What Astro Images Do

To the Editors:

I was of course pleased to see my last book reviewed [“Revelations from Outer Space,” NYR, May 21], and gratified at the generally favorable assessment of Cosmigraphics by astrophysicist Priyamvada Natarajan.

There are nevertheless a few elements of Dr. Natarajan’s piece that require comment. She writes that the book renders the history of cosmology in such a way that it seems to have “evolved seamlessly and without intellectual conflicts and ruptures.” In support of this she states flatly: “The trial of Galileo over his interpretation of the solar system isn’t mentioned.” But his trial under threat of torture, and his 1633 recantation of heliocentrism, is indeed addressed in some detail in Chapter 5, among other places. So for that matter is Giordano Bruno’s conviction, and burning at the stake in 1600; Copernicus’s reluctance to publish his findings until the last months of his life, due to fear of ridicule or worse; and numerous other examples of contention, schism, and dispute throughout the history of cosmological thought, reaching back as far as Aristarchus’s finding that the sun is at the center of the solar system, an idea overruled in favor of Aristotle’s geocentric design.

But my major demurral concerns Dr. Natarajan’s evaluation of my assertion that scientific discoveries are made within images. She labels this idea “Benson’s fantasy.” In fact, Cosmigraphics documents several examples of such discoveries. These include oceanographer Marie Tharp’s midcentury assembly of pointillistic sonar data in pictorial form, leading to her discovery of a chain of rift valleys running down the center of the mid-Atlantic ridge—substantial empirical proof of the plate tectonics theory. And cosmologist Richard Gott’s discovery of a wall of galaxies 1.38 billion light years long—1/60th the diameter of the visible universe—as he and researcher Mario Jurić used conformal projection techniques to plot thousands of galaxies from the Sloan Digital Sky Survey. Gott and Jurić became aware of the Sloan Great Wall while creating their astonishing logarithmic map of the cosmos over a decade ago. As with Tharp, cartographic convention was decisively and thrillingly upended as mapmaking led to discovery, rather than the other way around.

More recently, astronomer R. Brent Tully, a leading researcher into the astrophysics of galaxy clusters, perceived the full extent of a dynamic flow of some 30,000 galaxies, including our own—a motion due to an enormous gravitational well comparable to a terrestrial watershed. He did so while studying a graphic supercomputer simulation of a cube of space some 500 million light years across. As Tully confirmed to me months before his “Laniakea Supercluster” was unveiled in Nature last year, he and his collaborators’ discovery was only possible because of that visualization. Here as elsewhere, graphical representation came before discovery, permitting the discernment of structure, and thus meaning.

Throughout history, images have been critically important tools in our ongoing attempts to understand the cosmos. No less than words and mathematical equations, graphic representations bear knowledge, convey theories, and yes, lead directly to scientific discoveries.

If this is fantasy, let’s have more of it.

Michael Benson
Visiting Scholar
Center for Bits and Atoms, MIT Media Lab
Cambridge, Massachusetts

Priyamvada Natarajan replies:

Michael Benson’s comment raises an important question about what astronomical images do and what they are for. While we both agree that images are tools, where we differ is on what kind of tool they are and what part they play in the scientific process. Benson conceives of astronomical images as constitutive of scientific discovery. To extend this claim to all of science, as Benson does, makes what is an occasional rarity the rule. Graphical representation most often does not come before discovery but rather, after it. Visualization of previously acquired complex data of course might lead to further insights and discoveries, but images do not create meaning independently of scientific frameworks or models.

Take, for example, my work in astrophysics using what is seen—galaxy light—to map the unseen—dark matter. My research relies on images taken by the Hubble Space Telescope that record the systematic distorted shapes of galaxies produced by light bending due to the presence of dark matter. These images serve first and foremost as data, data that only become scientifically significant when interpreted through Einstein’s Theory of General Relativity. Einstein’s General Relativity, it is worth noting, was not inspired by these or any other images, it was derived from equations using the language of mathematics.

Images help scientists conceptualize our ideas and in particular, they play an important part for nascent ideas that require much more fleshing out in science. Visualizing data, especially utilizing computer tools to see three-dimensional renditions, does refine scientists’ understanding. Images, however, are more often than not important and critical post-facto representations of our current best-to-date understanding.

For instance, my colleague Brent Tully’s realization of the shape of the supercluster Lanikea did not arise from just the rendition on his computer. He has painstakingly gathered detailed astronomical data on the positions and speeds of galaxies in this entire region of the sky over decades. He knew of the existence of an excess of galaxies from his observational data before his image and computer visualization came into the picture. It was when he plotted his data and used a three-dimensional rendering that the shape, orientation, and scale of the supercluster emerged. Without the astronomical data that he had gathered prior, there would be no basis to explore and have figured out the existence of this incredible superstructure in our cosmic backyard.

Thus visualization is only part of the scientific process that includes other important prior steps. Images are not in general generative ab initio of ideas; they are often a consequence and do provide new conceptions. There are instances when they play a more pivotal part, such as the one I discussed featuring Kip Thorne and Interstellar: the equations that describe the bending of light around black holes and wormholes are well known. It is only after these known complex equations were rendered for the movie that Thorne noticed new additional effects that he and others had been unable to visualize earlier.