Galileo’s Dialogue: Four Days, Three Speakers, and the Argument He Got Wrong

Salviati does not dispute that bodies have three dimensions. He asks why three dimensions → completeness → perfection should follow necessarily. The weakness is the transition, not the conclusion. This is the move he repeats all book: Aristotle defined nature this way, therefore nature must behave this way is not an argument. The world has to be investigated, not deduced from definitions.

Salviati challenges the classification rather than the observation. Straight downward movement is matter moving toward an equilibrium arrangement; once that arrangement is reached, continued straight motion makes no sense. Circular motion can continue indefinitely without carrying a body away from its ordered position. The categories become straight movement establishes an arrangement, circular movement can preserve one — which no longer marks a difference in kind between heaven and Earth.

The argument hides an enormous assumption. That bodies fall toward the Earth demonstrates a tendency toward the Earth. It does not demonstrate that the Earth’s centre is the centre of everything. A stone could fall toward the Earth wherever the Earth happened to be.

Salviati attacks the observational premise directly: “the two new stars of 1572 and 1604, which were indisputably beyond all the planets”, plus “many comets generated and dissipated in places above the lunar orbit” (Day 1, p. 58), plus sunspots appearing and decaying on the Sun itself. Then the clever turn: Aristotle himself holds that sensory evidence should be preferred where experience contradicts reasoning, so accepting observed celestial change is the more Aristotelian position. The complaint is against blind Aristotelianism, not against Aristotle.

The reply is that parallax measurement, repeated telescopic observation and consistency between independent observers make the dismissal progressively harder to sustain. Day 3 returns to this with the 1572 nova in detail.

Salviati attacks the anthropocentrism. How could anyone know that enormous celestial bodies exist primarily for us? The argument confuses usefulness to humans with the purpose of the cosmos, and it argues from presumed purpose rather than from observation.

Now the offensive. If the Moon is unquestionably celestial and the Earth resembles it, the absolute division weakens. The resemblances are specific: both spherical, both opaque bodies lit by the Sun, both showing light and shadow, both with uneven surfaces, mountains and valleys, and comparable illumination patterns. The Moon starts to look less like a perfect celestial sphere and more like another world.

Telescopic light and shadow show irregularities, and the shadows near the terminator change exactly as they would if mountains rose above valleys (Day 1, p. 55). The reasoning is geometrical rather than merely visual: sunlight plus relief predicts a shadow pattern, and the predicted pattern is what appears. That breaks the equation celestial = perfectly smooth.

The illumination changes as the Sun’s angle changes, in exactly the way real relief would produce. Preserving a smooth exterior requires an increasingly artificial auxiliary hypothesis to explain shadows that a rough exterior explains directly.

When the bright crescent is visible, the dark portion is sometimes faintly visible too (Day 1, pp. 80, 103). The explanation is a chain: Sun to Earth to Moon — the Earth reflects sunlight onto the Moon. If that is right, the Earth behaves optically exactly like another celestial body, which is the conclusion the whole day has been building toward.

Everything in this day hangs off this one proposition. Every objection below attacks it, and the two general principles answer the whole class of objections rather than any single one. The claim itself is never demonstrated here — Day 2 is defensive, and its achievement is to show that the terrestrial evidence does not refute it.

Every day the Sun, Moon, planets and stars appear to travel east to west. Two explanations produce an identical appearance: the Earth stands still and the whole celestial sphere revolves around it every twenty-four hours, or the Earth rotates west to east and the heavens need no daily revolution at all. Salviati argues that moving one relatively small body is simpler than moving the entire universe daily.

Salviati concedes it outright: simplicity is not a demonstration. It establishes probability and elegance, not necessity. The concession matters because it shows the book distinguishing plausibility arguments from conclusive demonstrations — a distinction most of the objections on both sides ignore.

The known motions show a pattern: smaller orbit, shorter period; larger orbit, longer period. The Moon takes a month; Jupiter’s satellites lengthen outward; the planets lengthen outward. But on the traditional account the largest sphere of all, the stellar sphere, must whip round in twenty-four hours. That is disorderly. If the Earth rotates instead, the anomaly disappears.

The central conceptual move of the book. Before release the stone is already moving with the Earth — as are the tower, the observer and the air. On release it does not lose that horizontal motion; it combines the shared horizontal motion with the new vertical fall, and relative to the tower it comes down essentially straight. The observation everyone agrees on is preserved, and the inference from it collapses.

The stone already shares the ship’s forward motion, so it lands at the foot in both cases — which makes the ship an argument for the possibility of the Earth’s motion. Salviati also dismantles the mechanics: if the rock followed the ship, the effect “would have to be attributed to the air, and not to the impressed force” (Day 2, p. 174), which is not what the objector believes. The objection is inconsistent with its own physics.

The passage the whole day exists for. “Shut yourself up with some friend in the main cabin below decks on some large” ship, with “some flies, butterflies, and other small flying animals”, a bowl of fish, a bottle dripping into a vessel beneath (Day 2, p. 215). Observe everything at rest. Then have the ship move uniformly. Nothing changes: the flies do not pile against the rear wall, the fish swim as easily in every direction, the drops still fall into the vessel, jumping is no easier one way than the other. Experiments performed entirely inside a uniformly moving system cannot reveal that motion. The stipulation is exact and flagged in the text — you must be below decks, “for if this took place above in the open air”, the air would not follow the ship (Day 2, p. 218).

The same answer: the ball, the cannon and the air all share the eastward motion before firing, and the vertical motion is superimposed on it. No large displacement should be expected.

Salviati identifies the recurring error precisely: the opponents imagine the Earth moving but forget to give the Earth’s motion to the things standing on it. Gun, ball, target and air all share it, and the ball does not begin from cosmic rest.

The expected deviation is nothing like the gross displacement the naive version predicts.

The calculated difference is far too small for contemporary artillery to reveal, and the gun and target share the motion in any case. Embedded here is a methodological rule the book uses repeatedly: failure to observe an effect is evidence only if the predicted effect is large enough for your instrument to detect. That principle does more work in Day 3 than anything else on this list.

The atmosphere shares the Earth’s motion, and the bird begins with that motion already in it, so flying relative to the surrounding air requires no catching up. The bird’s own self-generated motion is distinguished from the shared terrestrial motion it possesses before it takes off.

A smoothly moving ship feels like a stationary one, and looking out makes the shore appear to move instead. Perception alone cannot distinguish the observer moving relative to the environment from the environment moving relative to the observer.

The supposed effect is given a mathematical treatment: it is not enough to say rotation is rapid, because whether material is extruded depends on the rotational radius, the speed, and the strength of the competing tendency toward the Earth. The inference from “rapid rotation” to “catastrophic ejection” does not follow.

Why should one body have only one natural motion? The Moon moves around the Earth while the Earth–Moon system could move around the Sun. What counts as a body’s motion depends on the system being considered.

A suggestive classification rather than a proof, and presented as such. The bodies that unquestionably move are those that do not shine with their own light; the Sun and fixed stars are luminous and are the ones held fixed. The Earth is dark and opaque like the planets. Assigning motion by physical resemblance puts the Earth with the planets, not with the Sun. Analogical, probabilistic, and offered at that weight.

Day 3 turns positive: instead of deflecting objections it offers evidence. Each observation below supports this claim, and the parallax objection attacks it directly — the only objection in the book that is never answered.

A new star appeared in Cassiopeia (Day 3, p. 326). If the heavens are incorruptible it should not have, so defenders tried to place it below the Moon. Parallax measurements put it enormously far away — in the celestial region. What is methodologically interesting is how the argument is conducted: rather than simply asserting better numbers, Salviati works through how the anti-nova calculations selected and rejected observations, arguing that Chiaramonti’s preferred measurements were internally less consistent than those supporting a distant nova. The target is selective use of evidence, not merely a rival result. The same opponent’s Anti-Tycho is picked over in Day 2 as well (pp. 287, 311).

The strongest observational argument in the book. Mercury and Venus never wander far from the Sun in angle, and Venus runs a full cycle of phases, appearing “horned when beneath the sun” and changing shape as the Moon does (Day 3, p. 389). It is “certain that Venus and Mercury must revolve” about the Sun (Day 3, p. 373). Simplicio is walked through the geometry himself: Venus cannot have an orbit enclosing the Earth, or it would reach opposition; cannot always sit between Earth and Sun, or the phases would not work; cannot always lie beyond the Sun, or it would never show a crescent. Therefore it circles the Sun.

It “cannot be comprehended that Mars does increase sixty”-fold on the traditional account (Day 3, p. 433). If Mars orbits the Sun on a larger orbit than the Earth’s, it is sometimes near and sometimes far, and a large variation follows directly. The same reasoning applies in weaker form to Jupiter and Saturn, whose proportional distance variation is smaller.

Mercury, Venus, Mars, Jupiter and Saturn all revolve about the Sun, which becomes the natural centre of the planetary revolutions. The question then reverses: why should the Earth be the one bizarre exception? This is a unification argument — Copernicus makes the Earth one member of a system, rather than a unique object requiring its own physics.

Planets sometimes appear to reverse against the stars. Geocentric astronomy reproduces this with deferents and epicycles — dedicated geometrical machinery. On a moving Earth it is simply what overtaking looks like: the Earth on a faster inner orbit catches and passes an outer planet, which appears to slide backwards, exactly as a slower train appears to when you pass it. The planet does not literally reverse. Retrograde motion stops being a special construction and becomes a consequence.

They establish that the Sun is not unchanging, that the Sun itself rotates, and — through the changing tilt of their tracks across the disc over a year — carry geometrical information about the relation between the Sun, the Earth and the orbital planes. Finocchiaro gives this an unusually large subsection, and it is the argument in Day 3 that most nearly bears on the Earth’s annual motion specifically.

If the stars are distant enough, the Earth’s orbital diameter is negligible against that distance and the shift falls below what any contemporary instrument could register. So no detectable parallax is not no parallax.

Bright objects acquire halos and rays that enlarge their apparent size, so naked-eye stellar discs are not measurements of physical diameter. Observing through a telescope, or through a restricted aperture, shrinks the apparent disc — which is what you would expect from irradiation rather than from a resolved surface.

The reply is that human imagination cannot set an upper limit on what God may create; calling the universe “too large” assumes human intuition constrains divine scale. The symmetry is exact and worth pausing on: Simplicio deployed divine power against the simplicity argument in Day 2 (p. 142), and here the same move is turned around against the objection from cosmic vastness. The same argumentative form serves both sides, which is a good sign it is settling nothing.

Gilbert’s magnetic philosophy is discussed near the end (Day 3, p. 464). Its relevance is indirect: it makes the Earth look like a physically organised body with real properties capable of maintaining orientation and motion, rather than a passive heap at the centre. It is a supportive analogy about what kind of thing the Earth is.

The attempted knockout. If the tides are produced by the compound of the daily and annual motions, then a terrestrial phenomenon requires a moving Earth — the physical demonstration Copernicanism had never had, and the one thing a Tychonic system could not accommodate.

Sea level rises and falls regularly; water also moves horizontally; and different places differ in both the strength and the timing of the effect. Something causes this.

The Earth has a daily rotation and an annual revolution. Because these combine differently through the daily cycle, a point on the surface undergoes a continually varying effective speed. Water in a moved container sloshes when the container changes speed — put “a floating ball” into “a basin of water” and move the basin (the analogy is set up at Day 3, p. 462 and applied at Day 4, p. 487). The ocean basins are the container; the compound motion is the change in speed; the tides are the slosh.

It reverses the whole book. Day 2 showed terrestrial phenomena do not disprove the Earth’s motion. Day 3 made it astronomically plausible. Day 4 would show a terrestrial phenomenon that requires it — the physical demonstration Copernicanism had never had, and the one thing a Tychonic system could not accommodate.

The rejection is explicit and it is a rejection on principle: there is no rope, no contact, no mechanical connection, and to Salviati the proposal resembles astrology and occult sympathy. Kepler’s lunar explanation is dismissed largely for that reason.

Tides differ enormously by location because oceans are not identical containers: depth, shape, length, orientation, coastlines, channels and the communication between seas all matter. This is physically sensible and modern tidal science agrees — basin geometry, resonance and bathymetry strongly modify local tides. The error is in the primary forcing, not in the recognition that geography shapes the outcome.

The reasoning is extended to air: if the Earth’s motions move water, they should have atmospheric consequences too, and there are “strong winds blowing continuously from the east” to point at (Day 4, p. 503), with a contrast drawn against occasional winds that “blow indifferently toward all” quarters (p. 510). Consistent extension of the mechanism — and wrong for the same reason.