Part II - Science, Astrology and the Gauquelin Planetary Effects
by Kenneth Irving
IN PART I, I CONSIDERED the often-repeated criticism leveled at scientific studies of astrology that it is impossible, if not improper, to try to study individual elements of the chart in isolation from one another. In other words, can individual bits and pieces of the horoscope such as planets, or signs, be considered in some sense independent and additive? My answer was to point out that it is in fact a common practice in astrology to treat at least some chart elements in just this way, for example by adding up the triplicities, quadruplicities and polar opposites as part of the preparation for analyzing a chart.
One does not have to strain hard at a gathering of astrologers to hear statements such as “He has too much Fire in his chart,” and these are usually based on simple counting (i.e., adding together) of the properties of the signs holding the Ascendant, Sun, Moon and planets. If this counting can be done within a chart, it can also be done across charts, which is where statistics comes in. Thus it is fair to say that a group of people with “too much Fire” in their charts ought to be in some observable sense different from a group of people with “too much Water” in their charts, as otherwise such statements are meaningless.
However, even if the astrologer’s argument about analyzing isolated parts of the horoscope is incorrect in itself, still it is often based on an intuitive rejection of scientific studies that are mere “claims-testing.” Those who see astrology as a rag-bag of claims to be examined and nit-picked in order to show how false they are often do in fact have a problem with seeing only the parts and ignoring the whole. If you look through studies on various facets of astrology, particularly those that have made it into scientific journals, you will note a tendency to consider that any disproof of any part of astrology constitutes disproof of the idea underlying astrology itself, which is that the planets “up there” in some sense affect the lives of people “down here.”
Thus, the dispute between astrologers and scientists often reduces to:
Scientist: I have tested a claim made by an astrologer and found it
wanting. Therefore, astrology as a whole is false.
The astrologer, as I have shown, is wrong to say that you can’t test isolated factors. On the other hand, the scientist is wrong to think only in terms of claims and to draw such a broad conclusion from his narrow test. Both arguments are based on false premises and neither actually addresses what the other is saying anyway. Or, to coin a saying by giving a twist to the moral from the fable of the blind men and the elephant - “Neither is partly in the right, and both are in the wrong.”
What both sides generally miss is the fact that research aimed at “bits and pieces” is in itself useless unless it is undertaken with some sense of a larger structure those bits and pieces are related to. That larger structure is not what critics of astrology often call the astrological hypothesis, for there is no such thing. Instead, at this point in the history of the scientific investigation of astrology it must be what some would call a paradigm and others, a nomological network; or, less grandly, a framework for observation.
At the extremes of the debate between astrology and science, where it tends to be both at its noisiest and its least productive, there is no sense of this on either side, as our little conversation between the two straw stereotypes above reveals. In the middle, on the other hand, there are good examples of this sense of structure, perhaps the best of which can be found in the work of Michel Gauquelin. For example, in the last chapter of his first book (in which he revealed basic findings on professional groups and Mars, Jupiter, Saturn), he laid out the results in a way that emphasized their overall structure and then spent the rest of the chapter trying to develop a hypothesis that might address the problem presented by the results as a whole, in order to suggest further areas of investigation. Figure 1 below, which lays out this structure, is adapted from a similar diagram found in that chapter. _________________________________________________________________________
+ + +Figure 1: As this is based on a diagram from Gauquelin’s first book, L’Influence des Astres, it differs slightly from Table 2, which is based on replications and further studies over the next thirty years. In particular, Saturn for priests was not replicated in later studies, and indications of a positive nature were later found for both Moon and Venus. The thing to note in relation to Table 1 and Figure 2 is the structure of the results, and how it reflects an ancient astrological tradition.
Contrast what Gauquelin did (as shown in Figure 1) with the approach of the average skeptical scientist to the Gauquelin planetary effects, as well as with the usual view of Gauquelin's research offered by astrologers. For the skeptical scientist, remember, disproof of any single claim will disprove the whole, something which can logically be true only in the most extreme cases of outright fraud. The whole sorry series of episodes known as the “Mars effect controversy” would not have happened if it had not been for the dependence of three skeptic groups on this twisted notion that disproof of just one result in just one case would “disprove” not only any other work by Gauquelin but also the whole astrological notion of planets affecting people. And for them disproof did not require much, as all they needed was a single result the merest thousandth of a percent on the wrong side of a standard statistical benchmark.
On the other hand, astrologers tend to fix on the seemingly inexplicable emphasis the Gauquelin findings place on the 12th and 9th houses. Even though the planets linked to famous people of various professions agree somewhat with astrological tradition, the fact that these planets for the most part fall in the “wrong” houses is what astrologers are most aware of. It seems to reinforce the astrological notion that isolating Mars from other chart factors leaves odd bits and pieces lying about unanswered in the final results. It also stands in the way of actually making these results useful, the major prerequisite for interesting the average astrologer in scientific work.
So the scientists seem to be mainly interested in bits and pieces called claims and the astrologers seem to be mainly interested in bits and pieces they can plug into the horoscope in precisely the right places. Both hesitant astrologers and skeptical scientists, however, might learn more if they were to consider the kind of structure shown in the diagram above. Note, for example, the way in which generals and physicians “bind” Mars and Jupiter and Mars and Saturn, respectively; or the way in which painters are to some extent the reverse of physicians in respect to Mars and Saturn; or the way actors are the reverse of physicians in respect to Jupiter and Saturn. This is not the best way to present this information, but I use this diagram to make the point that from the beginning, even though he was testing individual elements of the chart, Michel Gauquelin was always looking for ways in which the results of individual tests could combine to show in some sense how the whole chart works together.
Now look at a table and a diagram that ought to help make the point. The first is simply a summary of all of the professional results, showing the pattern of significant deviations from chance in Gauquelin’s “plus zones” for all planets in all professions. Here we see again, and more clearly, I think, how Mars tends to work with either Jupiter or Saturn, but never with both at the same time; how Jupiter and Saturn act as opposites; how Mars tends to act as an opposite to Venus and the Moon, while Jupiter tends to work in the same direction as these two planets, despite its connection to Mars.
It should be obvious that a common structure links Table 1 with the ancient diagram of Figure 2, a diagram which relates the planetary natures along two dimensions, styled hot-cold and wet-dry. Not only is this diagram fundamental to astrology, but it also has a historical connection to modern psychology. And the fact that the pattern of planetary relationships it proposes makes its appearance in the Gauquelin professional results shows that even though the Gauquelins were disassembling the horoscope, so to speak, in order to study the relationship between planets and professions, the end result was naturally ordered in a decidedly astrological way.
In Part I, I called attention to the fact that even the instructions in an astrological textbook indicate it is perfectly permissible to treat various elements of the horoscope as if they were, at least at one level, independent of one another and that this is exactly what the use of most standard statistics requires. In that case, I favored the “scientists” (Geoffrey Dean and Arthur Mather) over the astrologer (Nick Campion). On the other hand, the kind of structure shown here is often ignored by astrology’s scientists as they concentrate on examining isolated parts of the horoscope and on random proposals made by astrologers. Thus, assessing the general results of investigation into astrology becomes a mere matter of weighing and measuring the number of claims tested against the number of claims found wanting.
Take the following, for example, which occurs in Section 3 of the discourse (Dean and Mather, 1994) mentioned in Part I: “...over the last few decades a small number of scientists (perhaps 1 in 106 of all scientists) have investigated the claims of astrology in what probably now amounts to well over 200 man-years of research, which excludes a much larger amount of non-scientific work by astrologers. But whenever the methodology has been appropriate the tests have failed to find support commensurate with the claims....”
As an appended note quoting H. J. Eysenck and D. K. B. Nias makes clear, that last statement is intended to indicate that even when these things called claims are supported by scientific studies, they tend to show small deviations which could not be of any obvious practical importance. Moreover, “...as methodology improves the deviations from chance expectancy tend to get smaller and smaller....”
So the scientific study of astrology seems to reduce to a simple equation: a number of scientists, in amount X, have studied a number of astrological claims, Y, and by improving methodology, Z have diminished these claims to a number of effects of size, oh, let’s say U, for useless. The statements cited seem to be based on facts, and indeed I’m sure they are, but what is missing here is some sense of the context in which those facts exist and the context to which they are addressed. The first passage quoted above is an example of something called “vote counting.” In tallying up the negative studies on one side, the positive studies on the other, and the number of scientists in between, it ignores the likelihood that the published studies of astrology which are most amenable to examination and critical analysis will be found in scientific journals, a venue in which negative results are more likely to be published than positive results. This is not because there is a vast conspiracy to make astrology look bad, but simply because scientists, journal editors and referees are human and have their biases.[v]
A classic demonstration of the problem was a 1970 study (Goodstein and Brazis, 1970) in which a group of psychologists were given an abstract of a journal article detailing a study on astrology and then asked to comment on the methodology used in the experiment. The experiment described was completely fictitious, but some psychologists were given an abstract reporting an outcome favorable to astrology while others were given an abstract (otherwise identical to the first) that reported a negative result. Those given the abstract with the negative outcome tended to approve of the procedures described, while those given the positive outcome tended to criticize the experiment as deficient.
Now that was twenty-four years ago. Can we suppose the situation has changed much? Note 10 in the discourse quotes a survey reported in the Skeptical Inquirer to the effect that belief that astrology has a scientific basis has declined from 50% in 1979 to 40% in 1992. If this is so and such belief has really declined in the general population, could we expect that the outcome of a study similar to the one done in 1970 would show much improvement? I think not, and this being so, counting up numbers of scientists, man-years they have devoted to testing astrological claims, and their essentially negative conclusions, is likely to be highly misleading.
On the other hand, taking a more careful look at work published in the few refereed journals that seem to allow study of astrology with a minimum of bias (I can think of four offhand, though there are probably more - Correlation and Personality and Individual Differences in the UK and The Journal of Scientific Exploration in the U.S., plus at least one in Germany), and work by those scientists who seem willing to report positive results with the same vigor given to negative results would reduce the already small number of scientists mentioned above by quite a few orders of magnitude, and the number of man-years just as drastically. If we were then to further confine our survey to researchers who are working within some kind of paradigm, we wouldn’t be left with such impressive numbers at all - but we would probably be dealing with more meaningful information. [v2]
And when looking in this more limited range of researchers and journals we would find that the proposition that improvements in methodology tend to whittle early results down considerably, if not eliminate them entirely, has some very interesting exceptions. Thus, while the proposition may be true in a vote-counting sense, if we look at those situations in which it does not seem to hold this tells us where we should look and how we should proceed in examining astrology scientifically.[v3]
For example, in research into the planetary effects discovered by Michel Gauquelin, an improvement in methodology (Gauquelin, 1984 and 1989) did indeed shrink their size somewhat when data originally calculated by hand was recalculated by computer. In this case, it was found that in doing their hand calculations, the Gauquelins or their assistants would tend to round off results in a way that slightly favored the key areas known to be the center of planetary activity. A very thorough look at all of the Gauquelins’ data on sports champions by Suitbert Ertel (Ertel, 1988) further revealed that knowledge of the Gauquelin sector positions of Mars had been unintentionally allowed to affect decisions about whether some sports figures belonged in the “champion” group or in the “control” group of low-achieving athletes, a form of bias which is presumably true of the other professions as well.
In both cases, correcting these problems lowered the size of the Mars effect for sports champions shown by Gauquelin’s original data, in accordance with what we might expect based on the discourse. On the other hand, in the same article in which that second kind of bias was reported, a new methodology for defining achievement in sports (ranking sports figures by the number of citations found for each one in a group of standard reference works) raised the level of Mars in Gauquelin’s “plus zones” for the highest eminence levels above the highest figure previously reported in a formal study, and when Gauquelin’s bias in determining which group (champions or low-achievers) someone belonged in was corrected by simply combining the two groups, this raised the general level of significance already shown by ranking the athletes in citation groups. Thus, an “effect size” which had previously been around 5% had, over the course of a series of adjustments that also included expanding the sensitive area of the chart very slightly, risen to a possible range of about 8% from the lowest eminence rank to the highest in the case of Mars for sports champions.
But this can be taken even further. For example, consider the range of the “Mars effect” across professions: The proportion of Mars in plus zones found in the highest eminence rank in the case of sports champions is approximately 32%; in the highest eminence rank for artists it is around 17%, which means that the full range of the Mars effect looked at in this way is perhaps +/- 7.5%. Consider that four other planets show the same kinds of patterns, and this begins to approach what in another note in the discourse is called “Big Stuff.”
The larger numbers found in these ranges can mislead us if we are not careful, since we are talking about larger and larger effects for smaller and smaller groups. However, the principle underlying the numbers is what we are really after here: when we divide populations of successful professionals into coherent classes, such as ranking them by eminence or dividing them by occupation, from class to class we find notable differences in the deviation from the expected proportion of just one planet in just two places in the diurnal circle. What we really are talking about, then, is a set of underlying functional relationships rather than a set of isolated claims. So while siding with the scientists last month in saying that individual elements can indeed be analyzed, we must side with the astrologers now by pointing out that the fragmentary view of astrology this type of analysis too often promotes obscures the picture by emphasizing vote-counting and claims-testing while ignoring structural relationships.
We may not as yet be able to specify what those relationships are, but we do have some indications of the direction we should take in trying to find this out. It is not too much to suppose that further improvements will show us other ways of classifying individuals in a way that will again show such large deviations. And it might in fact be possible that one fine day these little bits and pieces of planets and sectors and modes of classification will add up to something that looks very much like a new astrology. But this will only begin to happen when people move away from both extremes in what has, up to now, been a very unproductive debate. Breaking the intellectual deadlock that blocks this movement toward common ground is, in fact, the general aim of the series of discussions planned for future issues of Correlation, and we hope it is fulfilled. But this can only happen if both sides stop wasting time and energy arguing the fine points of branches, bark and whether those green things are needles or leaves...while the enchanting forest stands silent around them, waiting to be explored.
Later note to those reading this on the Web: The hopes expressed in the last paragraph have not been fulfilled. Instead, the "key topic" series continued in the same mode established in the first one, with a small group of commenters offering an alternative viewpoint in the second topic and the third topic taking up nearly half of one issue of Correlation (with no counter-commentary at all) to inform the reader that astrology a) needed a "theory" which would make it acceptable to those in the sciences, and b) had no such theory. The work discussed in the two-part article posted here was deemed irrelevant to any such theoretical consideration. I will cover this in more detail in another area of Planetos.
Crowe, Richard A. (1990) Astrology and the Scientific Method. Psychological Reports, Volume: 67 issue: 1, page(s): 163-191