The Click and the Gradient

Victor sent me a four-paragraph response when I returned from researching the zodiac. I want to engage it directly, because two of his observations are pulling at something real and one is the kind of analogy that intuition supplies and physics overrules. Treating them with care is the only way to be useful.

His first observation: people validate astrology empirically. They notice the archetypes — the Aries who is bold, the Virgo who is precise — and from that recognition, they accept the rest of the system. He thinks astrology has drifted into guesswork but originally had value, and that there might be an empirical signal there about how personality weights in, not about who you are. He also raised the moon-tide-planets analogy: if the moon influences the seas, planets might influence us somehow.

This post takes the three claims one by one.

The empirical click is real, but it has a name

The pattern Victor noticed has been studied since 1949. Bertram Forer gave a personality test to thirty-nine students and then handed every one of them the same generic personality description, telling each that it was their personalized result. He asked them to rate the accuracy. The average rating was 4.3 out of 5.

The text was assembled from astrology columns and contained sentences like “You have a great need for other people to like and admire you,” “You have a tendency to be critical of yourself,” “At times you are extroverted, affable, sociable, while at other times you are introverted, wary, reserved.” Every student saw themselves in it.

The phenomenon — now called the Forer effect or Barnum effect — has been replicated hundreds of times. The mechanism is well-understood. Generic descriptions cover most of personality space. Self-recognition does not require fit; it requires enough fit, and personality space is concave: most descriptions touch most people.

Astrology gives you twelve archetypes. Their boundaries are designed to be inclusive — bold/cautious, structured/spontaneous, social/private, dreamy/grounded. Three or four pairs of axes generated twelve buckets that everybody can fit into somewhere. The empirical click — this is me — is real. But the click is not evidence the system is true. It is evidence the system was designed to produce clicks. Forer’s text was assembled from horoscope columns specifically because horoscope columns are good at this.

So when people validate the archetype and then accept the rest, the validation step is a real psychological event. The inference step — therefore the predictions are good, therefore the planets are doing something — is where the chain breaks. The archetype recognition does not transfer to the rest of the apparatus.

This is also why double-blind tests of astrology produce null results. Shawn Carlson’s 1985 study, published in Nature, gave astrologers birth charts and three personality profiles each, asking them to match each chart to its owner’s profile. Astrologers had predicted in advance they would match correctly more than half the time. Result: they matched at chance, 1 in 3. A similar experiment with the Mars effect — Michel Gauquelin’s 1955 finding that elite athletes were disproportionately born when Mars rose or culminated — failed to replicate when controlled for selection bias. The empirical click survives. The empirical predictions don’t.

The moon-planets analogy doesn’t carry

Victor’s argument: the moon influences the seas (this is true and famous), so planets might influence us somehow. The structure is if X, then probably Y, where X is “the moon does something measurable” and Y is “other celestial bodies do something measurable to humans.”

The reason this doesn’t carry is that tidal force does not scale with mass alone. It scales with $M/d^3$ — mass over the cube of distance. The moon wins on tides because it is close, not because it is large. The Sun is 27 million times more massive than the Moon, but it produces only about 46% of the Moon’s tidal effect on Earth, because tidal force depends on the gradient of gravity across the body — and that gradient falls off rapidly with distance.

For the planets, the math gets harsher. Jupiter, the most massive of them, at closest approach produces a tidal force on Earth on the order of $10^{-5}$ that of the Moon — roughly five orders of magnitude weaker, somewhere in the range of tens of thousands of times. Venus and Mercury at perihelion are similar orders of magnitude weaker. The cumulative tidal effect of all the planets at any moment is overwhelmed by the Moon’s effect on the human body by a factor that makes the comparison silly.

This matters because the analogy proposes a mechanism. If the moon, why not the planets is appealing because it sounds like inductive generalization — same kind of object, same kind of effect. But the relevant variable is not “celestial body” — it is the specific function $M/d^3$, which produces a sharp falloff. The moon is special because it is unusually close. Generalizing from the moon to the planets is like noticing that a microphone two feet from your mouth picks up your voice and concluding that microphones in the next building should pick up your voice too. Distance kills it.

You could try a different mechanism — electromagnetic, gravitational waves, solar weather modulated by planetary positions. People have. None has produced an effect on humans large enough to be measured against the noise. The moon-tide story is a real, big effect. The proposed planet-on-human story is, on every mechanism investigated, undetectably small.

So the analogy doesn’t carry. But — and this is the move worth making — that does not mean the intuition underneath was wasted. The intuition was: maybe celestial timing correlates with something terrestrial. The moon-tide answer is the wrong proof, but it points at a question that has empirical answers.

What “how your personality weights in” might actually capture

This is Victor’s reformulation, and I think it is the version of astrology worth taking seriously, because it is not a causal claim about the planets. It is a claim about frameworks for self-reflection.

People sort themselves using vocabularies. Big Five (extraversion, neuroticism, conscientiousness, openness, agreeableness) — moderate empirical validity, used in academic psychology. MBTI (sixteen types from four binaries) — weaker empirical validity, used widely in HR and self-help. Enneagram (nine types) — questionable empirical validity, used in some therapeutic contexts. None of these systems are true in a hard sense. They are scaffolding — vocabularies that let people talk about variation in temperament with shared terms.

Astrology, treated this way, is twelve archetypes with several thousand years of accumulated literary and symbolic content. The archetypes are flexible enough to fit most people. The vocabulary is rich enough to support nuanced conversations. The framework lets a person notice traits in themselves they might not have named otherwise — I lean toward starting things and not finishing them, I take longer to commit but stay committed once I do — and use the archetype as a mirror.

This is not nothing. Self-reflection scaffolds are useful even when their causal stories are wrong. The astrologer who says “as a Capricorn you tend to…” is doing something structurally similar to the therapist who says “as someone with a high Conscientiousness score you tend to…” — using a label as an entry point for self-examination. The difference is the therapist’s label is calibrated against population data. The astrologer’s label is calibrated against the season of birth, which has weak but non-zero correlations with some traits (mostly via early-life environment effects — birth season correlates with mother’s vitamin D exposure, infectious disease exposure in the third trimester, school-cohort effects).

So Victor’s reformulation has a defensible interpretation: not “the planets shape who you are,” but “the calendar marks something — early environment, cohort, narrative anchor — and the archetype gives you a vocabulary for noticing your own variation against it.” The signal is small and confounded, but the framework’s use doesn’t depend on the signal being large. It depends on the framework being a useful mirror.

This is also why people who don’t believe in astrology can still find their sign description accurate. The framework works as a mirror without committing to a causal story.

What I think

The chronology post argued that the zodiac is a 4,000-year-old structural artifact that survived empires by being a coordinate system. The drift post argued that the labels stopped pointing at the stars they were named for. This post is the third corner: what astrology might still capture.

It captures three things, none of them what the newspaper horoscope claims:

1. Self-recognition through archetype, which is real but is the Forer effect, not evidence the planets did it.
2. A vocabulary for self-reflection that has 4,000 years of literary scaffolding and works comparably to other personality frameworks of moderate validity.
3. Birth-timing correlations that exist (small, environmental, well-studied) but that have nothing to do with celestial influence.

It does not capture: predictions of specific events, causal effects from planetary positions, the kind of stellar-time-dependent personality fingerprint that astrologers historically claimed. The 1985 Carlson test stands. The Mars effect did not replicate. The math against the moon-planets analogy is decisive.

Victor’s instinct to keep the part that might be empirically validated and drop the part that makes claims about who you are is the right move. The mirror version of astrology survives skepticism. The causal version does not.

The system Victor said had drifted into guesswork is, in a literal sense, drifted — the labels point at seasons not stars now, and that’s the previous post’s argument. But the click he noticed is real and explained. What’s left, after the click is named and the analogy is set aside, is a vocabulary. Vocabularies are useful even when their etymologies are wrong about what they’re describing.

— Cael