Technological Escalation and the Exploration Model of Natural Science
(1) Our cognitive competence is well accounted for by our evolutionary niche in the world's scheme of things. (2) The development of inquiry in natural science is best understood on analogy with exploration -- to be sure, not in the geographical mode but rather exploration in nature's parametric space of such physical quantities as temperature, pressure, and field strength. (3) The technology-mediated exploration at issue here involves an interaction between us humans and nature that becomes increasingly difficult (and expensive) as we move ever farther away from the home base of the accustomed environment of our evolutionary heritage. The course of scientific progress accordingly involves a technological escalation -- an ascent to successively higher levels of technological sophistication that is unavoidably required for the production of the requisite observational data.
Section 1.-- Accounting for our cognitive competence
How is it that we humans are actually so competent in coping in the domain of cognitive complexity? How is it that we possess the intellectual talent to create mathematics, medicine, science, engineering, architecture, literature, and other comparably splendid cognitive disciplines? What explains the immense power of our intellectual capacities?
To be sure, at a level of high generality the answer is relatively straightforward. Basically, we are so intelligent because this is our place in evolution's scheme of things. Different sorts of creatures have different ecological niches, different specialties that enable them to find their evolutionary way along the corridor of time. Some are highly prolific, some very hard, some swift of foot, some difficult to spot, some extremely timid. Homo sapiens are different. For the evolutionary instrument of our species is intelligence -- with everything that this involves in the way of abilities and versatilities. Thus if we weren't so intelligent, we wouldn't be here as the anatomical creatures we are. We have all these splendid intellectual capacities because we require them in order to be ourselves.
Of course it's not all just a matter of fate's lottery bringing intelligence our way. Evolution's bio-engineering is the crucial factor. Bees and termites can achieve impressive prodigies of collective effort. But an insect developed under the aegis of evolution could not become as smart as a man because the information-processing requirements of its lifestyle are too modest to push its physical resources to the development of intelligence.
Intelligence are an inherent concomitant of our physical endowment. Our bodies have many more independently movable parts (more «degrees of freedom») than do those of most other creatures.<1> Foot note 1_1 This circumstance has significant implications. Suppose a system with n switches, each capable of assuming an ON or OFF position. There are then 2[[sup_n]] states in which the system can find itself. With n=3 there are only 8 system-states, but with n doubling to 6 there are already 64 states. As a body grows more complex and its configuration takes on more degrees of freedom, the range of alternative possible states expands rapidly (exponentially). Merely keeping track of its actual position is already difficult. To plan ahead is more difficult yet. If there are m possible states which the system can now assume, then when it comes to selecting its next position there are also m choices, and for the next two there are m[[sup_2]] alternatives overall (ignoring unrealizable combinations). So with a two-step planning horizon the 3-state system has 64 alternatives while that 6-state system has 4096. With a mere doubling of states, the planning problem has become complicated by a factor of sixty-four.
The degrees of freedom inherent in variable movement over time are pivotal considerations here. The moment one walks upright and begins to develop the modes of motion that this new posture facilitates -- by way of creeping, running, leaping, etc. -- one has many more factors of physical movement to manage.
Considerations of this sort render it evident that a vertebrate having a more highly articulated skeleton, with many independently operable bones and bone-complexes, faces vastly greater difficulties in control and manipulation -- in what military jargon calls «command and control.» Versatile behavior involves more complex management. Physically more versatile animals have to be smarter simply because they are physically more versatile.
We are driven to devising greater capabilities in information acquisition and processing by the greater demands of the lifestyle of our ecological niche. The complexity of our sophisticated surveillance mechanisms in the context of friend-or-foe identification is an illustration. We can observe at a considerable distance that people are looking at us, discriminating minute differences in eye orientation in this context. The development of our sophisticated senses with their refined discrimination of odors, colors, and sounds is another example. Environmental surveillance is crucial for our lifestyle. We have to know which feature of our environment to heed and which can safely be ignored. The handling of such a volume of information calls for selectivity and for sophisticated processing mechanisms -- in short, for intelligence. Not only must our bodies be the right size to support our physical operations and activities, but our brains must be so as well.
The complexities of information management and control pose unrelenting evolutionary demands. To process a large volume of information nature must fit us out with a large brain. A battleship needs more elaborate mechanisms for guidance and governance than a row boat. A department store needs a more elaborate managerial apparatus than a corner grocery. Operating a sophisticated body requires a sophisticated brain. The evolution of the human brain is the story of nature's struggles to provide the machinery of information management and operative control needed by creatures of increasing physical versatility. A feedback cycle comes into play -- a complex body requires a larger brain for command and control, and a larger brain requires a larger body whose operational efficiency in turn places greater demands on that brain for the managerial functions required to provide for survival and the assurance of a posterity. As can be illustrated by comparing the brain weights of different mammalian species, the growing complexities and versatilities of animal bodies involve a physical lifestyle whose difficulties of information processing and management requires increasingly powerful brains. How one makes one's living matters: insect-eating and fruit-eating monkeys have heavier brains, for their size, than leaf-eating ones do.<2> Foot note 1_2
Here then is the immediate (and rather trivial) answer to our question: We are as intelligent as we are because that is how we have had to evolve to achieve our niche in nature's scheme of things. We are so smart because evolution's bio-engineering needs to provide those smarts for us to achieve and maintain the lifestyle appropriate to our ecological niche.
But there remains the problem of why evolution would take this course. Surely we didn't need to be that smart to outwit the saber toothed tiger or domesticate the sheep. Let us explore this developmental aspect of the matter a little.
The things we have to do to manage our lifestyle must not only be possible for us, they must in general be easy for us (so easy that most of them can be done unthinkingly and even unconsciously). If our problem-solving resources were frequently strained to the limit, often groaning under the weight of difficulty of the problems that they are called on by nature to solve in the interests of our lifestyle, then we just wouldn't have achieved this modus operandi.
For evolution to do its work, the survival problems that creatures confront have to be by and large easy for the mechanisms at their disposal. And this fundamental principle holds just as true for cognitive as for biological evolution. If cognitive problem-solving were too difficult for our mental resources, we wouldn't evolve as problem-solving creatures. If we had to go to as great lengths to work out 2+2 as to extract the cube root of a number, or if it took us as long to discriminate 3- from 4-sided figures as it takes to discriminate between 296 and 297-sided ones, then these sorts of issues would simply remain outside our repertoire. The «average» problems of survival and thriving that are posed by our lifestyle must be of the right level of difficulty for us -- that is, they must be relatively easy. And this calls for excess capacity. All of the «ordinary» problems of one's mode of life must be solvable quickly in real time -- and with enough idle capacity left over to cope with the unusual.
A brain that is able to do the necessary things when and as needed to sustain the life of a complex and versatile creature will remain underutilized much of the time. To cope during times of peak demand, it will need to have a great deal of excess capacity to spare for other issues at slack times. And so, any brain powerful enough to accomplish those occasionally necessary tasks must have the excess capacity to pursue at most normal times various challenging projects that have nothing whatsoever to do with survival.
These deliberations resolve the objection that evolution cannot explain our intellectual capacities because we are a lot smarter than evolution demands -- that, after all, nature does not quiz us on higher mathematics or theoretical physics. What is being maintained here is not the absurd contention that such disciplines as such are somehow an evolutionary requisite. All that is being said is that the capacities and abilities that make such enterprises possible are evolutionarily advantageous -- that evolution equips us with a reserve capacity that makes them possible as a side-benefit. The point is that an intelligent creature whose capacities do not allow of development in these abstract directions just isn't smart enough to pass evolution's examinations in other matters -- that is, would not be able to make intelligence its evolutionary specialty after all.
The brain/computer analogy once again proves helpful in this connection. Very different things can be at stake with being «simple»: the simplicity of «hardware» involved with comparatively less complex computers is one sort of thing, while the simplicity of «software» at issue with comparatively less complex programs is something quite different. And there are clearly tradeoffs here: solving problems of the same level of difficulty is generally easier to program on more sophisticated (more complex) computing machines. Something of an inverse relationship obtains: greater machine complication can make the actual use of the machine easier and less demanding. And this circumstance is reflected in the fact that a creature which makes its evolutionary way in the world by intelligence requires a rather powerful brain.
To be sure, evolution is not, in general, over-generous. For example, evolution will not develop creatures whose running-speed is vastly greater than what is needed to escape their predators, to catch their prey, or to realize some other such strictly utilitarian objective. But intelligence and its works are a clear exception to this general rule, owing to its self-catalyzing nature. With cognitive artifacts as with many physical artifacts, the character of the issues prevents a holding back; when one can do a little with calculation or with information processing, one can in principle do a great deal. Once evolution opens the door to intelligence, it gets «the run of the house.» When bio-design takes the route of intelligence to secure an evolutionary advantage for a creature, it embarks on a slippery slope. Having started along this road, there is no easy and early stop. For once a species embarks on intelligence as its instrument for coping with nature, then the pressure of species-internal competition enters as a hot-house forcing process. Intelligence itself becomes a goad to further development simply because intelligence is, as it were, developmentally self-energizing.
The result of the preceding deliberations is straight-forward. Intelligence is the evolutionary specialty of homo sapiens. If we were markedly less smart than we in fact are, we would not have been able to survive. Or rather, more accurately, we would not have been able to develop into the sort of creatures we have become. Intelligence constitutes the characteristic specialty that provides the comparative advantage which has enabled our species to make its evolutionary way into this world's scheme of things. We are so smart because this is necessary for us to be here at all.<3> Foot note 1_3
In the course of deploying our intelligence on the world about us we arrived ultimately at the project of natural science. Gradually our natural curiosity got the better of us and we began to push the project of inquiry beyond the level of actual need.
Section 2. The exploration model of scientific inquiry
In cultivating scientific inquiry, we scan nature for interesting phenomena and grope about for the explanatory useful regularities they may suggest. As a fundamentally inductive process, scientific theorizing calls for devising the least complex theory structure capable of accommodating the available data. At each stage we try to embed the phenomena and their regularities within the simplest (cognitively most efficient) explanatory structure able to answer our questions about the world and to guide our interactions in it. But step by step as the process advances, we are driven to further, ever greater demands arise which can be met only with an increasingly more powerful technology of data exploration and management.
In theory, a prospect of unending scientific progress lies before us. But its practical realization is something else again. One of the most striking and important facts about scientific research is that the ongoing resolution of significant new questions faces increasingly high demands for the generation and cognitive exploitation of data. Though the veins of cognitive gold run on, they become increasingly difficult -- and expensive -- to mine.
In developing natural science, we humans began by exploring the world in our own locality, not just our spatial neighborhood but -- more far-reachingly -- our parametric neighborhood in the space of physical variables such as temperature, pressure, and electric charge. Near the «home base» of the state of things in our accustomed natural environment, we can operate with relative ease and freedom -- thanks to the evolutionary attunement of our sensory and cognitive apparatus -- in scanning nature with our unassisted senses for data regarding its modes of operation. But in due course we accomplish everything that can be managed by these straightforward means. To do more, we have to extend our probes into nature more deeply, deploying increasing technical sophistication to achieve more and more demanding levels of interactive capability. We have to move ever further away from our evolutionary home base in nature toward increasingly remote observational frontiers. From the egocentric standpoint of our local region of parameter space, we journey ever more distantly outward to explore nature's various parametric dimensions in the search for cognitively significant phenomena.
The appropriate picture is not, of course, one of geographical exploration but rather of physical exploration -- and subsequent theoretical systematization -- of phenomena distributed over the parametric space of the physical quantities spreading out all about us. This approach in terms of exploration provides a conception of scientific research as a prospecting search for the new phenomena demanded by significant new scientific findings. As the range of telescopes, the energy of particle accelerators, the effectiveness of low-temperature instrumentation, the potency of pressurization equipment, the power of vacuum-creating contrivances, and the accuracy of measurement apparatus increases -- that is, as our capacity to move about in the parametric space of the physical world is enhanced -- new phenomena come into view. After the major findings accessible via the data of a given level of technological sophistication have been achieved, further major findings become realizable only when one ascends to the next level of sophistication in data-relevant technology. Thus the key to the great progress of contemporary physics lies in the enormous strides which an ever more sophisticated scientific technology has made possible through enlarging the observational and experimental basis of our theoretical knowledge of natural processes. A homely fishing analogy of Eddington's is useful here. He saw the experimentalists as akin to a fisherman who trawls nature with the net of his equipment for detection and observation. Now suppose (says Eddington) that a fisherman trawls the seas using a fishnet of two-inch mesh. Then fish of a smaller size will simply go uncaught, and those who analyze the catch will have an incomplete and distorted view of aquatic life. The situation in science is the same. Only by improving our observational means of trawling nature can such imperfections be mitigated.<4> Foot note 1_4
This idea of the exploration of parametric space provides a basic model for understanding the mechanism of scientific innovation in mature natural science. New technology increases the range of access within the parametric space of physical processes. Such increased access brings new phenomena to light, and the examination and theoretical accommodation of these phenomena is the basis for growth in our scientific understanding of nature.
Section 3.-- Technological escalation: an arms race against nature
Natural science is fundamentally empirical, and its advance is critically dependent not on human ingenuity alone but also on the ongoing enhancement of our technologically facilitated interactions with nature. The days are long past when useful scientific data could be had by unaided sensory observation of the ordinary course of nature. Artifice has become an indispensable route to the acquisition and processing of scientifically useful data. The sorts of data on which discovery in natural science nowadays depends can be generated only by technological means. The discoveries of today cannot be made with yesterday's equipment and techniques. To conduct new experiments, to secure new observations, and to detect new phenomena, an ever more powerful investigative technology is needed.
The pursuit of natural science as we know it embarks us on a literally endless endeavor to improve the range of effective experimental intervention, because only by operating under new and heretofore inaccessible conditions of observational or experimental systemization -- attaining extreme temperature, pressure, particle velocity, field strength, and so on -- can we realize situations that enable us to put knowledge-expanding hypotheses and theories to the test. As one acute observer has rightly remarked: «Most critical experiments [in physics] planned today, if they had to be constrained within the technology of even ten years ago, would be seriously compromised.»<5> Foot note 1_5
This situation points toward the idea of a «technological level,» corresponding to a certain state-of-the-art in the technology of inquiry in regard to data-generation and processing. This technology of inquiry falls into relatively distinct levels or stages in sophistication -- correlatively with successively «later generations» of instrumentation and manipulative machinery, which are generally separated from one another by substantial (roughly, order-of-magnitude) capacity improvements in regard to such information-providing parameters as measurement exactness, data-processing volume, detection sensitivity, high voltages, high or low temperatures, and the like.
Physicists often remark that the development of our understanding of nature moves through successive layers of theoretical sophistication.<6> Foot note 1_6 But scientific progress is clearly no less dependent on continual improvements in strictly technical sophistication:
Some of the most startling technological advances in our time are closely associated with basic research. As compared with 25 years ago, the highest vacuum readily achievable has improved more than a thousand-fold; materials can be manufactured that are 100 times purer; the submicroscopic world can be seen at 10 time higher magnification; the detection of trace impurities is hundred of times more sensitive; the identification of molecular species (as in various forms of chromatography) is immeasurably advanced. These examples are only a small sample.... Fundamental research in physics is crucially dependent on advanced technology, and is becoming more so.<7> Foot note 1_7Without an ever-developing technology, scientific progress would cease. The discoveries of today cannot be advanced with yesterday's instrumentation and techniques. To secure new observations, to test new hypotheses, and to detect new phenomena, an ever more powerful technology of inquiry is needed. Scientific progress depends crucially and unavoidably on our technical capability to penetrate into the increasing distant -- and increasingly difficult -- reaches of the spectrum of physical parameters in order to explore and to explain the ever more remote phenomena encountered there.
The instrumentalities of scientific inquiry can be enhanced not only on the side of theoretical resources but preeminently on the side of technology of observational and experimental intervention. Pioneering scientific research will always operate at the technological frontier. For revealing here further «secrets» nature inexorably exacts a drastically increasing effort in to the acquisition and processing of data. This accounts for the recourse to more and more sophisticated technology for research in natural science.
No doubt, nature is in itself uniform as regards the distribution of its diverse processes across the reaches of parameter space. It does not favor us by clustering them in our accustomed parametric vicinity: significant phenomena do not dry up outside our parochial neighborhood. And phenomenological novelty is seemingly inexhaustible: we can never feel confident that we have got to the bottom of it. Nature always has fresh reserves of phenomena at her disposal, hidden away in those ever more remote regions of paramative space. Successive stages in the technological resources of scientific inquiry accordingly lead us to ever-different views about the nature of things and the character of their laws.
The salient characteristic of this situation is that, once the major findings accessible at a given level of sophistication in data-technology level have been attained, further substantial progress in any given problem area requires ascent to a higher level on the technological scale. Every data-technology level is subject to discovery saturation: once the potential of a given state-of-the-art level has been exploited, not all our piety or wit can lure the technological frontier back to yield further significant returns at this stage. Further substantive findings become realizable only by ascending to the next level of sophistication in data-relevant technology. But the exhaustion of the prospects for data extraction at a given data-technology level does not, of course, bring progress to a stop. Rather, the need for enhanced data forces one to look further and further from man's familiar «home base» in the parametric space of nature.
The requirement for technological progress to advance scientific knowledge has far-reaching implications for the nature of the enterprise. For the increasing technological demands that are requisite for scientific progress means that each step ahead gets more complex and more expensive as those new parametric regions grow increasingly remote. With the progress of science, nature becomes less and less yielding to the efforts of further inquiry. We are faced with the need to push nature harder and harder to achieve cognitively profitable interactions. The dialectic theory and experiment carries natural science ever deeper into the range of greater costs. We thus arrive at the phenomenon of technological escalation. The need for new data forces us to look further and further in parametric space. Thus while scientific progress is in principle always possible -- there being no absolute or intrinsic limits to significant scientific discovery -- the realization of this ongoing prospect demands a continual enhancement in the technological state of the art of data extraction or exploitation. Given that we can only learn about nature by interacting with her, Newton's third law of countervailing action and reaction becomes a fundamental principle of epistemology. Everything depends on just how and how hard we can push against nature in situations of observational and detectional interaction. As Bacon saw, nature will never tell us more than we can forcibly extract from her with the means of interaction at our disposal. And because this extraction can only be realized by ever deeper probings, this state of affairs has far-reaching implications for the perfectibility of science. The impetus to augment our science demands an unremitting and unending effort to enlarge the domain of effective experimental intervention. That there is «pay dirt» deeper down in the mine avails us only if we can actually dig there. New forces, for example, may well be in the offering, if one able physicist is right:
We are familiar, to varying degrees, with four types of force: gravity, electricity, the strong nuclear force that holds the atomic nucleus together and the weak force that brings about radioactive decay by the emission of electrons.... Yet it would indeed be astonishing if . . . other types of force did not exist. Such other forces could escape out notice because they were too weak to have much distinguishable effect or because they were of such short range that, no matter whether they were weak or not, the effects specifically associated with their range were contained within the objects of the finest scale that our instruments had so far permitted us to probe.<8> Foot note 1_8But, of course, such weak forces would enter into our picture of nature only if our instrumentation were able to detect them. This need for a constant enhancement of scientifically relevant technology lies at the basis of the enormous increase in the human and material resources needed for modern experimental science. Frontier research is true pioneering: what counts is not just doing it but doing it for the first time. Aside from the initial reproduction of claimed results needed to establish the reproducibility of reproducibility of results, repetition in research is in general pointless. As one acute observer has remarked, one can follow the diffusion of scientific technology «from the research desk down to the schoolroom»:
The emanation electroscope was a device invented at the turn of the century to measure the rate at which a gas such as thorium loses its radioactivity. For a number of years it seems to have been used only in the research laboratory. It came into use in instructing graduate students in the mid-1930's, and in college courses by 1949. For the last few years a cheap commercial model has existed and is beginning to be introduced into high school courses. In a sense, this is a victory for good practice; but it also summarizes the sad state of scientific education to note that in the research laboratory itself the emanation electroscope has long since been removed from the desk to the attic.<9> Foot note 1_9In science, as in a technological arms race, one is simply never called on to keep doing what was done before. An ever more challenging task is posed by the constantly escalating demands of science for the enhanced data that can only be obtained at the increasingly costly new levels of technological sophistication. One is always forced further up the mountain, ascending to ever higher levels of technological performance -- and of expense. As science endeavors to extend its «mastery over nature,» it thereby comes to be involved in a technology-intensive arms race against nature, with all of the practical and economic implications characteristic of such process.
The exploration of nature's parametric space confronts us with the reality of physical limits: particle velocities in accelerations are limited by the speed of light, temperatures in low temperature research are limited by absolute zero, vacuums are limited by condition of emptiness, temperatures by the cosmic boiling point of the big bang. And such limits amount to resistance barriers. With every step we take towards them every time we move from where we are to 10% closer yet -- we find it exponentially more difficult to take yet further steps as the technological demands for further progress grow increasingly massive.
The enormous power, sensitivity, and complexity deployed in present-day experimental science have not been sought for their own sake but rather because the research frontier has moved on into an area where this sophistication is the indispensable requisite of ongoing progress. Nature's inherent complexity means that in science, as in war, the battles of the present cannot be fought effectively with the armaments of the past.<10> Foot note 1_10
University of Pittsburgh
Department of Philosophy