidueGap ∀a≥8∀JKm (size condition, one-sided) Möbius sums
relation between the centred residue and the actual carry two_mul_actualOddHalfCenteredLift_eq_terminal_sub_trueCarry ∀a≥8∀q (size condition, fit bound 2(H+q+2)≤4q) Möbius sums
divisibility of the three-term totient difference two_mul_totient_dvd_fixedRankSecondDifference ∀a≥4∀j>0(j2≤2a) totient differences
equality of two finite rational approximations renormalizedResidueAgreement_of_twice_le H>0, D≥2H; no separation hypothesis for this equality Möbius sums
a Mersenne divisor need not divide a tail difference idCoeff_mersenneModulus_does_not_annihilate_tailShift ∀K,q(K>0,K<q,q∣2K−1) Möbius sums
finite numerator separation implies nonintegrality scaleFullTarget_miss_of_lambert_projected_num_gap ∀H,D fixed, given uniform numerator-gap bound Möbius sums
periodic-weight Lambert dichotomy irrational_or_bpow_mul_eq_intCast_intWeightedErdosSeries_periodic ∀b≥2∀m>0∀w (w(n+m)=w(n)) Möbius sums
periodic weight with nonnegative divisor coefficients irrational_intWeightedErdosSeries_periodic_of_coeff_nonneg_of_frequently_ne_zero b≥2, m>0, w(n+m)=w(n); cw(n)≥0 and cw(n)≠0 arbitrarily far out Möbius sums
zeros on one progression for the character modulo four intWeightedCoeff_periodFourSignWeight_eq_zero_of_mod_four_eq_three cχ4(n)=0 for every n≡3(mod4); the weight is fixed Möbius sums
positivity of the translated LCM tail difference actualLcmTailDiff_shift_pos ∀a≥8∀J(J+(a+6)<2⋅2a), no rationality hypothesis Möbius sums
Farey window K=240 denominator bound gap_check_window_1_240_le_79639646646701375323355774875831053

∀q:N,0<q≤
79,639,646,646,701,
375,323,355,774,875,831,053

rational intervals
denominator lower bound S≠p/q, q≤Q0 tsum_totient_div_pow_two_ne_ratCast_of_den_le_79639646646701375323355774875831053 ∀p:Q,p.den≤Q0 rational intervals
finite LCM certificates covering thresholds a0≤6 powerTwoActualLcmShortArithmeticKillSupply_through_six ∀a0≤6∃aL Möbius sums
contiguous diagonal certificate band through t≤82 ErdosProblems.Skip.LadderT67.exists_diagonalKill_le_82 ∀t≤82∃L; no t=83 certificate claimed LCM diagonal
residue certificates for 1≤h≤8, N=12, L=16 certifiedKill_all_small ∀h∈[1,8], C h 12 16 (direct evaluation) finite binary sums
denominator exclusions for 1≤h≤16, binary exponent 14 totient_series_ne_rat_of_den_dvd_pow_two_mul_mersenne_upto_sixteen ∀h∈[1,16], denominator exponent 14 (direct evaluation) finite binary sums
the second-difference test uses more depth at (h,N)=(1,8) totient_tail_rank_two_kill_sound_but_not_shallower_cell first-difference certificate at L=8; no second-difference certificate for 1≤L≤8, and one at L=9 finite binary sums
listed diagonal certificates with t≤64 certifiedKill_diagonal_all_imported t∈{1,2,3,4,5,7,8,9,11,13,16,17,…} (28 explicit values) LCM diagonal
LCM certificates at a=4 and a=6 lcmDiagonalArithmeticKill_two_pow_four a=4, a=6 (concrete, formally checked) Möbius sums
square-CRT coprime-factor examples: zero / nonzero differences squareCRT_clean_block_can_vanish witness n=52 (vanishes); witness n=27 (nonzero) totient factorisation

certificate criterion (proved implication; unproved premise) irrational_totient_series_of_certificate_supply ∀h≥1∀N0≥0∃N≥N0∃L, ChNL finite binary sums
period-multiple criterion (equivalent; unproved condition) irrational_totient_series_of_multiple_certificate_supply ∀h0>0∀N0∃m>0∃N≥N0∃L multiples of periods
LCM-diagonal criterion (unproved condition) irrational_totient_series_of_lcm_diagonal_certificate_supply ∀t0∃t≥t0∃L LCM diagonal
LCM-grid certificate criterion (unproved premise) irrational_totient_series_of_lcm_cone_window_kill_supply ∀t0∃t≥t0∃q>0∃m∃L LCM grids
LCM nonintegrality criterion (unproved condition) irrational_totient_series_of_lcm_diagonal_nonintegrality_supply ∀t0∃t≥t0, R(2Ht)−R(Ht)∉Z real tail differences
finite-grid residue criterion (unproved premise) irrational_totient_series_of_lcm_cone_nonflat_supply for every t0, data t≥t0,L,Q satisfying Theorem 6.57 finite LCM grids
first-harmonic norm criterion (unproved premise) irrational_totient_series_of_first_harmonic_norm_gap ∀h>0∀X0∃X≥max(X0,1)∃L, size condition 16(2X+h+L+2)≤2L binary-weighted series
four-tail residue criterion (unproved premise) irrational_totient_series_of_primeJumpSharpKill_supply ∀t0∃t≥t0∃p,L>0 binary-weighted series
short LCM-window criterion (unproved premise) PowerTwoActualLcmShortArithmeticKillSupply ∀a0∃aL, a0≤a∧L<2⋅2a Möbius sums
LCM-tail separation criterion (unproved premise) irrational_totientSeries_of_actualLcmOrbitSeparationSupply cofinal 1/32+error separation at prescribed odd depths Möbius sums
five endpoint criteria (unproved premises) irrational_totientSeries_of_flexibleActualTerminalDominanceSupply one of the five quantified conditions in Proposition 9.14 Möbius sums
proved rational counterexample to bound, parity and aperiodicity exists_totientParity_arbitrarilyManySeparatedCarry_rational_countermodel ∃c:N→N, c≤6, c≤n, c≡φ(mod2), arbitrarily-separated blocks binary-weighted series
proved mod-4 prime construction (CRT and Dirichlet) exists_prime_deltaTotient_four_mul_mod_four_two ∀h>0∀B∃primep>B Möbius sums
sufficient mod-4 condition (unproved premise) irrational_totientSeries_of_cofinal_modFourPulseSurvivorKill the displayed mod-four residue condition at arbitrarily large scales Möbius sums
proved two-adic prime construction exists_prime_totient_twoAdic_pulse_divisors ∀K≥2∀H>K∀B∃cofinally many primesp>B divisibility and linear algebra
proved implication from eventual integrality eventual_integral_tailDiff_has_cofinal_twoAdic_half_pulse ∀K≥2, given the stated two-adic prime construction and eventual integrality divisibility and linear algebra

About this paper

Authorship and AI use. Will Cook built and directed the research infrastructure and maintains the public release. He reviewed claims when he could. AI agents did most of the research and drafting. Cook did not independently verify every claim.

Cite and contact. Cite this paper by its title, author and date above, with its PDF; cite the earlier sources it uses for a mathematical result. For software, use the release citation and give the commit used. Contact Will with questions or corrections.

Prefer the manuscript?

Equations typeset from the exact TeX. Rendered from the LaTeX manuscript source with file fingerprint sha256:2be13b340f9dafc4. Its file fingerprint matched the published source record, so the PDF above, the LaTeX, and this page are one manuscript. Redeploying the site regenerates this page from whatever the public repository holds at that moment.