Pattern Recognition & Zeta Reticuli
by Carl Sagan & Steven Soter
Astronomy, December, 1974
"The Zeta Reticuli Incident" is very provocative. It claims that a map, allegedly shown on board a landed extraterrestrial spacecraft to Betty Hill in 1961, later drawn by her from memory and published in 1966, corresponds well to similar maps of the closest stars resembling the sun based on stellar positions in the 1969 Gliese Catalog of Nearby Stars. The comparison maps were made by Marjorie Fish using a three dimensional physical model and later by a group of Ohio State University students using a presumably more accurate (i.e., less subjective) computer generated projection. The argument rests on how well the maps agree and on the statistical significance of the comparison. Figure 1 [not available here] show the Hill map and the Ohio State computer map with connecting lines as given in the Astronomy article. The inclusion of these lines (said to represent trade or navigation routes) to establish a resemblance between the maps is what a lawyer would call "leading the witness". We could just as well have drawn lines as in the bottom of Figure 1 to lead the other way. A less biased comparison of the two data sets, without connecting lines as in Figure 2, shows little similarity. Any residual resemblance is enhanced by there being the same number of points in each map, and can be accounted for by the manner in which these points were selected. The computer star map includes the sun and 14 stars selected from a list of the 46 nearest stars similar to the sun, derived from the Gliese catalog. It is not clear what criteria were used to select precisely these 14 stars from the list, other than the desire to find a resemblance to the Hill map. However, we can always pick and choose from a large random data set some subset that resembles a preconceived pattern. If we are free also to select the vantage point (from all possible directions for viewing the projection of a three dimensional pattern), it is a simple matter to optimize the desired resemblance. Of course such a resemblance in the case of selection from a random set is a contrivance - an example of the statistical fallacy known as "the enumeration of favorable circumstances".
The presence of such a fallacy in this case appears even more likely when we examine the original Hill drawing, published in The Interrupted Journey by John Fuller. In addition to the prominent points that Betty Hill connected by lines, her map also includes a number of apparently random dots scattered about - evidently to represent the presence of background stars but not meant to suggest actual positions. However, three of these dots appear in the version of the Hill map used in the comparison, while the others are absent. Thus some selection was made even from the original Hill map, although not to the same extent as from the Gliese catalog. This allow even greater freedom to contrive a resemblance. Finally, we learn from The Interrupted Journey that Betty Hill first thought she saw a remarkable similarity between her UFO star map and a map of the constellation Pegasus published in the New York Times in 1965 to show the position of the quasar CTA-102. How many star maps, derived from the Gliese catalog or elsewhere, have been compared with Betty Hill's before a supposed agreement was found? If we suppress information on such comparisons we also overestimate the significance of the result.
The argument on "The Zeta Reticuli Incident" demonstrates only that if we set out to find a pattern correlation between two nearly random data sets by selecting at will certain elements from each and ignoring others, we will always be successful. The argument cannot serve even to suggest a verification of the Hill story - which in any case is well known to be riddled with internal and external contradictions, and which is amenable to interpretations which do not invoke extraterrestrial intelligence. Those of us concerned with the possibility of extraterrestrial intelligence must take care to demand adequately rigorous standards of evidence. It is all too easy, as the old Chinese proverb says, for the imprisoned maiden to mistake the beating of her own heart for the hoof beats of her rescuer's horse.
Steven Soter is a research associate working under the advisement of Carl Sagan, director of Cornell University's laboratory for Planetary Studies.