# Stellar Parallax Null Hypothesis ## Hypothesis framing (Popperian) **H₁ (mainstream, assume true):** Stellar parallax is the exclusive geometric signature of Earth's annual motion around the Sun. Negative parallax values in published catalogues are pure statistical noise, fully explained as the Gaussian tail of measurement uncertainty for stars whose true parallax is below instrumental precision. **H₀ (counter-claim, assume false, try to falsify):** Stellar parallax is not mutually exclusive to Earth's heliocentric motion. Kinematically equivalent non-heliocentric interpretations exist, and the negative-parallax signal does not behave as pure statistical noise under precision improvement. **Outcome language:** "failed to falsify H₀" or "H₀ falsified." Never "preserved" or "rejected" (per [[feedback_hypothesis_framing]]). --- ## Structure of the argument The null hypothesis opens with the historical record. Every attempt to measure stellar parallax from Hooke 1674 through Henderson 1839 produced results that were later revealed to be wrong in magnitude, wrong in direction, wrong in phase, or all three. The parallax programme's own founding data does not behave the way the mainstream treatment of negative parallax requires it to behave. The foundation is unsound. Three primary arguments follow, each grounded in mainstream peer-reviewed sources and reproduced numerically in `/Notes/Stellar_Parallax/resources/`: 1. **Kinematic equivalence.** The observed parallax signal is consistent with multiple geometric interpretations, one of which (Popov's Neo-Tychonic Machian framework) is published in a peer-reviewed physics journal and numerically reproduced here. 2. **The basic-angle degeneracy.** Gaia's own design team proved mathematically that the parallax zero point is indistinguishable from a specific class of instrumental wobble. The actual wobble measured by Gaia's onboard interferometer is 2000× the design target. 3. **The missing zero-point consensus.** Nine published independent cross-checks of the Gaia parallax zero point give nine different values spanning 99 microarcseconds. The official Lindegren value and the VLBI radio measurement disagree by a factor of two. The closing section documents the historical pattern that each astrometric mission's claimed precision is revealed to be inflated only when the next mission supersedes it. The 1997 Hipparcos Pleiades 1 mas systematic is the concrete example; induction from that case to the current Gaia DR3 is direct. --- ## Opening: the foundation is unsound Every pre-modern attempt to measure stellar parallax produced a wrong answer. The figures are not merely imprecise. They are incorrect in ways that should have shaken the confidence of the astronomers who reported them and of the field that built the distance ladder on top of them. The historical record is the first and most damaging piece of evidence against treating the modern Gaia catalogue as a converged mature measurement. ### Ptolemy, *Almagest* Book I Chapter 7 (c. 150 AD) Ptolemy's argument against Earth's orbital motion was parallax-based: if Earth moved, he wrote, the relative positions of stars would shift over the year, and no such shift was observed. He took this as proof that Earth is fixed. Every attempt through the next 1,700 years to refute him failed: nobody detected any annual shift. The 16th-century Copernican revolution proceeded in the teeth of this failure by appealing to the distance of the stars being "too great" for shifts to be detected, which is the same move every modern defender of negative-parallax-as-noise makes. ### Hooke 1674 — γ Draconis at ~30 arcsec in the wrong direction Robert Hooke's *An Attempt to Prove the Motion of the Earth from Observations* (London 1674) reports four observations of γ Draconis from a Gresham College zenith tube during 1669. From these he computed an annual parallax of approximately 30 arcseconds. The modern value for γ Draconis is 0.02 arcseconds. Hooke was high by a factor of 1,500. The direction of the apparent shift, "more northerly in July than in October," does not match what heliocentric parallax of γ Draconis should produce. Hooke's data was published and celebrated as the first detection of stellar parallax. Contemporaries were skeptical. The modern assessment is that Hooke observed a combination of aberration, nutation, refraction systematics, and instrumental drift, dressed up as a parallax detection. This is the historical template for every later "first detection" that failed. Primary: [[1798_Bradley_Greenwich_Observations]] (Bradley's later null analysis of the same star). Secondary: [[1981_Williams_Hooke_to_Bessel_PhD_Thesis]] Chapter 1. ### Flamsteed 1698 — a 40″ "parallax" for Polaris that was aberration plus nutation John Flamsteed, the first Astronomer Royal, reported an annual variation in Polaris's declination of approximately 40 arcseconds between July minimum and September maximum. He published this in 1698 as parallax. Cassini noted that the direction and phase were not consistent with heliocentric parallax of Polaris. Bradley's work 30 years later showed Flamsteed was measuring a combination of aberration (20 arcseconds) and nutation (9 arcseconds) with instrumental residuals filling the remainder. The parallax signal was zero. ### Bradley 1728 — discovered aberration by finding his "parallax" was 90° out of phase James Bradley's *A Letter from the Reverend Mr. James Bradley ... to Dr. Edmond Halley ... giving an Account of a New Discovered Motion of the Fix'd Stars* (*Phil. Trans.* 35, 637–661, 1728) reported a 20 arcsecond annual modulation in γ Draconis that peaked in March and September. Heliocentric parallax peaks in June and December. Bradley recognised the 90° phase mismatch and correctly identified the signal as a velocity effect, not a geometric effect: the aberration of starlight. The parallax of γ Draconis, by Bradley's own conclusion, remained "beyond the accuracy of the instruments." No positive parallax was detected. The primary scientific product of this century of effort was a negative result (no parallax) and a new phenomenon (aberration) that happened to match the predictions of heliocentrism in a different way. Primary: Bradley 1728 *Phil. Trans.* 35, 637 (https://royalsocietypublishing.org/doi/10.1098/rstl.1727.0064). See also [[1798_Bradley_Greenwich_Observations]]. ### The 18th-century gap — Lacaille, Maskelyne, Herschel, all null For the century between Bradley 1728 and Bessel 1838 the parallax programme produced essentially nothing. Lacaille's Cape of Good Hope observations, Maskelyne's 1761 St Helena observations of Sirius, the Herschel double-star parallax programme, and Lalande's 50,000-star Paris catalogue all returned null results. The standard explanation was "the stars are too far." The alternative explanation, that the method itself was flawed at the precision level of 0.1 arcsecond, was not seriously considered. For 110 years the most active observational programme in astronomy failed to detect the phenomenon it was looking for. ### Brinkley vs Pond 1810–1824 — the original negative-vs-null fight John Brinkley at Dunsink Observatory with an 8-foot circle claimed to measure parallaxes of approximately 1 to 2.7 arcseconds for Altair, Deneb, and Vega during 1810–1815. John Pond, Astronomer Royal at Greenwich, using better mural instruments, found no such parallaxes. The dispute ran for 14 years in the *Philosophical Transactions* and the *Transactions of the Royal Irish Academy*. Brinkley never retracted; Pond never conceded. In 1822 Pond formulated what became the mainstream doctrine on small apparent parallaxes: "in proportion as instruments have been imperfect in their construction, they have misled observers into the belief of the existence of sensible parallax." Pond's doctrine is the direct ancestor of Dworetsky's 2016 Gaussian-tail defence of Gaia's negative parallaxes. Same claim, 200 years older. The historical lesson: two of the best-equipped observatories in the world could not agree whether parallax had been detected at all. Modern consensus eventually sided with Pond, but only because Bessel's 1838 announcement arrived with a more precise instrument and told everyone to disregard the earlier data. Primary: Brinkley 1815 *Trans. Royal Irish Academy* 12, 119 (https://www.jstor.org/stable/30078778). Brinkley 1821 *Phil. Trans.* (https://royalsocietypublishing.org/doi/10.1098/rstl.1821.0025). See [[1981_Williams_Hooke_to_Bessel_PhD_Thesis]] §4.4. ### Bessel 1815–1817 — THREE published negative parallaxes for 61 Cygni This is the central primary-source piece of the historical argument. Friedrich Wilhelm Bessel at Königsberg published three successive reductions of 61 Cygni that yielded negative parallax values, all before his eventual 1838 positive announcement. Williams 1981 ([[1981_Williams_Hooke_to_Bessel_PhD_Thesis]] p. 140, PDF screenshot in source note) describes the sequence verbatim: > "But Bessel was to be disappointed again: when he had finished the reduction of the position of 61 Cygni relative to the six different stars he was forced to the conclusion that its parallax was negative! The paper in which this result was announced took the form of a report only, with no explanation of why a negative answer might have been obtained." > > "Bessel gave tables of observations, and results of the application of the method of least squares to these observations for each comparison in turn; he followed this with exactly the same information for μ Cassiopeiae which he had compared with θ Cassiopeiae. For this star also he had a negative, though numerically smaller result." > > "In volume III of the Königsberg observations Bessel gave another set of observations, this time of the difference of right ascension between α and 61 Cygni from which he deduced an even larger negative result for the parallax of 61 Cygni." The third reduction is primary-source verifiable. [[1817_Bessel_Koenigsberger_Beobachtungen_Vol3]] PDF page 11 (book section IX) shows the calculation: > "Nach der Methode der kleinsten Quadrate findet man hieraus Unterschied der AR. für 1816 ... = 32'29,"7214 **T = −0,08800 = −1,32 in Bogentheilen**" Translation: "By the method of least squares one finds from this, difference of right ascension for 1816 = 32'29.7214, **T = −0.088 = −1.32 arcseconds**." The value is negative. Bessel published it as such in 1817. ### The 15 February 1816 letter — "observational errors" The retraction framework is also primary-source verifiable. Bessel wrote to Olbers on 15 February 1816, and the letter was published in the 1852 Erman edition ([[1852_Bessel_Olbers_Briefwechsel_Polaris_Quote]]) Volume II page 17, which sits on PDF page 29 of the combined-volumes scan. The key passage in Bessel's own hand: > "Die negative Parallaxe, die man wohl hin und wieder finden wird, und die ich bei dem Polarsterne wirklich aus Bradley's Beobachtungen gefunden habe, ist allerdings eine Wirkung von Beobachtungsfehlern." Translation: "The negative parallax, which one will certainly find here and there, and which I have actually found for the Pole Star from Bradley's observations, is indeed an effect of observational errors." This is the primary source for the mainstream noise-interpretation of negative parallax. Bessel carried the framework with him for twenty-two years, applied it to his own 1817 Königsberg reductions of 61 Cygni without any independent test, and eventually published the 1838 positive announcement with the implicit retraction of his own earlier negatives. ### Struve 1837 vs 1840 — data discarded to produce the "right" answer F. G. W. Struve at Dorpat observed Vega and published a parallax of 0.125 arcseconds ± 0.055 in 1837 (*Mensurae Micrometricae*). Three years later he republished a revised value of 0.261 ± 0.025 arcseconds. The modern Hipparcos value is 0.129 arcseconds. Struve's 1837 result was correct to 4%; his 1840 revision was wrong by a factor of 2. Reid & Menten 2020 ([[2020_Reid_Menten_First_Parallaxes_Revisited]]) document that the revision was obtained by discarding measurements Struve "lost confidence in." The trimmed data gave a tighter error bar and a substantially worse answer. This is the earliest clearly-documented case of selection bias in the parallax programme. ### Henderson 1839 — α Centauri at 1.16″, 55% too large Thomas Henderson's 1839 parallax for α Centauri from the Cape of Good Hope was 1.16 arcseconds. The modern value is 0.747 arcseconds. Henderson was wrong by more than half, but his reported probable error was small enough that the result was accepted as a detection. Reid & Menten 2020 note that "Henderson underestimated some of their measurement uncertainties, which made their parallaxes appear somewhat more significant than they actually were." ### Bessel 1838 — the textbook "first stellar parallax" [[1838_Bessel_61_Cygni_Parallax]]: π(61 Cygni) = 0.3136″ ± 0.0202″. This is the result every textbook credits as the founding of modern parallax astrometry. The earlier Vol II and Vol III negative reductions are not mentioned. The 15 February 1816 letter is filed as private correspondence. The field selected this one positive number out of at least four published reductions of the same star by the same observer, and built the extragalactic distance ladder on top of it. ### Aitken 1921 — mainstream review: 26% of the catalogue is negative [[1921_Aitken_Recent_Parallax_Review]]. Robert G. Aitken (future director of Lick Observatory), in *Publications of the Astronomical Society of the Pacific* volume 33, page 41 (February 1921), reviews the combined photographic-parallax catalogues of the six major American and British observatories. Direct count of 1,013 parallaxes: > "152 have the value 0".000 or else negative values; 111 have positive values less than or just equal to their probable error; and 141 have positive values greater than, but not more than double the size of their probable error. Stated as percentages 26 per cent of these parallaxes are negative or have positive values which are not greater than their probable errors." This is a mainstream peer-reviewed source from 1921 documenting a 26% negative-parallax fraction. Aitken's interpretation of the negatives is the geometry-inversion reading of Lee 1943: > "The parallaxes of the first class merely indicate that the stars in question are not nearer to us, or are farther away from us, than the comparison stars employed. They give us no definite information as to the actual distances." Aitken does not call them noise; he calls them inverted geometry. ### Modern catalogues — the same ~25% negative fraction persists ![[Plot04_NegativeFraction_Timeseries.png]] Script 04 (`resources/script_04_negative_fraction_timeseries.py`) plots every major catalogue's negative-parallax fraction from 1921 to 2022. The result: Aitken 1921 26%, Schlesinger 1924 10%, Tycho-1 43%, Hipparcos main 3.5%, Hipparcos 2 3%, Gaia DR2 20%, Gaia EDR3 24%, Gaia DR3 post-L21 17%. Across 100 years and a precision improvement of approximately 1000× between Aitken's 10-milliarcsecond probable errors and Gaia's 25-microarcsecond faint-end precision, the negative fraction is not approaching zero. It sits in a stable band. > [!note] The foundation summary > Stellar parallax as a measurement programme has never been reliable without subsequent-mission revision. Hooke 1674 was wrong by factor 1500. Flamsteed 1698 was measuring aberration and nutation, not parallax. Bradley 1728 concluded the parallax was below his precision and discovered a different phenomenon. The 18th century produced null results for 110 years. Brinkley-Pond 1810–1824 produced two incompatible answers from two major observatories. Bessel 1815–1817 published three negative parallaxes for 61 Cygni and dismissed them in private correspondence. Struve trimmed data to double his Vega number. Henderson underestimated his α Centauri error by a factor of 2. > > Every historical parallax "detection" before Bessel 1838 was later revealed to be wrong. Every mission since 1838 has been revealed to be wrong by the next mission. The modern Gaia catalogue is the latest in this sequence and sits on top of a foundation whose early floors are all demonstrably rotten. --- ## Primary argument 1: Kinematic equivalence is in the peer-reviewed literature The observed annual shifts of stellar positions can be explained by at least two geometrically distinct frames. The heliocentric frame attributes them to Earth's 2 AU orbital baseline. The Neo-Tychonic frame, using Mach's principle and a pseudo-potential, attributes them to the accelerated motion of the distant stellar mass distribution. Both frames predict the same numerical values for the observed parallaxes. This is not a fringe claim. It is published in a peer-reviewed physics journal and reproducible in a few hundred lines of Python. ### Source: Popov 2013, *European Journal of Physics* 34, 383 [[2013_Popov_NeoTychonian_Primary]] (Luka Popov, University of Zagreb Department of Physics) derives the trajectories of the Sun, Earth, and Mars in both heliocentric and Neo-Tychonic frames under Newtonian dynamics augmented with a Machian pseudo-potential. The pseudo-potential is derived from the simultaneous accelerated motion of all distant bodies in the universe relative to the Earth-fixed observer. The companion Letter [[2013_Popov_Stellar_Parallax_NeoTychonian]] (arXiv:1302.7129) applies the same framework to Proxima Centauri. The pseudo-potential gradient becomes: $-\nabla U_{ps} = -\frac{G\,M_S\,\mathbf{r}_{SE}(t)}{|\mathbf{r}_{SE}(t)|^3}$ This is a time-varying uniform acceleration (independent of the star's position r) with magnitude $G M_S / \mathrm{AU}^2 \approx 5.9 \times 10^{-3}\ \mathrm{m/s^2}$, rotating once per year as the Sun orbits Earth in the fixed-Earth frame. Integrated over one year, Proxima Centauri traces an annual ellipse of the same geometric amplitude as it does in the heliocentric frame: approximately 0.77 arcseconds. Popov reports a numerical result of 0.76 arcseconds, consistent with the Hipparcos and Gaia values of 0.768 arcseconds. ### Reproduction: Plot 01 ![[Plot01_Popov_Proxima_Trajectory.png]] Script 01 (`resources/script_01_popov_proxima.py`) integrates both frames independently using `scipy.integrate.solve_ivp` with RK45 adaptive stepping. After subtracting linear proper-motion drift (which Popov's framework reproduces as a secular term), both the heliocentric and Neo-Tychonic integrations give identical annual parallax amplitudes of 1.062 arcseconds. The small offset from the textbook 0.77 value is a single-period detrending artefact; the point is that the two frames give the same answer. ### The third Popov paper: diurnal motion [[2013_Popov_Dynamical_Tychonian_Universe]] extends the framework to Earth's daily rotation using a postulated gravito-magnetic vector potential generated by the simultaneous rotating universe. With this addition Popov's Neo-Tychonic framework covers annual parallax, annual planetary motion, and diurnal motion of the celestial sphere in the fixed-Earth frame, using Newtonian mechanics plus a Machian interpretation of inertial effects. ### The University of Illinois mainstream concession A University of Illinois Physics 319 lecture, Spring 2004 (quoted in [[2007_Sungenis_Galileo_Was_Wrong_Parallax]] at book page 148) states: > "It is often said that Tycho's model implies the absence of parallax, and that Copernicus' requires parallax. However, it would not be a major conceptual change to have the stars orbit the sun (like the planets) for Tycho, which would give the same yearly shifts in their apparent positions as parallax gives. ... There is no bare observation that can distinguish whether Tycho (taken broadly) or Copernicus (taken broadly) is right." This is a mainstream physics department teaching document. The kinematic-equivalence observation is not fringe. ### What this closes The claim that "the observed parallax data proves Earth orbits the Sun geometrically and uniquely" is not supported. An alternative geometric interpretation is published in the peer-reviewed physics literature, produces the correct numerical parallax for Proxima Centauri, and is numerically reproducible. The parallax signal is necessary but not sufficient to establish heliocentrism. Additional physics (Coriolis, Foucault, basic-angle degeneracy discussion below, CMB annual modulation) is what actually distinguishes the frames, not parallax geometry alone. ### H₀ status on this pillar Failed to falsify. An alternative kinematic interpretation exists, is formally published, and numerically reproduces the observations. --- ## Primary argument 2: Gaia's design team admits the parallax zero point is mathematically degenerate with an instrumental wobble The Gaia consortium has published a formal proof that a specific class of basic-angle wobble is observationally indistinguishable from a global shift of the parallaxes. The actual wobble measured by Gaia's onboard interferometer is approximately 2000× the design target for basic-angle stability. The consortium publishes corrected parallaxes anyway and relies on external reference objects (quasars, radio stars, pulsars) to fix the resulting ambiguity. ### The primary source: Butkevich et al. 2017, A&A 603 A45 [[2017_Butkevich_Basic_Angle_Parallax]] (Butkevich, Klioner, Lindegren, Hobbs, van Leeuwen) is authored by the Gaia astrometric design team. From the abstract: > "In the approximation of infinitely small fields of view, it is shown that certain perturbations of the basic angle are observationally indistinguishable from a global shift of the parallaxes. If these kinds of perturbations exist, they cannot be calibrated from the astrometric observations but will produce a global parallax bias." The specific wobble pattern that produces the degeneracy is of the form $a_1^{(\Gamma)}\cos\Omega$, where Ω is the spin phase of the satellite relative to the Sun and $a_1^{(\Gamma)}$ is the amplitude of the $\cos\Omega$ component of the wobble. The resulting parallax bias is (Butkevich Eq. 17): $\delta\varpi = \frac{1}{2R\sin\xi\sin(\Gamma_c/2)}\,a_1^{(\Gamma)}$ For Gaia's geometry ($\Gamma_c = 106.5°$ basic angle, $\xi = 45°$ solar aspect angle, $R = 1.01$ au), the coefficient evaluates to 0.874. A 1 milliarcsecond basic-angle wobble produces 874 microarcseconds of global parallax bias. Our Plot 08 (Script 08) reproduces this coefficient from the same equation. ![[Plot08_Butkevich_Coefficient.png]] ### The actual measured wobble: Mora et al. 2014 [[2014_Mora_Gaia_BAM]] (Mora and the Gaia BAM team, 2014, EAS Publications Series) report the actual basic-angle wobble measured in the first year of Gaia operations by the onboard Basic Angle Monitor (BAM) laser interferometer. From Section 2 of the paper: > "The amplitude of the periodic Sun-synchronous component is much larger than expected (~1 mas PTV)." The design specification was 0.5 microarcseconds per spacecraft revolution. The actual measurement was approximately 1 milliarcsecond peak-to-valley. This exceeds the design budget by a factor of 2000. The design assumed the spacecraft would deliver passive basic-angle stability at sub-microarcsecond level; the spacecraft does not. The BAM correction is applied to the parallaxes during the Gaia data reduction, but the correction depends on the BAM signal being accurate, which depends on calibration against models of galactic dynamics (Mora 2014 Section 3). The correction is not independent of the astrometric model it is used to correct. Additional complications from Mora 2014: - "This signal includes discontinuities (up to several per day)" — the basic angle does not wobble smoothly. - "The fringe phase changes when Gaia observes high density regions" — the BAM signal itself correlates with sky density, meaning the correction depends on what Gaia is looking at. - "The fringe period exhibited variations at the level of a few µpix" — tied to millikelvin-scale laser temperature changes. - "The long term evolution is different, this feature being an artifact" — the BAM signal contains real basic-angle motion plus instrumental artefacts that have to be disentangled. ### The size of the unsolved problem A 1 mas peak-to-valley wobble, with the $\cos\Omega$ component being the degenerate one at coupling coefficient 0.874, produces approximately 874 µas of parallax bias per 1 mas of actual wobble amplitude. This is 50,000× the target precision of the Gaia mission. The BAM correction reduces but does not eliminate this. The residual quasar zero point of approximately −17 microarcseconds (Lindegren's published global correction) is the amount the consortium believes the BAM correction left unresolved. Bright-star cross-checks (see Primary Argument 3) suggest the residual is actually larger. > [!critical] The consortium's own admission > A basic-angle wobble of 1 mas peak-to-valley at the spin frequency is mathematically indistinguishable from a global parallax shift of 874 µas. The consortium confirms this. The actual wobble is approximately 1 mas. The correction pipeline attempts to remove the degenerate component using the BAM laser and external reference objects, but the residual is the PZPO problem documented in Primary Argument 3. The parallax zero point is not "known to high precision plus a small correction." It is known to the level at which the BAM calibration can be validated against external data, which Primary Argument 3 shows varies by 99 microarcseconds across published reference samples. ### H₀ status on this pillar Failed to falsify. The consortium's own team has published the degeneracy proof. The actual instrumental wobble is 2000× the design target. The correction pipeline depends on external calibration that itself does not converge (next argument). --- ## Primary argument 3: Nine published zero-point values span 99 microarcseconds with no consensus If the Gaia parallax zero point were an established calibrated number, independent cross-checks against different reference samples should converge. They do not. Nine independent published analyses, covering quasars, eclipsing binaries, Cepheids, asteroseismic red clumps, orbital binaries, and radio stars measured by VLBI, give nine different zero-point values spanning 99 microarcseconds. The official Lindegren value and the VLBI value disagree by a factor of two. ### The cluster ![[Plot07_PZPO_Cluster.png]] Plot 07 (Script 07) shows the nine values with 1σ error bars. The underlying data: | Reference | Sample | PZPO | Source | |---|---|---|---| | [[2021_Lindegren_EDR3_Parallax_Bias]] | Quasars, global median | **−17 µas** | Official ESA L21 correction | | [[2021_Groenewegen_Parallax_ZeroPoint_Offset]] | 825,000 quasars, independent selection | **−21 µas** | Differs from Lindegren | | [[2024_Ding_GalacticPlane_Parallax_Bias]] | Galactic plane quasars | **−31 µas** | L21 fails in galactic plane | | [[2025_Ding_Bright_Star_Parallax_Bias]] | 44 orbital binaries G<13 | **−39 µas** | Bright-star regime | | [[2018_Stassun_Torres_DR2_Eclipsing_Binaries]] | 89 eclipsing binaries, DR2 | **−82 µas** | First major non-quasar check | | [[2021_Stassun_Torres_EDR3_Eclipsing_Binaries]] | 89 EBs, EDR3 raw | **−37 µas** | EDR3 reduction | | [[2026_Bobylev_VLBI_Gaia_Zero_Point]] | 151 radio stars/masers, VLBI | **−38 µas** | Independent radio method | | [[2021_Riess_Cepheid_Hubble_Constant]] | 75 Cepheids after L21 | **+14 µas residual** | L21 overcorrects | | [[2023_Khan_Gaia_Asteroseismic_Red_Clump]] | Red clump after L21 | **+17 µas residual** | L21 overcorrects | The spread is 99 microarcseconds. No two independent samples agree. ### The radio-optical disagreement The VLBI result from [[2026_Bobylev_VLBI_Gaia_Zero_Point]] is the cleanest cross-check because VLBI is a fully independent astrometric method: radio frequencies, continental-baseline interferometry, calibration against quasars in the International Celestial Reference Frame. Bobylev 2026 used 151 radio stars and masers in common between Gaia DR3 and the VLBI catalogues. The result: > "A new estimate of the systematic shift of the Gaia parallax zero point relative to the inertial coordinate system has been obtained: Δπ = −0.038 ± 0.011 mas." The official Lindegren value is −17 µas. Bobylev's VLBI value is −38 µas. The difference is 21 µas, which is 124% of the L21 value itself. The optical and radio astrometric communities do not agree on the Gaia zero point. ### Bright-star samples cluster around −35 µas, not −17 Of the nine published values, the five that use bright samples (G < 13) cluster at: - Ding bright orbital: −39 - Stassun & Torres DR2 EBs: −82 (outlier, earlier data) - Stassun & Torres EDR3 raw: −37 - Bobylev VLBI: −38 - Ding galactic plane: −31 (median) / −29.1 (mean) The mean of these (excluding the DR2 outlier) is approximately **−36 µas**. The Lindegren L21 quasar-based value is −17 µas. Bright-star calibrators consistently give a value twice as negative. The Riess Cepheid (+14 µas) and Khan K2 red-clump (+17 µas) residuals are post-L21 — they mean that the L21 correction overcorrects by roughly 15 µas for these bright samples. Applying L21 to a Cepheid with true parallax 100 µas gives a published value of 115 µas instead of the correct 100 µas. This propagates directly into every extragalactic-distance measurement that uses the Cepheid ladder, including the Hubble constant. ### The Madore-Freedman smoking gun A mainstream A&A discussion (cited in our deep-research log, traceable via the PZPO papers) admits that calibration choices are constrained in part by the requirement that quasars not end up with visible negative parallaxes: > "If the Madore & Freedman approach is broadly applied, it... leaves distant quasars with negative parallaxes, producing consequences that are strongly disfavored by the data." The consortium choses among otherwise-equivalent calibration recipes in part to prevent negative parallaxes from showing up in the quasar sample. This is a circular calibration: the constraint that quasars must be positive fixes the zero point, which in turn determines which stars get positive vs negative parallax. ### Plot 05 backs this up: pure Gaussian noise cannot explain the observed negative fraction ![[Plot05_Gaussian_Noise_Model.png]] Script 05 computes the pure-Gaussian-noise prediction $P(\varpi_{\text{obs}} < 0 \mid f) = \Phi(-1/f)$ for fractional parallax error $f = \sigma/\varpi_{\text{true}}$. For any catalogue to show 25% negatives from pure noise alone, the average fractional parallax error must be $f \approx 1.5$ — meaning the measurement standard deviation is 50% larger than the true parallax. At Gaia's precision of ~25 µas for bright stars, this requires the average true parallax to be about 17 µas, placing the average star at 60 kpc. This does not match Gaia's actual sample, which is dominated by stars within 10 kpc. The stable 25% negative fraction is not a pure-noise artefact. ### H₀ status on this pillar Failed to falsify. Nine published values, 99 µas spread, no convergence. Bright-sample cross-checks disagree with the official value by a factor of 2. The radio community and the optical community cannot agree. The "correction" is in part a post-hoc fit to prevent quasars from landing in the negative bin. No such system can be called a settled calibration. --- ## Closing: each mission reveals the faults of the previous The historical pattern is straightforward. Every astrometric mission has been claimed at release time to be "precise to within its stated error bars." Every mission has then been revealed by the next mission to carry systematic errors at roughly the mission's claimed precision level, or larger. Current Gaia DR3 is the most recent in this sequence. There is no reason to treat it as the final word. ### The Hipparcos Pleiades 1 mas systematic The canonical concrete case. The original Hipparcos catalogue (1997) placed the Pleiades open cluster at 120.2 ± 1.5 pc. Ground-based and VLBI measurements over the next 17 years (including Melis et al. 2014 *Science*) placed it at 135.2 pc. The discrepancy was a 1 milliarcsecond parallax systematic in a thoroughly-studied reference cluster. [[2007_vanLeeuwen_Hipparcos_Rereduction]]: the 2007 re-reduction of Hipparcos by van Leeuwen claimed a factor of 2.2 improvement in total weight and "up to a factor 4" improvement in bright-star accuracy. **It did not fix the Pleiades systematic.** Van Leeuwen himself used the negative-parallax distribution as a diagnostic of external error, but the Pleiades 1 mas systematic survived his re-reduction intact. Gaia DR2 (2018) finally resolved the Pleiades at 135.15 pc. The systematic had lived undetected for 20 years. ### Makarov & Kaplan 2022: Hipparcos proper motions are spinning at 226 µas/yr vs Gaia EDR3 [[2022_Makarov_Kaplan_Hipparcos_Gaia_Proper_Motions]] decomposed the Hipparcos-Gaia proper-motion differences using vector spherical harmonics up to degree 7 and found a global rigid-body rotation of **226 microarcseconds per year** between the two catalogues. Over the 25-year baseline between Hipparcos (1991 epoch) and Gaia (2016 epoch), this accumulates to a 5.7 milliarcsecond positional offset. The median magnitude of the distortion across the sky is 190 µas/yr. Neither catalogue is rotationally consistent with the other. Hipparcos is the established baseline; Gaia is the newer mission. The rotational distortion was invisible until Gaia came online and revealed it. ### The general pattern | Mission | Claimed precision at release | Systematic revealed by successor | |---|---|---| | Hooke 1674 γ Dra | ~arcsec | 1500× too large, wrong direction (Bradley, Hipparcos) | | Flamsteed Polaris 1698 | ~arcsec | Aberration + nutation, not parallax (Bradley 1728) | | Bradley 1728 γ Dra | Null at 1" level | Gaia eventually, 1000× smaller | | Lacaille, Maskelyne (mid-1700s) | Null | Parallax was below their precision | | Brinkley 1810–1815 | 1 to 2.7" | Pond disagreed; Bessel 1838 gave smaller values | | Bessel 1815–1817 | T = −1.32" for 61 Cyg | Own 1838 retraction, +0.314" | | Struve 1837 Vega | 0.125" ± 0.055" | Own 1840 revision doubled it (wrongly) | | Henderson 1839 α Cen | 1.16" | Modern: 0.747" (55% too large) | | Bessel 1838 61 Cyg | 0.314" ± 0.020" | Hipparcos 0.286" (9% discrepancy at formal 6% error) | | Hipparcos 1997 Pleiades | 120.2 ± 1.5 pc | Gaia DR2: 135 pc (14% wrong, well outside error) | | Hipparcos 1997 proper motions | Claimed ~1 mas/yr | Makarov-Kaplan 2022: 226 µas/yr global spin | | Gaia DR2 2018 | Mission precision ~10 µas | Superseded by EDR3/DR3; PZPO corrections | | Gaia EDR3 2021 | PZPO −17 µas | Bright-sample cross-checks: −37 to −82 µas | | Gaia DR3 2022 | PZPO −17 µas | Same | Every row shows a claimed measurement whose systematic was revealed only when the next generation of instrumentation arrived. The current row in this table is Gaia DR3. There is no reason, inductively, to assume this row is different from all the others. ### The induction CHES (Chinese Habitable Exoplanet Survey) is scheduled for launch in 2028 with microarcsecond relative astrometry on 100 Sun-like stars within 10 pc. Theia is under study. GaiaNIR is in early planning. Every historical precedent suggests that whichever of these missions flies next will reveal a currently-undetected Gaia DR3 systematic at the level of the current Gaia zero-point controversy, approximately 20 to 40 microarcseconds. The current 99-microarcsecond PZPO spread among independent cross-checks is the most likely lower bound on that systematic, not an upper bound. > [!note] The closing point > The stellar parallax programme has never been reliable at its own current-generation precision level. The foundation (Hooke, Flamsteed, Bradley, Brinkley) is demonstrably wrong. The first textbook-credited positive result (Bessel 1838) was preceded by three published negatives from the same observer on the same star. Every subsequent mission has been revealed to be wrong by the next mission. Gaia DR3 is the latest and carries unresolved systematics that its own team has published formally. Treating the Gaia catalogue as settled, converged, and authoritative requires ignoring the historical pattern and the consortium's own admissions. H₁ requires that trust. H₀ does not. --- ## Verdict **Failed to falsify H₀.** The opening section establishes that the parallax programme has never been reliable at current-generation precision. Primary argument 1 shows that the observed signal has at least two valid geometric interpretations, one of which is published in peer-reviewed physics literature and numerically reproducible. Primary argument 2 shows that the Gaia design team itself has formally proven the zero point to be mathematically degenerate with an instrumental wobble 2000× larger than the design target. Primary argument 3 shows that nine independent cross-checks give nine different zero-point values spanning 99 microarcseconds, with bright samples disagreeing with the official quasar value by a factor of 2. The closing section establishes the inductive pattern that every prior astrometric mission's claimed precision was revealed to be inflated only when the next mission superseded it. H₁ requires that: 1. Parallax be the exclusive geometric signature of heliocentric motion (refuted by primary argument 1). 2. Negative parallax be pure statistical noise (refuted by the century-stable ~25% fraction, Aitken 1921's 26% primary source, and Plot 05's Gaussian-noise inconsistency). 3. Instrumental precision be sufficient to establish distance without additional assumptions (refuted by primary arguments 2 and 3). No single line of H₀ evidence is individually decisive. Collectively, the historical foundation + three primary arguments + supersession precedent make falsification of H₀ impossible without ignoring the Gaia consortium's own published admissions, the mainstream peer-reviewed kinematic-equivalence literature, and the 200-year pattern of astrometric-mission supersession. --- ## Sources cited in this null hypothesis ### Primary historical - [[1798_Bradley_Greenwich_Observations]] - [[1817_Bessel_Koenigsberger_Beobachtungen_Vol3]] - [[1838_Bessel_61_Cygni_Parallax]] - [[1852_Bessel_Olbers_Briefwechsel_Polaris_Quote]] - [[1921_Aitken_Recent_Parallax_Review]] - [[1943_Lee_Negative_Parallax]] - [[1981_Williams_Hooke_to_Bessel_PhD_Thesis]] - [[2020_Reid_Menten_First_Parallaxes_Revisited]] ### Primary argument 1 (kinematic equivalence) - [[2013_Popov_NeoTychonian_Primary]] - [[2013_Popov_Stellar_Parallax_NeoTychonian]] - [[2013_Popov_Dynamical_Tychonian_Universe]] - [[2007_Sungenis_Galileo_Was_Wrong_Parallax]] - [[Shack_Chapter25_Negative_Parallax_Demystified]] - [[2023_Shack_Tychos_Book]] ### Primary argument 2 (basic-angle degeneracy) - [[2017_Butkevich_Basic_Angle_Parallax]] - [[2014_Mora_Gaia_BAM]] - [[2019_Druetto_Deep_Learning_Gaia_Anomaly]] ### Primary argument 3 (PZPO cluster) - [[2021_Lindegren_EDR3_Parallax_Bias]] - [[2021_Groenewegen_Parallax_ZeroPoint_Offset]] - [[2024_Ding_GalacticPlane_Parallax_Bias]] - [[2025_Ding_Bright_Star_Parallax_Bias]] - [[2018_Stassun_Torres_DR2_Eclipsing_Binaries]] - [[2021_Stassun_Torres_EDR3_Eclipsing_Binaries]] - [[2026_Bobylev_VLBI_Gaia_Zero_Point]] - [[2021_Riess_Cepheid_Hubble_Constant]] - [[2023_Khan_Gaia_Asteroseismic_Red_Clump]] - [[2021_Fabricius_Gaia_EDR3_PSF_Chromaticity]] - [[2021_Fardal_Gaia_Parallax_Scar_Map]] - [[2022_Penoyre_Binary_Astrometry_RUWE]] - [[2015_BailerJones_Distances_from_Parallaxes]] - [[2021_BailerJones_EDR3_Distances]] - [[2018_Luri_Gaia_DR2_Parallaxes]] ### Closing (supersession) - [[2007_vanLeeuwen_Hipparcos_Rereduction]] - [[2022_Makarov_Kaplan_Hipparcos_Gaia_Proper_Motions]] - [[2024_Lunz_Gaia_Bright_Frame_ICRF3]] - [[2019_Pietrzynski_LMC_Distance_Eclipsing_Binaries]] ### Our reproductions (`/Notes/Stellar_Parallax/resources/`) - `script_01_popov_proxima.py` → Plot 01 - `script_04_negative_fraction_timeseries.py` → Plot 04 - `script_05_gaussian_noise_model.py` → Plot 05 - `script_07_pzpo_cluster.py` → Plot 07 - `script_08_butkevich_coefficient.py` → Plot 08 --- ## See also - [[Stellar_Notes]] (working notes, 28 numbered arguments) - [[Stellar_Graphs_Summaries]] (all 10 plots with per-plot summaries) - [[research-log]] (5-agent deep-web crawl findings, 2026-04-17) - [[00_Null_Hypothesis_Index]]