Occaam's Razor, Cheap Heuristic

Occaam’s razor is often used as a cheap heuristic+: all else being equal the simpler of two arguments should be selected. A famous exception is provided by the Shapley-Curtis Debate in astronomy between Harlow Shapley (Princeton) and Heber Doust Curtis (UVA), which took place on April 26, 1920. This debate and others show that Occaam’s Razor often fails in decision-making under uncertainty.

M64: The Black-Eyed Galaxy

Messier 64 (M64, NGC 4826) is the famous Black Eye galaxy, sometimes also called the "Sleeping Beauty galaxy." The conspicuous dark structure is a prominent dust feature obscuring the stars behind. An observer situated with this dust cloud would only be able to see other galaxies by looking up and down along the axis of the galaxy.

In an unexpected twist, recent observations have shown that the gas in the outer regions of this photogenic spiral is rotating in the opposite direction from all of the stars!

Credit: NASA and The Hubble Heritage Team (AURA/STScI)
Acknowledgment: S. Smartt (Institute of Astronomy) and D. Richstone (U. Michigan) STScI is operated by the Association of Universities for research in Astronomy, Inc. under a contract with the National Aeronautics and Space Administration.

Shapley argued that spiral nebulae such as the Andromeda galaxy were nearby gaseous clouds analogous to what we now call globular clusters, and like these clusters they orbited the only galaxy, the enormous Milky Way Galaxy, as satellites. Curtis placed our Sun near the center of a relatively small Milky Way Galaxy, and argued that the Milky Way and Andromeda were merely two of many galaxies in the universe separated by vast distances. Telescopes had not yet reached the level of sophistication required to discern the details of the spirals or to provide accurate measurements of their distances and rotational speeds. Both eminent astronomers in this debate were working with fragmentary and sometimes faulty data.

Shapley used estimates that Andromeda rotated .02 seconds of arc per year. If Andromeda were as far away as Curtis claimed, that meant rotational speeds faster than the speed of light: an impossibility. Shapley argued the apparent high rotational and recessional velocities for the spirals were simply the result of the spirals being close to the Milky Way galaxy. Also, if Andromeda were a separate galaxy it should have photometric and spectral characteristics similar to the Milky Way, but the Milky Way was apparently much less bright than Andromeda. Key to Shapley’s argument, most spirals seemed to avoid the Galactic plane; they were found preferentially near the Galactic poles. Shapley argued that the Milky Way exerted a dark repulsive energy on the nearby spirals and there would be fewer spirals along the galactic plane because they were forced farther away by the extreme breadth of the Milky Way along the galactic plane.

Curtis countered that many of the spirals showed a dark thin band in their midplane, and that if the Milky Way had the same dark band and we were situated in that band, we would only see spirals looking out along our own Galactic pole. At the time, the idea that a galaxy’s brightness was due to its core and not its spiral arms was unknown, as was the fact that our position in the Milky Way was on one of its spiral arms giving us a reduced measurement of the Milky Way brightness. Curtis could not explain the erroneous high recessional velocities and rotational speeds measured for some of the spirals.

The debate ended in a draw. Shapely had the simpler argument but relied on an assumption of a dark energy of unknown origin to repulse the spirals. Curtis had what turned out to be the correct argument. Andromeda, and indeed all of the spirals were galaxies similar to the Milky Way and located far away. But Curtis needed two assumptions (dark band and our location in that band in the Milky Way galaxy) and could not explain some of the phenomena (the incorrectly calculated rotational speed of the spirals, and the apparent differences of the photometric and spectral characteristics of Andromeda). Occaam's razor failed in this major scientific debate: the simpler argument was simply incorrect.

This discussion was excerpted and reworded from Physical Universe: An Introduction to Astronomy