Life Transcending Physics and Chemistry

Michael Polanyi was an important thinker (1891-1976) whose work I appreciate. His essay Life Transcending Physics and Chemistry was published in 1967. I have not found it anywhere online.

It argues that the properties of life, that is, the arrangement of the molecules making up the chemical constituents of organic substances, are not completely determined by the properties of physics and chemistry. This is understood in the sense that there are very many possible stable compounds which can exist for a given set of constituents, but of which only one (or a few) will work at all to support life. So, on the one hand, the laws of chemistry and physics support and enable the configurations for life, but, on the other hand, they do not require the particular configurations we find. This is my understanding of his essay.

Here then is the article.


Life Transcending Physics and Chemistry


Michael Polanyi

Chemical and Engineering News, August 21, 1967, pages 55-66

(page 55)

[article synopsis]

Form and junction in an object may not be explicable in terms of the laws that govern the properties of its atomic constituents. Both animate and inanimate systems can illustrate this proposal.

DNA (deoxyribonucleic acid} may determine the boundary (the morphology) of a biological system. But the form and function of the resulting biological system cannot be explained by the laws governing its parts. An example of this principle may be found in administrative hierarchies. Here, a higher authority governs lower levels while relying on the autonomous workings of these lower levels.

A similar irreducibility may be found in machines. Their design, shape, and operation are comprehensive features not due to physical and chemical forces. A description of a machine in physical or chemical terms would result in a topography of atoms and molecules unique to the subject. It could not identify the machine as belonging to a class of machines based on certain operational principles.

Biological systems, like machines, have, therefore, functions and forms inexplicable by chemical and physical laws. The argument that the DNA molecule determines genetic processes in living systems does not indicate reducibility. A DNA molecule essentially transmits information to a developing cell. Similarly, a book transmits information. But the transmission of the information cannot be represented in terms of chemical and physical principles. In other words, the operation of the book is not reducible to chemical terms. Since DNA operates by transmission of (genetic) information, its function cannot be described by chemical laws either.

The life process is essentially the development of a fertilized cell, as the result of information imparted by DNA. (Transmission of this information is nonchemical and nonphysical, and is the controlling factor in the life process) The description of a living system therefore transcends the chemical and physical laws which govern its atomic constituents.

[article body begins]

The discovery by Watson and Crick of the genetic function of DNA (deoxyribonucleic acid), combined with the evidence these scientists provided for the self-duplication of DNA, is widely held to prove that living beings can be interpreted, at least in principle, by the laws of physics and chemistry. Barry Commoner has queried this view by citing evidence to show that the self-duplication of DNA is not proven ("Science and Survival"). But the latter point, though important, is not in fact decisive. For even if we granted the self-duplication of DNA, this would not show that living beings can be represented in terms of physics and chemistry. It would, for example, not offer a possible physical-chemical explanation of human consciousness.

Moreover, Commoner disregards the fact that, viewed by his own criteria for the reducibility of biological processes, his challenge of the self-duplication of DNA presents no difficulty to the reduction of genetics to physics and chemistry. For Commoner shares the view held by most biologists, that so far as life can be represented as a mechanism, it is explained by the laws of inanimate nature, and there is nothing in the observations he cited that could not be ascribed to some yet undiscovered mechanical operations.

For my part, I differ from Commoner and from most biologists, by holding that no mechanism — be it a machine or a machinelike feature of an organism — can be represented in terms of physics and chemistry. This principle precludes the possibility of biology ever becoming a molecular science and thus leads to Commoner's conclusion on grounds very different from his. But my principle has so far been accepted by few biologists and has been sharply rejected by Francis Crick, who is convinced that all life can be accounted for by the laws of innanimate nature ("Of Molecules and Men"). Thus my conclusions conflict both with Crick's view, as with the grounds on which Barry Commoner takes an opposite view. If I am right, both these positions must be radically revised.

(page 56)

My account of the situation will seem to oscillate in several directions, and I shall set out, therefore, its stages in order.

I shall show that:

  • Commoner's criteria of irreducibility to physics and chemistry are incomplete; they are necessary but not sufficient conditions of it.
  • Machines are irreducible to physics and chemistry.
  • By virtue of the principle of boundary control, mechanistic structures of living beings appear to be likewise irreducible.
  • The structure of DNA, which according to Watson and Crick controls heredity, is not explicable by physics and chemistry.
  • Assuming that morphological differentiation reflects the information content of DNA, we can prove that the morphology of living beings forms a boundary condition which, as such, is not explicable by physics and chemistry (the suggestion arrived at in the third item).

Let me proceed on these lines now.

Commoner's incompatible views

Commoner holds two views (A and B) that are incompatible. His view A, shared by a number of scientists who acknowledge the existence of irreducibility, is that when the joint presence of parts shows features which cannot be observed in the isolated parts, these features are not explicable by the laws governing the separate parts. Hence, when the parts are governed by the laws of physics and chemistry, their joint entity is considered to be irreducible to physics and chemistry.

His view B, long since predominant among all scientists, is that the explanation of living functions in terms of a mechanical model amounts to explaining them in terms of physics and chemistry. These two views are incompatible, for machines — or living functions operating mechanically — are entities the characteristic features of which are absent in their separate parts. I hold both view A and B to be erroneous. I shall be first concerned with A — Commoner's criteria of irreducibility.

Criteria of irreducibility

Let me show that the presence of joint properties not observed in the isolated parts does not prove irreducibility to terms of the laws controlling the parts. Take the fact that the sun is a sphere. Its separate parts are not spheres; nor does the law of gravitation speak of spheres. But the mutual gravitational interaction causes the parts of the sun to form a sphere. The same law causes the planets to move on elliptic paths around the sun.

Physics is rich in examples of comprehensive features of a system that cannot be observed in the isolated parts of the system. Crystals are marvelously ordered aggregates. Their particles are arranged in the pattern of one out of 230 space groups, which are not observed in the separate atomic parts. But these patterns can be derived from the interaction of their component parts. Snell's Law says that a beam of light passing through a medium of variable refractive powers will take the path along which it reaches its endpoint

(page 57)

in the shortest time, and this comprehensive feature is derived from the laws that determine the curvature of the path at any single point in space.

These comprehensive features, which cannot be observed in any single particle or any pair of particles nor, for the case of Snell's law, at any point of a beam, are all computable by the mathematical integration of the laws observed in the isolated components. And this is true also for the theory of superconductivity, which Commoner quotes as an example for the emergence of irreducible principles. It is but an integrated form of laws applying to parts or pairs of parts.

Clearly, holistic systems which can be computed from the laws of physics do not point to the existence of irreducible biological principles. They have often suggested on the contrary that all organized functions of living beings might well be explicable in terms of physics and chemistry. Whenever the complex order of morphogenesis evoked the thought that there was manifested here a principle not present in inanimate nature, the answer came that this order of life might well be derived from physical laws, as the order of crystals is derived from them. Wolfgang Koehler has likewise used the harmonious distribution of the electric current in a system of parallel conductors as an example for suggesting that we should explain gestaltlike perception and its corresponding neural processes, by a physical equilibration of their parts.

Machines are irreducible

But while we have met so far no irreducible holistic systems in inanimate nature, we find such systems among inanimate artifacts, like machines.[*] Machines seem obviously irreducible, since they have comprehensive features that are not due to a spontaneous integration of physical and chemical forces. They do not come into being by physical-chemical equilibration, but are shaped by man. They are shaped and designed for a specific purpose, which they achieve by the interaction of their characteristic parts working in accordance with distinctive operational principles. But

  • [footnote] In my "Personal Knowledge," pages 390-92, I have pointed out that the entire system of conceptions based on randomness (such as the ideas of chance, the theory of gas pressure and temperature, and all thermodynamics) is unspecifiable in terms of atomic physics, and that hence we must recognize randomness as a comprehensive principle above mechanics, quantal or macro dimensional. I know of no other pair of inanimate principles, both valid (not complementary) where one of them (the "higher" one) excludes the application of the other, even while its operations are based on the latter. We may note also that boundaries of inanimate systems established by the history of the universe, are found widely in geological, geographic, and astronomic domains and that their information content per unit of matter is very much less than that of a living thing.

(page 58)

No text on this page.

(page 59)

I must argue this conclusion in detail, if I want it to carry conviction.

Try to describe a machine in physical-chemical terms. A complete physical-chemical topography of my watch — even though the topography included the changes caused by the movements in the watch — would not tell us what this object is. On the other hand, if we know watches, we would recognize an object as a watch by a description of it which says that it tells the time of the day, by hands sweeping round a face, marked by the hours of the day. We know watches and can describe one only in terms like "telling the time," "hands," "face," "marked," which are all incapable of being expressed by the variables of physics, length, mass, and time. The impossibility is of a logical kind, similar to that by which a poorer deductive system cannot define the terms of a richer one. For example, propositional calculus cannot define arithmetics.

But we can still sharpen this argument. A physical-chemical topography of my watch might make it possible, at least in principle, to identify this particular watch as an object. But it would fail to identify it as a watch, for it is incapable of defining a class of watches, as needed for assigning the watch to that class. To realize more clearly what I mean, suppose you had invented the watch and made one, and you applied for a patent specifying the watch you made, by its precise physical-chemical topography. Your patent would protect only the manufacture of an exact replica of your model. Your competitor could circumvent your patent by merely displacing one single atom of the patented topography. Only the principles underlying the operations of the watch in telling the time could specify your invention of the watch effectively, and these cannot be expressed in terms of physical-chemical variables. (This is the refutation of Commoner's view B.)

Now, from machines let us pass on to books and other means of communication. Nothing is said about the content of a book by its physical-chemical topography. All objects conveying information are irreducible to the terms of physics and chemistry.

Up to a point, we can transfer what has been said of machines to machine-like aspects of living beings. Take some examples from the higher animals — their organs of circulation, breathing, digestion, secretion, and thermal regulation. Think of their anatomy and of the way they operate in performing their functions. None of these conceptions can be defined in terms of physics and chemistry. In his treatise, "The Structure of Science," Ernest Nagel has attempted to eliminate the purposive character of physiological functions by describing these as mere events that happen to be beneficient to the organism, without purposively serving this benefit. But the fact still remains that a process can be regarded as a biological function only if it does benefit the organism. This remains its essence, as much as it is the essence of a machine to serve a purpose acknowledged by its designer. We can see this from the fact that the concept of functional disorders interfering with normal achievements applies equally both to living and inanimate mechanisms.

Moreover, despite his reductionist claims, Nagel admits irreducibility of vital functions by using their biological names for talking about them. He must do so, for the mere shape of a living being defeats any physical-chemical definition and this is true throughout the anatomical features of life. And again, even supposing we did produce a mathematical expression for the shape of one living specimen, including all its anatomy at one particular moment, the formula would not cover its changes, due to growth and decline and it would of course fail even more widely to cover the variety of specimens belonging to one species and to cover at the same time the innumerable future additions due to accrue to a species in the course of time.

Living systems seem irreducible, too

But might this barrier to physical-chemical reduction not prove to be temporary? Before Newton, the geometrical shapes of the sun and of the planetary paths appeared to be ultimate facts, not reducible to any laws governing their ultimate particles.

(page 60)

Might machines and machinelike aspects of living things not be shown one day to result from the working of physical or chemical laws?

We can exclude this for machines. Our incapacity to define machines and their functions in terms of physics and chemistry is due to a manifest impossibility, for machines are shaped by man and can never be produced by the spontaneous equilibration of their material. But morphological structures are not shaped by man; could they not grow to maturity by the working of purely physical-chemical laws?

To answer this question, we must first deal with a more general problem. Let us go back and ask how it is possible for machines to be controlled by two independent sets of principles. For the material of the machine is subject to the laws of physics and chemistry, while the shape and the consequent working of the machine are controlled by its structural and operational principles. The solution is found by remembering that no given material system can be wholly determined by the laws of physics and chemistry.

The laws of physics are given in terms of differential equations which determine a definite system only within a set of fixed conditions. The spherical shape of the sun, the elliptic

(page 61)

paths of the planet, the trajectory of a beam of light covered in a minimum of time — each of these arrangements arises under the dual control of a differential equation working within the bounds of a particular set of conditions. Laplace thought we would know all that can be known in the world, if we knew the course of its atoms. But for this he required a complete map of atomic positions and velocities to start with. Physics is dumb without the gift of boundary conditions, forming its frame; and this frame is not determined by the laws of physics.

The laws of chemistry have similar limitations. We can demonstrate chemical change by pouring a solution of reagents into a container and setting the concentration and temperature as required. Generally, to have a definite chemical process, we must frame it by boundary conditions not fixed by the laws of chemistry.

We speak of such boundaries as "fixed conditions" rather than "controlling principles," for their intervention, though indispensable, is not highly significant. This is different for a machine. The boundary conditions of the physical-chemical changes taking place in a machine are the structural and operational principles of the machine. We say therefore that the laws of inanimate nature operate in a machine under the control of operational principles that constitute (or determine) its boundaries. Such a system is clearly under dual control.

The relationship between the two controls — the devices of engineering and the laws of natural science — is not symmetrical. The machine is a machine by having been built and being then controlled according to principles of engineering. The laws of physics and chemistry are indifferent to thsse principles; they would go on working in the fragments of the machine if it were smashed. But they serve the machine while it lasts; machines rely for their operations always on the laws of physics and chemistry.

Returning now to living beings, we may start by observing that to speak of life as something to be explained by the laws of physics and chemistry is strictly speaking absurd, for physical and chemical processes do not determine by themselves any finite system. (We must ask what the boundary conditions are within which physics and chemistry do explain biotic phenomena. The answer is found in the fact that biochemistry and biophysics are always concerned with processes that have a bearing on an existing organism. These sciences seek to determine the chemical and physical principles on which the organism relies for its operations. Any chemical or physical study of living things that is irrelevant to the working of the organism is no part of biology, just as the chemical or physical studies of a machine must bear on the way the machine works, if it is to serve engineering.

It is this basic principle of biology that the physiologist, J. S. Haldane, insisted upon throughout his philosophical writings.

I have written elsewhere at some length about higher principles that govern the working of lower laws on which they rely for operating a system. Administrative hierarchies are common examples of a higher authority governing lower levels, while relying on the autonomous workings of these lower levels. Hierarchies formed by successive levels of the organism have been described similarly. My own theory expands the structure of hierarchic levels to the relation between biological principles and the laws of physics and chemistry. Biological principles are seen then to control the boundary conditions within which the forces of physics and chemistry carry on the business of life. This dual action of a system is said to work by the principle of boundary control.

But is it not conceivable that an organism developing from a fertilized cell might shape the boundary conditions of the developed organism, without itself being subject to such boundaries? One may reply that a machine

(page 62)

that manufactures machines for a factory produces them within its own boundaries, as set to the machine by its operational principles. The embryo producing the biological boundaries of maturity works likewise within its own embryonic boundaries. Indeed, no biological process ever takes place in an unstructured medium; at least not in the world today. But I must yet show that this is necessarily so.

Boundary conditions for DNA

The next step takes us back all the way to the claims made currently for DNA as an explanation of life in terms of chemistry. I shall assume here the currently prevailing view that DNA determines altogether the outcome of embryonic growth, including the whole design of the final organism. Does this not face us with the fact that a pure chemical compound controls supreme biological functions? Where is then the boundary condition which controls this chemical effect?

One might be tempted to reply that a DNA molecule produces nothing by itself, its genetic program being initiated within the richly structured framework of a fertilized cell, and that subsequently DNA controls morphogenesis within a steadily developing framework. But this is not to the point. For it is DNA itself that introduces within its chemical structure a pattern that acts as a controlling framework to the ensuing generative process. This is what is meant by saying that DNA controls the genetic development of an organism by transmitting to its cells a quantity of information that induces in them an equivalent amount of organic differentiation. Where, then, is the boundary condition?

Remember then our earlier conclusion that a book, or any other object bearing a pattern that communicates information, is essentially irreducible to physics and chemistry. It would follow that we must refuse to regard the pattern by which DNA spreads information as part of its chemical properties. Its functional pattern must be recognized as a boundary condition located within the DNA molecule. This is what I shall try to demonstrate.

The boundary conditions forming a machine have two interrelated aspects: They consist in a distinctive structure sustaining a purposive operation. The various functions of a living organism are similarly sustained by its structure known as its morphology. A written or printed text functions by its structure alone, without generating motion; it acts passively by being read. A plant or an animal, recognizable by its shape, its pattern, and its coloring may be said to transmit information likewise passively, by being seen. The boundary condition generating this function consists in the case of such a plant or animal in its typical appearance, its morphology. If DNA is regarded as bearing a pattern that forms part of an organism and as transmitting information through this pattern, then such a pattern is to be classed likewise as a morphological feature of the organism, and hence be irreducible to terms of physics and chemistry.

By the same token, any chemical compound bearing a complex structure and transmitting thereby substantial information to its neighborhood must be irreducible to physics and chemistry in respect of this particular feature. Let me show this.

All chemical compounds consist of atoms linked in an orderly manner by the energy of chemical bonds. But the links of a compound forming a code are peculiar. A code is a linear series of items which are composed, in the case of a chemical code, of groups of atoms forming a chemical substituent. In the case of DNA, each item of the series consists of one out of four alternative substituents. In an ideally functioning chemical code — to which I shall limit myself — each alternative substituent forming a possible item of the series must have the same mathematical chance of appearing at any point of the series. Any difference of alternative chances would reduce the amount of information transmitted, and if there were a chemical law which determined that the constituents can be aligned only in one particular arrangement, this arrangement could transmit no information. Thus in an ideal code, all alternative sequences being equally probable, its sequence is unaffected by chemical laws, and is an arithmetical or geometrical design, not explicable in chemical terms.

The conception of a purely geometrical structure in a chemical compound is exemplified by the configuration of two optical antipodes which have the same statistical probability. A chemical synthesis of any particular compound of this kind tends to produce both antipodes in equal quantities. They can then be separated, as Pasteur separated them, by letting them crystallize and picking out the pure forms by their differently shaped crystals. But this discrimination is not due to chemical forces. A chemical discrimination of the two antipodes can be brought about by using optically active reagents or solvents, but this leaves open the problem as to the origin of these optically active partners. To prepare chemically a compound that is one out of millions of equally probable DNA alternatives would produce, along with it, about equal amounts of each of these millions of alternatives. Moreover, the task of separating the desired compound from the others would present many times over the kind of problem presented by the chemical separation of optical antipodes.

One could build a chain of substituents arranged according to any desired pattern by adding each consecutive link of the series separately. But the resulting pattern would not be the product of chemical forces, and if it functioned as a code transmitting information, this would be the information which we had imparted to the pattern by the sequence of operations that built it up. Such a code might be made to transmit the words of a national anthem, or to serve as a message for military secrets, neither of which could be regarded as the product of chemical forces.

Another way of reaching the same conclusions is by considering the theory advanced by a number of neurophysiologists, that the nervous system registers the memory of a habit acquired by an organism in the structure of its RNA molecules. This is called the fixation of experience by

(page 63)

No text on this page.

(page 64)

RNA in the manner of a tape recording, and this analogy illustrates what I am saying here. The information imparted to a molecule of this type is received and held by it in a way similar to that in which a tape recorder would do this. The pattern of its traces is the pattern of the impacts in which the message was embodied. The pattern can no more be derived from the laws of physics and chemistry when engraved in an RNA molecule than it can be when inscribed on a tape or, for that matter, on the surface of a rock.

Morphology is the framework

We can now lend greater precision and force to the conclusion that morphological features are the boundary conditions of physical-chemical laws in living things and thus are not accountable by these laws, on which they rely for their functions. The functional structure of machines, products of man's designing and shaping, manifestly represent boundary conditions imposed on the laws of inanimate nature to press them into the service of a technical purpose. We have before us now the corresponding process that shapes the morphological development of the germ cell. What happens here, according to the theory of Watson and Crick, is that DNA transmits to the developing body of cells a quantity of information in an equivalent amount of organic differentiation. And thus it follows that the shape and structure of living beings has the structure of an information.

When this structure reappears in an organism, it is a configuration of particles that typifies a living being and serves its functions; at the same time, this configuration is a member of a large group of equally probable (and mostly meaningless) configurations. Such a highly improbable arrangement of particles is not shaped by the forces of physics or chemistry. It constitutes a boundary condition, which as such transcends the laws of physics and chemistry.

This brings the vital shaping of offspring by DNA into consonance with the shaping of a machine by the engineer. The manufacturing of a machine also represents a distinctive distribution of matter not due to the working of physical-chemical forces and it, too, forms the characteristic boundary conditions of the system in question. We can see now more clearly why such shaping of boundaries may be said to go beyond a mere "fixing of boundaries" and establishes a "controlling principle." It achieves control of the boundaries by imprinting a significant pattern on the boundaries of the system. Or, to use information language, we may say that it puts the system under the control of a non-physical-chemical principle by a profoundly informative intervention.

We can see then also why a selection of the crystals of one optical isomer from a mixture of the crystals including both antipodes is a nonchemical procedure. It is that because it produces a significant distribution of matter not determined by the laws of chemistry. And we can note that the same holds obviously for the construction of a polymer chain by joining a preselected sequence of substituents.

I have accepted here the view of Watson and Crick that the information content of an organism equals the information conveyed by its DNA. But I doubt we have as yet a quantitative measure of the information — or negentropy, as Schrödinger has called it — of a living thing. The currently discussed question, whether the information content of the developed organism does not exceed that of its DNA, seems therefore not ripe for evaluation in the light of the principles I have put forward here.

Finally, a word on the way the boundary conditions controlling physical-chemical processes in an organism may have come into existence from inanimate beginnings. The question is whether or not the logical range of random mutations includes the formation of novel principles not definable in terms of physics and chemistry. It seems very unlikely that it does include it. This is the ground on which emergence has been defined in my recent writings, including my book "The Tacit Dimension."

But the problem of evolution lies beyond my subject here. When I say that life transcends physics and chemistry, I mean that biology cannot explain life in our age by the current

(page 65)

workings of physical and chemical laws.

In retrospect

The moment one succeeds in proving that machines cannot be explained in terms of physics, this appears so obvious that one wonders whether something so trivial could have ever been overlooked, and if it has been, what use there can be in forcing it to our attention now.

It may seem unbelievable, but it is yet a fact, that for 300 years writers who contested the possibility of explaining life by physics and chemistry argued by affirming that livings things are not, or not wholly, machinelike, instead of pointing out that the mere existence of machinelike functions in living beings proves that life cannot be explained in terms of physics and chemistry. In the late 17th century, we find the Neo-Platonist Cudworth, and likewise the naturalist John Ray, opposing the view that life can be explained in terms of matter in motion, by affirming that living beings are not machines. And 200 years later, Driesch and his supporters fought for the recognition that life transcends physics and chemistry, by arguing that the powers of regeneration in the sea urchin embryo (discovered by Driesch) were not explicable by a machinelike structure. Up to this day one speaks of the mechanistic conception of life both to designate an explanation of life in terms of physics and chemistry, and an explanation of living functions as machineries — though the latter excludes the former. The term "mechanistic" is in fact so well established for referring to these two mutually exclusive conceptions, that I am at a loss to find two different words that will distinguish between them.

My own experience may suggest an explanation for this strange story. During the past 15 years, I have worked on these questions, achieving gradually stages of the argument presented in this paper. These are:

  • Machines are not formed by physical-chemical equilibration.
  • The functional terms needed for characterizing a machine cannot be defined in terms of physics and chemistry.

Dr. Michael Polanyi, 76, has a long, distinguished career in both the physical and social sciences. He has earned M.D. and Ph.D. degrees, and has been awarded many honorary degrees. He is an outstanding teacher, has taught at many universities, and has authored nine books on such diversified topics as science, economics, and philosophy.

Born in Budapest, Hungary in 1891, Michael Polanyi graduated from the University of Budapest in 1913 as a doctor of medicine. From there, he went on to study chemistry at the Technische Hochschule in Baden, Germany. He became an army medical officer in 1914, but was struck down with diphtheria shortly after joining. While convalescing in 1915, he wrote a thesis in physical chemistry for his Ph.D. from the University of Budapest.

Dr. Polanyi taught at the University of Budapest in 1919 and, in 1920, began teaching at the Kaiser Wilhelm Institute, Berlin. In 1929, he was appointed life member of the institute, but he resigned four years later, to accept the chair of physical chemistry at Victoria University, Manchester, England. In 1948, he exchanged his chair in physical chemistry for one in social sciences at Manchester. In 1950, he began a series of visiting professorships and fellowships which he has continued to this day. He has taught at the universities of Aberdeen, Keele (Staffordshire), Oxford, California (Berkeley), Chicago, Duke, Stanford, Wesleyan (Middletown, Conn.), and Yale.

This year, Dr. Polanyi has lectured at the University of Chicago and at Oxford. He returned to the U.S. from England this month and expects to give the opening address at the meeting of the American Psychological Association, in Washington, D.C.

Dr. Polanyi was elected a foreign member of the National Society of Science, Letters, and Arts (Naples) in 1933, a fellow of the Royal Society (England) in 1944, and a member of the Max Planck Society (West Germany) in 1949. He is a foreign honorary member of the American Academy of Arts and Sciences (Philosophy Section) and the International Academy of the Philosophy of Science. In addition, he holds Hon. D.Sc. degrees from Princeton, Leeds, Notre Dame, and Manchester as well as Hon. LL.D. degrees from Aberdeen and Wesleyan University.

Dr. Polanyi was a founder-member of the Society for the Freedom of Science in 1938. He was also a member of the International Executive of the Congress for Cultural Freedom and chairman of the Hamburg Conference of the Congress for Cultural Freedom (1953). He lectured at the International Congress for the Philosophy of Science in Jerusalem (1964).

(page 66)

  • No physical-chemical topography will tell us that we have a machine before us and what its functions are.
  • Such a topography can completely identify one particular specimen of a machine, but can tell us nothing about a class of machines.
  • And if we are asked how the same solid system can be subject to control by two independent principles, the answer is: The boundary conditions of the system are free of control by physics and can be controlled therefore by nonphysical, purely technical, principles.

Turn next to living things. Of the points that apply to machines the first point fails to apply to living beings. For it is not obviously clear that living things are not formed by mere physical-chemical equilibration. And at this point, strangely enough, the discovery of DNA, which is so widely thought to prove that life is mere chemistry, provides the missing link for proving the contrary. The theory of Crick and Watson, that four alternative substituents lining a DNA chain convey an amount of information approximating that of the total number of such possible configurations, amounts to saying that the particular alignment present in a DNA molecule is not determined by chemical forces. And the additional theory, that the information of a DNA molecule is embodied in the morphology of the corresponding offspring, assures us of the fact that this morphology is not the product of a chemical equilibration, but is designed by other than chemical forces. This is the step that my present paper adds to my earlier paper arguments.

This story of long gradual progress toward proving my point shows, in my view, that such proof is not easily established. Add to this the widespread, erroneous arguments which those who share my basic views have put forward — as exemplified by my critique of Commoner's argument. It may appear then less surprising that these views have not gained firm ground so far, and that the confusion of the past centuries continues to prevail. The effort of dispelling this confusion may also appear worthwhile then.


  1. Commoner, Barry, "Science and Survival," Viking Press, New York, 1966, Chapter 3, "Greater than the Sum of its Parts," and writings of the author quoted by him.

  2. Crick, Francis, "Of Molecules and Men," University of Washington Press, Seattle, 1966.

  3. Polanyi, Michael, "The Tacit Dimension," Doubleday, New York, 1966; Routledge and Kegan Paul, Ltd., London, 1967, pp. 87-91.

Previous publications of the author containing much of the present argument include:

"Personal Knowledge: Towards a Post-Critical Philosophy," Routledge and Kegan Paul, Ltd., London, and University of Chicago Press, Chicago (1958), pp. 328-35; "The Study of Man," Routledge and Kegan Paul, Ltd., London, and University of Chicago Press, Chicago (1959), pp. 47-52; "The Scientist Speculates," Ed. L. J. Good, G. P. Putnam's Sons, N.Y., and Wm. Heinemann, Ltd., London (1962), pp. 71-78; "Science as a Cultural Force," Ed. Henry Woolf, Johns Hopkins Press, Baltimore (1964), pp. 54-76; Rev. Mod. Phys., 34, 601 (1962); Encounter, London, 24, No. 5 (1965); Brain, London, 88, Part IV, 799 (1966); Philosophy, London, 40, 369 (1966).

You'll only receive email when they publish something new.

More from Framing Life