Brian D. Josephson/2/

Niels Bohr's arguments indicating the non-applicability of quantum methodology to the study of the ultimate details of life given in his book "Atomic physics and human knowledge" conflict with the commonly held opposite view. The bases for the usual beliefs are examined and shown to have little validity. Significant differences do exist between the living organism and the type of system studied successfully in the physics laboratory. Dealing with living organisms in quantum-mechanical terms with the same degree of rigour as is normal for non-living systems would seem not to be possible without considering also questions of the origins of life and of the universe.


This paper is dedicated to three great thinkers who have insisted that the world is not quite the straightforward affair that our successes in describing it mathematically may have seemed to suggest: Niels Bohr, whose analyses of the problem of explaining life play a central role in the following discussion; Erwin Schroedinger, erudite student of the relationship between mind and matter, at the London Celebratory Centenary Conference of whom the talk on which this paper was based was originally given/3/; and David Bohm, the modern doubter or sceptic to whom this issue of Foundations of Physics is dedicated, who has devoted much of his scientific career to the demonstration of the invalidity of a number of orthodox opinions about nature.

Most physicists regard physics as being the most fundamental of sciences (in the sense that natural phenomena of all kinds can, in principle, be explained in terms of its fundamental laws, i.e. in terms of quantum mechanics). But a minority have not accepted this doctrine of universality. Schroedinger(1), for example, wrote "Without being aware of it and without being rigorously systematic about it, we exclude the Subject of Cognizance from the domain of nature that we endeavour to understand ... mind could not cope with this gigantic task (of erecting the objective outside world of the natural philosopher) otherwise than by the simplifying device of excluding itself."

Since Schroedinger wrote these words, the cognitive sciences (e.g. psychology, neurophysiology and artificial intelligence) have conjointly developed a appreciable degree of understanding of the "ourselves" that the physical sciences have ignored. However, the fragmentation about which David Bohm has written(2) has served to keep these two kinds of studies apart, and physicists have largely avoided discussion of the deeper issues (within the official terms of reference of their subject, at any rate) . Furthermore, because of the presumed fundamental nature of physics, the life sciences and even the cognitive sciences (transpersonal psychology excepted) have tended to adopt the mechanistic view of life that characterises physics, a view that tends to ignore all the subtleties of personal experience, and to comprehend psychological processes in their most superficial aspects only. On the basis of such facts Bohm(3) concluded, as did Schroedinger before him, that the modern mechanistic approach to the study of nature may cover no more than a tiny area in a much vaster field.

The present paper is an attempt to put this belief on a more objective footing. It is based largely on Niels Bohr's discussion of the possibility of applying quantum theory to the understanding of life(4). Bohr argued that the disturbance that would be caused to a living system, were one to attempt to determine its state with sufficient accuracy as to be able to understand completely all the details of its functioning, would probably, by virtue of the Heisenberg uncertainty principle, be sufficient to injure or even kill it. He concluded on this basis that in all probability living systems fall outside the domain of phenomena describable in complete detail by quantum mechanics.

This argument has been curiously neglected by physicists, and by scientists in general. It appears that this neglect has as its principal basis a kind of vote-taking process, in which the numerous successes of orthodox methodology (ultimately based on quantum mechanics) in application to the study of living systems on the one side are set against the single opposing voice of Bohr on the other. Such grounds for deciding the truth fall rather short of what might reasonably, within the context of science, be regarded as adequate.

In what follows, I shall in essence be elaborating Bohr's argument by filling in significant details. The problem regarding the position he took seems to be not that defenders of the completeness point of view have good arguments to set against those of Bohr, but rather that they are simply for a number of reasons disinclined to believe in his conclusions. As a result, much of the paper will consist of examining the various reasons that are commonly given for disbelief, and showing them to be inadequate.

Section 2 describes the essence of Bohr's argument, particular attention being given to demonstrating how the differences in conditions of observation between the those that pertain to the typical physics experiment and those that pertain to the typical experiment in biology make the quantum theory applicable in a straightforward way in the former case but not in the latter. The differences that are described are not essentially those between living and non-living systems; the relevant feature is the difference between the controlled situation of a quantum experiment (i.e. an experiment in accordance with the particular terms of reference of the quantum theory of measurement(5,6)) and the general situation occurring in nature. Section 3, based on a hypothetical example of scientific development that has close analogies to the actual situation in the case of quantum mechanics, provides insight into the way in which physics acquired for itself a theory that provided an illusory vision of completeness of descriptive power. Then section 4 reviews a number of standard objections to the idea that living systems do not fit in a straightforward way into the quantum mechanical scheme, while finally the concluding discussion of section 5 attempts an overall evaluation of the situation in which physics appears to find itself as a result of the difficulties that have been shown to be involved in reconciling the quantum approach and the phenomenon of life.



Predicting the future development of a system in classical mechanics is in principle straightforward. We simply measure the current values of the parameters of the system, and then feed these as initial conditions into the equations of motion in order to determine the subsequent behaviour as a function of time. In quantum mechanics the situation is less straightforward. The Schroedinger equation provides a suitable analogue to the equations of motion, but in determining the initial condition the complication arises that according to quantum mechanics measurement of a system unavoidably disturbs it. We cannot find out what the system we are experimenting on actually is without changing it. Bohr's argument is then (as has already been noted) that a living system would be seriously disturbed if were one to attempt to determine its state with sufficient accuracy as to be able to understand completely all the details of its functioning.

But, one may well ask, this argument seems to be equally applicable to non-living systems as well. How is it that we seem to be able to do experiments at the quantum level on non-living systems and not be troubled by the situation depicted in this argument? Bohr talked in general terms about the "different conditions of observation" that were involved in the two cases. The precise significance of this point will now be spelt out in detail.

We shall divide up the situations that occur in a physics experiment into the following three types:

(i) the situation where the system we are dealing with is sufficiently macroscopic that the effects of the disturbance can be ignored within the context of the experiment.

(ii) where the disturbance caused by measurement is, in effect, the actual object of study of the experiment. An example is the Stern-Gerlach experiment, where one studies the statistics of the way a polarised particle beam is decomposed into distinct sub-beams when it is subjected to a measurement process that consists of passing it through a region of space containing an inhomogeneous magnetic field.

(iii) where the disturbance caused by the process of measurement can simply be ignored, because the process of measurement (which may equally be thought of as a filtering process) has only the role of preparing the system for the experiment, so that interest is therefore focussed entirely on the post-measurement, post-disturbance situation. As a simple illustration, consider the high energy physics experiment. Here the accelerator or other particle source produces a beam of indeterminate composition. However, the measuring device (e.g. a bubble chamber in conjunction with a magnetic field) enables a particular particle type to be associated with each observed event. The disturbing effect produced by the measurement process is simply that of collapsing the indeterminate combination of particle states in the source into a sequence of individual particle states, of which each has become well-defined as a result of the observation. As far as the experimenter is concerned, these well-defined states are the initial states that he has to know in order to compare theory and experiment, rather than the original initial states from the source. As a result, the "disturbance by measurement" need not be taken into account at all.

These three cases can be characterised as follows: in case (i), the effect (disturbance due to measurement) is negligible; in (ii), it can be calculated; and (iii) the disturbance is not relevant. But these three cases do not exhaust all the possibilities. Bohr's case of concern is an example of the following quite different situation:

(iv) We would like to be able to predict how a given system would develop with time if it were left undisturbed. The system concerned is sufficiently sensitive to disturbances that an attempt to identify its current state with sufficient accuracy to be able to make the desired prediction would seriously interfere with the result of the experiment.

The circumstance of the difference between case (iv) and the other cases discussed provides the reason for putting aside as being irrelevant, when considering the question of whether the disturbance involved is a genuine problem in biology or not, the fact that in the case of the physics experiment it is not found to be a problem. Once one has disposed of this particular objection to Bohr's argument, one seems to be forced to the conclusion that Bohr's point about quantum mechanics and life was a perfectly valid one. However, the issue will be examined in more depth, taking into account a number of other perspectives on the problem, later on in the course of the paper.

Distinguishing the case (iv) situation (for which making predictions presents a problem for quantum mechanics) from the other cases described, we may state our main conclusions now:

(a) Without any inconsistency, one may regard the standard theory simply as a specialised piece of formalism that applies, as a special case only, to those particular systems that have been prepared with the aid of the usual measuring/filtering procedures (cases (i) to (iii)). There is no obvious reason for identifying the category of circumstances where quantum theory can make precise predictions with the category of all conceivable natural phenomena that can be investigated scientifically.

(b) The fact that quantum theory has been demonstrated, in many applications in the fields of both science and technology, to describe nature with very high accuracy, does not in itself provide us with a reason for regarding the theory as giving a complete account of all natural phenomena. Again, these successes are based on the situations of categories (i) to (iii) only.

(c) Finally, it is likely that, just as proposed by Bohr, living systems in general constitute one of the kinds of system to which quantum mechanics cannot be applied in a routine way.



It has been argued in the preceding section that, contrary to the usually held views, serious objections exist on grounds of principle to the idea that quantum mechanics constitutes a comprehensive theory of all natural phenomena. In this section an informative analogue will be given that helps to show how this contradiction between the generally held beliefs and what is suggested instead by a careful examination of the processes of prediction in quantum mechanics came historically into being. It involves a hypothetical alternative development of geometric optics on another planet, carried out by scientists who possess vision in monochrome only. In this imaginary alternative, workers in the field successfully discover Snell's law for the refraction of light, but owing to their lack of colour vision they do not gain the Newtonian insight that natural light consists of a mixture of lights of different colours. Further research leads them to discover the existence of chromatic aberration, but because of their lack of colour vision they perceive it only as a blurring that affects their ability to take high-precision measurements. They then discover that if the light is first passed through certain kinds of materials (which Earth scientists know as colour filters) then the blurring is considerably reduced. This discovery gives geometric optics a new lease of life, and Snell's law acquires the status of an exact law, that explains the behaviour of all possible configurations of lenses and prisms in respect to the passage of filtered light. By this stage, natural light is considered by the geometric optics practitioner as being an imperfect form of light, that lacks any real interest as far as the modern physicist is concerned.

The analogy involved with quantum mechanics is between the filtering process that makes a "good light source" for the geometric opticians and the "preparation process" that creates well-defined systems for the quantum physicist, while natural light provides the analogue to life. The overriding concern, in the case of quantum mechanics, with the possibility of being able to make precise predictions of the properties of atoms (where no adequate predictions could be made before) led to the adoption, as in the imaginary optical case, of a definition of experiment that was not all-inclusive. But the fact that, within this stated domain, the theory seemed to work universally led to an erroneous extrapolation to the idea that the theory was truly universal.

The optical analogue possesses the weakness as an analogue that in practice a simple superposition process permits the properties of natural light under processes of refraction to be inferred from those of filtered light. But at a deeper level the analogy is still relevant, for two reasons. Firstly, superposition is not an exact result; it is only an approximation, which is accurate only as long as the light intensities are such that non-linear effects can be neglected. And secondly, the question "What is life?" is a much more difficult one to answer satisfactorily than is the analogous question for optics, "What is natural light?" The initial conditions that would have to be assumed in order to describe life and its evolution in quantitative terms remain a matter of speculation (see section 5).



This section will address various attempts to argue that living systems do not really suffer from the kind of considerations raised by Bohr. Some of these seem to depend critically on the assumption that just because at the present time biology seems to be able to ignore those specifically quantum-mechanical aspects of nature such as indeterminacy that have to be taken into account in the case of physical systems, these aspects of nature will never be relevant in the subject of biology. This is a short-sighted view, since whether it is necessary or not, for any particular category of system, to take into account typical quantum effects is a function not only of the type of system concerned but also of the techniques that are available at the relevant time for investigating such systems. Quantum effects such as indeterminacy cannot be expected to be irrelevant in biology for all future time.

In the same way, there is no real justification for the assertion that understanding of the behaviour of an organism may not in fact require such accurate knowledge of the wave function of the organism that it is impossible to obtain such knowledge without the serious disturbances to the system described by Bohr. While possible in principle, such a state of affairs seems improbable in practice. Important properties of inanimate systems frequently depend on the microscopic details of the wave function: there seems to be no good reason for supposing that biosystems will be any different from inanimate ones in this respect.

A somewhat different argument is based on the many observed features of living organisms can indeed be explained adequately in terms of the physical and chemical properties of their constituents. If these physical and chemical properties can be explained in their turn in terms of quantum mechanics, one gets the impression, extrapolating a little, that life itself can be reduced to quantum mechanics. The defect in this argument lies in the way it fails to distinguish between explanations that are necessarily approximate and calculations that under ideal conditions are in principle exact. As an idealisation, a perfect experimental technique would allow the creation of an experimental system that corresponded exactly to a particular well-defined theoretical model, whose behaviour, if unlimited computing power were available and if the Hamiltonian that described time evolution were known exactly, could be calculated with arbitrary accuracy. This perfect computability depends crucially on the fact that one is concerned with a situation corresponding to case (iii) of section 2. Living systems in their natural state, on the other hand, correspond to case (iv).

A final argument depends on the presumed continuity between life and non-life. From the two assumptions, that there is little difference between a very large molecule and a very small organism, and that the properties of large molecules can be calculated rigorously from quantum mechanics given sufficient computing power, it would appear to follow that the properties of a living organism can be calculated on the basis of quantum mechanics.

The situation here is almost exactly the same as that just discussed. Once again the argument neglects the importance of the "differences in the conditions of observation", in other words the differences between cases (iii) and (iv), In the case of the molecule, the preparation process creates specific molecules, whose excitation spectrum, if relevant at all, can be inferred with adequate accuracy for the purposes concerned from a suitable model, and so the properties of the system of interest can in principle be calculated from first principles. In the case of a living organism, however, the preparation conditions are not under this degree of precise control, and even if the molecular structure could be determined, the electronic structure would be uncertain, the uncertainty possibly being unimportant for a very small organism such as a virus, but very large for larger organisms.



The arguments that have been given above have served to confirm from a number of viewpoints the existence of difficulties in applying quantum mechanics to general situations occurring in nature. A variety of commonly held beliefs to the effect that no difficulties in principle exist in applying quantum mechanics to living systems have been shown to be naive. This section will be concerned with a number of deeper issues.

The problem that has been discussed is the difficulty in defining the state of certain kinds of system, including living organisms, in order that the quantum formalism can be applied to them for the purpose of making predictions of future behaviour. This unpredictability is not, it should be noted, simply equivalent to the standard kind of quantum unpredictability involved in situations such as the spontaneous decay of a radioactive nucleus. Biosystems are in general less unpredictable in their behaviour than are radioactive nuclei, since often laws of behaviour for organisms can be found by empirical means, and these laws can extrapolated so as to make predictions of the future behaviour of the same organism, and of other organisms of the same type.

The important question is that of whether legitimate means exist (means that are well-defined and are not arbitrarily created on an ad hoc basis to fit the given situation) for deriving empirical laws such as these on the basis of fundamental principles alone. The difficulty becomes apparent if one imagines that a quantum model has been devised that can reproduce all the observed characteristics of a given organism. Unless there were something specifically non-arbitrary about the choice of model, an evident circularity would be involved in any claim to have predicted the observed properties on the basis of the quantum theory.

Perhaps there is a way of defining an ensemble of all possible life forms (or an equivalent many-worlds state vector), which would contain any particular life-form within it. If it existed, one could perhaps compute all the general laws of behaviour of any given organism. Even on such a basis, deriving the behaviour of an organism from first principles would have required consideration of life as a whole, a consideration unnecessary for the understanding of the behaviour of matter in the physics laboratory.

Anthropic principle considerations(7) suggest that there is an intimate relationship between life and the conditions prevailing at the origin of the universe. Deriving any such ensemble of life-forms must therefore involve consideration of origins. Hartle and Hawking's "wave function of the universe" (8) was an attempt to produce a model of reality that had no arbitrary boundary conditions in it at the origin of the universe. But in a discussion of his point of view Hawking(9) admits that his universal wave function cannot describe every detail of reality as we perceive it; it does not in fact provide a "description of everything". Thus, as an attempt to rule out ab initio subtleties of nature that science would be unable to define on the basis of its equations alone, the "wave function of the universe" idea seems to fail; the idea is as consistent with entities such as David Bohm's postulated "deeper meanings" as it is with Hawking's own anti-mystical views.

Just as in the case of the analogue discussed in section 3, orthodox physics has placed all its bets on a particular set of beliefs as to the correct format for scientific descriptions of nature, ignoring the manifest lack of generality of the methodology which has been a consequence of its insisting on retaining such a format. If the attempt to derive life and meaning on the basis of quantum physics meets intractable problems, it may be appropriate to reconsider the strategy of science, and attempt to work at the same time in the opposite direction. This strategy is apparent in some of the recent writings of David Bohm and other authors(3,10-12). This line will be pursued in more detail in a subsequent paper.



/1/ Published in Found. Phys. 18, 1195-204 (1988).

/2/ Cavendish Laboratory, Madingley Road, Cambridge CB3 OHE, England.

/3/ at a supplementary lecture, not included in the Conference Proceedings.



1. E. Schroedinger, What is Life; Mind and Matter (Cambridge University Press, Cambridge, 1967), Chapter 3 of Mind and Matter.

2. D. Bohm, Wholeness and the Implicate Order (Routledge and Kegan Paul, London, Boston and Henley, 1980), Chapter 1.

3. D. Bohm, Unfolding meaning (Ark, London and New York, 1987), Chapter 3.

4. N. Bohr, Atomic Physics and Human Knowledge (Wiley, New York, 1958), Chapters 2 and 3.

5. E. Wigner, Amer. J. Phys., 31, 6-15 (1963) (reprinted in J.A. Wheeler and W.H. Zurek, Quantum Theory and Measurement (Princeton, Princeton, 1983), section II.4).

6. P. Dirac, The Principles of Quantum Mechanics (Oxford University Press, Oxford, 1958), Chapter 2.

7. J.D. Barrow and F.J. Tipler, The Anthropic Cosmological Principle (Clarendon, Oxford, 1986).

8. J.B. Hartle and S. Hawking, Phys. Rev. D28, 2960-75 (1983).

9. R. Weber, Dialogues with Scientists and Sages (Routledge and Kegan Paul, London and New York, 1986), Chapter 11.

10. C.N. Villars, Psychoenergetics 5, 129-39 (1983).

11. H.P. Stapp, Found. Phys. 12, 363-99 (1982).

12. M. Conrad, D. Home and B. Josephson, in Microphysical Reality and Quantum Formalism, G. Tarozzi and A. van der Merve, eds. (Reidel, Dordrecht, to be published).