PTOLEMY'S FAILINGS

(JHA,xi 1980, p. 133-135)

The Crime of Claudius Ptolemy. Robert R. Newton (The Johns Hopkins University Press, Baltimore and London, 1977). Pp. xiv + 411. £15-75.

Whatever various generations of astronomers and of historians of astronomy have thought and shall come to think of Ptolemy, his Almagest (or Syntaxis, as Newton prefers to name it) is bound to remain a grand focus in the history of astronomy. Here a thousand years' work of Babylonian and Greek astronomers evolved into a comprehensive account, in geometrical language, of all the motions in the sky. Here the stage was set for another fifteen centuries of efforts, on the part of Indian, Arabic, and Latin astronomers, to maintain control with these motions. So any major change of view concerning the development of any part of the history of astronomy from 800 B.C. to A.D. 1600 must also add new facets to our picture of Ptolemy and of his work. Almost as a stocktaking of Ptolemy, the 1970s saw the publication of three monographic writings on his astronomy, viz, O. Neugebauer's authoritative account of the contents of the Almagest in its context of the mathematical side of the rest of ancient astronomy,1 O. Pedersen's Almagest companion volume designed to guide also the student of Latin astronomy up to the Renaissance,2 and R. R. Newton's book with a scientist's announcement of zero reliability for the observations Ptolemy claims to have made himself as well as for many of the observations that he attributes to other astronomers. All of the three works are indispensable on my shelf with standard books of reference, in the case of Newton's book for two reasons: first because of its highly acute and reliable analysis of the observational and theoretical data in the Almagest, and of their mutual interdependence, and second because I disagree completely with the author in his judgement of Ptolemy as a deliberate criminal in his activity as an astronomer.

In a series of publications from the period 1969-76 Newton has investigated the non-gravitational accelerations within the solar system. To this end he compared positions of the heavenly bodies, as reported in Ancient and Medieval observations, with "gravitational" positions drawn from calculations back in time with the most accurate modern ephemeris programs. On this plan of scientific research he had to evaluate the reliability of the old observations, as far as possible on a priori considerations of their historical context alone. Both from such evaluations and from the lack of coherence with earlier and later observations, Newton soon came to distrust the authenticity of the observational data in the Almagest. Therefore, you find spread in several of his publications from the period 1973-76 chapters and sections devoted to documentation of the thesis that all of Ptolemy's claimed observations that he used and that can be tested prove to be fraudulent. In the book under review he collects, in a single cohesive source, all of this material together with further evidence to support the more far-reaching thesis that the fraudulent data result from a deliberate plan, on Ptolemy's part, of hoodwinking his fellow scientists and scholars. It is true that Ptolemy, time and again, derived parameters known a priori from earlier astronomy by using observations ostensibly fabricated to yield the result he wanted. Instances of this procedure have been suspected, and in some cases documented, by earlier astronomers and historians of astronomy. But only Newton's thorough analysis reveals its full scope, demonstrating that it pervades the entire structure of the Almagest. You may like or dislike Newton's habit of accompanying his conclusions with statistical estimates of their confidence level. In some cases equally probable premises for the statistics lead to other levels, but never to invalidation of the partial conclusion in question. However, I do not understand the sense of multiplying such confidence levels for the partial conclusions into a combined×and triflingly low×probability of Ptolemy's lack of criminal character.

To be sure, I do not find Ptolemy guilty at all, for I do not imagine him striving to solve the particular problem that Newton tacitly presupposes to be the ideal main concern of Ptolemy's, viz, to initiate a new doctrine of astronomy based on fresh data. On the contrary, I think that Ptolemy spent his efforts in synthesizing the results of earlier astronomy rooted back into the work of his Babylonian colleagues older than himself by a millennium. This includes the necessity of preserving the continuity with basic parameters that had worked well for centuries and had, therefore, deservedly grown canonical. In current philosophical jargon×as coined by I. Lakatos×you may think of these parameters as the 'hard core' of Babylonian as well as of Greek astronomy. But while the Babylonians, by their arithmetic schemes, aimed at predicting series of discrete events of a certain kind, say Full Moons and lunar eclipses, the Greeks in their turn dramatically broadened the scope of astronomical science. Their geometric models were designed to establish a general and continuous correlation between any given date and the directions of sight towards individual heavenly bodies, say, the Sun and the Moon. Clearly this represents a theoretically 'progressive problem-shift'. But to ensure that the new programme would yield virtually the same results for Full Moons as the meritorious Babylonian schemes, the Greeks apparently chose to build into the geometric models some of the basic Babylonian parameters. In cases of conflict between the resulting calibration of the models and contemporary observational data, produced by new observational techniques the perfection of which could still be doubted, the right choice×'right' in modern retrospect×may well have been anything but obvious. The fact that Ptolemy, according to Newton's analysis, backed the wrong horse does not make for a crime. The crime thesis provoked me to spend a couple of years sorting out elements of the said 'hard core',3 and that is why the present review has been unduly delayed.

Like Newton's earlier writings, this book with its huge accumulation of numerical data makes for a well-organized work of reference. Its thirteen chapters subdivided into ninety-six sections are carefully tabulated, and separate lists guide the reader to any of the sixty figures or of the thirty-nine tables. The text is set in unjustified typescript.

It is impossible to survey here the wealth of arguments treasured up in the book, but as an appetizing example let me refer to the particularly beautiful demonstration, in chap. 9, that the coordinates of the Almagest star catalogue stem from an Hipparchian star list, now lost, but originally based on genuine observations of star longitudes and latitudes. This does not clash with H. Vogt's well-known analysis that the Almagest star coordinates do not originate specifically in Hipparchus's "Commentary on Aratus and Eudoxus". The core of Newton's argument hinges only on the statistical distribution of the fractional parts of a degree in the coordinates. For the latitudes the distribution in all probability reflects genuine observational data produced by an instrument with integral degree markings only, or perhaps also with half-degree markings. For the longitudes one may create a distribution with the very same characteristics, but only after subtraction of 40'. This renders it probable that the Almagest longitudes came about by the addition of 40' and, say, 2° to longitudes observed by the same instrument as the latitudes.

University of Aarhus
KRISTIAN PEDER MOESGAARD

REFERENCES
1. O. Neugebauer, A history of Ancient mathematical astronomy (New York, 1975).
2. O. Pedersen, A survey of the Almagest (Odense, 1974).
3. K. P. Moesgaard, "The Full Moon serpent: A foundation stone of Ancient astronomy", Centaurus, xxiv (1980), 51-96.