ON BECOMING DIGITAL THINKERS

Jerry Harkins*

On the morning of April 6, 1964, I presented a carefully researched, closely reasoned report on a question posed by the Head of the Department of Mechanical and Technical Equipment of the United States Army Engineer School at Fort Belvoir, Virginia.   Three months earlier, he had asked whether the military field computers of the future would employ analog or digital technology.  Now I was addressing him and his top aides armed with a 3,000-word text and a deck of overhead projector slides.  I had interviewed experts and learned about the kinds of decisions made by battlefield commanders.  I drew up a list of the kinds and sources of the information they needed and their special requirements for portability, ruggedness, reliability, and practicality.  The take-away was that analog computing was likely to be the dominant technology for the foreseeable future. 

Twenty-four hours later IBM introduced System 360 thereby rendering my analysis antiquated.  Luckily the Colonel was an engineer with a sense of humor.  He recited the immortal words of Robert Burns:  “The best laid schemes o’ Mice an’ Men / Gang aft agley, / An’ lea’e us nought but grief an’ pain, / For promis’d joy!”

Not that anyone could operate a 360 or any other digital machine outside a large air conditioned environment or, especially, inside a Vietnamese rice paddy.  But as the first general purpose solid state computer, it pointed the way to the future of everything.  Apple’s hand held smart phone is much faster and has far more computing power than did the 360 but is nonetheless its direct descendent.  For IBM the 360 represented the third generation of digital computing, a technology that had evolved in the 1930’s and that was represented in the popular press by ENIAC developed for the Army and introduced in 1946.  Designed by scientists at the University of Pennsylvania, ENIAC weighed 30 tons and employed 17,468 vacuum tubes.  Its successor was the even bigger MANIAC designed in Los Alamos and introduced in 1952.

Digital technology works by reducing everything to discrete bits and bytes and using its vast power and speed to bull through them to a solution.  “Discrete” is the key word.  Every part of every problem is represented by a series of 1’s and 0’s corresponding to the on and off conditions of solid state devices such as transistors.  This is the meaning of the term “binary,” the either/or condition in which there is nothing in between.  The world, or at least the world we perceive with out unaided senses, is not like that.  There are (or at least our brains think there are) infinite shades of gray, spectra of color, nuances of meaning, subtleties of expression and gradations of touch, taste, sight and sound.  The digital world can simulate these seamless transitions but cannot duplicate them.  A common example is digitized music which can reproduce a performance to an impressive exactitude.  Still, an experienced listener cannot ultimately be fooled.  Of course, the same listener would not be fooled into thinking an analog recording was anything but a recording in part because the old technology introduced extraneous noise.  Still, the sound of vinyl is often heard as more “natural” and more “human.”  Digital sound is often perceived as too perfect, too cold, too engineered.  No one has ever heard anything in the real world quite so detached.

Before proceeding, we must pause to recognize quantum theory and its offspring which suggest that in the subatomic world matter and energy do indeed exist in discrete packets.  These exotic phenomena are the true building blocks of the sensible universe.  Thus, elegance seems to demand that all creation must be discrete at heart.  If so, it would follow that reality can be better represented by digital than analog.  At least if digital is taken as discrete or discontinuous.  A black and white photograph made on film is composed not of infinite gradations of gray but of eleven discrete tones.  Our eyes are capable of much finer discriminations—a visit to any paint store will confirm there are more than fifty shades of gray—but the number is not infinite.  No matter, that is not the way we perceive things.  Our minds fill in the gradations.  Hamlet says, “…there is nothing either good or bad, but thinking makes it so.”  In a similar manner, there is nothing digital or analog but perception makes it so.  And from the time we came down out of the trees, our senses have suggested that, for the most part, the world is continuous or analog.

True binaries or opposites like “on” and “off” do exist:  positive and negative in the algebraic sense, up and down in the physical sense, absent and present, yes and no, clockwise and counterclockwise, left and right.  But there is a rich class of antonyms that constitute more complex dualities, things that are not absolute opposites but are contrasting in varying degrees.  The known and the unknown.  Good and evil.  Order and chaos.  Male and female.  Long term and short term.  Individual and community.  Self and other.  In each case, there is an element of tension, vigorous or feeble, between the members of the pair.  Good and evil are certainly opposed but there are gradations of both.  Not only are these dualities not opposites but, in some cases, they depend on each other.  “Male” is without meaning except in relation to “female.”  Yet both exist independently.  And both manifest themselves across a single, fairly wide spectrum of gender characteristics.  The common requirement to check a box for “male” or “female” is almost without meaning.  Facebook provides 58 choices for its members including “neither,” “other” and “pan.”  If these don’t work you can provide a customized descriptive.  Imagine Plato attempting to discuss the pure idea of “male.”  Anything he might say must ultimately include reference to “female.” Like yin and yang, each is half of an idea yet both can stand alone.

Rather than think of such ideas as “opposites,” I prefer to call them “complementarities.”  They are fundamentals of the human experience.  To take another example, very early in life we come to the realization of “self” and “other.”  The recognition is simultaneous; “self” implies “other” and vice versa.  Except for people with multiple personality disorder, the distinction between self and other is absolute—binary—but they are not opposites.  “We” addresses a community made up of multiple “self” and “other” pairs.  As we mature, we learn to subdivide “self” into “me” and “mine” which are different but again not opposite.

The real world of our senses is complex, ambiguous and subtle.  An analog computer can target artillery fire very accurately and fast enough for most situations.  But the genius of digital technology is to make it seem simple while doing so with much greater precision. Deep Blue is an IBM supercomputer that can play chess at the grandmaster level and Watson is an even more advanced breed that can beat the best Jeopardy players in the world.  But, in spite of decades of development, neither can carry on a real conversation with a human being.  You can ask Apple’s Siri for restaurant recommendations.  She will make reservations for you and call a taxi to get you there.  But don’t ask her to discuss the differences between Beaujolais and Burgundy or even choose between Pinot gris and Pinot noir. 

In 1950, Alan Turing published a paper purporting to answer the question “Can machines think?”  By the criteria he established for the famous Turing Test, the unequivocal answer was that there was no reason to suppose they would not someday do so.  Within a few years, researchers at MIT developed ELIZA, an artificial intelligence system that simulated psychotherapy sessions between a computer/therapist and a human patient.  The developers knew it was therapeutically primitive but were surprised by how readily it was accepted by its human users.  It seemed to pass the Turing Test even if many felt it was due to the simplicity of the test rather than the sophistication of the software.  Today’s “chatterbots” are less ambitious but much improved in terms of conversational realism.  Still, anyone who has ever used a computerized Help Line knows they are sometimes efficient but often hateful substitutes for knowledgeable humans.

The question is still being asked as to whether digital computers will ever reach a stage where they can realistically mimic the human brain.  It would be foolhardy to suggest such progress is impossible.  However, if it ever does come about, the computer in question will probably employ some hybrid technology.  Not to worry.  At present, such verisimilitude seems unlikely for economic as well as technical reasons.  For one thing, the product development emphasis remains on traditional device speed and capacity both of which put a premium on compression of data.  For example, when music or video is streamed, the result is lower fidelity than that delivered by CD’s or DVD’s.  Each successive generation of Apple’s iTunes application has delivered slightly lower fidelity which most users don’t notice or ignore.  The quality of digital photographs taken by smart phones has increased with each generation of the technology but is still lower than that of either conventional film or high end digital technology.  Cost and convenience are dramatically improved and even users who notice the decline in quality are happy to make the compromise.  Digital material is also easier to manipulate.  The old saw, “A photograph doesn’t lie,” was never guaranteed given the ability of darkroom technicians and airbrush artists to modify what the camera saw.  But programs like Photoshop make it easy to alter an image completely.  Whether the intent is benign or malicious, this level of control alters the nature of truth itself.  A viewer cannot know anything about what the camera actually saw or how the photographer felt about it.

This is an ancient problem.  Writing and printing, music, photography and other “extensions” of human senses (to borrow Marshall McLuhan’s locution) are all languages.  The great semanticists of the 1930’s knew that language acts to constrict thought in order to serve the interests of communication.  Poetry, for example, is a compromise between logic and emotion effected through the intermediation of vocabulary, grammar and other conventions.  It proceeds from mental imagery we refer to with words like inspiration, intuition and imagination, processes that cannot be quantified.  Moreover, every language treats the constriction process differently so that the architecture of a language accounts for a significant part of the world’s cultural diversity.

Language is the central tool of knowledge but it also burdens us with an unavoidable problem.  Except for the possibility of telepathy, there is no way for people to communicate anything other than the most basic ideas without using twice-translated symbols.  A speaker or writer chooses a word to convey an idea that is not a word and the listener or reader translates that symbol into his or her own mental construct.  Both run the risk of treating the word as though it is the idea it refers to.  But really it is merely an undisciplined coding system that is more or less analogous to the way we think. 

Digital language is altering culture not only because numbers are more abstract than words and binary numbers are more abstract than continuous numbers, but also because it is insidious.  The 1’s and 0’s are hidden from view and the vast majority of people are unaware of them.  To a noticeable extent, it has already made us digital thinkers.   As a society, we have become data-driven.  We assess our scholarship by counting the number of times our publications are cited in other publications, our visual culture by the prices investors pay for art at auction and our spirituality by the frequency of church attendance.  We follow opinion polls assiduously at a time when societal dispositions are conspiring to significantly reduce their validity and reliability.  We reduce a sport like baseball to a vast number of sums and ratios and believe religiously that the data can help us predict the future.  But such numbers are merely imperfect surrogates for the things they purport to measure.  They are never the referents themselves.

The idea that Galileo’s dispute with the Holy Inquisition was about heliocentricity is simplistic at best.  It was rather a debate about the nature of reality.  Galileo, of course, was right about the earth’s rotation around the sun because he had observed and measured it.  But he was wrong in believing that the laws of nature are written in mathematics or, if not wrong, he was merely expressing an ingenious but unsupportable opinion.  To me, mathematics is a human construct.  It is often quite useful but Pi is only a useful ratio.  Indeed, the “laws of nature” are nothing more than artifacts.  There is nothing to suggest that “life” in some parallel universe must be carbon-based or that the hare will always outrun the tortoise.  Pythagoras was able to prove that the expression:

A² + B²  = C²

defines the relationship between the hypotenuse and the legs of a right triangle.  His discovery turned out to be one of the most useful in the history of science.  But it is not a right triangle.  It derives from one aspect of the triangle.  Similarly, 261.625565 hertz describes one important aspect of Middle C but is not music.

The human race has turned a corner in its evolution.  Until recently, energy was what Daniel Bell called the “strategic and transforming resource” of the developed world.  It was energy that, when harnessed and applied to raw materials by technology, added value to them.  If you date the industrial revolution from James Watts’ first practical steam engine in 1776, it lasted about 200 years.  Today energy has been displaced as the great value adder by information harnessed and applied by digital technology.  The information revolution is different from anything that preceded it.  For one thing, information is the only resource that cannot be depleted.  The more we use it, the more it grows and the more valuable it becomes.  For another, it is ubiquitous, instantly available to anyone anywhere and acting as though it wants to be free.  For better and sometimes for worse, it both enables and compels the global economy.

In 1964, I had a conversation with a scholarly officer at the Command and General Staff College at Fort Leavenworth, Kansas.  He showed me a sandbox configured to display the terrain and troop dispositions around Cemetery Ridge near Gettysburg, Pennsylvania on the morning of July 3, 1863.  Twenty or so students were supposed to re-think Lee’s strategy and revise Pickett’s tactics accordingly.  The Army wanted to computerize the exercise so that each student’s responses could be critiqued in real time.  Given the sheer number of variables and the fluidity of the situation, it would be hard to imagine a more complex problem whether it was considered from the perspective of a commander on the ground or a computer programmer in a software laboratory.  The textbook answers had failed Lee, Longstreet and Pickett.  Although the Confederate attackers outnumbered the Union defenders by more than 2 to 1, their artillery could not see the enemy through the smoke and their infantry occupied the lowest part of an undulating field.  The software was supposed to assess each decision every student made and to display the results in the same amount of time it would have become clear to the battlefield commanders in 1863.  Today’s most sophisticated computer games are nowhere near as complex and I am not sure today’s digital computers could handle the challenge.  Given that feedback time was meant to correspond to real time, an analog system would have a better chance especially if a student came up with a brilliant new solution.  The analog machine “thinks” more like a creative human being. 

It reassures me to believe that neither technology could whisper the only good solution in Lee’s ear:  Don’t do this;  get the hell out of here.

Brave new world indeed!  There is an important advantage to be gained by thinking digitally or numerically and that is it may nudge us away from the scandalous scientific illiteracy that besets our society.  If, as we think, language is the springboard of culture, constant exposure to the language of science is sure to ease our anxieties about it.  The caveat is we must recognize that science, like every other approach to knowledge, employs a metaphoric language.   For all our history we have thought, reasoned, imagined and even dreamed in metaphors.  Dealing with their imprecision, their fuzziness has served us well.  It is important that we disconnect our young people from their smart phones long enough to teach them to use the brains they were born with.

_______________

*Jerry Harkins is a statistician who, long ago, spent two years as a member of the Army Corps of Engineers rising to the rank of Specialist Fourth Class, the same pay grade as a corporal.  In addition to regular turns at KP and guard duty, he was a member of a small unit of displaced draftee academics that undertook special projects related mostly to automated learning systems.  He is well aware of the epistemological problem posed by quantum weirdness even if he remains stubbornly unable to understand anything else about it.