«People and Events Behind the Fractal Images When asked to write this article I, without space limitation, have unleashed a flood of recollections ...»
People and Events Behind the Fractal Images
When asked to write this article I, without space limitation, have unleashed a flood of
recollections about some men and some ideas involved in the art of fractals, including
both art for art’s sake and art for the sake of science. A few of these recollections may
even qualify as history, or perhaps only as what the French call le petite histoire. As
some readers may already know, history for me is forever a part of the present.
The Prehistory of some fractals-to-be: Poincaré. Fricke & Klein, and Escher To begin, while fractal geometry dates from 1975, it is in many ways important to know that a number of shapes now called fractals have been known for a much longer time.
But surprisingly few had actually been drawn before the computer era. Most were selfsimilar or self-affine, and represent the artless work of the draftsmen on the payroll of science publishers. Also, there are renditions of physical and simulated Brownian motion in the book by Jean Perrin, Les Atomes, and William Feller’s Introduction to Probability.
These renditions have helped me dream in fruitful ways (as told in my 1982 book Figure The Fractal Geometry of Nature p. 240), but they are not beautiful. Fractals-to-be-occur in the work of Fatou and Julia circa 1918, but they led to no illustrations in their time.
However, Poincaré’s even earlier works circa 1890 do include many sketches, and two very different nice stories are linked with illustrations that appeared shortly afterwards, in the classic book Vorlesugen ber die Theorie der automorphen Functionen, Fricke & Klein 1897. This book’s text and its preface are written by Robert Fricke, but we read on p. vi of the book that the great Felix Klein, “a teacher and dear friend,” seems to have graciously consented to having his name added on the title page.
The illustrations became even more famous than the text. They have been endlessly reproduced in books on mathematics, and, for better or worse, have affected the intuition of countless mathematicians.
A tenacious legend claims that students in industrial drawing at the Technische Hochule in Braunschweig, where Fricke was teaching mathematics, drew these figures as assignment, or perhaps even as an exam. Unking words have been written about some of the results.
In fact, I have done my share of detailing the defects of those figures which claim to represent the fractal-to-be limit sets of certain Kleinian groups (leading some to wonder which of Fricke’s students should b failed posthumously). These two dubious figures were drawn with the help of the original algorithm of Poincaré, which is very slow, too slow even for a computer. However, my paper in The Mathematical Intelligencer, M 1983m, has given an explicit and quick new algorithm. The comparison is summarized in Figure The Fractal Geometry of Nature, page 179. As was to be expected, the actual shape is by far the more detailed and refined of the two. But this is not all: against all expectations, it is not necessarily perceived as more complicated. I feel it is more harmonious, and can be comprehended as a whole, therefore it is perceived as far simpler than the clumsy old pictures. However, a famous mathematician (15 years my senior) has expressed dismay at seeing the still vibrant foundation of his intuition knocked down by a mere machine.
Of wider popular interest by far are Fricke’s drawings of “hyperbolic” tessellations,” the reason being that they have become widely popular behind diverse embellishments of Morits C. Escher, as seen, for example, in the book Image The World of M.C. Escher. Many people immediately perceive some “obvious but hard to describe” connection between Escher and fractals, and it is good to know that these tessellations are indeed closely related to fractals.
In fact, they were knowingly triggered by Poincaré, as is well documented by H.S.M. Coxeter in his 1979 Leonardo paper. Having seen some of Escher’s early work, this well-known geometer wrote to him and received the following answer: “Did I ever thank you...? I was so pleased with this booklet and proud of the two reproductions of my plane patterns!... Though the text of your article [in Trans. Royal Soc. Canada, 1975] is much too learned for a simple, self-made plane pattern-man like me, some of the illustrations.... gave me quite a shock... Since a long time I am interested in patterns with “motives” getting smaller and smaller till they reach the limit of infinite smallness... but I was never able to in which each “blot” is getting smaller gradually from a centre towards the outside circle-limit, as [you] show... I tried to find out how this figure was geometrically constructed, but I succeeded only in finding the centers and radii of the largest inner-circles. If you could give me a simple explanation..., I should be immensely please and very thankful to you! Are there other systems besides this one to reach a circle limit? Nevertheless,... I used your model for a large woodcut.” This was his picture “Circle Limit I,” concerning which he wrote on another occasion: “This woodcut Circle Limit I, being a frist attempt, displays all sorts of shortcomings.” In his reply, Coxeter told Escher of the infinitely many patterns which tessellate a Euclidean or non-Euclidean plane by black and white triangles. Escher’s sketch-boks show that he diligently pursued these ideas before completing Circle limites II, III, and IV. He wrote: “In the colored woodcut Circle Limit III, most of the defects [of Circle limit I] have been eliminated.” In his Magic Mirror of M.C. Escher (1976), Bruno Ernst wrote: “best of the four is Circle Limit III, dated 1959... In addition to arcs placed at right angles to the circumference (as they ought to be), there are also some arcs that are not so placed.” [Now going back to Coxeter] “In fact all white arcs ‘ought’ to cut the circumference at the same angle, namely 80 degrees (which they do, with remarkable accuracy). Thus Escher’s work, based on his intuition, without any computation, is perfect, even though his poetic description of it was only approximate.” The reader is encouraged to read Coxeter’s paper beyond these brief quotes. But an important lesson remains, and deserves to be restated. The Coxeter pictures, which made Escher adopt the style for which he became famous, hence eventually affected the aesthetics of many of out contemporaries, were not the pure creation of an artist’s mind.
They came straight from Frick & Klein, they were largely inspired by Henri Poincaré, and they belong to the same geometric universe as fractals. Also note that the preceding story is one of only two in this paper to involve a person who had been professionally trained as an artist.
The Fractal Mountains of R. F. Voss
My next and very pleasant task is to tell how I met Voss and some other people important to the story of the Art of Fractals.
During the Spring of 1975, Richard F. Voss was hopping across the USAA in search of the right job. He was soon to become Dr. Voss, on the basis of a Berkeley dissertation whose contents ranged from electronics to music, without ever leaving the study of a widespread physical phenomenon (totally baffling then, and almost equally baffling today), called 1/f noise. Other aspects of this noise, all involving fractals, were favorite topics of mine since 1963, and my book Figure Les objets fractals, which was based on a generalization of 1/f noise from curves to surfaces. One of the more striking parts of Voss’s thesis concerned (composed) music, which he discovered had many characteristics involving 1/f noises. He had even based a micro-cantata on the historical record of Nile River discharges, a topic dear to my heart.
Therefore, Voss and I spoke after his job-hunting talk at IBM, Yorktown, and I offered him a deal: come here and let us play together; something really nice is bound to come out. He did join the Yorktown low-temperature group and we soon became close co-workers and friends. Contrary to what is widely taken for granted, he never joined my tiny project, and he spent the bulk of his time on experimental physics. Nevertheless, his contribution to fractals came at a critical juncture, and it has been absolutely essential.
First, we talked about writing a book on 1/f noise, but this project never took off (and no one else has carried it out, to my knowledge, to this day). Indeed, each time he dropped by, to try and do something together, he found me involved with something very different, translating and revising Figure Les objects fractals. The end product came out in 1977 as Figure Fractals. There were many graphics problems in its preparation. Voss ceaselessly inquired about what Sig Handelman and I were doing, and kept asking whether we would consider better ways. Then he found a sure way of obtaining our full attention: he conjured a computer graphics system where none was supposed to exist, and brought along pictures of fractals that were way above what we had been dealing with until then. They appeared in Fractals, which is why the foreword describe him as the coauthor of the pictures in that book.
Color came late to Yorktown, where it seems we fractalists continued to be the only ones to use demanding graphics in our work. We first used color in my next book, the 1982 Figure Fractal Geometry of Nature. In late 1981, the text was already in the press, but the color pictures had not yet been delivered to the publishers. The film recorder we were using was ours only on a short lease, and this fact and everything else was conspiting to make us rush, but I fought back. Since ‘the desire is boundless’ (Figure FGN, p. 38), I fought hardest for the success of the Fractal Planetrise on the book’s jacket. It was soon refined to the point of what (by the standards of the day) was perfection, but this was not enough. Just another day’s work, or another week’s, I pleaded, and we shall achieve something that would not need any further improvement, that would not have to be touched up again when the “lo-fi” graphics of the day were to be replaced by the “hi-fi” graphics. To my delight, Voss was also a perfectionist.
Fractal illustrations had started as wholly ulitarian; the perceived beauty of the old ones by Jean-Louis Oneto and Sig Handelman was an unexpected and unearned bonus.
Buy by 1981 their beauty had matured and it deserved respect, even from us hard scientists, and it deserved time. Many people have, since those days, showed me their fractal pictures by the hundreds, but I would have been happier in most cases with fewer, more carefully worked out ones.
Everyone experiences wonder when observing Voss’s pictures, and “to see [them] is to believe [in fractal geometry].” Specialists also wonder how these pictures were done, because, without ever drawing specific attention to the fact, Voss has repeatedly conjured technical tricks that were equivalent to computer graphics procedures that did not officially develop until much later. This brings to mind a philosophical remark.
Watching Voss the computer artist, and Voss the physicist at work for many years had kept reminding me of the need for a fruitful tension between the social and the private aspects of being a scientist. The only civilized way to be a scientist is to engage in the process of doing science primarily for one’s private pleasure. To derive pleasure form the public results of this process is a much more common and entirely different matter. The well-known danger is that, while dilettante, is a term of contempt. While not a few individuals profess to be serious scientists, yet many are motivated primarily by personal enjoyment of their work, very few could provide what I view as the only acceptable evidence of “serious dilettantism.” This demonstrates a willingness and, perhaps, even a compulsion to leave significant portions of one’s best work unpublished or unheralded – knowing full well that one could claim no credit for them. This may be easiest for the scientific giants; lars Onsager was a legend in this respect. On the other hand, every scientist has been the active or the passive witness of episodes when one could not or would not work in a field without becoming thoroughly uncivilized. The true test, therefore, arises when civilized behavior is neither easy nor impossible. On these, and other stringent grounds, I view Dick Voss (as graphics expert and as physicist) as being one of the most civilized serious scientists in my wide acquaintance.
Old Films: Competing with the Good Lord on Sunday
What about films? We were ill-equipped to produce them, having only an exhausted film recorder (a tube-based Stromberg Carlson 4020) at our disposal. In 1972 with Hirsh Lewitan, however, we did prepare a clip on the creation of fractal galaxy clusters, using the Seeded Universe method. Then, in 1975, with Sig Handelman, we added a clip on fractal mountains: the landscape later used as Plate 271 of Figure The Fractal Geometry of Nature emerged slowly from the deep, then rotated majestically (or at least very slowly), and finally slipped back under water. Everyone spontaneously called this the Flood Sequence. By a fluke, the highest altitude achieved at two distinct points, and a programming flaw stopped the Flood when these points were still visible. Delighted, I indulged in commenting that my fractal model of relief had predicted that there were two tips to Mount Ararat, not one... until an auditor straight from Armenia reported very dryly that this fact was well-known to everyone in his country.
The Galaxy Clustering and the Flood sequences, taken together, were nicknamed Competing with the Good Lord on Sunday. They soon came to look out-of-date and pitifully primitive, but now they look good again: they are of historical interest...
In the Flood, the observer simply moved around a landscape without zooming.
The same was true in the animation of one of Voss’s data bases, done by R. Greenberg Associates for an IBM commercial clip.