This is a guest post by Hal Morris.
For our pre-technical ancestors, the clockwork at the bottom of the material world was so clothed in messiness that hardly a trace of it appeared on the surface. But you could say that three exposed bits collectively formed a Rosetta stone to the mathematical language of nature: a thrown rock, a pendulum, and the solar system, revealed by the night sky. The last had to be viewed from such a difficult angle that reams of tables, centuries worth of exact observations, and a huge advance in mathematics were required to see it, but it was there to be seen.
The concept of machine pervades our culture, and has driven many philosophical debates for centuries.
For example, it is often argued that living organisms, or the human mind, are “ultimately just machines”. I.e. underlying all the messy organic complexity of the world’s surface is a level at which things function with mechanical or mathematical precision. Sometimes it is then too blithely concluded that this proves we can eventually replicate anything, including the human brain.
But if “everything is a machine”, it can’t contribute anything to any argument because it doesn’t distinguish anything from anything else.
The fact is, we have intuitions about what a machine is, namely a man-made object the exact size, shape, and other properties of whose parts are precisely specified and realized so as to behave, or respond to manipulation, in precise and predictable ways.
In nature, without high-tech tools, one finds virtually nothing like this. To find mechanical behavior in a non man-made earthly environment with pre-modern eyes, we have to start with systems simpler than what we normally think of as a machine. A good example is a rock traveling through the air after it is thrown. Once set in motion, there is such predictability about the trajectory of the rock that a skilled person can make it strike in a certain place from 100 feet away or farther.
The business of throwing rocks has two parts: first the thrower has an aim to hit a target, and does whatever it is that sends the rock into the air, and second, the rock flies along a predictable path until it is stopped by something (hopefully the target).
A good account of the first part was completed in the 17th century by Newton, “standing on the shoulders of giants”. As for what goes on between having the aim, and a human hand’s release of the rock, we are still connecting the dots
The path of a flying rock or cannonball, as science tells us today, is simple and predictable (to some approximation) because we have managed to almost isolate just two forces out of the usually messy stew of nature. We can almost say that the rock’s path is determined by (a) its initial velocity, including direction, and (b) the earth’s gravity, and the interplay between these two is indeed simple, and reducible to a formula, as Newton showed.
Now the caveman rock-thrower could not work out even the “easy part” (the one Newton solved); but the right-brain, a kind of analog computer which deals with spatial relations in a nonverbal way, was solving it in a way. We know there is another complicating force, wind resistance, but it is strongest with light objects like feathers, and negligible, to a useful approximation for rocks and other compact heavy bodies.
Machines are built on this principle of isolating forces, so that those affecting any one part are extremely simple to understand and control. Forces like friction, usually in conflict with our design goals, are minimized partly through using polished, precisely straight or precisely round, or otherwise precisely, regularly, fashioned surfaces, and arrangements. Once set in motion, we know which way our mechanical contraption will go; or if we want it to go a certain way, we know how to make this happen. Where do we find anything of such regularity in nature?
Our rock in free flight may be just the thing. Being a system that only exists for a second, it seems a stunted sort of “machine” at best. Yet much of what Galileo learned about nature came from studying such trivial mechanical systems as a compact body in passive free fall or flight.
Humankind has tried for millennia to make sense of the night sky. Early observers noticed exact arrangements of stars, or constellations seeming to glide across the night sky, locked in one large configuration, with the moon slowly passing among them. But a handful of lights in the sky wandered through the fixed background making strange loops, if one takes the trouble of mapping them night after night. These tantalized humans, who rightly intuited something important behind the irregular regularity of the planets. Mostly all they got were some concepts of cyclic time, useful for agriculture, and astrology. But it drove people to keep meticulous records, from Babylonian times, always looking for something more.
In the 17th century, with the help of ancient record keepers, and more recent astronomy, Isaac Newton revealed the whole celestial machine from rocks to planets in a new mathematical language. As Alexander Pope wrote “Nature and nature’s laws lay hid in the night. God said, Let Newton be! and all was light!”, and science was never the same. He showed that the rock, in its parabolic path, and the moon in eternal orbit were two instances of the same thing, and that the earth and those wandering dots in the sky were as well, in orbit around an unimaginably large sun, and he mathematized their movements
The movement of the planets appeared slow (only due to their vast distance), and nightly plotting of their position against the fixed stars revealed a sense of their motion in an appropriate time scale (days). But there was no way of discovering the laws of motion of earthly objects in their brief free flights, so as to make any comparison without a way to measure seconds, and smaller units of time. The pendulum with its one moving part was the next best thing to a natural machine. Galileo discovered its ability to “count time”. Some inexact clocks existed, so seconds were at least thinkable. Soon after Galileo, it was discovered that a 39.1 inch pendulum marked off a second with each swing, and clock technology began to improve rapidly.
Galileo noted the more fundamental fact that a pendulum could go on swinging back and forth for days, and even as its speed decreases and its swing shortens, the length of time for each swing remains almost exactly the same. To explain this in terms of physics is quite complex. The formula given by Wikipedia is:
We’ve noted that any kind of mechanical regularity is hard to spot in natural objects, or in any object as simple as a pendulum. We have to look very hard to find it, and in doing so, we develop a habit of focusing on the regular., and taking it for granted.
We see the regularity that we do, in fact, wherever we look, not so much because we “discovered” the regularities of nature. It is more because we rearranged the substances of our world into regular, governable, objects of a sort not found in nature. We have done this so well that it is hard to imagine not finding mechanical regularity all around us.
The invention of the clock, and discovery of clock-like regularity in the universe were needed to launch modern physics, and the mathematical sciences that came in its train, and we can’t overstate the impact that had in creating a true scientific attitude to nature.
But some respectable disciplines begin with less perfect and well-polished touchstones.
Any science requires, and is shaped by, a tractable domain, and the progress of science is measured in the tractability we have found or engineered with our observation, theories, and instrumentation. Just what qualifies, and how we learn to see a tractable domain, is more a matter of art than logic.
The parable of the blind men and the elephant, generally leads to a dead end of ignorance, but if our blind men were a bit more clever and persistent, we might have a parable of a portal into a tractable domain.
The story comes from India. One version, in Wikipedia follows:
Six blind men were asked to determine what an elephant looked like by feeling different parts of the elephant’s body. The blind man who feels a leg says the elephant is like a pillar; the one who feels the tail says the elephant is like a rope; the one who feels the trunk says the elephant is like a tree branch; the one who feels the ear says the elephant is like a hand fan; the one who feels the belly says the elephant is like a wall; and the one who feels the tusk says the elephant is like a solid pipe.
If only our blind men had been used to the customs and incentives of science, comparing notes, arguing this interpretation and that, they could have eventually made sense of the critter, perhaps making a clay model of an elephant that one person could get their hands around. Of course if they are no good at listening to each other, or lack persistence and/or the right sort of discipline this won’t happen.
Earlier, I argued that the decipherment of a tiny number of “natural machines” pointed the way towards solid analytic science. Now I’d like to suggest a different metaphor that finding the touchstone of a science is like finding and comprehending invisible elephants (rather than make all scientists blind, let’s for the moment imagine elephants as invisible).
What does it look like when we are failing to find our elephant? Maybe one man really is grasping an invisible pillar, another an invisible tree branch, or hand fan, and another pushing on a wall. No wonder their observations don’t add up. If they still try to force their observations to add up to one thing, something like the theory of humours (see below) may result.
So what does a science finding its elephant look like? There should be some convergence of observations when the blind men work together effectively. Maybe four men are saying “this is like a pillar”, and they can tell by listening they are close to each other, and reach out to grasp each other’s hands, and get a sense of where each man is. Maybe all link hands to discover that the pillars are in a rough square. Someone again says this is like a rope, and they wave their hands around until the one with the “rope” finds he’s roughly equidistant from two of the “pillar” men. And on and on. Someone says “this is like a creek”, and the others say “You’re too far away, come back over here.” And as long as they stay in contact with the object, certain observations occur repetitively, and all the relations between the observations begin to add up to something. Maybe someone bounces a basketball their way, one grasps it and says “Aha! something new!” but they soon realize it isn’t part of the thing they’re trying to understand. It is “irrelevant data”, or “noise” as in Nate Silver’s The Signal and the Noise: Why So Many Predictions Fail – but Some Don’t.
After recognizing one specimen, where might invisible elephantology go? In seeking more elephants, one might carry around a clay model of one, some instruments (as simple perhaps as tape measures) for measuring the circumference of a leg, or height and length; one might discover related phenomena, such as elephant footprints, leading to a methodology of following them to see where elephants go and congregate, and so on.
When observations add up and complement each other, and it becomes increasingly clear which observations belong to the new object of study, and which do not, we can say there is a tractable domain. I’ll make no attempt to define tractability, but hope I have somewhat illustrated it. Until the 20th century, medicine was largely intractable. We had only glimmerings of understanding here and there that could not be worked into any sort of whole. Failure to admit this — wishful thinking — led to systems like that based on the “humours” (or 4 basic fluids supposed to account for the body’s workings and malfunctions) which lead to inappropriate bleeding and purging, and sometimes even stopping up things that shouldn’t be stopped up, all on the theory that too much or too little of some “humour” caused a given syndrome.
A tractable domain means one has a good sense of the thing that is under study – and graspable techniques that can lead one deeper into its understanding; techniques likely to be very specific to the domain. It might be invisible elephants, or equally invisible atoms, whose properties can only be known via more complex and roundabout ways than merely looking. The idea that all scientific methodology can be summarized by one “method” is more often heard in disciplines that are trying hard to be scientific, such as the social sciences, than in those that have actually produced a wealth of regularities.
Consensus that we are talking about the same thing (from different angles) is a sign that we may have found our elephant. In 17th century physics it was the Earth, Moon, planets, and Sun and it lead to powerful general principles. In the 18th century, various manifestations of electricity presented many novel mysteries; for example, sparks generated by various electrical machines looked like tiny lightning bolts, but could they actually be the same thing? Benjamin Franklin is supposed to have answered this, on one of the first of many steps towards drawing electrical phenomena out of the mists and and into tractability.
Many of the most interesting and important results of mature science must be reached through interaction, and sometimes hot debate between different disciplines or specialties, and the ultimate arrival at consensus.
This essay was partly inspired by an argument with someone who insisted experiments supporting AGW (human contribution to global warming) were clearly not following “the scientific method”. His “scientific method” was received from a psychology professor some decades ago, as selecting a hypothesis, choosing a test group and a control group, and so on. I.e. we get to science by collecting inductive evidence for true-false questions.. This generic methodology is, I think, symptomatic of an aspiring science (like psychology, at least several decades ago) that has too little sense of the shape of its object of study — its invisible elephant.
I’ve tried to suggest that a hallmark of a mature science is a methodology that reflects or fits the object studied. The kind of interaction that pulls an object out of obscurity will, as the unique shape of the object emerges, call for new techniques for answering questions of a sort no one ever dreamed of before. No single method tells us we are being scientific. I don’t see how one can judge a science as mature and truth-generating without familiarity with other sciences, and getting into the substance and details of the science Perhaps all sciences use the textbook “scientific method” on occasion, just not exclusively. But generally, in assessing the solidity of a science, we must look at both the methodologies and the manifestations of the object(s) of study, and the processes by which they interact, and there is no simple rule for how we judge that.
Cosmologists, to understand the structure of the universe, turn optical, radio, and nowadays highly computerized telescopes to the sky so it can be scrutinized bit by bit. They say “this looks like one of those” (galaxy, nebula, certain kind of star, or black hole), and if on closer and closer observation, it looks more and more like one of those, they are kind of relieved, and kind of disappointed, because this isn’t the day they see something that nobody else knows about. If they can’t identify a thing, they will try to do “whatever it takes” to arrive at some understanding. The spectroscopy that discovered helium on the sun (see below) before it was found on earth, and showed the sun to be mostly hydrogen and helium shows us there are other kinds of stars with heavier elements. Such are some of the methods of cosmology.
The late 18th and early 19th centuries were a heyday of exploration and of natural philosophy, and the scrutiny, analysis, description, collecting, and classifying of everything on the surface of the earth, or accessible via digging. James Hutton and Charles Lyell convinced the scientific world that the earth’s surface had evolved for millions of years, that much of the of rock found on land was formed as silt in the sea, and, joined with countless other layers formed by various processes formed a sort of rock parfait , with each deeper layer formed at an earlier time. The layers, representing “deep time”, often contained fossils of plants and creatures that grew stranger the deeper and older the layers. This study is called stratigraphy, a part of geology. The classical “elephant” might be a geological record, in rock, of hundreds of millions of years in some region, revealed by instances of identical “earth parfait” found hundreds of miles apart, the strata being the same not just in color and consistency, but down to the kinds of sea-shell fossils found in a given layer.
The 19th century was also the century of chemistry, and the “elephants” were new elements. The isolating and understanding of oxygen, followed by hydrogen, led to a new idea of pure substances, or elements. Various metals were among the first elements identified, through the heating of ores to remove oxygen, and over time, more complex processes. By around 1860, it was discovered that each element gave off light only in a certain set of frequencies (the prism, since Newton, was shown to split mixed light into component frequencies). This lead to the discovery of helium as a major component of the sun, long before it was ever found on earth. The methods were a kind of grab bag, evolving as greater insight was gained, as can be learned (with pleasure) by reading Oliver Sacks’ Uncle Tungsten: Memories of a Chemical Boyhood, which is both memoir and history of science.
Set pieces like the trial of Galileo occupy and excite our imagination, and were used to promote the importance of the Enlightenment and shedding of old orthodoxy. The individual vs the greatest ideological power of the time makes a great story. But it also flatters our tendencies to see great matters from a “God’s-eye view”; the attitude of hero vs the universe (and world of men), often great for scientific puzzle solving, but sometimes destructive to hard-won and well deserved trust. No ideology has a monopoly on this sort of narrative. Liberals have had Bertolt Brecht’s Galileo, Vincent Van Gogh and they used to have Columbus; others favor Dirty Harry, Jack Bauer and Anthony Watts of the “Watt’s Up With That” anti climate change website, all viewed by their fans as standing alone with their truths against the establishment.
But Newton, as he said, stood on the shoulders of giants, not to mention hundreds of humble sky watchers making their notations night after night. Science took its first big steps towards benefiting society in the age of salons and establishments like the Royal Society, with its “Transactions” the first real scientific journal, and simple reproducible experiments presented for groups to comment on.
Science still on occasion needs the unique visionary who finds a new elephant, or finds out that everybody else had the elephant upside down. Sometimes they can’t make themselves understood, and suffer frustration, but, at least in the hard sciences, a unique and true vision in the end wins the greatest rewards, such as the the Nobel Prize. Scientific culture is such that if nearly everyone is mistaken, and one person can demonstrate this, they may be controversial for a while, but will be lionized in the long run (sometimes posthumously).
Scientific consensus has been confused with group-think, but is really a matter of many scientists putting maps, table, graphs, observations and experiments together and after much wrangling coming to approximate agreement about what they add up to.
A discipline may be so vast that no one scientist fully understands, and like ordinary mortals, they need screening methods to determine who to trust and what is solidly known. In some cases, as in discovery of the “top quark”, a study with 451 co-authors as noted in Paul Thaggards’ 1997 “Collaborative Knowledge” (Noûs, 31: 242-261), no one person can follow the entire demonstration, so trust is essential, though it should be earned in an atmosphere of skepticism.
One harsh criticism aimed at Global Warming is that if there is a consensus, it is just that sort of authoritarian “group think” that Galileo confronted in the Inquisition, or that mainstream climatologists have abandoned “the scientific method”. Some who say they have a case against AGW maintain that they are practicing the true “scientific method” and since their demonstrations or studies aren’t adopted, someone has to be rejecting the scientific method.
There is a wonderful case study in Miriam Solomon’s book Social Empiricism (Bradford Books, 2001), of how over several decades, a consensus was reached on continental drift i.e. plate tectonics. (It is also covered more vividly and with less of an overt Social Epistemology emphasis in Naomi Oreskes 2003 Plate Tectonics: An Insider’s History Of The Modern Theory Of The Earth).
At the end of the 19th century, Eduard Suess an Austrian paleontologist and geologist proposed that India, South America, Africa, Australia, and Antarctica had all been connected in the distant past, based on similarities in stratigraphy, and the occurrence of certain fossil plants from particular prehistoric times only on these continents. India, not even being a separate continent but broadly connected to the rest of Asia makes this especially hard to explain. The part of India studied was called Gondwana, and Suess thus called the joined region Gondwanaland. He did not think of movement of continents, but of long vanished land bridges. Others in the past had speculated about some or all of these continents having been physically joined and later dispersed, based on the way they seemed to fit together, which is now the accepted theory.
Continental drift was first proposed by a German geologist and meteorologist named Alfred Wegener in 1912, but the theory didn’t go very far for a couple of decades. In the 1930s, geologists studying the formation of mountain ranges became interested in paleomagnetism, the study of variations in the Earth’s magnetic poles in the distant past, which induced slight magnetism in volcanic rocks as they solidified, oriented with the current state of the poles. Actually, although these changes were first thought to represent “wandering” of the poles, later evidence showed rocks formed at the same time on different continents pointed in different directions, as if one or both had rotated since these rocks formed. Reconstruction of the putative Gondwanaland, and movement of continents as they separated suggested a rotation consistent with this. While paleomagnetists were pursuing this argument, some in other disciplines not, finding any confirmation of the astonishing “wandering continents” hypothesis ridiculed them as “paleomagicians” (Solomon p104). The data could be very confusing and contradictory indeed, until it was realized that the poles frequently reversed polarity throughout geological time (even as recently as approx 10,000BCE). Before recognition that the record is full of 180 degree reversals, looking for subtle rotations showed inadequate grasp of the “elephant”.
Such research was halted by WWII, but the most dramatic piece of evidence for drift came as a byproduct of the war and subsequent cold war, which lead to extensive mapping of the ocean bottom to facilitate games of hide and seek between submarines. Sometimes called “magnetic striping”, it was the discovery of ridges in the ocean floor from which magma or molton volcanic rock seemed to ooze, and “stripes” on either side of the ridges of volcanic rock which were older the further they were from the ridge, and whose polarity was that of the Earth’s at the time of their creation. These alternating polarities showed clearly that the ocean crust was moving away from the ridges, and as other research placed dates on these reversals, provided a timetable of the spreading.
It remained only to see these ridges as one of three kind of boundaries between ocean plates: divergent (just described), convergent (ocean floor descending under the edges of continents), and transform (plates sliding past each other) and a comprehensive theory, called Plate Tectonics emerged in the mid 1960s.
The next group to support the theory were seismologists who found correlating data in the ages of volcanic islands formed on either side of a ridge, the furthest away being the oldest, and also in the patterns of earthquakes in the vicinity of convergent boundaries.
As the picture took shape, there were many interdisciplinary conferences on the new findings, attended by some in the lagging disciplines (stratigraphy, and paleontology, ironically the source of some of the first clues), and within a few years these disciplines began adding to the evidence for continental drift.
The picture can be seen in this animation of the last 200M years, in which India seems to fly across the southern seas, and you can imagine it slamming into Asia to form the world’s highest mountain range.
This kind of process is really what is meant by scientific consensus. It is not groupthink, but different specialists deciding whether their field supports a theory or not, and as long as one specialty holds out, those who feel convinced of the theory remain uncomfortable, wondering if the recalcitrant fields, whose work they don’t understand so well, know something they don’t. Once scientists in all these fields regularly produced discoveries supporting and helping elaborate the theory, there was a consensus. Scientists will naturally and I think rightly hesitate to confirm a thesis when their own discipline should provide confirmation but so far, has not.
As long as a proposed theory leaves one or more major disciplines skeptical, general science publications will tend to say the matter is controversial. It is generally only when the last discipline(s) are convinced that these publications say there is a scientific consensus.
I began to read science seriously around the time when continental drift became a settled question in science, when plenty of older books and articles still referred to it as controversial.
I’ve been a casual observer of the global warming debate over the decades, but the process looked very similar (to me) to that of continental drift, moving from “this is a theory that some scientists have” to “still controversial” to “scientific consensus”. As with continental drift, the sources of evidence are diverse, from layers of trapped atmospheric bubbles in ancient ice strata to meteorology to seismology to chemistry and to computer model building to the physics of the planet and the solar system.
The picture is very difficult to assemble, especially to see a tendency on top of several well known sources of climate oscillation. Any single paper or piece of evidence for global warming should not convince one, and some bear only tangentially on the main lines of argument. Scientists look for many demonstrations from many different angles, and keep looking even after they are convinced, at the risk of finding something contradictory. Likewise, any single anti-AGW argument/demonstration is not going to turn the world around. It has never been the case that the most recent paper determines what is believed in a field. If there is a contradiction then either the new study is flawed or based on “noise” (the usual case) or the old ones are wrong in some way.