The Pyramid as Memory Hypothesis: The Hill That Rose from the Waters, Made into a Pyramid — Pyramid Texts, Utterance 600

Description

Series information (English)

CORRECTIONS

This version of The Pyramid as Memory corrects errors found during a full re-verification of the paper against its own sources, its own tables and the analysis scripts deposited with it, including the reference list read entry by
entry. The hypothesis is unchanged.

Three of the corrections bear on figures a reader may have reused. A fraction-system control table could not be reproduced from the fraction sets it named and has been withdrawn. The paper contradicted itself on its central
count, reporting ten of fourteen targets in one section and eleven of thirteen in others; ten of fourteen is correct, and three further statistics have been brought into line with the deposited output. And the Sothic cycle was given as
1,460 civil years where it is 1,461, which puts the perimeter relation outside the tolerance and withdraws it as a correspondence. Readers who cited the previous version for eleven of thirteen, for the rival fraction-system scores,
or for the Sothic correspondence should not carry those figures forward.

Four passages attributed claims to sources that do not contain them, and a negative control deposited with the previous version was never reported in it. All of the above is itemised in the technical information accompanying this
version. The errors are mine. The earlier version remains available under its own version record for transparency; it should not be cited as current.

-- Sefy Levy

Technical info

What has not changed. The hypothesis is unchanged. The reading of the Giza complex as the permanent architectural encoding of the Zep Tepi system, the four-element landscape correspondence with the Richat origin landscape, the Wedjat fraction encoding of the six exterior dimensions, the three-tier structure of claims, and the predictive structure all stand. The corrections restate figures that had been superseded, correct attributions, align the chronology with a single stated source, remove the language of pre-registration where no pre-registration existed, and rebuild the reference apparatus.

The correspondence count and the reported statistics. The scored result is ten of fourteen targets, not eleven of thirteen. The target list carries fourteen entries and four of them are misses: the sidereal month, the synodic month, Mercury, and the Sothic cycle. The previous version was internally inconsistent on this figure, reporting ten of fourteen in the system-specificity section while reporting eleven of thirteen in the jackknife and the Monte Carlo summary. Three further statistics did not match the output of the analysis script deposited with the paper: the denominator margin at the 0.20% tolerance, given as +6 at p = 0.002, is +5 at p = 0.005; the Egyptian unit fraction head-to-head score, given as 10 of 14, is 11; and the tolerance profile was reported on a ten-target independent set where that set holds eleven. The base sweep ranking is unchanged in substance but now states its figure, 4.2 per cent, rather than rounding it to four. Every figure in this version has been checked line by line against the deposited output. These are differences between the paper's prose and the script's output rather than corrections to the procedure, which is unchanged; the one exception is the Sothic cycle, described next, where a target value itself was wrong.

The Sothic cycle. The cycle is 1,461 Egyptian civil years, equivalent to 1,460 Julian years. The previous version used 1,460 civil years, which conflates the two counts. The perimeter relation, 1,760 cubits x 53/64 = 1,457.5, is 0.240% from 1,461 civil years and therefore falls outside the +/-0.2% threshold used throughout. It is recorded as a miss and excluded from the scored count. It is retained in the discussion of the Sirius cluster, where it is now identified as a near-miss rather than a correspondence.

Pre-registration. The paper previously described its confirmatory analysis as frozen and pre-registered and referred to a pre-registration protocol. No pre-registration was deposited with a registry. The analysis specification was fixed in advance and published alongside the paper, and the text now says that instead. Nothing about the procedure changed; only the description of it.

The stellar negative control. The paper now reports the control that tests the objection most directly threatening its result: whether the same six dimensions, under identical rules, also reach stellar positions. They do not. The pyramid reaches five of thirty-one stellar targets against a chance expectation of 4.83 at p = 0.543, where the calendrical suite scored by the identical procedure returns ten of fourteen against 6.25 at p = 0.023. The control was deposited with the previous version and its output, but the paper did not report it, while a later paper in the series cited this paper for it.

Supplementary files. The deposited analysis scripts are unchanged in every calculation. Their only edits are to comment and label text: the description of the specification; three output labels that read "max-independent 8" where the set holds nine targets and the script prints 6/9; and the stellar control's docstring, which stated that four stars are circumpolar at the construction epoch and contribute no rising azimuth where five are, Dubhe included, as the thirty-one-target count already assumed.

Citing this work. Readers who cited the previous version for the eleven-of-thirteen count, for the denominator margin of +6 at p = 0.002, for the Egyptian unit fraction score of 10 of 14, for the tolerance profile on a ten-target set, for the Sothic cycle as a correspondence, for the Sothic cycle at 1,460 civil years, for the Seven Sages as the Edfu Building Texts' account of the first sacred mound, for Hoffman et al. 1986 on the Hierakonpolis revetment, for Ritter 2003, or for the dating of the Giza complex to c. 2560-2510 BCE, should not carry those citations forward. Readers who went to the previous version for the stellar negative control will find it reported here for the first time. The earlier version remains available under its own version record for transparency; it should not be cited as current.

REVISION CHANGES

Categories of change, not an itemised list.

Figures restated

    The correspondence count, from eleven of thirteen to ten of fourteen, throughout the paper including the jackknife, the Monte Carlo summary and the null-model discussion
    The denominator margin at +/-0.20%, from +6 (p = 0.002) to +5 (p = 0.005)
    The Egyptian unit fraction head-to-head score, from 10 to 11 of 14
    The base sweep ranking, from "top four per cent" to "within the top 4.2 per cent" of the 526 candidate bases
    The tolerance profile restated on the independent eleven-target set rather than a ten-target set
    The Sothic cycle target, from 1,460 to 1,461 civil years, in the body and in both summary tables
    The Sothic relation restated as 0.240% from target and recorded as a miss, in the body and in Prediction 7
    The fraction-system control test, where the prose reported eleven Wedjat correspondences against its own table's ten

Claims attributed to sources that do not contain them, corrected

    The Edfu Building Texts as an account of the Seven Sages who survived the destruction of a homeland in the Far West; the passage describes a primeval island and the Great Primaeval Mound, and is now cited to Reymond 1960 at the page
    The Hierakonpolis temple mound as enclosed by a roughly circular revetment on the authority of Hoffman et al. 1986; now a sand mound encased in a stone revetment, reinterpreted by Friedman 2008
    Ritter's challenge to the wedjat derivation, dated 2003; the chapter is 2002
    The Royal Cubit's proximity to pi/6 attributed in part to Petrie's survey; Petrie's own determination from the Great Pyramid is larger, and the claim now rests on the cubit rods with Rossi 2004 and Zapassky et al. 2012
    "Thoth's restoring portion" as an Egyptian formulation; now given as the 1/64 that the modern account assigns to Thoth
    Pyramid Texts quotations re-pointed from Allen 2005 to Allen 2015, the second edition, with utterance and page locators

Dates corrected

    The Unas pyramid and its Pyramid Texts, from c. 2345 BCE to c. 2375-2345 BCE, in seven passages
    The Giza complex, from c. 2560-2510 BCE to c. 2580-2477 BCE on the long-reign chronology
    The external architecture of Giza, from c. 2589-2560 BCE to c. 2580-2540 BCE
    The Predynastic grammar window, from c. 3700-3100 BCE to c. 3800-3100 BCE
    Nabta Playa, from c. 7000-4800 BCE to c. 8300-3400 BCE
    The transmission interval, from 540 years to roughly 500 years
    Herodotus, from c. 450 BCE to the mid-fifth century BCE, and the comparative table's range restated accordingly
    The Amenpanufer confession, from approximately 1,200 years after Khufu to some fourteen centuries
    A chronological conventions section added, following Shaw 2000 and, for the First Dynasty, the Bayesian radiocarbon chronology of Dee et al. 2013

Language of pre-registration removed

    "The pre-registered primary test" as a section heading
    The pre-registration protocol, now the analysis protocol, published rather than deposited, fixing rather than freezing the target list
    The frozen confirmatory analysis, in four passages
    The threshold described as frozen in the pre-registration protocol
    The contrast between post-hoc and pre-registered search, restated as a protocol fixed in advance
    Eighteen instances in total; one use of "pre-specified" is retained, where it describes the statistical procedure rather than a deposited file

Scope and attribution narrowed

    Prediction 8 rewritten: it previously rested on the Sothic cycle as a cubits correspondence that disappears in metres, which the restated count does not support, and now rests on the anomalistic lunar month alone, with the falsification condition narrowed to match
    The Sirius cluster identified as three arithmetical relations of which one falls inside the multiplication and division test, the 360-day relation being addition-based and the Sothic relation outside tolerance
    The abstract's statement of evidence sources, which omitted mathematical analysis, cultural memory theory, archaeometry and comparative architecture
    Terminology aligned with the rest of the series: Papers I-III are now Corridor I-III, and the cattle cosmogram is now the herd for the dead
    Forward citation to a later paper in the series removed from the closing line

The fraction-system control table, replaced

    The table comparing the Wedjat set against a small-denominator set (1/2, 1/3, 1/4, 1/5, 1/6) and a Babylonian sexagesimal subset (1/6, 1/12, 1/30, 1/60) could not be reproduced. Its Wedjat column is correct, but its rival columns report 6 and 8 of 14 where those fraction sets, applied to the same six dimensions under the same operations and tolerance, reach 1 and 0; allowing all subset sums gives 2 and 1, and allowing pairwise sums and differences of dimensions gives 3 and 1. The column totals also disagreed with their own ticks. The table has accordingly been replaced with the per-target result from the deposited analysis script: every target, the single Wedjat operation that comes closest to it, the error, and whether it falls inside the tolerance
    Before that replacement the same table was corrected in three respects: a sidereal month row added, the table having carried thirteen rows for a fourteen-target test; the Sothic row restated at 1,461 years; and the small-denominator total corrected to 6 of 14 against its own six ticks
    The surrounding paragraph rewritten to match, with the per-target claims about which rival system misses which planet removed, and the rival-system question pointed to the system-specificity test, where each rival is downsampled to the Wedjat set's exact size and the figures are reproducible
    The claim that the Wedjat system "outperforms comparable alternative fraction systems" on this evidence, withdrawn from this paragraph

Additions

    A new subsection, Domain Specificity: The Stellar Negative Control, in the statistical section after the Monte Carlo and system-specificity tests: eleven stars precessed to the construction epoch under Vondrak et al. 2011, converted to declination, rising azimuth and culmination altitude at Giza, with the obliquity, Giza's latitude and the circumpolar limit added to give thirty-one angular targets; scored against the same six dimensions, sixty-three fractions, two operations and +/-0.2% tolerance, with a 200,000-trial shadow-angle control at seed 42
    The five stellar targets that do reach tolerance, with their formulae and errors, and the observation that they include Sirius, Canopus and Alnitak
    The same control re-scored at every year of the 2580-2540 BCE long-reign window, where the count ranges from three to six hits against a chance expectation of 4.83 and is never significant
    The one-year sensitivity at the tolerance edge, where Canopus's declination error moves from 0.198% to 0.202% and the count falls from five to four
    The code and output identified as Part 4 of zep_tepi_calculations.py, formerly the standalone stellar_control.py
    Dee et al. 2013, the Bayesian radiocarbon chronology the First Dynasty dates rest on, which the text cited and the reference list omitted
    Citations attached at the point of claim where the source was listed but the claim carried none: Schmidt 2006 for the Gobekli Tepe enclosures, Dietrich et al. 2012 for cult and feasting, Wendorf & Schild 1998 and Malville et al. 2008 for the Nabta Playa circles and alignments, and Peet 1930 for the tomb-robbery papyri behind the Amenpanufer confession

Reference apparatus rebuilt

    Two entries recovered from after the Artificial Intelligence Use Disclosure, where they had been stranded outside the reference list, including the precession model the astronomy depends on
    A running footer line removed from the body of the reference list, where it had been captured as a paragraph
    Eleven entries moved to the correct field heading, among them a palaeoenvironmental paper under pyramid construction, an ancient-DNA paper and a grain-metrology paper under archaeoastronomy, and a Richat geochronology paper under star shafts
    Three headings added: Saharan Palaeoenvironment, Chronology (holding Dee et al. 2013 and Shaw 2000, the two sources the new chronological conventions section rests on), and Non-Peer-Reviewed Sources
    Grey literature consolidated under the new non-peer-reviewed heading: an impact-structure database that had been shelved under a heading marked peer-reviewed, four November 2025 news reports that had been nested inside a peer-reviewed section, and two government press releases that had been shelved under cultural memory with no label
    Every heading re-sorted alphabetically, with single-author entries placed before multi-author entries of the same surname
    Self-citation letters corrected, 2026e and 2026f to 2026h and 2026i
    Entries completed where a subtitle, publisher order or ISBN was missing, and a page locator added where a source carries two people of the same name
    Five entries no claim in the paper points at, removed: a Dendera temple monograph and a Dendera encyclopaedia entry, a 19th-century history of astronomy, an excavation annual report, and a paper on the geodynamic setting of African basins
    Four entries that had no citation attached, given one at the point of claim: Monod 1975 for the three prehistoric sites recorded inside the Richat structure, Reimold & Koeberl 2014 and Koeberl et al. 2005 for the ages of the African impact structures in the elimination set, and El-Daly 2005 for the modern case that medieval Arabic writers engaged systematically with Egyptian antiquity
    van de Loosdrecht 2018, previously listed but uncited, now cited for the role it actually plays: the Pleistocene North African population anterior to the window, against which the continuity reported by the three Holocene studies is measured. It is not a fourth node of the same period, and the sentence says so
    The Jerusalem Post and WAM reports of the November 2025 corridor announcement, previously listed but uncited, now cited with the other two at the point where the announcement is reported
    The Italian ministry announcement cited under its Italian name as well as in translation, so the in-text citation and the reference entry match
    Lluch & Philip 2003 moved to the non-peer-reviewed heading with its venue stated, the Cahiers de l'AARS imprint describing a reading committee but no review procedure and placing articles under the sole responsibility of their authors; the entry remains cited, for the Tililit rock art
    The Dendera Temple heading removed, both its entries having gone; this paper's only Dendera claim is the temple-orientation sentence, which is cited elsewhere
    The qualifier removed from the heading Richat Archaeology (Peer-Reviewed), every work in the list being peer-reviewed apart from the entries under the non-peer-reviewed heading
    The heading Geological Formations for Elimination Table renamed African Impact Structures, this paper containing no elimination table

Abstract

    Restated for the correspondence count and the evidence sources, and a landscape-scale paragraph added; otherwise unchanged, the hypothesis being unchanged

Notes