While it is true that most practicing astrologers do not really care for the search of scientific evidence definitively proving astrology, it is also true that most capable scientists do not care at all for astrology as such. Not surprisingly, neither helps the research on the subject. In the current state of affairs, it is entirely unclear how to distinguish a good astrologer or astrological theory from a bad one. As a result, there is a growing body of studies of unclear validity, which is however sufficient in the sustaining of skeptical negativism, while most astrologers wouldn't really care about any results of any tests in the first place. The atmosphere is a one of mutual hostility and various forms of denial on either side. Since this is not a constructive approach, I intend to try, as both an astrologer and a social scientist, to meet currently unresolved critical issues with both theoretical analysis and practical proposals.
LET'S LEARN FROM WHAT WORKED
Among all of the scientific tests conducted so far, the most successful was definitely Gauquelin's Mars Effect study, which was replicated repeatedly and only stopped working, as it appears to be the case, in the samples collected from the 50s onward. After all of the supposed statistical artifacts were shown not to be causing the result, the only problem left is objectively the fact that it stopped working at some point in time. This lead the skeptics to conclude it is indeed an artifact, caused by people misreporting the time of birth precisely because of the planetary positions being known and some being somehow preferred. This is currently officially a sufficient explanation of why it didn't prove astrology, however I am really unsure where does that confidence come from.
First of all, there is no proof of the exact strength of the effect, even if it was the case, which clearly makes it speculative; secondly, assuming this took place to a significant degree, it is not the only potentially relevant variable that changed in mid-twentieth century, for instance there is no reason to expect that a link between any astrological aspect and any objective phenomenon will remain exactly the same forever or anywhere, just like it is commonly accepted in the whole of social sciences; and thirdly, the same effect didn't show in the samples of general population from before 50s, as it did in the samples of elite sportsmen, which to me still sounds like it would require the parents to somehow know what their children will grow up to be.
Let's look at the practical problems and their potential solutions, both outlined in comparison of what Gauquelin did right and others did wrong:
1) Objectivity and personal knowledge problem
It is an important thing to understand how objectivity is problematic here, and so I will explain it from the standpoint of astrological theory, or at least the best approximation of it that I am capable to produce. Astrology is, as a discipline, entirely qualitative in what it describes and how. That is a fundamental difference between it and behavioral social science, the paradigm which uses quantitative statistics to prove things about people, groups and behavior - the paradigm of all of the tests done so far. It doesn't mean that they are mutually entirely incompatible, but the difference must be accounted for in the design of a study.
As it has been the case most often, the studies done so far tend to assume that astrology is supposed to predict objective phenomena, such as IQ, occupation, marital status, etc., measured entirely by incidence (how often something objective happens out of all cases). As Gauquelin's studies show, astrology may predict objective incidence, but, and I cannot emphasize this enough, it is not supposed to. Astrological observations directly describe qualitative states, such as why would someone do something or how would they experience it, not what exactly they eventually decide to do, or what exactly will happen to them to cause the predicted qualitative state. The genius of Gauquelin's choice of variables, in my opinion, is that a success, though itself an objective phenomenon, is very closely bound to the issue of personal qualities.
Conversely, the as of yet still unpublished Dean's comparative study of time twins, often cited as the definitive proof against astrology, uses mostly objective outcomes as the basis of comparison, which, as I have explained above, are not even supposed to work within the framework of astrological theory. There certainly are astrologers who think that they can directly predict objective traits or events using astrology, and there could be a way to do so better than chance, however the studies typically show them to be overconfident, and therefore not realistic about it or sufficiently skilled.
The opinion based studies frequently show also the general inability of people to identify their own birth chart. This leads me to propose to try and go in the opposite direction and eliminate personal opinion from the studies as much as possible, unless an adequate test of astrological ability or self-knowledge can be devised specifically for the purposes of the testing of astrology. For the same reason, the other kind of comparison used by Dean in his phantom study - the opinion of parents or teachers about the subjects' personality and behavior - is not expected to be of any merit either, if the body of research into the gullibility and misjudgment of people is to be taken into consideration, as it definitely should be, even when it goes against the skeptical position, for a change.
2) Sampling and random averaging problem
The next fundamental difference between what Gauquelin did and most other studies is the way in which he worked with samples. The standard practice in behavioral social sciences is to use random samples of people, emphasizing all that is average about them. In terms of astrological theory, inner qualities of people should show the better (more often) on the outside (as behaviors or events), the more extreme they are. It is quite logical and straightforward assumption, and relatively easy to incorporate into research design, yet most scientists have not chosen this route. If you compare Mars Effect between the sample of successful sportsmen and a sample of general population of average people (in terms of success in sport), not only is the effect stronger among sportsmen, but gradually grows with their eminence - the better they are, the stronger the effect.
That is a clear indication that the emphasis in research design should be given to situations, where extreme qualities are linked to extraordinary behaviors or outcomes, rather than to common behaviors of people unremarkable in terms of personal quality. What is especially to be avoided is to sample people with both averaged out natal charts and objective behavior, such as the common ill-advised sampling using only students of the same age, intelligence level and interests, like Carlson (1985) did in his famous study, to the point where even the individual results of CPI tests of the subjects were indistinguishable. In principle, it is a way how to substantially reduce the vagueness of the astrological claims subjected to experimental verification, and with this increase in precision, increase in significance of observed effects is to be expected.
Some astrologers would argue that astrology is too complex for such high precision predictions to be made, while some skeptics would say the same, only in their case because of how vague astrology is. I would address both by saying that yes, there is a lot of nuance to be accounted for, but ultimately, as Gauquelin's tests imply, there are ways to control for the variability in outcome that astrology theoretically allows. First of all, Gauquelin in all likelihood didn't start from astrological theory in terms of meaningful interpretation of what rising or culminating Mars should mean. Rising or culminating planets are simply something that is clearly observable and is supposed to be more significant than other positions of planets.
That in itself is enough to try to look for some correlation of something objective with it. It could have turned out to be something else than professions, or more interestingly, different professions in relation to the same planets, or vice versa. Which is not to say the meaning wasn't there, but skeptic not versed in astrology would not appreciate that Mars linking to sportsmen is in accordance with traditional astrological lore, which didn't have to be the case. Today, different professions may be linked to the same planet, as the social status of sport changes in time or between cultures. Within astrology, the eternal qualitative principle is specifically that Mars is related to a strongly competitive activity - maybe businessmen fulfill that role better today, or pro gamers. That is how one should go about it from meaning to observation.
Which finally brings me to the next step - the elite samples do not have to be constructed on the basis of a shared objective trait, they could be constructed on a shared rare and/or extreme astrological configuration instead. Since both natal chart and behaviors are objectively observable, they can be compared equally well even if the thing shared in the sample is a notable similarity in horoscope. The complete lack of appreciation for this approach by the researchers so far shows precisely the lack of understanding of astrological theory and its nuances, necessary for truly viable research design. An example of this would again be the time twin study apparently conducted by Dean, where over 2 000 subjects were involved, all in the sign of pisces.
If you are an astrologer who sees where I'm going with this, you may be rolling your eyes right now. If you are a skeptical statistician, you really can't. If you are looking for exact objective similarities between people born at the same time, the sign of pisces is probably the worst choice in the whole zodiac. If you are looking for similarities in behavior or physical attributes, it is much more logical to go for signs of fire or earth, action or materialism, rather than the most mysterious, transcendental and weak-willed sign of them all. A nuance completely lost on non-astrologers, or astrologers who don't really analyse what they're doing. You could also compare on the basis of what elements are missing from the chart, unlike anyone until now. Knowing what should be more detectable outwardly on the basis of astrological theory is really important, not just an excuse, especially when you are selecting a sample.
3) Eternal principles in changing times problem
One of the fundamental assumptions of science in general is that you can replicate initial conditions for an experiment, or more precisely, that changing quality of time from moment to moment is not an issue. When a physicist tests for laws, he or she is looking, invariably, for eternal, immutable laws. As even a brief glimpse into the conventional social sciences shows, psychological or social laws do not seem to work that way, which is why theories in social sciences typically specify a range within which they are supposedly valid - age limits, time periods, localities, etc. For a social science theory to work for decades in numerous countries, like in Gauquelin's study, that's not bad, actually.
Astrology very much assumes that times change (never repeat exactly, in fact), even though qualitative principles behind it all don't change, but it is not an insurmountable obstacle in terms of research controls. In astrology, the quality of a moment in time or a time period is objectively defined by the observably changing positions of planets in signs or houses, which means that they can be selected and compared. For example, the slow moving outer planets remain more or less the same for years (if you generalize them as being in particular signs), making them easily controllable in short term, while precise (within 5 degrees) angular distance aspects (mainly 0, 90, 120 and 180 degrees) between planets are clearly much stronger (and therefore significant) than other angular distances of planets in a horoscope. Partial combinations do even repeat, and certainly individual planets periodically return to each degree of the horoscope.
What should be done about it? At minimum, this should be noted - when the experimental comparison takes place in terms of where planets are relatively to each other in the charts of subjects of the experiment and in those of times of events that are analysed. Attempts at replication of various hypotheses should also be made ideally at times when some more essential planetary aspects repeat. It will never be a perfect and complete match, but there are numerous ways in which smaller or greater matches can be differentiated. The hypotheses to be tested should also theoretically account for how planetary alignments are supposed to affect them.
For instance, if a study focuses on how Mars affects people, it could be a good idea to only include people into the sample who were born when Mars was strong, perhaps in the sign it rules, aries, and not in a weakening conflict to other planets. It is not too much to control. Just like Gauquelin focused on planets rising or culminating, which are the strongest positions for planets to be within the 24 hour cycle, especially for Mars rising (since the first house is the house of aries, ruled by Mars). The Mars here really conforms to astrologically predictable interpretation, which gives me hope it can be done for other planets individually or aspects of planets as well. At least in synchronous studies, rather than diachronic or longitudinal ones.
The time variable is really fundamental in astrology, and if anything may be the physical causal mechanism behind it, it would most likely be something to do with time itself. Much like randomizing computer algorithms are not really random, but instead hiding an intrinsic ordering structure, which may even be based on some system of meaning - "when" something happens can have a "why" to it, even assuming that the system is somehow self-assembled, like human language. Which leads me to another fundamental unprovable assumption of science today - that meaning is not a natural property, and that, somehow, subjective experience of consciousness is not real in a physical sense. I am not saying this assumption is definitely false, but you need to suspend it when you are trying to test astrology.
The research design needs to assume that time itself has a universally meaningful quality to it, and then be devised from there. As modern quantum mechanics suggest, time may also flow in both directions, so you shouldn't necessarily assume that astrological cause has to precede the outcome in the conventional direction of time. In terms of what it means in practice, creative tests can be designed which account for the more exotic implications of astrological theoretical foundations. Test may focus on comparison of experience rather than behavior. The replication may actually be planned in advance as several independent studies of one hypothesis done at the same time, instead of in a sequence. If it is in a sequence, try to predict how planetary configuration should affect the latter experiments. Prediction may also be tested on several independent groups, some of which would receive major stimulus after the study is conducted, not before or during. Just whatever you do, always mind the time.
That's enough for now. I will present more problems and propose more solutions next time.
LET'S LEARN FROM WHAT WORKED
Among all of the scientific tests conducted so far, the most successful was definitely Gauquelin's Mars Effect study, which was replicated repeatedly and only stopped working, as it appears to be the case, in the samples collected from the 50s onward. After all of the supposed statistical artifacts were shown not to be causing the result, the only problem left is objectively the fact that it stopped working at some point in time. This lead the skeptics to conclude it is indeed an artifact, caused by people misreporting the time of birth precisely because of the planetary positions being known and some being somehow preferred. This is currently officially a sufficient explanation of why it didn't prove astrology, however I am really unsure where does that confidence come from.
First of all, there is no proof of the exact strength of the effect, even if it was the case, which clearly makes it speculative; secondly, assuming this took place to a significant degree, it is not the only potentially relevant variable that changed in mid-twentieth century, for instance there is no reason to expect that a link between any astrological aspect and any objective phenomenon will remain exactly the same forever or anywhere, just like it is commonly accepted in the whole of social sciences; and thirdly, the same effect didn't show in the samples of general population from before 50s, as it did in the samples of elite sportsmen, which to me still sounds like it would require the parents to somehow know what their children will grow up to be.
Let's look at the practical problems and their potential solutions, both outlined in comparison of what Gauquelin did right and others did wrong:
1) Objectivity and personal knowledge problem
It is an important thing to understand how objectivity is problematic here, and so I will explain it from the standpoint of astrological theory, or at least the best approximation of it that I am capable to produce. Astrology is, as a discipline, entirely qualitative in what it describes and how. That is a fundamental difference between it and behavioral social science, the paradigm which uses quantitative statistics to prove things about people, groups and behavior - the paradigm of all of the tests done so far. It doesn't mean that they are mutually entirely incompatible, but the difference must be accounted for in the design of a study.
As it has been the case most often, the studies done so far tend to assume that astrology is supposed to predict objective phenomena, such as IQ, occupation, marital status, etc., measured entirely by incidence (how often something objective happens out of all cases). As Gauquelin's studies show, astrology may predict objective incidence, but, and I cannot emphasize this enough, it is not supposed to. Astrological observations directly describe qualitative states, such as why would someone do something or how would they experience it, not what exactly they eventually decide to do, or what exactly will happen to them to cause the predicted qualitative state. The genius of Gauquelin's choice of variables, in my opinion, is that a success, though itself an objective phenomenon, is very closely bound to the issue of personal qualities.
Conversely, the as of yet still unpublished Dean's comparative study of time twins, often cited as the definitive proof against astrology, uses mostly objective outcomes as the basis of comparison, which, as I have explained above, are not even supposed to work within the framework of astrological theory. There certainly are astrologers who think that they can directly predict objective traits or events using astrology, and there could be a way to do so better than chance, however the studies typically show them to be overconfident, and therefore not realistic about it or sufficiently skilled.
The opinion based studies frequently show also the general inability of people to identify their own birth chart. This leads me to propose to try and go in the opposite direction and eliminate personal opinion from the studies as much as possible, unless an adequate test of astrological ability or self-knowledge can be devised specifically for the purposes of the testing of astrology. For the same reason, the other kind of comparison used by Dean in his phantom study - the opinion of parents or teachers about the subjects' personality and behavior - is not expected to be of any merit either, if the body of research into the gullibility and misjudgment of people is to be taken into consideration, as it definitely should be, even when it goes against the skeptical position, for a change.
2) Sampling and random averaging problem
The next fundamental difference between what Gauquelin did and most other studies is the way in which he worked with samples. The standard practice in behavioral social sciences is to use random samples of people, emphasizing all that is average about them. In terms of astrological theory, inner qualities of people should show the better (more often) on the outside (as behaviors or events), the more extreme they are. It is quite logical and straightforward assumption, and relatively easy to incorporate into research design, yet most scientists have not chosen this route. If you compare Mars Effect between the sample of successful sportsmen and a sample of general population of average people (in terms of success in sport), not only is the effect stronger among sportsmen, but gradually grows with their eminence - the better they are, the stronger the effect.
That is a clear indication that the emphasis in research design should be given to situations, where extreme qualities are linked to extraordinary behaviors or outcomes, rather than to common behaviors of people unremarkable in terms of personal quality. What is especially to be avoided is to sample people with both averaged out natal charts and objective behavior, such as the common ill-advised sampling using only students of the same age, intelligence level and interests, like Carlson (1985) did in his famous study, to the point where even the individual results of CPI tests of the subjects were indistinguishable. In principle, it is a way how to substantially reduce the vagueness of the astrological claims subjected to experimental verification, and with this increase in precision, increase in significance of observed effects is to be expected.
Some astrologers would argue that astrology is too complex for such high precision predictions to be made, while some skeptics would say the same, only in their case because of how vague astrology is. I would address both by saying that yes, there is a lot of nuance to be accounted for, but ultimately, as Gauquelin's tests imply, there are ways to control for the variability in outcome that astrology theoretically allows. First of all, Gauquelin in all likelihood didn't start from astrological theory in terms of meaningful interpretation of what rising or culminating Mars should mean. Rising or culminating planets are simply something that is clearly observable and is supposed to be more significant than other positions of planets.
That in itself is enough to try to look for some correlation of something objective with it. It could have turned out to be something else than professions, or more interestingly, different professions in relation to the same planets, or vice versa. Which is not to say the meaning wasn't there, but skeptic not versed in astrology would not appreciate that Mars linking to sportsmen is in accordance with traditional astrological lore, which didn't have to be the case. Today, different professions may be linked to the same planet, as the social status of sport changes in time or between cultures. Within astrology, the eternal qualitative principle is specifically that Mars is related to a strongly competitive activity - maybe businessmen fulfill that role better today, or pro gamers. That is how one should go about it from meaning to observation.
Which finally brings me to the next step - the elite samples do not have to be constructed on the basis of a shared objective trait, they could be constructed on a shared rare and/or extreme astrological configuration instead. Since both natal chart and behaviors are objectively observable, they can be compared equally well even if the thing shared in the sample is a notable similarity in horoscope. The complete lack of appreciation for this approach by the researchers so far shows precisely the lack of understanding of astrological theory and its nuances, necessary for truly viable research design. An example of this would again be the time twin study apparently conducted by Dean, where over 2 000 subjects were involved, all in the sign of pisces.
If you are an astrologer who sees where I'm going with this, you may be rolling your eyes right now. If you are a skeptical statistician, you really can't. If you are looking for exact objective similarities between people born at the same time, the sign of pisces is probably the worst choice in the whole zodiac. If you are looking for similarities in behavior or physical attributes, it is much more logical to go for signs of fire or earth, action or materialism, rather than the most mysterious, transcendental and weak-willed sign of them all. A nuance completely lost on non-astrologers, or astrologers who don't really analyse what they're doing. You could also compare on the basis of what elements are missing from the chart, unlike anyone until now. Knowing what should be more detectable outwardly on the basis of astrological theory is really important, not just an excuse, especially when you are selecting a sample.
3) Eternal principles in changing times problem
One of the fundamental assumptions of science in general is that you can replicate initial conditions for an experiment, or more precisely, that changing quality of time from moment to moment is not an issue. When a physicist tests for laws, he or she is looking, invariably, for eternal, immutable laws. As even a brief glimpse into the conventional social sciences shows, psychological or social laws do not seem to work that way, which is why theories in social sciences typically specify a range within which they are supposedly valid - age limits, time periods, localities, etc. For a social science theory to work for decades in numerous countries, like in Gauquelin's study, that's not bad, actually.
Astrology very much assumes that times change (never repeat exactly, in fact), even though qualitative principles behind it all don't change, but it is not an insurmountable obstacle in terms of research controls. In astrology, the quality of a moment in time or a time period is objectively defined by the observably changing positions of planets in signs or houses, which means that they can be selected and compared. For example, the slow moving outer planets remain more or less the same for years (if you generalize them as being in particular signs), making them easily controllable in short term, while precise (within 5 degrees) angular distance aspects (mainly 0, 90, 120 and 180 degrees) between planets are clearly much stronger (and therefore significant) than other angular distances of planets in a horoscope. Partial combinations do even repeat, and certainly individual planets periodically return to each degree of the horoscope.
What should be done about it? At minimum, this should be noted - when the experimental comparison takes place in terms of where planets are relatively to each other in the charts of subjects of the experiment and in those of times of events that are analysed. Attempts at replication of various hypotheses should also be made ideally at times when some more essential planetary aspects repeat. It will never be a perfect and complete match, but there are numerous ways in which smaller or greater matches can be differentiated. The hypotheses to be tested should also theoretically account for how planetary alignments are supposed to affect them.
For instance, if a study focuses on how Mars affects people, it could be a good idea to only include people into the sample who were born when Mars was strong, perhaps in the sign it rules, aries, and not in a weakening conflict to other planets. It is not too much to control. Just like Gauquelin focused on planets rising or culminating, which are the strongest positions for planets to be within the 24 hour cycle, especially for Mars rising (since the first house is the house of aries, ruled by Mars). The Mars here really conforms to astrologically predictable interpretation, which gives me hope it can be done for other planets individually or aspects of planets as well. At least in synchronous studies, rather than diachronic or longitudinal ones.
The time variable is really fundamental in astrology, and if anything may be the physical causal mechanism behind it, it would most likely be something to do with time itself. Much like randomizing computer algorithms are not really random, but instead hiding an intrinsic ordering structure, which may even be based on some system of meaning - "when" something happens can have a "why" to it, even assuming that the system is somehow self-assembled, like human language. Which leads me to another fundamental unprovable assumption of science today - that meaning is not a natural property, and that, somehow, subjective experience of consciousness is not real in a physical sense. I am not saying this assumption is definitely false, but you need to suspend it when you are trying to test astrology.
The research design needs to assume that time itself has a universally meaningful quality to it, and then be devised from there. As modern quantum mechanics suggest, time may also flow in both directions, so you shouldn't necessarily assume that astrological cause has to precede the outcome in the conventional direction of time. In terms of what it means in practice, creative tests can be designed which account for the more exotic implications of astrological theoretical foundations. Test may focus on comparison of experience rather than behavior. The replication may actually be planned in advance as several independent studies of one hypothesis done at the same time, instead of in a sequence. If it is in a sequence, try to predict how planetary configuration should affect the latter experiments. Prediction may also be tested on several independent groups, some of which would receive major stimulus after the study is conducted, not before or during. Just whatever you do, always mind the time.
That's enough for now. I will present more problems and propose more solutions next time.
No comments:
Post a Comment