Speed-Error Cross-Correlation Dating of Ancient Star Catalogues, with Application to the Almagest

This paper introduces the SESCC method, which dates ancient star catalogues by minimizing the cross-correlation between stellar proper-motion speeds and positional residuals, and applies it to the Almagest to provide statistical evidence supporting a pre-Christian Hipparchan origin over a later Ptolemaic one.

The Gist

Imagine you find an old, dusty map in an attic. The map shows where the stars were, but the paper is yellowed, and the handwriting is from a time long ago. You want to know: When was this map actually drawn?

For centuries, historians have argued over the most famous star map in history: The Almagest, written by the ancient Greek astronomer Ptolemy around 137 CE. Ptolemy claimed he drew it then. But many suspect he actually copied an older map drawn by a predecessor named Hipparchus, who lived about 250 years earlier.

The problem is that stars don't stay still. They drift slowly across the sky, like cars on a highway. If you look at a star's position today and compare it to where Ptolemy said it was, the difference (the "residual") tells you a story. But that story is messy because of two things:

1. The Drift: The star's own movement (proper motion).
2. The Shift: The entire sky seems to rotate slowly over time (precession), which changes the "zero point" of the map.

This paper introduces a clever new detective tool called SESCC (Speed-Error Signals Cross-Correlation) to solve this mystery. Here is how it works, explained simply:

1. The "Speed vs. Mistake" Detective Game (Latitudes)

Imagine you are trying to guess when a photo of a moving crowd was taken.

• The Idea: If you take a photo exactly when the event happened, the mistakes in your photo (blurry edges, wrong positions) are just random noise. They don't care how fast the people were running.
• The Problem: If you guess the wrong date (say, 100 years too early), the people who were running fast will have moved a huge distance from where you think they should be. The faster they ran, the bigger the mistake.
• The Method: The author looks at every star in the Almagest. He checks: "Do the stars that move the fastest have the biggest errors?"
  • If YES: You guessed the wrong date. The fast stars have drifted too far.
  • If NO: You guessed the right date. The errors are just random noise, unrelated to speed.

By testing every possible year from 600 BCE to 1900 CE, the method finds the specific year where the "fast stars" and "big mistakes" stop being connected. This is the True Epoch.

2. The "Relative Distance" Trick (Longitudes)

Dating the North-South position (Latitude) is easy because the whole sky doesn't shift up or down. But the East-West position (Longitude) is tricky because the whole sky rotates (precession). It's like trying to guess when a photo was taken while the camera itself is slowly spinning.

To fix this, the author invented SESCC-pairs.

• The Analogy: Imagine you are in a moving train. If you try to measure how far you are from the station, the train's speed messes up your calculation. But if you measure the distance between two passengers sitting next to each other, the train's speed doesn't matter! They move together.
• The Method: Instead of looking at one star's position, this method looks at the distance between pairs of neighboring stars.
  • If the whole sky shifts (precession), both stars shift by the same amount. The difference between them stays exactly the same.
  • This cancels out the "spinning camera" problem perfectly, allowing the author to date the map without needing to know the exact rules of how the sky rotates.

3. The Results: Who Drew the Map?

The author tested this method on two known maps first (Tycho Brahe and Ulugh Beg) to make sure it worked. It did perfectly.

Then, they applied it to the Almagest:

• The Latitude Test: The "Speed vs. Mistake" game suggested the map was drawn around 49 BCE.
• The Longitude Test: The "Relative Distance" trick suggested the map was drawn around 165 BCE.
• The Verdict: Both methods agree that the map was likely drawn before the birth of Christ. This strongly supports the idea that Ptolemy didn't draw it himself in 137 CE, but rather copied an older map made by Hipparchus (who lived around 127 BCE).

4. The "Smoking Gun" Evidence

The paper also found a funny clue in the numbers themselves, like finding a fingerprint on the map.

• Ancient astronomers wrote numbers in a base-60 system (like minutes and seconds).
• Hipparchus used a tool that naturally created numbers ending in 15 minutes or 45 minutes (quarter-hours).
• The Clue: The Almagest's North-South numbers (Latitudes) still have lots of these "15 and 45" endings.
• The Twist: The East-West numbers (Longitudes) almost never have them.
• Why? Ptolemy admitted he added a specific correction (2 degrees and 40 minutes) to update the longitudes for his own time. When you add that specific number to a "15" or "45," it mathematically turns into a "55" or "25," destroying the original "quarter" pattern.
• Conclusion: The fact that the North-South numbers kept their "Hipparchan fingerprint" while the East-West numbers lost it proves that Ptolemy copied the original map and only tweaked the East-West numbers.

Summary

This paper is like a high-tech time machine for star maps. By using a new mathematical trick to ignore the "noise" of the universe's rotation, the author proved that the famous Almagest star catalogue is actually a "reprint" of an older Hipparchan original, not a fresh observation by Ptolemy. It's a victory for open science, using modern computers to solve a 2,000-year-old historical mystery.

Technical Summary

1. Problem Statement

The primary challenge addressed is determining the true epoch of observation for ancient star catalogues, specifically the Almagest (attributed to Ptolemy, c. 137 CE). While Ptolemy claims the data is contemporary, historical analysis suggests the positions may derive from earlier observations by Hipparchus (c. 127 BCE), adjusted using Ptolemy's erroneous precession constant.

Existing dating methods face significant limitations:

• Bulk Methods: Previous approaches (e.g., Dambis & Efremov) rely on linear regression between proper-motion speeds and positional residuals for a selected subset of stars. These have been shown to lack statistical significance in Almagest data and fail to recover known epochs in other catalogues (e.g., Ulugh Beg).
• Longitude Dating: Dating via ecliptic longitudes is complicated by precession (which shifts all longitudes uniformly) and potential global offsets (systematic errors in the zero-point). Methods relying on absolute longitude residuals conflate temporal signals with these static offsets.

2. Methodology: SESCC and SESCC-pairs

The author introduces SESCC (Speed-Error Signals Cross-Correlation), a new method that minimizes the cross-correlation between stellar proper-motion speeds and positional residuals.

A. SESCC for Ecliptic Latitudes

• Principle: At the true epoch ($T_0$), the residual error ($\delta\beta$) between the catalogue and modern positions is random noise, independent of the star's proper motion speed ($|\mu_\beta|$). At any incorrect epoch, the residual contains a systematic term proportional to the speed and the time difference ($|\mu_\beta| \cdot |T - T_0|$).
• Statistic: The method calculates a dot product across the entire catalogue (no star selection):
  $$C(T) = \sum_{i} |\mu_{\beta i}| \cdot |\delta\beta_i(T)|$$
• Advantages:
  • Implicitly weights stars with high proper motion (high temporal information) while down-weighting slow stars (noise).
  • Requires no linear modeling or subset selection.
  • Immune to precession, as precession does not affect ecliptic latitudes.

B. SESCC-pairs for Ecliptic Longitudes

• Principle: To eliminate the effects of precession and global longitude offsets, the method uses pairwise differences between neighboring stars rather than absolute positions.
• Statistic: For pairs of stars $(i, j)$ within a proximity threshold:
  $$C_p(T) = \sum_{(i,j)} |\mu_{\lambda i} \cos \beta_i - \mu_{\lambda j} \cos \beta_j| \cdot |\delta\lambda_{ij} - \delta\lambda_{ij}^{mod}(T)|$$
• Algebraic Invariance: If all longitudes are shifted by a constant $\epsilon$ (due to precession correction or instrumental error), the difference $\delta\lambda_{ij} = (\lambda_i + \epsilon) - (\lambda_j + \epsilon)$ remains unchanged. Thus, the method is immune to global offsets by construction.
• Implementation: Uses J2000 inertial frame positions to ensure precession does not re-enter the calculation. Excludes stars with high parallax (e.g., $\alpha$ Centauri) to avoid non-linear proper motion biases.

3. Validation

The methods were validated against catalogues with known epochs and synthetic data:

• Tycho Brahe (1576–1597 CE):
  • Latitudes: Estimated $1570$ CE.
  • Longitudes: Estimated $1547$ CE (Bootstrap median $1552$ CE).
• Ulugh Beg (1437 CE):
  • Latitudes: Estimated $1177$ CE (affected by spatially variable systematic errors).
  • Longitudes: Estimated $1452$ CE (within 15 years of the true epoch), demonstrating the superiority of the pairwise approach for longitudes.
• Synthetic Catalogues: Generated with Gaussian noise matching Almagest error levels.
  • Hipparchan Epoch (-127 BCE): 85% of simulations yielded a pre-Christian minimum.
  • Ptolemaic Epoch (+137 CE): Only 20% of simulations yielded a pre-Christian minimum.
  • The method successfully distinguished between the two epochs and remained invariant under global longitude offsets of $\pm 6^\circ$.

4. Application to the Almagest

The method was applied to the 1,022 secure entries in the Almagest.

• Latitude Results (SESCC):
  • Point estimate: -49 BCE.
  • Bootstrap distribution: 74% of resamples yielded a pre-Christian minimum.
  • This is inconsistent with a Ptolemaic origin (which would yield ~20% pre-Christian) and consistent with a Hipparchan origin.
• Longitude Results (SESCC-pairs):
  • Point estimate: -165 BCE.
  • Bootstrap distribution: 74% of resamples yielded a pre-Christian minimum.
  • The convergence of independent latitude and longitude methods on the same statistical conclusion (74% pre-Christian) strongly supports a pre-Christian origin.

5. Independent Corroboration: Sexagesimal Fractions

The paper provides a secondary, non-dynamical line of evidence based on the arithmetic structure of the data:

• Observation: Quarter-degree fractions ($15'$ and $45'$) appear in 24.1% of latitude entries but only 0.7% of longitude entries.
• Mechanism: Ptolemy explicitly added $2^\circ 40'$ to Hipparchan longitudes to correct for precession (using his constant of $1^\circ$/century).
  • Mathematically, adding $40'$ to a quarter fraction ($15'$ or $45'$) results in $55'$ or $85'$ ($1^\circ 25'$).
  • When rounded to the nearest $10'$ or $15'$ (the Almagest's standard), these values never land on a quarter-degree fraction.
• Conclusion: The near-total absence of quarter-fractions in longitudes is the deterministic consequence of Ptolemy's arithmetic adjustment, while their presence in latitudes (which he copied unchanged) confirms the original data source was likely Hipparchus.

6. Key Contributions and Significance

1. New Methodology: Introduces SESCC and SESCC-pairs, the first longitude-based dating method that is algebraically immune to global offsets and requires no assumptions about precession constants.
2. Resolution of Historical Debate: Provides robust statistical evidence that the Almagest star catalogue is based on Hipparchan observations (c. 127 BCE) rather than contemporary Ptolemaic observations.
3. Independence of Coordinates: Demonstrates that both ecliptic latitudes and longitudes independently point to a pre-Christian epoch, overcoming previous criticisms of single-coordinate methods.
4. Arithmetical Proof: Links the statistical dating results to the physical structure of the data (sexagesimal fractions), offering a coherent explanation for the "missing" quarter-fractions in longitudes.
5. Limitations: The precision is limited to approximately $\pm 200$ years due to the limited number of high-proper-motion stars in the ancient catalogue, and the latitude method shows sensitivity to spatially systematic errors (as seen in the Ulugh Beg validation).

Conclusion: The paper concludes that the Almagest catalogue is a Hipparchan work, adjusted by Ptolemy for precession, rather than a direct observation by Ptolemy himself. The convergence of dynamical dating (SESCC) and structural analysis (sexagesimal fractions) provides a compelling case for this historical revision.