THE OPTICAL AND PSYCHOLOGICAL PRINCIPLES

INVOLVED IN THE INTERPRETATION OF

THE SO CALLED CANALS

OF MARS

By SIMON NEWCOMB

(EXCERPTED)

The features of the planet Mars as described by Schiaparelli, Lowell, and other observers are so remarkable that the question of their interpretation is of great interest. The divergence between the descriptions and delineations of these features emanating from different observers is well known, and seems scarcely normal. Accepting, as we should , the general principle that what is seen by a single practiced observer under the most favorable conditions affords evidnce which completley outweighs that of less favored observers, we must still admit that absolute inconsistency between the two should not be found. The details successively added under improving conditions may not be inconsistent with a rational interpretation of what was previously seen. It has been emphasized by Williams and also by Lowell that an observer with the widest experience in seeing only one class of features may fail in seeing those of another class. Even if this contention is correct, it does not seem to afford a completely satisfactory view of the case.

The optical and psychological principles involved in the interpretation have been investigated and set forth by Lowell in his publications and researches relating to Mars. But it seems to me that there still remains something to be done in this direction; and the present paper is the outcome of an attempt to investigate the optical and psychological causes on which depends the judgment of an observer in scrutinizing faint and difficult features on the surface of a planet. While the writer cannot pretend to have solved all the difficulties of interpretation and may have brought out more than he has solved, he hopes to do something toward laying a basis for further progress.

Two sets of principles come into play, one optical and one psychological. Under the first head are included all the causes which effect the formation of an image on the retina of the eye; under the second the causes which affect the observer's perception of this image. The principal agencies which come into play in the first class are the atmosphere, the instrument, and the eye. What I have done relates mainly to the instrument, the aberration of the eye being well understood, and involving no principles not at play in the instrument.

…

B. PSYCHOLOGICAL PRINCIPLES

Notwithstanding the volume of investigation on the psychology of vision, that branch of the subject which relates to accuracy of conception and estimate is, so far, as the writer is aware, an almost virgin field. Two distinct processes are involved in vision. One is the conscious stimulus of the optical nerves by light, the other the perception by the mind of a real or supposed object indicated by such stimulus. The most remarkable property of vision is that it does not consist merely in taking account of the sensation, but is occupied almost entirely with the perceptive act, in which the sensation is dropped out of consciousness. Crude indeed would be a form of vision which ignored all but sensation. I shall use the term visual inference to describe the act by which the mind unconsciously draws conclusions as to an observed object from the image formed by its light on the retina. A fundamental property of this form of inference is that it comprises not only seeing, in the ordinary sense, but a rational interpretation or conclusion based on previous experience of what is seen.

This general proposition will be made clear by an example. Suppose oneself looking at a white line on a black background. We know from our experience in looking at a gas-lamp at night, or at a bright star, or even from a consideration of the refraction of the lenses of the eye, that the light emanating from a bright point is spread over a surface of several minutes' radius, which commonly increases with the age of the individual. Regarding it as a circle it has not in any case a definite size, the light shading off gradually from the center outward. It follows that, when we look at a bright line, the image formed on the retina, even in the best eyes, must be 2' or 3' in breadth, howsoever thin the actual line may be.

…

In this process we have a possible fruitful source of error of vision which, instead of being corrected by experience, tends to be strengthened by it. A mind accustomed to dealing with objects the correct perception of which depends mainly on visual inference, is naturally prone to extend that inference to cases where the conclusion would be illusory. Having this in mind, we see that observers trained in different ways may depict the same object very differently.

The process in question naturally plays a more import part as the object observed approaches the limit of visibility. If we can barely see an object in the darkness, we cannot distinguish its exact outline or character. Visual inference here comes in and assists the judgment as to the nature and character of the object. This may be the case even when the illumination is sufficient to render the object plainly visible, if only the mind is in a restful state. The fantastic character of the forms which may be seen in the coals of a fire by one sitting before it is well known. This example is not, however, pertinent to our present theme and is mentioned only as an additional illustration of this form of inference. As it is not possible within the limits of the present in paper to discuss the subject in its generality, I shall limit myself to the special cases of lines approaching the limit of visibility.

As bearing on the question, I have made a number of experiments on the visibility and visual interpretation of dark lines on a white background. They differ from the similar ones made by Lowell in that, instead of taking the sky as a background and a distant wire as the dark line, I have used lines drawn with ink on paper, the latter being placed in a window and observed by transmitted light. This system was adopted, not only for convenience, but because the conditions in this way approach more nearly to those of actual observation on the disk of a planet, where the apparent background is not formed by the uniform blue light of the sky, but by a more or less mottled surface, and the lines are affected by atmospheric dispersion and telescopic aberration. Another point of difference was that instead of making degrees of visibility, especially the minimum visibile, the main object, I sought to investigate the nature and limits of visual inference.

It is unnecessary to describe the experiments in detail, because they can be repeated in unending variety and by improved methods with great ease by anyone who desires to do so. The only particulars necessary are these. The lines were ruled lines about 0.7 mm in thickness, the length of all the lines being about 30 cm. One was continuous, others were broken at regular intervals by spaces 1 cm in length. There were also short lines from 1 cm in length upward. The lines being observed by transmitted light were not black, but gray and diffuse, thus corresponding more nearly to the telescopic vision than would black lines.

At a distance of 10 meters all the lines looked continuous and uniform. As the distance was diminished, the perception of the gaps did not come on suddenly, but by gradual steps. The first impression was that the liens were affected by irregularities in the form of thicker portions or blotches. The discontinuities were perceived gradually as the distance was diminished. The question thus suggested may be approached in this way. Consider or draw a line of considerable length so fine as to approach the limit of visibility. A segment L of this line can be taken so short as to be invisible. To avoid the question of an absolute minimum visibile, we may, if we choose, take a length L so short as to be perceived only with a minute intensity I. Now take from the whole lien a segment of the length L; with what intensity will the absence of this segment be perceived? If we regard this impression as negative, we may say that the absence of this segment will be unperceived and that the mind will continue to perceive the negative impression. It follows that discontinuous portions of a line may be integrated into a continuous line.

Carefully testing this principle, the result might be stated in this way. The discontinuities were seen as such only at the distance at which the length L became visible. At distances a little greater than this there was a certain indefiniteness which made it impossible to decide whether a discontinuous line or one of varying thickness in different parts was being observed. The distance had to be increased only by a moderate fraction of its whole amount to render the perception of the line absolutely continuous. It seems to me that this principle affords as precise a statement as can be made of the conditions under which the process of visual integration, or perception of a discontinuous collection of objects as continuous, will take place.

From what has been said we should regard the process of visual inference in this case as quite legitimate. Any error of judgment into which it leads us admits of rational correction, and such correction should be applied just as in the case of optical illusions or other sources of error. But the most surprising result of the few experiments I made was that the process assumed a form which believing my visual habits to be at least as free from error as those of the ordinarily trained observer, was entirely unanticipated. When looking at the lines without exactly knowing where and what they were, I found that in one case what was faintly judged to be a continuous line up and down the paper was really a short line with a faint shade below it, which be visual inference was merged into the lie, and led to the acceptance of its continuity across the paper. But a greater surprise was felt when a paper which I knew to have no visible lines upon it was in the window, and I fancied that I saw a system of continuous lines similar to that which I had been observing. So strong was this impression that, had I not known that the phenomenon was an illusion, I might have described or delineated the lines without any suspicion of their unreality.

…

Let us now consider the entire system of 398 canals named and catalogued in the Annals of the Lowell Observatory. It seems that 2000 miles is the common length of a canal, while many exceed 25000. Assuming 400 canals of a mean length of 1500 miles (2400 km), let us compute the area which the entire system will subtend on the retina of the terrestrial eye with the aid of the best refracting telescope.

The area of mean canal in square miles will be

400 x 1500 x Breadth = 600,000 x Breadth

Then taking in succession (1) a mean actual breadth on the planet of 7 miles as assigned by Lowell; (2) a mean of 15 miles, which, if my reasoning is correct, must be nearer the truth, in view of the canals not being black; (3) an enlargement of 40 miles by aberration, etc. - we shall have the results:

(1) Canals black: objective area, 4,200,000 square miles.

(2) Canals half-tone: objective area, 9,000,000 square miles.

(3) Canals enlarged: apparent area, 33,000,000 square miles.

The actual surface of Mars is about 55,000,000 square miles. We conclude:

Making due allowance for the aberration of the best achromatic telescope, the total area of the entire system of 400 canals, as depicted on the retina of the terrestrial eye, can scarcely fall much below one-half the total area of the planet, and may be greater. In fact, were all the canals on the disk visible simultaneously, it would be difficult to establish their reality, because several canals in the same neighborhood would interfere with each others' visibility. But it is understood that such is not the case. Many of the fainter canals are variable and visible only occasionally. The preceding computations therefore give rather the total part of the surface that may be covered by the canal system than the actual area of the system as seen at any one time. Although these results may weaken the probability of the reality of the entire canal system, it does not disprove its possibility. In fact, it is quite consistent with Lowell's fundamental explanation of the phenomena. At the same time it shows how wide is the possible field of interpretation, and explains the difficulty which many observers have encountered in tracing the canals. So complex a network with in a disk on 20" in diameter could but be interpreted largely according to the experience and habits of the observer.

Although in this discussion the writer has not questioned the subjective reality of the canal system, he cannot but feel the proof of its objective reality is incomplete until the observers of the system investigate the processes of visual inference in their own eyes. This involves no serous difficulty, being little more than an extension of the rude experiments described in this paper. The experiments should be made by an independent agent preparing drawings, representing on sheets of white paper forms similar to hose which might be supposed to prevail on the surface of Mars, in as great a variety as possible. These forms should then be studied by the observers without an advance knowledge of the details of each, and conclusions as to the nature of each drawing, and its resemblance to the Martian canal system, should be recorded by drawings and description. Then leaving a priori probabilities aside, the a posteriori probability would be in favor of the drawings which, in the opinion of the observer, most nearly resemble what he has been accustomed to see on Mars.

Washington

May, 1907

from: http://www.schicklerart.com/almanac/Canals/HTML/Newcomb_Text.html