French, Tuesday Night at Velis 402 -- French Appendix

"The French Appendix" by Lane Slate, excerpted by permission from

A Brief History of the French Academy of Improvisational Graphics

© PGC Press 2005; abridgement and annotations © 2006 L.S.

Café de Graphique

The story of the Academy starts 60 years ago in Paris soon after World War II. At the time a night club called Café de Graphique was enjoying great popularity, being much the hot bed of drinking, thinking, and viewing of local art on the walls. The Café in turn was greatly enjoyed by one of its frequent patrons, Jean Paul Lagrange.

Jean Paul was a young philosopher/mathematician studying at the Sorbonne. He found the atmosphere of the Café de Graphique a welcome respit from the rigors of his classwork. In particular, he liked the dynamic patterned imaging of the Abstract Films shown on Tuesday movie nights. JP especially liked the works by Oskar Fischinger and Len Lye. And then by chance the management of the Café decided to take advantage of another artistic trend of the time; they started having music nights on Wednesdays with live American jazz. Well, JP and his friends at the Café immediately loved the jazz.

So it came to pass that the same group of young French intellectuals were exposed to the two different art forms in close juxaposition, jazz and abstract film. With plenty of wine and spare time, JP and the artsy crowd with which he was hanging began to envision a synthesis of the two. Improvisation was in the air and they started to think seriously about live imaging ... jazz ... visual jazz ... something ... who knows, perhaps a live art form based on improvising moving abstract images?

Thus was born the Cercle de Graphique Improvisé, whose purpose was to explore exactly that artistic possibility. More wine and more spare hours later, the equipment hackers of the circle      like all performing graphics outlaws before and since, utilizing the technology available to them at the time, in their case rotating glass disks, viscous colored gels, and an old movie projector for a light source      had put together a visual instrument with which they gave impassioned if crude visual performances.


John Paul's Theoretical Task

Interestingly enough, while thoroughly digging how the improvised graphics looked, at the same time some members of the circle wanted to know what it all meant      Ah, the French!      and so John Paul Lagrange was conscripted to become their theory guy. He was flattered and not unwilling.

And for which task he actually had the pedigree as well as the inclination and the training. JP's remote ancestor was the mathematician Joseph Louis Lagrange, famous for his 1772 solution to a restricted version of Newton's three body problem (which solution has very practical consequences today for designing the orbits of satellites). With this heritage, and bolstered by the support of the Cercle, young Lagrange totally embraced his role of theoretician. He began looking around for other conceptual frameworks      they had to be French of course      which could serve as elements in the framework of his endeavor. And he found two.

Element one: Poincaré,  Algèbre Visuelle (Visual Algebra)

The first framework was the Algèbre Visuelle of Henri Poincaré. The Algèbre was a virtually unknown monograph written by the famous geometrician in 1910, two years before his death. Jean Paul was very familiar with Poincaré's work, not only because the latter was a great mathematician, but also because Poincaré had once claimed to have solved the general Newtonian three body problem. Just to be clear, this was not a wimpy restricted version like JP's ancestor Joseph Louis had solved, but rather the whole enchilada, the general problem. And so the Lagrange family was, and still is, enormously pleased when it turned out Henri had made a mistake and that his results were in fact erroneous. Only rappers have bigger egos than mathematicians, and even dead mathematicians can still be players (reputation is a family legacy, and protecting it a family project, passed down from generation to generation).

The Algèbre, more successful than Poincaré's failed three body solution though unfortunately not nearly as well known, came about because at day's end and tired of non Euclidean geometry and the relativity of space, Henri then immersed himself in the visually rich absolute 3D space of his town and times (in total conflict with his professional endeavors, and so? Even as Hume could question causality during the day in his study, but not at night when drinking wine and playing billiards with his friends). Thus comported, Henri wandered the streets of the city in retreat from the formal, just looking at stuff. He concentrated on relishing the images of 1908 Paris and leaving his geometrician's mind far behind. Which artifice worked for some time. He enjoyed of course the picturesque urban vistas, and also, especially, the paintings, soaking in early Cubism from the works of Picasso and Braque.

After a while, despite his best efforts to the contrary, Poincaré's analytical mind did creep back in. But with resolution, he simply would not allow it to deny or distort the visual imagery he found so delightful. Instead Poincaré used the framework geometrical to hold and appreciate the particular perceptual experience of each work while at the same looking for something in common behind them all. And thus was born the Algèbre Visuelle.

John Paul Lagrange was ecstatic when he discovered this monograph in the dusty basement archives at the Sorbonne. Of particular use to JP was the notation Poincaré invented for his visual equations. Poincaré used these equations to rigorously describe his favorite paintings in terms of the visual objects he saw therein and the interrelations among those objects. The unique notational technology of the visual equations permitted one to include the objects themselves in the equations, even as digits can be mentioned in the formulas of traditional algebra. And, not surprisingly given Poincaré's love hate affair with closed systems, at the center of the Algèbre was a paradoxically indeterminate formalism, which conflicted mechanism would later serve Lagrange's purpose very well indeed.

[ LS: Poincaré's visual algebra is truly an algebra on and of images! It is not just some scheme to use diagrams for teaching symbolic algebra; rather, it is a powerful para geometric formal tool for describing the structure, relation, quantity and tonal quality of regions in a visual field. And, because the objects of description can be themselves included in the equations, the equations are sometimes stunningly beautiful simply as visual objects on the page. Henri explored the capabilities of this tool off and on during the remaining two years of his life with mixed results. He did succeed in using the system to solve for image unknowns across a series of paintings. He also tried various formal definitions for emotional qualities in a visual work, ultimately failing (of course?) but in very interesting ways. Further explanation would be off topic and space does not permit; perhaps I will someday get the opportunity to explain the wonders of these still revolutionary concepts in another article. ]

Element two: Merleau Ponty,  Vécu corps comme instrument (Lived body as instrument)

The second French conceptual framework, actually more of a viewpoint, that JP commandeered to his cause arose out of an investigation carried out by the philosopher Maurice Merleau Ponty. In 1938 Merleau Ponty did a careful phenomenology of "Le corps vécu comme instrument d'expression artistique" (The lived body as instrument of artistic expression). Developed in close collaboration with a ballerina, that painstaking philosophical exploration was an examen minutieux (and sadly is, like the viz Algèbre of Poincaré, also little known today). Merleau Ponty's final report on the subject was published but once, just before the War, in Épistémologie Esthétique, an obscure journal printed on the Left Bank.

JP heard about the report through another member of the Cercle and tried very hard to find a copy. Unsuccessful in this venture, he then took a more direct approach and went to see the famous thinker himself, who was teaching at the Sorbonne at the time. JP came away from the interview not only with a copy of the essay, but also encouragement in his project (and some say the influence went the other way as well, M P later using some of JP's visual intuitions for his reflections on Semiotics in Signes, but JP himself denies this).

In any case, that copy of M P's "le corps comme instrument" essay (one of the last extant) has long occupied a place of honor in the Académie archives at Versailles. And the concepts therein have long been central to JP's thinking. One of M P's key observations was that only in performance does the lived experience of the body most fully deny detachment of mind from body, idea from material. Not tainting the thesis at all was the gossip about M P, the dancer, and the exact kind of performances he was investigating. Thus, as with Poincaré's work, in the writing of Merleau Ponty JP found much that would be of future use to him. Perhaps most important was the notion that in passionate performance mind and body are not two; this insight would later define the meditative practice at the core of the improvisational training for the Académie's Knights Visual. And JP likewise took to heart a direct corollary of not two: the centrality of consciousness as experienced in and through the body, and as such the proper starting point for philosophical investigations.

Element three: the Circle,  L'impératif d'improvisation graphique (The graphic improvisation imperative)

After this bit of homework, with two conceptual allies in hand (not only French, but as well fellow habitués at the Sorbonne, past and present), for a third and final constituent, JP chose not a formalism or a viewpoint, but rather in fact a conceit from the original genesis of the project: the simple artistic mandate to improvise visually, taken directly from the manifesto of the Cercle de Graphique Improvisé.


The Three Body Synthesis

So now young Lagrange was ready to proceed with his task of uncovering the meaning of graphique improvisé. It was at this seminal point in the project that genetics and/or family tradition kicked in. Young Lagrange saw the unresolved situation, so desperately in need of a synthesis, as a three body problem. Only for Jean Paul, the three bodies were not the Earth, the Sun, and the Moon      as they had been for his famous ancestor, Joseph Louis      but instead the visual equations from the Algèbre Visuelle, Merleau Ponty's description of the lived body as instrument, and the Cercle's graphic improvisation imperative.

Then for weeks Jean Paul pondered and sweated and envisioned. And finally, in an amazing moment of clarity he found the solution (or, at least one solution) which would bring the three elements into harmonious ensemble. At that moment, afraid the insight might as suddenly fade, JP feverishly started writing what would become his opus: a treatise describing precisely the dynamics of the three elements and how they played together.

Several important themes emerged as JP worked his thesis of synthesis; these motifs would then shape the next fifty years of his work and life.

Embodied visual performing consciousness

The first theme, following Merleau Ponty, was basically a viewpoint: start with the human's consciousness as experienced in and through the body. And this engendered a subtle yet profound change in JP's mission statement. Originally charged by the Cercle to "find the meaning of improvised graphics," under the influence of M P, JP then redefined that task to be "describe the phenomenon of a human improvising graphics"      and in so doing, totally transformed the project.

As a philosopher, JP did not have to approach the subject in the same way as the other members of Cercle. In fact, rather, it was his obligation as a philosopher not to accept other people's defintion of the problem. And so his view was quite different than the artists who came at it images first, or the equipment hackers whose first concern was the instrumentation. Instead, JP's investigation began in and with the consciousness of a human performer already in the act of graphic improvisation. Now remember that what philosophers like JP call "investigating a phenomenon," and which enterprise they tend to describe as if they were wandering around a piece of real estate and taking careful notes on what they observe, well, see, for lay people that kind of "investigating" looks an awful lot like just sitting at a desk in their study and thinking. So, in any event, as JP undertook this investigating/thinking/imagining, it was thanks to the viewpoint from M P that the phenomenon under scrutiny was "embodied visual performing consciousness." This choice of focus would serve JP very well as we will see, allowing his philosophical results to remain useful for over fifty years across many different changes in fashion both visual and technical. As Alan Kay says, "point of view is worth 80 IQ points" (plus or minus, I think, depending; JP picked a plus).

Manipulation precedes structure

JP's second theme was a direct consequence of the first. Starting his observations with the performer hot in the act of visual creation      and it is worth noting that in his "observing" (again, for the rest of us, a lot like just imagining), JP pictured that the performer could generate not merely the limited images supported by the Cerlce's crude instruments at the time, but in fact all the rich moving imagery displayed in Abstract Films on movie nights      so anyway, in observing this very rich act of visual creation, what JP found foremost in the performer's experience was the moment when creator and creation are still one. What is present most immediately in the performer's experience is the performing activity, the performer doing something, but then a split moment later, it gives itself as something visual, and thus the essence of graphic performing activity is revealed as doing something visual  shaping what is seen; manipulating graphics; flowing the stuff visuel. In the instant of that primordial visual time of "first light," the flowing is at the fore, and then appears the what: something visual is flowing.

And only then, subsequently, does careful examination of the flowing as visual manipulation allow the something to come forth and further reveal itself, and then that something presents itself as visual structure. JP using the method phenomenological found in the (act of) manipulating visual structure parts and wholes being created and shaped and moved and disappeared. And still more careful observation of the flowage would uncover finer and finer structures (Ah, the French! Ah, recursion!), leading JP to later say, "The first question is manipulation; ask it correctly and what is being manipulated will naturally show itself." This insight (and second major theme) is often summarized in the slogan "La manipulation précède la structure" ("Manipulation precedes structure"). And, as always, insisting performer and performed are not two, JP called this "une structure vécue" ("a lived structure"). He would then use the visual equations of the Algèbre Visuelle of Poincaré to describe this structure. Lived manipulation producing yet not separate from lived image structure.

[ LS: It is worth noting what did not appear in these investigations: the physical performing instrument. In his careful delineation of embodied "visual performing consciousness," JP saw that tools of this kind, performing tools of expression, disappear for the user in the flow of use. Two points need to be made here. First, disappearing in use did not mean unimportant. And second, from this viewpoint, "physical performing instrument" included the human body as well as constructued external mechanisms. Although the physical performing instruments disappear during use, yet because of the `embodied' carried down from M P, those instruments were always still there and simply awaiting description, both moving arms and fingers and eyes to control/perceive as well as equipment manufactured to manipulate/display. Which would serve JP so very well later when it came time to design exercises and meditations and devices for the Knights Visual. ]

Thus for Jean Paul the basic strategy of his work developed naturally after adopting first the human centric viewpoint of Merleau Ponty. With his investigation thus orientated, JP proceeded in the prescribed direction and described the terrain as it opened up: starting with the performer (the human), then finding the performing (visual manipulation), and lastly the performed (structured images). Leaving only remaining to be determined the finer structure of the performed images, which he would finally try to delineate using the tool of Poincaré (modulo a few little glitches which later turned out to be not so little, finally ripening into the very pesky Problèmes de Poincaré).

Text v graphics, don't!

In undertaking that determination of detailed image structure      an examination just after visual manipulation but just before "what"      JP the careful observer did not find yet text or graphics, but simply visual stuff. The final wrinkle in the final theme of JP's synthesis was the fundamental non distinction between text and graphics. That is, JP chose not to differentiate between text and graphics (a dubious distinction at best and one that usually fails under close scrutiny). Instead in the ms he ambiguously refers to a discernible visual chunk as a "text graphic image" which may in turn recursively be composed of other text graphic images until finally at the bottom are monochromatic visual "regions" (which may happen to be just dots or strokes in a Kanji or Hebrew or Arabic character form, or who knows? ... one observer's character is another's ink blot; Romanji text is just a family of 26 little squiggles plus layout rules.). JP thus avoided the distinction between text and graphics, or at the least deferred it for as long as possible      and then would make it most begrudgingly, restricted to concrete situations on a case by case basis, and for sure never absolutely once and for all independent of context.

Or so he hoped. But when he then tried to apply the tool of Poincaré for rigorous characterization of image structure, JP encountered a small problem in leaving the text graphic distinction to last: that was exactly the distinction which Henri had made first! Remember, Poincaré's original domain was painting, from which domain he had intentionally omitted text like objects. Undaunted, JP fumed and schemed and re read many times the basic formulation of the Algèbre Visuelle. At last he found a workaround, albeit a very unsatisfying one. It was basically an ingenious but dirty hack (of course Jean Paul wouldn't use the term "hack;" nonethless he would certainly admit to the method being far from clean.) The hack involved region grammars, relative sizing exceptions, and a subtle paradox in Henri's formal visual axioms. And so, using the workaround for the first of Les Problèmes de Poincaré (at the time JP, while he knew it as a problem, didn't yet realize its severity, nor that there were more in store), JP finally had the analytical technology to complete his painstaking portrayal of the fine structure of performed images using the Algèbre.

In this way, Lagrange was able to describe not only the visuality of what would be considered the content proper of abstract films, but also the title and introductory credits. When building ontologies (and phenomenological ontologies are no different), generality is power. JP wanted his ontology to handle all the imagery seen by the audience on film nights at the Café de Graphique, from first flicker to last (he went of course for the General Solution, not wanting to settle for some sissy "restricted version"      rappers and philosopher/mathematicians, hubris tempered by chutzpah, ya gotta love 'em).



S&M completed! "Une Structure Vécue Provenant de la Manipulation Improvisée pour des Images de Texte Graphique"

So, elements well in mind, themes well in hand, and still well powered by his original unifying epiphany (and the wine and women of Café de Graphique), John Paul Lagrange slaved away at his researches. Months later the tome of synthesis was at last finished, smoothly combining elements and themes in a deep way where they were not integrated but alloyed. He called it "Une Structure Vécue Provenant de la Manipulation Improvisée pour des Images de Texte Graphique" (A Lived Structure Arising Out of Improvised Manipulation for Text Graphic Images). JP then printed up a dozen or so copies and handed them around. And, believe it or not, most of the Cercle were still hanging in there with him, mostly. Meaning they all read the very dense Structure et Manipulation ms (or S&M as it was affectionally known within the crowd) and some of them could actually discuss the finer points with JP. Once again, Ah, the French!

The rest of the story is standard fare for a French (or any) intellectual movement. Petty squables and personal prejudices, publicly recast as major conceptual issues, begin to shape and reshape the theory. In 1951 Lagrange feuded with the main Cercle over sound and realism, declaring both to dilute the "graphique improvisés authentiques." Now it is true that, in support of his position, JP supplied some nicely reasoned arguments. But it certainly helped to motivate those arguments that his off again on again girlfriend at the time was a professional musician, and that her on again off again ex boyfriend, the problematic third body in what Jean Paul considered to be strictly a ménage à deux, was a popular painter of the visual realist variety.

So, not coincidentally at all, it was exactly at this point in time that JP had a further revelation, suddenly discovering that neither sound or realistic images had a place in the authentic improvisational graphics. The rest of the Cercle, however, didn't buy this abrupt conceptual shift, and so JP and the group parted theoretical ways.

The Villa and the Académie

Over the next few months relations worsened, and finally in 1952 Lagrange split from the Cercle geographically as well as philosophically, departing Paris and moving to nearby Versailles. Thus JP got to leave town in a huff and yet still be only 15 klicks from his favorite night clubs. Once in the new venue he purchased a large villa (family money), and then established the Académie de Graphique Improvisé, part school and part cult (in the good sense). Which Académie was then and is now still based upon the ideas in his unpublished S&M opus. Life and aesthetic practice at the Villa were organized around the central doctrine of S&M, "Dans les graphiques vécus est la vérité de ce moment" ("In lived graphics is the truth of this moment"). Said doctrine of course modified by the hasty post S&M decree      so essential to Jean Paul and so frivolous to most everyone else      "Sans musique, sans réalisme!" ("No music, no realism!").

[ LS: Upon re reading what I have written, I now see a whole in this narrative. We could call it the How to get from a description of a phenomenon to the founding of a cult whole. It works like this: once you, as John Paul in 1950's France, have a careful description of embodied visual performing consciousness such as presented in "A Lived Structure Arising Out of Improvised Manipulation for Text Graphic Images", then naturally that description leads to, necessitates, and yields the design of exercises, meditation, devices and framework for visual improv. Why? Because the whole project arose from an informal contract with the Cercle, which itself had a defining relation with the graphic improvisation imperative. Because the Cercle was founded in Paris shortly after WWII. Because the leap from description to practice, although perhaps not justified in the abstract, is completely justified in the existential nexus of action which was the mind set of post War France: the resistance; a time of action rewarded; they wanted to DO IT! Inspired by but not content with JP's compelling account of "what improvised graphics meant," the Cercle insisted that JP apply his phenomenology to the actual improvising of graphics and he all too willingly acquiesced ...
                phenomenological description   visual improv theory/meditation/practice         ]


Decline of the Knights Visual

Unfortunately in the years following the founding of the Académie, improvised live graphics never quite caught the public's fancy. Jean Paul et al in the Villa were much dismayed, as was the Cercle back in Paris. In both cases fervent expectations were dashed, optimistic hopes squashed. Eventually the Cercle withered away, and the Café de Graphique even stopped showing Abstract Films on movie night.

JP and his gang, however, soldiered resolutely on. What came next were decades of relative quietude and obscurity for the Académie. Cloistered on the outskirts of Versailles in the Villa, the true believers behind its walls kept the faith of orthodox text graphic improv. As the years passed, the Chevaliers et Chevalières Visuels (Knights Visual, male and female) of the Académie continued to diligently train the ever dwindling number of accolytes in the approved theory. And also train them, of course, in the established practice of visual improv, using the finest technology available to them at the time while simultaneously decrying each one, "Insuffisant, insuffisant." Knights of Resignation, simply marking time, modding theory, and valiantly making do with ever so clumsy instruments. And always patiently awaiting the instrumentalité appropriée (the proper instrumentality, in the Krell sense). Where was the performing graphics equipment that had the spontaneous fludity of the jazz saxophones from JP's youthful nights at the Café de Graphique?

During this long period of melancholic lurking, JP continued to refine and embellish the underlying formalism of his visual improv theory/meditation/practice. The desk work was interspersed with occasional unsuccessful quests for better instrumentality, like his brief forays into overhead projection and video synthesis. On one such excursion there occurred a most significant incident, an incident with long lasting and almost fatal conseqences for the Académie. In 1967 JP saw a demonstration of computer graphics. Before the demo, influenced by media hype, JP thought that the long awaited instrumentality for live graphic performance had finally come. He got the hopes of himself and everyone else beaucu up and there was dancing in the Villa.

Followed by extreme disappointment and sorrow when JP actually saw a demonstration. He discovered that the computer graphics of the time was completely unsuitable for the performing of images in real time. JP said afterwards, "It is like a player piano      takes them two months to punch the holes for two minutes of canned entertainment" (the punched cards they used then only supported this scornful comparison). Always one to take things personally, JP felt betrayed by computers. He shunned them from that point forward, and forbad their presence in the Villa and their use by members of the Académie.

The prohibition against all things digital lasted for the next 30 years and greatly diminished Académie recruitment. Isolated from this important technology, the Knights Visual became irrelevant to the very people they needed to attract. Over the decades even Jean Paul's fervor cooled and his theoretical progress slowed. The doctrine became frozen; what had been bleeding edge in the 50's and hip in the 60's became obscure in the 70's and then pathetically arcane in the 80's. New blood diminished and membership drooped.

Behind the walls of the Villa the spark of enthusiasm burned lower and lower, but still refused to extinguish. Esprit de corps visuel forever. The Knights Visual in the upper echelon of the Académie proved themselves steadfast if not fanatical. They stayed the course as best they could, even as the situation became increasingly grim. Finally, however, things got so bad that the death rate among the aging old guard exceeded recruitment of new members. By 1994 the end seemed inevitable.


The Graphics-Lisp trojan

But then in its eleventh hour the Académie was saved by a few bits of subterfuge. A young computer scientist named Louis Baudelaire applied for an apprenticeship. He was from l'École Polytechnique Centre de Mathematiques Appliquees near Paris, loved graphics, and played jazz guitar. Baudelaire was accepted into the Villa (he never mentioned the computer connection when he interviewed, his first bit of subterfuge; and he promised to leave always the jazz separate from the graphics, which he honors to this day).

Once inside the Villa, Baudelaire embarked on an ambitious mission of multiple stages and subversive intent. The first stage was to study JP's visual improv theory/meditation/practice in depth, including of course reading and understanding the lengthy S&M ms. He also undertook the even more difficult task of getting to know JP, who was a cantankerous old idealogue. But lonely as well it turned out. And so Baudelaire's sincere and strenuous efforts to learn JP's ideas, more successful than any apprentice in years, warmed the aging art cult leader to the youngster. After several months, Baudelaire had insinuated himself into the personal acquaintance of JP and gained some measure of friendship with the aging patriarche.

And then it came to pass that one day Louis Baudelaire saw the opportunity to initiate stage two of his mission. At the time he and JP were discussing some problems with Poincaré's conception of the domain of visual equations as defined in the Algèbre Visuelle. Conceived for the purpose of describing paintings, Henri's system had a decidedly graphical bias. This bias had all along caused technical problems for JP, who instead wanted to describe images from a more radically primordial viewpoint, the one grounded way back at "first light" before the distinction between text and graphics.

Under Poincaré's system, in order to formally treat text in anything like a fair fashion, to let it be even a second class player in the theory, JP had for the past fifty years been forced to use his hack (the one involving region grammars, relative sizing exceptions, and the visual axiom paradox). Yeah, the hack did mostly get the job done, but it was ugly as well as clumsy, and it pissed him off every time he had to use it. Over the intervening years since JP first came up with the hack, his irritation had only increased with Poincare's intial short sightedness in neglecting text. What had also increased, as JP was forced to use the hack again and again, was his realization of the loss of utility. Increasingly he ran into graphic situations for which there were visual equations he wanted to write, should be able to write, but could not write. The hack was both inelegant and functionally hobbling. In fact, the status of Herni's inital mis sight in the sans text matter had been elevated to a major problem, referred to around the villa as Le Problème de Poincaré (at that late date JP still thought of it as the only problem, even though he had also been tripping over the other problem for fifty years, but not formally recognizing it      like jmc said of the old Italian numbering sytem, "Zero is a good idea as the Romans experienced but never learned.").

Denying the text graphic rift

So on that fateful day, in the throws of textual problems, Louis casually remarked that he knew of a possible solution which would allow text and graphics to be handled together evenly. And then he brought up the work of American computer scientist John McCarthy, who Louis said, was like JP a mathematician and logician. Plus, inventor of the Lisp programming language, but once again Louis neglected to mention the digital connection. Instead he simply presented McCarthy's work as an elegant mathematico logico notation for describing patterns of textual symbols (it also slipped Louis's mind to explain that the same work was a powerful axiomatic system based on the lambda calculus of Alonzo Church which in turn had been first designed to investigate computability; but hey, why complicate the situation?). So Baudelaire focused only on McCarthy's notation, touting it as a "tool for thought" that could easily be extended to visual objects. The ploy Louis used was that since McCarthy's system already had the "text", all JP had to do was add the " graphic"      using visual equation notation and other tools from the Algèbre Visuelle to rigorously define "graphic"      and then JP would have a way to handle the "undifferentiated text graphic images" at the heart of S&M. Élémentaire, right?

Well, as it happened, this ploy and extension were actually feasible. McCarthy's original work was, amongst other things, a framework for thinking about the manipulation of symbolic expressions. That framework included an elegant notation and a few simple rules for processing lists of textual symbols (LISP stands for LISt Processing, but of course he didn't mention that either). Rather, Louis merely pointed out that the framework could be used to handle the descriptive heavy lifting in the formal/logical parts of JP's visual improv theory/meditation/practice. One simply had to generalize McCarthy's system to deal with lists of visual objects and then included textual symbols as one kind of visual object. This point was both true and also pandered to JP's rabid text graphic agnosticism, the frustration of which obsession was of course exactly the crux of Le Problème de Poincaré as formulated and experienced by JP.

But, JP asked, Wouldn't they then simply be trading a graphical bias for a textual one?

Infiniment au contraire! said Louis, who had been waiting for this very objection. One beauty of McCarthy's system was that it also provided an extremely elegant method for deferring distinctions. In McCarthy's case, the distinction was between text atoms and lists. McCarthy handled the problem by simply and arbitrarily lumping these two dramatically different things together into the category "symbolic expression," and then having his system happily operate on those expressions without knowing or caring what kind (of symbolic expression) they were ... that is, until as such time it needed to care, in which case only then did it distinguish. So, proposed Louis, they could use McCarthy's strategy to formally instantiate the deferred text versus graphic distinction      visual late binding, does it get any better than that?!      which JP had proposed long ago, but which had proven so troublesome for the concepts and system of Poincaré.

Visual Equations meet the Lambda Calculus

JP was intrigued; being a formalist himself, he was quick to grasp the possibilities of McCarthy's approach. He withheld further judgement while Louis took the time to show him the details of the notation and concepts of the new system. But after a few days of explanation and examples, he began to appreciate the power of McCarthy's system. He could see the elegance of the ideas and how beautifully they complemented his own. JP confirmed that Baudelaire's instincts had been correct: there was a good mapping between JP's careful description of visual structure de la manipulation and McCarthy's patterns of maniplation for textual expressions. Plus, an unexpected bonus for JP (although forseen by Baudelaire) was that for McCarthy manipulation of textual symbolic expressions included the domains of algebra and calculus! This was because numbers were a kind of symbol in McCarthy's system, and equations represented as lists of symbolic expressions became themselves objects for manipulation. Now, algebra and calculus were two tools for which JP had both a hereditary foundness as well as a practical use, being as they played a key role in his system (it was after all Poincaré's Algèbre Visuelle; and as for calculus, stay tuned). Thus was JP convinced to give the new approach a shot, and so he and Baudelaire set about recasting the Poincaréan formal grounding of JP's visual improv theory/meditation/practice in terms of McCarthy's notation and rules, while at the same time retaining Henri's gorgeous notation for visual equations, as well as his insights into the finer granularity of all matters graphique.

JP was at first challenged      it might well be time after fifty years to revisit the formal underpinnings      and then invigorated      it was indeed time after fifty years to rethink the formal underpinnings! The project went better than even Baudeliaire had initially imagined, and as their venture proceeded, JP became more and more excited.

As a first step, they created a text graphic version of Poincaré's algebra, meaning that they re axiomized and re formulated his system so that Henri's ever so expressive visual equations could cleanly describe text objects as well as graphic ones. And, the domain of this enhanced descriptive capacity included the visual equations themselves, which ability nicely complimented the ability of the original Algèbre Visuelle to mention visual objects inline. Making everything cleanly text graphic took the kinks out and thus added power. Plus, now the descriptive reach was reflective, which feature was to pay off large theoretical and practical dividends when our story gets to the next part (calculus!).

Surprisingly, the new version of Poincaré's algebra was simpler than the original. Louis and JP next undertook, using the new Algèbre Visuelletg (as JP fondly coined it), to rewrite the set of visual equations at the formal heart of JP's system. Once again, the new version of the system (this time, JP's S&M system) was simpler than the original. The dirty text hack had over time inserted its diseased tentacles everywhere; simply making the Algèbre text graphic had cleaned up many of the foundation equations for JP's system in one fell swoop. Which made that system easier to understand, even for JP. Which in turn, within this more perspicuous framework, allowed Louis and JP to find inconsistences in JP's original conception. And by correcting the inconsistences, to then discover natural extensions to the formal grounding of visual improv theory/meditation/practice which increased its power without increasing complexity (holy grail of math heads everywhere and  when, cha cha cha).

Calcul Visuel

Finally, if that all wasn't enough, for the concluding payoff, JP was able to finish the formal definition of his "Calcul Visuel" (Visual Calculus), an idea with which he had struggled for many years but only now had the conceptual technology to complete. And, once completed, JP then saw how important the calculus was to his project, and how its absence had crippled his life's work. Without it, he had been hamstrung and hogtied from in front for many years, just as he had been by the Poincaréan anti textism. Except that in the case of this impediment, JP had never formally recognized it! Far more insiduous, the theoretician's ultimate nightmare. Only when gazing at last upon the finished formalization of his calculus did JP finally realize the extent of the damage      only in solving it did he at last see the full impact of the second and most malignant of Les Problèmes de Poincaré.

Bluntly stated: Poincaré was anti dynamic as well as anti text! His concepts and notation had been born out of and designed for a world of static images      duh, paintings. But the soul of improvised graphics is dynamic imaging, properly the domain of JP's theory. So he had a fundamental mismatch between tool and domain. However, there was also a traditional dirty hack to get around this mismatch, a hack so culturally accepted that it was not even seen as a hack let alone dirty, and JP had used it without thinking. The hack was the approximation of continuous functions using summations of many small discrete units, the standard method of Newton's calculus for dealing with continous rates of change. Therefore when JP was faced with the phenomenon of dynamic imagery, and the tool he had was a visual agelbra of static images, he automatically represented the continuously changing visuals with a sequence of static images, each differing slightly from the last. The tender trap he never saw was made even more inviting and inevitable by the fact that his original visual corpora were the Abstract Films shown on movie night, themselves simulating motion at 24 static frames per second.

And, what's so wrong with this culturally endorsed workaround? Exactly this: it uses jerky continuity in place of continuous continuity. It substitutes stairs for a ramp, and then says if the steps are small enough you'll never notice the difference (tell that to someone sliding down a discrete approximation of a banister). Jerky continutity was "the settling for less that got a lot," and, in many applications, really didn't seem to make much difference. But if the goal is to provide a formal description for the improvising of graphics, then the description itself should facilitate the immediate apprehension of imaging dynamics, should itself allow one to throttle the flux and feel the flow smoothly course and twist in one's grasp. The description itself should put the user directly in touch with the phenomenon. Which JP's Calcul Visuel truly did.

Dynamic visual equations on the move

How? it might well be asked. That question is easy to answer when you have a Calcul Visuel "equation" in front of you; otherwise, almost impossible. JP's insight and invention was dynamic dynamic equations, a visual represention of visual dynamics that was itself moving. You read his equations by watching them. Using the Algèbre Visuelletg, JP was one day writing visual equations about visual equations of rates of visual change and suddenly it seemed that the equations themselves were changing in front of him. Even as Poincaré's algebra had included visual objects in the equations which described them, JP in that instant saw that visual equations describing imaging dynamics should themselves be dynamic as well as visual      well, more accurately, what JP immediately saw was the equations moving. Only after the fact did he analyze the character and power of the (necessarily) vizomatopoeic tool he had discovered.

[LS: JP's visual calculus and its dynamically dynamic equations is a fascinating invention. Although somewhat off topic, a few further points need to be made.

At that time, in 1994, only JP could "see" the equations move (whether as a hallucination or in his mind's eye is not clear). Whereas Louis could not see them move, try as he might, and much to his dismay. By whatever means JP did it that first time, his ability to do so was undoubtably engendered by almost fifty years of conscious intellectual struggle with the formalisms involved      and by almost fifty years of burning subconscious need to see his intuitions formally realized and reified.

For as long as he had been using Herni's Algèbre, JP had always intuited that the tool was already wrong for his project in the sense that the proper focus was not the structure of static images, discretely and sequentially left behind after a performance, but instead the very structure of the imaging itself, the flowage in artist consciousness      what JP sought was the live coursing of the raging stream, not the dry channel cut in the dirt and left behind after the rain.

And then when he saw the equations move, JP knew that finally he had the right tool in the Calcul Visuel. A tool that gave algebraic rigor to his pre formal intuitions about continuous imaging dynamics which had been underlying S&M from the very start; a tool arising out of visual equations about visual equations, reflective and recursive as all good tools are; a tool that put the user directly in touch with the phenomenon it represented. In short, a tool that was both experiential and formal      the only kind of tool that could do the job JP had needed done all those years.

It's worth mentioning that in 1995, after months of hard work, Louis was at long last also able to see the equations move. And since then, there is the occasional Acolyte who is able to see the motion as well. To do so takes a special combination of perserverance and deep understanding of traditional math concepts, and yet also the willingness to let go of those concepts      to finally use the traditional framework against itself in order to defeat its fnordly hold on the mind and allow the acolyte to leave behind jerky continuity and directly apprehend continuous continuity in all its fluxing splendor. Louis calls this ability to see the equations move unaided the "flux capacity"; JP calls it La Transe de Calcul Visuel.

And today, of course, with the proper viewer application (more on this below), all that mental mumbo jumbo can be by passed and anyone with a computer can see the equations of the Calcul Visuel move. In fact, I would love to show you right now, dear reader      there is plenty of room in the margins, if only the margins could play equations! Fermat's last movie. Darn paper.

In conclusion, I have seen JP's wonderful Calcul Visuel in action (and, no, the equations don't move for me, yet, so I must use the viewer like most everyone else). This amazing tool puts the user directly in touch with the phenomenon it is "about": the dynamics of structural imaging manipulation which are at the heart of the Academié visual practice. And because it is formal as well as experiential, the Cacul is also essential to any computational engine which supports and abets that practice (lots more on this below). I hope that someday I will get the opportunity to offer a far deeper description of these revolutionary concepts in another article.   ]


Their Winter of profound content

The Winter of 94 95 passed in long hours of exhilarating work. There was now no doubt that McCarthy's notation was indeed a "tool for thought," just as Baudelaire had promised. And, the underlying concepts proved to be as useful as the notation. Utilizing both concepts and notation together resulted in first the text graphic ification of Poincare's Algèbre, which then lead in turn to JP's Calcul Visuel and concurrently the major rewrite of JP's fifty year old system of visual equations. In fact, so thorough was the rewrite that JP began to call the new/old work S&Mjmc to acknowledge the critical role played by John McCarthy's ideas in the revitalized undertaking (the project was raised both to and with the power of jmc's ideas). Whereas Baudelaire secretly referred to it as S&M 1.5 (vive la differénce between math guy and computer guy).

Thus, in terms of the Lagrange family tradition, a fourth body (jmc) had been introduced into the system, and JP had then found a way to incorporate, synthesize and tropicalize. Which finally produced the new, more elegant version of JP's massive system of visual equations (where elegant meant simpler yet at the same time more powerful; again, the holy grail for math heads and hackers alike). Elegant also meant looked cool on paper, thanks to the new and very compact notation, as can be seen in this somewhat blurry image of one visual equation from the original handwritten ms of Winter 94 95:

When they neared the end of the rewriting, JP, who although he had been boycotting computers for thirty years, was yet not unfamiliar with the basic principles of computation and began to suspect that this new expression of his theory might be very compatible with representation in a machine. Among other clues, at one point JP had recognized the tell tale spoor of Church's lambda calculus (as already mentioned, first invented to poke at the very idea of computability), which was one of the foundations of McCarthy's work.

The mechanical visual bookkeeper

Up to this point Baudelaire had wisely remained silent about any use of actual digital hardware in the project. But when JP himself finally brought up the subject, Baudelaire cautiously ventured the opinion that maybe there was a chance that a computer might be able to play some small role ... perhaps as a mechanical bookkeeper, say, just to verify that the many parts of JP's formal theory didn't violate the selfsame rules of that theory. And, Baudelaire said, it might not be too much work to transcribe the existing McCarthyized pen and paper version of JP's formalisms, which they had created in the last six months, into a machine usable form. This statement was somewhat disingenuous on Baudelaire's part, since in fact the pen and paper version was already computation ready except for syntax, and of course the inevitable inconsistencies in a system of mathematics that size (it ran seventeen pages of closely spaced hand written equations). However, those inconsistencies would be caught (for the most part) as the system was entered into the computer, which was of course exactly the overt purpose of the task as proposed to JP by Baudelaire.

And so the two men actually started a computer implementation of JP's new/old S&Mjmc system using the Lisp programming language      but "only for the theoretical benefits," Louis assured JP. At this time, in the Summer of 1995, the machine Baudelaire chose for the task was an Intel 133MHz Pentium powered PC (and, a very big deal: first computer in the Villa, ever!) running Debian GNU/Linux, the X11 window system, and Austin Kyoto Common Lisp (aka GNU Common Lisp).

[ LS: Lisp is so cool. Neal Stephenson says it's the only computer language that is beautiful.

He's absolutely right. First, the code just looks so durn clean and purty. Simplest yet most powerful syntax ever.

Second, the logic is durn clean and purty too. Paul Graham explains thusly: "In 1960, John McCarthy ... did for programming something like what Euclid did for geometry. He showed how, given a handful of simple operators and a notation for functions, you can build a whole programming language. He called this language Lisp, for `List Processing,' because one of his key ideas was to use a simple data structure called a list for both code and data." [from: "The Roots of Lisp,"]

Alan Kay has his own take on the logical beauty of lisp: "Yes, that was the big revelation to me when I was in graduate school      when I finally understood that the half page of code on the bottom of page 13 of the Lisp 1.5 manual was Lisp in itself. These were `Maxwell's Equations of Software!' This is the whole world of programming in a few lines that I can put my hand over." [from: "A Conversation with Alan Kay",] .

On a personal note, I have built a text graphic editor for performing called vmacs. vmacs is based on the original pen and paper formalization of JP's work done by Baudelaire and JP in 1995, which I copied from the actual ms when I was allowed access to the Académie archives in 1998. Going back to the original hand written system of visual equations was of great technical advantage to me in the vmacs enterprise, as I explain below. ]

Fairly early on the project      which JP thought would be a major undertaking, some completely new and complicated computer "implementation" kind of thing      instead turned out to be a relatively simple transcription of the many hand written pages. At that point JP realized that he had been duped. It had been too easy: the seventeen pages of pen and paper equations were the Lisp program (just without the parentheses).

[ LS: Yup. And, it was also too easy from a good Lisp technology point of view. Doing the translation in the most obvious way did indeed produce a Lisp program, but a horribly inefficient one (see discussion below). However, regardless of efficiency, it was also the first use of what came to be the vizlisp programming language. Or, the vizlisp dialect of Lisp, depending on your point of view. Since Lisp is so easy to customize, what start as extensions often end up as languages. Whatever you call it, vizlisp has since been re implemented by both Adobe and myself, and I hear there is a GNU version in the works. ]

{ AM: see vizlisp entry in the Glossary for an example of vizlisp code. }


At this point Louis confessed his subterfuge, there was yelling but not too much, and JP got over it. He was already hooked and looking forward to seeing the thing in action. He thought of it as a computer assisted "visual proof checker" for the logical foundations of his visual improv theory/meditation/practice, since that was the payoff that the younger man had promised. Hence work continued and the computer implementation went well; so well in fact that it was all too soon a victim of its own success. The startlingly easy and rapid accomplishment of stage two in Baudelaire's subversive agenda quickly led Louis to prematurely embark on stage three, with disastrous result.

But before those gruesome details, let us first, along with JP and Louis, savor the inspiring victory of stage two. Using Lisp, in a short time the two had a working model of JP's theory running on their computer. Which delighted the artful old codger no end. He could see the theory "work" and actually prove visual theorems. What they had done was to first build a graphical calculator programmed in vizlisp to instantiate Poincaré's visual algebra (it was of course, the tg version). Next they used that running version of visual algebra to then instantiate (OK, OK, implement) JP's S&M system. So that finally JP could enter text and graphic queries into a simulation of the S&M system, it would crunch for a while, and then return text and graphics in response according to his theory. Way cool.

[ LS: It is worth noting that all the features of vizlisp as we know it today were present by this time, but it is not clear exactly when in the minds of those first users it actually emerged as a different kind of animal, distinct from textual Lisp. That's the thing about Lisp: it's so easy to extend that signficant boundaries may be crossed without noticing it at the time. In their case, when did they realize that it wasn't just another graphics package for Lisp with screen painting subroutines, but instead a full blown paradigm shift into a radically flipped way of thinking, that they had gone from computing with (textual) symbolic expressions to computing with text graphic forms? When did they realize that the latter in fact subsumed the former, and that graphics were at long last first class objects enjoying all the benefits of full computational status, that graphic objects could be programs which manipulated other graphic objects, that visual programs were visual data ... in short, that indeed they had made graphics do Lisp? Well, we can only wonder, and JP and Louis can't or won't say. Do they lack the memory or the will? Or perhaps the problems lies with us, and such questions are really just an expression of a romantic longing, like wanting to be there on the day the first sentence was uttered, a quest both poignant and yet also in its essence philosophically flawed and thus unsatisfiable. ]

In addition, way cooler, using the ability of vizlisp to manipulate visual equations as visual objects, they were finally able to create a crude simulator to "play" the dynamic dynamic equations of JP's Calcul Visuel! The good news: for the first time someone besides JP could apprehend the little buggers in vivo. And the bad: crude is an understatement. The visual representation of the dynamics for the simplest equation representing the simplest move in beginner's improv practice      dessiner, la rotation, efface (draw, spin, erase)      took overnight to compute, and then an hour to run the display list. For a two second riff!

Nevertheless, inspiringly heady stuff for the duo.

But now hanging over the project like a sword on a string was the ever more obvious stage three, the possibility that the visual theory engine could next be turned into a functioning visual performing instrument. The sword of actual graphic improvisation, surely mightier than the computational pen. JP and Louis could envision it so very clearly      the working stage two was already the detailed plan for stage three      their remaining task was merely to execute the plan. And like so many visionaires before and since, they mistook a clear view for a short distance (as Paul Saffo says), with tragic consequences.

VizLisp disappoints

Making an actual tool for performance by building on and extending the stage two engine they had at the time turned out to be a task of overwhelming difficulty. The Lisp incarnation of JP's system, as programmed so directly (and easily) from the original pen and paper version, was simply much too slow. And as for the hydra headed user interface of the "Lagrangian visual calculator and proof checker" (one way of describing their working system): complete disaster. The interface was adequate to laboriously construct a precise text graphic form which could be submitted to the visual theory engine (the engine would then chug away for a very long while and finally return a text graphic answer). But as a performing instrument, that interface, no way! After a tortured period this third stage of the digital conversion began to falter, then stall, and by Winter of 1995 96 the project had fallen into profound necrostasis. The implementation of the instrument proper was just too hard, especially for Louis who was working alone as the only programmer at the Académie (of course, there were no other programmers thanks to JP's thiry year prohibition on all things computer) .

[ LS: Louis is a nice guy and very smart, but not the best LISP hacker; vmacs is written in Lisp and has more than enough graphitrons for visual performing. ]

Such was the state of things at the Académie Villa in the Summer of 1996 when Flashmaster Grandé shone the beam seen 'round the world. FmG's grand and masterful flash of light on a wall in San Jose ignited the Chevaliers et Chevalières Visuels in Versailles. Visual improvisation was finally in the public eye and in demand. And, computer technology was the enabler. Thanks to FmG´'s vivid demonstration, JP and his gang at the Villa quickly saw that a digital tool could indeed be the performance instrument they had been seeking for almost forty years, that the graphic sword was real, that the benefits of a computer implementation were more than theoretical      they were practical in the extreme. At long last their instrumentality had come! JP and Baudelaire took up again the struggle toward digital embodiment of the lived structure de la manipulation as so elegantly heralded in the original S&M ms.

VizLisp disappoints again

But, unfortunately, renewal of inspiration did not reduce difficulty; their task remained intractable. The computer version of JP's system they had at the time was large and ran slow. True, it was adequate for the visual calculator and proof checker of stage two. However, in order to satisfy the needs of performing graphicists, the software needed to run very much more quickly (by at least three orders of magnitude!).

Plus they needed a way to generate a different personal user interface for each performer. JP was adamant that they honor one of his central obsessions: in performance his system was actually a different theory for each individual (the French perfume thing).

Baudelaire was forced to re examine the design choices he had made in building the stage two system, especially the one of computer programming language, a crucial decision that could make or break the enterprise. It was tempting to continue with Lisp, where they already had a large amount of functioning code that ran stage two just fine. It's really hard to throw away something that works and start over from scratch. But finally Baudelaire had to admit that Lisp might not be appropriate for stage three. First, it was just too slow; and second, he didn't have a clue as how to do the different user interface for each performer trick (which technically was not a programming language issue but rather a conceptual GUI problem, very far removed from Baudelaire's strength in computer science, which was why he didn't recognize his category mistake). By early Fall of 1996, the visual performing instrument project was again horribly stuck.

[ LS: in fairness to Lisp, I think the problem was that Baudelaire severely underutilized the technology. Louis was and is one of the mathematcio logico type computer scientists (at university he specialized in "recursive function theory"). So when he took the survey class in programming languages at l'École Polytechnique, he deeply grokked the essence of Lisp and it blew his mind (he "got" the eval lecture). And Lisp was absolutely the correct choice as a formalism to express JP's framework of visual image dynamics (and in so doing beef up the rigor of that framework).

The easy, rapid (and fun!) transcription of JP's original system of visual equations into vizlisp is stark testimony to the absolute correctness of that choice. But in such a task, the easy and obvious route as seen by logico mathematico types like JP and Baudelaire is not for all purposes the best route (transcription rather than translation was a clue). Yes, they indeed got something up and running very quickly. And that was good. The "working model" of stage two as in "actually functioning" was amazing and motivationally essential to the whole project. But that same code base then became their worst enemy when attempting to build the performing instrument of stage three.

Grasping the formal idea of Lisp is not the same as being a great Lisp hacker, a crafty programmer able to get the last erg of performance out of the machine. Lisp is notoriously difficult for that; the power it gives at the level of conceptualization it can take away at the level of efficiency. Lisp is a wonderful tool for thought, but a price is paid when it is used as a tool for construction      unless you are one of that rare breed of Lisp hackers with the verticality to go from the conceptual down to the bits and then back up again.

Unfortunately, the lineage of Louis Baudelaire was not quite that rare. And so he severely underutilized the techology. Hence the massive stuckitude.

In fact, it was just to avoid that stuck and take full advantage of Lisp that I went back to the orginal system of visual equations when I implemented vmacs. I have a little more verticality than Louis, and so was better able to translate JP's high level visual equations into efficient working code, thus preserving conceptual power while at the same time getting drastically better performance. Of course, the inexorable march of Moore's Law helped me out a little as well 8-) . ]



Adobe to the rescue

At this point once again serendipity of time and place played a crucial part in the Académie saga. FmG´'s Summer illumination event in Saint James Park, and subsequent transformation of the walls nocturnal in San Jose, had occurred just four blocks from the world headquarters of Adobe Systems. The younger employees very quickly picked up on the new visual art form and the offices were abuzz with talk of "throwing beam" and "vizjammin''. The buzz percolated up the org chart and finally reached the very top: John Warnock, co founder of Adobe. The timing was auspicious. Desktop publishing had been very very good to Adobe; Warnock was looking for new worlds to conquer. And John had acquired a taste for jazz from his days hanging out with Alan Kay in the late 60's at the University of Utah Computer Science Department (Kay was a jazz guitarist as well as a computer scientist). So performing graphics strongly hit two different chords in Warnock simultaneously. He began to conceive of an Adobe application to support this new art form.

Thus in late Fall of 1996 two complimentary forces 6,000 miles apart were ripe for coupling. Adobe had the programmers and the software to do fast graphics; the Académie had the vision and the framework ready to implement. It remained simply for the two to come togaether. Which union took place thanks to another French connection, the Baudelaire family. Louis's uncle, Patrick Baudelaire, had worked with Warnock at PARC back when in the 70's it was still Xerox.

Quagmired again during his second try at using Lisp to build a visual performance instrument for JP, Louis happened to think of Adobe's speedy yet flexible Display PostScript system. Hmmm, it might just work. A call to uncle Patrick got a referral. And then a phone call to Warnock in San Jose met with an eerie reception. After Louis's first few sentences, John basically said he had been waiting for the call (though he had never heard of the Académie; few had in the States). Non plussed by this reaction only a bit, Baudelaire went on to briefly describe the Académie's mission and the role they thought computers could and should play. Warnock loved what he heard and suggested they get together in person. Since JP's health at the time precluded travel, Warnock was on the next plane to France. He met with JP and young Baudelaire in the Villa on a fine autumn day.

The meeting went well to say the least. At the level of ideas, Warnock loved JP's theory/meditation/practice of visual improv. And at the level of the practical, Warnock was of course familar with Lisp and said Adobe could implement JP's vizlisp dialect directly in Display PostScript. The result would be a language conceptually true to JP's ideas      enough so as to run all of their existing high level vizlisp code without change!      and yet also efficient enough to build a lightning fast performing instrument in vizlisp. The combination of speed and philosophical congruence was no accident: Display Postscript was Forth like if not based on Forth (ruthlessly speedy) which in turn was Lisp like if not based on Lisp, all of which simply meant that the essential underlying computation technologies were in place, optimized for quick, and ready to go; chief amongst them were efficient stacks and a recursive frame of mind, which is to say, "Lisp friendly."

So regarding the insides of their joint venture visual instrument, the two parties were in agreement. JP's structure from manipulation, as "mocked up" in the working vizlisp code of the visual calculator, would provide the framework for the visual improv engine at the heart of the application Warnock had in mind. He was already calling it Visual Jazz Master.

Visual Jazz Master, derailed

But on the outsides, where the human touched the visual engine in order to control it, big disagreement. Warnock thought the user interface should be standardized so that an experienced Visual Jazz Master user could walk up and play any copy of the application on any computer. The position was not unreasonable from the commercial software point of view. Whereas, from that same point of view, the prospect of each user generating their own unique interface from the ground up, not just "customizing" or "skinning" but totally different in look and feel      very unreasonable! In response, from the traditional Académie point of view, JP was aghast, horrified, and extremely angry to boot. Every user forced into the same interface relationship with the general theory engine?! "C'est impossible!!". It was a deal breaker.

Any chance of a partnership seemed dead, so Warnock departed the Villa and headed to the airport.

However, after Warnock had left, Baudelaire began thinking. Then he had an idea. Computer software is infinitely malleable (again, the somewhat cavalier view of the theoretical computer scientist). Therefore, one standard interface, or many personal interfaces with each user building their own      that didn't have to be an either or choice. There was a technically and commercially viable way to do both.

Visual Jazz Master, back on track with GIG plugin

Plugins, additional modules which are "plugged in" to an application in order to modify its behavior, had been around for a long time. Adobe as a company was particularly hospitable to plugins (even those written by third parties). But a simple plugin would not do what JP wanted; what he wanted and needed instead was a way for each Apprentice to make their own interface from the ground up, not to merely customize the behavior of the standard interface.

In order to meet this need, Baudelaire's insight was an interface generator plugin: an affordance which would allow each user to then make their own interface (within which things like normal plugins operate). We're talking the big meta here folks, up a level. So to describe, we retreat to meta-phors: re landscape the playground, not just throw in a new merry go round; change the game, not just add a few rules. And, yeah, OK in this case Baudelaire's more theoretical view did pay off in the GUI domain. He thought of the interface generator plugin as a "mechanical enumerator of possible interface worlds from an idiosyncratic kinesthetic specification." Because he understood both JP's S&M theory (better than anyone except JP himself), and the vizlisp system (better than anyone including JP), Baudelaire knew that within the basic framework for text graphic manipulation there could be specified an indefinite number of ways for humans to get their hands on the visual flow. The vizlisp system was in fact infinitely malleable (the payoff for proper logico mathematico grounding: you can cavalierly make big deep changes and they work!).

Louis called Warnock on his cell and described the idea. Warnock got it immediately, liked it very much, and had the limo turn around. Louis then undertook the more lengthy process of explaining the interface generator plugin notion to JP, who finally understood and also liked it. Baudelaire had just finished this elucidatory task by the time Warnock returned to the Villa.

And so the user interface issue was happily resolved for all concerned. Adobe would write a special Académie interface generator for Visual Jazz Master which could then be used to create a multiplicity of "interface worlds," each one wholely unique. Adobe would also teach Académie members how the interface generator worked, disclose initial and all subsequent source code, which would be written in high level user accessible vizlisp, and agree to keep the interface generator functioning and updated for the lifetime of VJM, however long that might be.

Once the technical design issues had been solved, a commercial licensing deal was quickly hammered out. Exact details are confidential, but it is said the Académie did very well: royalties for 2005 were rumored to be around €650,000. Current sales are still mostly to buyers outside the US, and increasing steadily as performing graphics gains popularity on the Continent.

And as for the actual instrument of graphique improvisé      well, within six months of Warnock's visit to Versailles, Adobe had delivered beta versions of both the basic Visual Jazz Master application and what came to be called the Gallic Interface Generator plugin (GIG). Because the heart of VJM was JP's structure for manipulation engine and Calcul Visuel, at the deepest level the software was congruent with general Académie theory and practice. And thanks to the GIG plugin, each Apprentice could grow their own unique interface to the general theory engine. An interface which expressed their personal version of the general theory, to be expressed yet again in each performance, finally and concretely manifested in the flow of visual imaging improvised by that person at that unique time and place.

Oh yeah, and, one more thing. The Calcul Visuel player. Bundled with Visual Jazz Master is a complete vizlisp programming environment, including intepreter, compiler, editor (just VJM itself), and debugger (again just VJM itself). In fact, most of the higher levels of VJM are written in vizlisp, and the sources are included in the bundle (Adobe keeping their agreement). Using the Adobe implementation of vizlisp      in beta, already four orders of magnitude quicker than the one he had written      Baudelaire was able in an afternoon to code up a new viewer for Calcul equations, and then everyone in the Villa could see them move (the only good fnord is a fnord thwarted). At last there were sufficient graphitrons to run vizcalc movie forms. Not surprisingly, the viewing of same turned out to be a very useful meditation for Apprentices. At both conceptual and emotional levels, grasping the dynamics, feeling the coursing of the imaging flux      and, with today's Academié touch panels that touch back, at the kinesthetic level as well      hey, vizpilgrims, it just don't get any better than that. Vive la machine du Calcul Visuel!

Visual perfume

Thus, finally, after lurking in virtual secret for all those years, as visual technology advanced yet did not satisfy, biding their time whilst perfecting their practice and their theory, the wait for instrumentality was over. The long suffering Knights Visual at last had their tool, Knights of Infinite Resignation no longer. And the Académie Villa in Versailles had a computer room, and an interface shop in the basement for building ensembles d'entrée. As graduates began to perform in the graphics bars and visual salons of Paris      amazing and delighting audiences wherever they went      the Académie soon acquired its justly deserved reputation for producing imagists capable of magical visual feats.

So that is the story of how an attempt to bring together three different French intellectual viewpoints      Poincaré's Algèbre Visuelle, Merleau Ponty's phenomenology of the lived body as artistic instrument, and the visual improvisation imperative of the Cercle de Graphiques d´Improvisational      was cast as a three body problem and then solved by the young Jean Paul Lagrange in 1947. Which solution, kept alive for almost fifty years at the Académie Villa in Versailles, and then profoundly augmented by the fourth body of McCarthy's work, was and is to this day a theory of visual improvisation that is also a meditative practice that is also the algorithmic framework for a computational imaging representation and manipulation system that finally emerges as the organizing principle which generates a different and unique user interface and set of input devices tuned to each individual practitioner of that meditation ...

Which is simply to say that the whole Gallic mess still lovingly and powerfully interacts with each performer's own visual chemistry just like good French perfume.



Concluding UnAmerican PostScript

Even now Adobe is one of the few US companies to put any credence in the performing graphics art form, let alone invest money in developing products for it. There are some other notable exceptions to the glaring indifference from the American hardware/software industry: Apple and their Mini Walkabout, Crate's "Big Throw" Street Projector, Fender's Stratoraster stylus, the Wu Tang Clan's revitalization of Wang computers. And can't forget the Performing Graphics Company of South Palo Alto, constructors of the justly famous gunslinger keyboard belts and thumb pick styli. [ LS: full disclosure      I am a principal in PGC. ]

Whereas in Europe, hardware, software, and general fan support for performing graphics is rampant and growing every day: Momo/Philips, Porsche, Schnell Komputronz AG, Thomson's Laboratoire Visuel de Technologie, The CERT program in Futurology and Synthesis; the obscenely lavish annual Performing Graphics festival in Nice; etc, etc.

[ LS: Finally, you should know, dear reader, that JP is still very much alive and doing well for a guy born in 1924. I hear his health is good and his spirit indomitable as always. And, when I visited the Villa in the summer of 1998, it turned out that I had just missed a visit by Flashmaster Grandé who had come all the way from San Jose, Calif to talk with JP. Now that must have been one cool meeting of the minds! I'm still trying to get a full account of the details; when I do, I will write them up and add to the next version of this History. ]

French Appendix Appendix      pages from S&Mjmc

[ LS: The fuzzy image above shows the handwritten notation from the original S&M ms of Winter 94 95. I snapped the photo on a reading table in the Académie basement archive; sorry about the quality and the thumb in frame. The notation in the picture is very close to what is still in use at the Académie today. This notation had come to be called "blackboard vizlisp" by the time I visited in 1998. Blackboard vizlisp uses the syntax and concepts of McCarthy's handwritten "blackboard notation" for Lisp, but includes functions on visual objects as well the ability to use actual visual objects in the code. Thus visual objects are employed both as elements of the notation and as visual literals in order to introduce visual data into a program (of course viz structures for notation and quoted vizliterals are just flip sides of the same old Lisp "programs are data" hit song      If visual programs are data, and data be the food of computation, play on; give me excess of it.).

For example, in the first function definition above, , , and , are all vizliterals to be tested against. In addition, the Kanji form is used as a variable name      just for the hell of it, apparently, because the token could have been used instead of and the definition would still represent exactly the same algorithm. Also see here use of as a block delimiter. And finally, note that the condensed and mathematically catholic blackboard vizlisp notation displayed on this page is in contrast to the more verbose and parenthetical stylings of "keyboard vizlisp." ]

{ AM: again, there is an example of kbd vizlisp code under the lisp entry in the Glossary. }

[ LS: A lot people have been bugging me to explain exactly what the above code does      often with the implication that it really doesn't do anything. Well, in fact, what it does is evaluate forms which are drawn/written in the "chargol" language, forms like the one on the next page. Chargol forms are known as "r chars", for recursive character; r chars are composed of arms (strokes) which in turn are composed of chars some of which may be recursive, and so on, r chars all the way down. Interpretation of chargol forms is a cooperative process between two functions, chareval and armeval which call each other in an alternating schema. chareval always gets first crack at a chargol form; if the char is atomic, chareval simply looks it up on an association list. If not, chareval applies the function designated by the name shape of the big char to the result of calling armeval on each of the arms that delineate it. armeval assumes the first char in an arm is either the name shape for a function, or a lambda form defining a function; it then applies that function to the result of calling chareval on the rest of the chars in the arm. And so it goes, back and forth, all the way down. Also defined at the bottom of the page are some helper funs. ]

[ LS: ... yeah, yeah, OK, and now you want to know what happens when you actually evaluate this particular form, right? Well, I haven't as yet gotten around to transcribing the code for the chargol interpreter into vizlisp and then running it on the form, but from JP's notes, he envisoned the form as both the ideograph for and complete contents of a future spiritual library to be housed at the Académie. Apparently the library will be related to all aspects, implications and ramifications of his theory/meditation/practice of visual improv. It's hard to tell from his notes, which are both vague and terse.

One phrase I can just barely make out is "Visual Tao, the flow of heaven." Or maybe that's "Visual Door to flow from heaven" (complementing my poor French is JP's poor penmanship, plus some water damage; the basement gets very damp during the Winter).

However, I do know that Tien, the big shape, means "heaven," "mind," and/or "spirit", so the Visual Tao translation might make sense. And then presumably the arms both describe and are branches of the library, with then deeper r chars in the arms being the ideographs for and complete contents of actual volumes (turtles all the way). And maybe some of the r chars are not only library content, but instead also include data base functions for navigating and interpreting the more idiosyncratic sub branches (smart and helpful meta turtles ); this seems likely as I see that the lower horizontal arm begins with a lambda form, presumably defining a function of use in the specialized processing required by that sub branch of the library.

That's about as much as I can say about r chars and how they work, except for a couple of odd annotations JP left in the margin. One reads "... for the first time in philosophy, the name IS the thing!". Not sure what that means. And even more cryptically, "Il prend un homme dur pour faire un lambda tendre." ("It takes a tough man to make a tender lambda."      ?! that can't be right; my French is insufficient.) ]




{ AM: Finally, much thanks to Lane Slate for providing this excerpt of his History, with annotations. I met Lane only recently, after finshing the first draft of my account of the events on that Tuesday. What prompted our meeting was my realization that both Sal and Liz were using this "vmacs" thing as their visual performing instrument. I had known previously that Sal used something called vmacs, but never investigated further (her genre being so far from my own, I simply appreciated her art as a viewer and a fan and didn't look into the tool behind the images). However, after seeing Liz in action during our blend, I asked her what she was running. When Liz said vmacs, the same sfwr as Sal, my curiosity was piqued and I had to investigate further. At which point I discovered the guy behind vmacs, Lane Slate. We met, got on famously, and he agreed to provide a Cliff Notes version of his History for the French Appendix. Coincidentally, it turns out that Lane lives a few blocks from where I grew up in South Paly, but I never knew him at the time. We might have met eventually, except that I moved out of the neighborhood in 1988, long before the Visual Singularity event in 1996 and any real interest on my part in things digital.

Thanks also to the whole gang at Veli's for putting up with my mood shifts and incessant questions during the writing of this account.

Alan Morgan, typing at a back table in Veli's Graphics Bar, Oakland, CA, 2006


© Fred Lakin 2005, 2014