Complementarity and the Nature of Empirical Knowledge
While Bohr's failure to convert Einstein to his point of view was a
significant disappointment to him personally, the nature of their disagreement
could be put to profitable issues, as both men spoke the language of physics.
However, the failure of complementarity to draw any appreciable audience
among philosophers was for Bohr a cause of bewilderment and frustration.
This was because Bohr and the philosophers hardly spoke the same language.
Bohr's statement in his last interview: "I think that it would be reasonable
to say that no man who is called a philosopher really understands what
is meant by the complementarity description" really says it all.
It would perhaps be helpful to consider in general the central issue
in philosophy over which communications between Bohr and the "philosophers"
broke down. This issue divides those who hold scientific theories are attempts
to describe the phenomena we experience as the empirical evidence of the
behavior of the objects behind these phenomena and those who don't. The
former hold the view called "realism" while their opponents are called
"anti-realists". However, it is hardly the case that all defenders of realism
in science defend the same form of realism. Indeed, the issues that separated
Bohr and Einstein were between two alternative forms of realism. Therefore,
it is convenient to distinguish "classical realism" from that of "complementarity
The classical realist holds that the success of the theories in classical
physics legitimizes the claim that phenomena are a consequence of the objects
having properties corresponding to the terms used to characterize physical
systems in classical mechanics. From the classical realist point of view,
the concept of the state of the system defined in terms of classical parameters
may be regarded as referring to properties possessed by an independently
real object, which causes the phenomena that confirm theoretical representations
of the system. Bohr's complementarity realism denies this claim of an independently
real object, but offers the possibility of an alternative form of knowledge
of the domain lying behind the phenomena.
Just as there are various forms of realism, so are there various forms
of anti-realism. The dominant one during Bohr's life was that of "instrumentalism",
the view that theoretical terms serve only as constructs enabling predictions
concerning phenomena observed in specific circumstances. Another form of
anti-realism was known as "phenomenalism", the assertion that the only
reality with any content is that of phenomena, and therefore statements
about a reality lying behind the phenomena are meaningless. Both these
views have been incorrectly imputed to Bohr.
Bohr did not see the issue between himself and the philosophers as an
ontological question about the nature of physical reality. It was important
for Bohr to discover whether or not complementarity provided an "objective"
description of phenomena. In his last interview, Bohr complained that the
philosophers "did not see that it [complementarity] was an objective description,
and it was the only possible description. So therefore the relationship
between scientists and philosophers is a very curious one".
Bohr saw the misunderstanding between himself and the philosophers as
an epistemological issue of the requirements for an "objective description".
Complementarity, as a conceptual framework for describing nature, is meant
to be a rational generalization of the framework of classical mechanics.
Since classical mechanics is not an epistemology per se, by analogy
neither is complementarity. Nevertheless, the advent and subsequent dominance
of mechanism had repercussions in epistemology, and the success of classical
mechanics was considered a major datum for the traditions of rationalism
and empiricism, as well as Kant's critical philosophy. Therefore it is
natural to ask, if complementarity is a general framework by which we are
able to better understand the description of nature in natural science,
what does it tell us about the nature of scientific knowledge?
The Relationship Between Complementarity and Epistemology
Classical mechanism provides a directive to look for theories and descriptions
of a certain kind, which can be accessed as adequate or acceptable descriptions.
Mechanism provides an "ideal" for the description of nature by stating
which accounts of phenomena can be considered as acceptable claims of scientific
knowledge. This classical ideal gives a standard against which prospective
theoretical descriptions can be measured. Mechanism not only provides a
framework of concepts such that they have specific meaning in the description
of nature, but it also serves
a directive function orientating the progress of science towards a
projected goal of describing an observed phenomenon in accord with its
ideal of description.
The development of quantum mechanics was conditioned by the ideal of
classical physics. Since these standards were, which Einstein said, not
to be regarded as changeable, quantum theory was necessarily incomplete.
These standards imply that a completely objective description is one that
determines properties possessed by an independent reality by representing
the object as a physical system isolated from any other object. If these
properties are represented by classical parameters, then it follows that
the inability of quantum theory to represent the object of its descriptions
in well-defined classical states means that it does not fulfill what the
classical framework considers acceptable scientific method.
Bohr developed his complementarity to change the standards for this
"ideal" description so that quantum theory would be evaluated as presenting
an acceptable, consistent and complete description of atomic systems. Therefore,
the switch from mechanism to complementarity implies a change of the standards
of what defines acceptable science. Complementarity changes the significance
of the terms of description of objectivity from what Einstein demanded.
Bohr's epistemological lesson begins with the recognition that empirical
knowledge must ultimately be be based upon experiences which are the property
of a perceiving subject, but in scientific knowledge we proceed from this
subjective starting point to an objective account of experienced phenomena.
For Bohr, "objectivity" is a property of the descriptions of phenomena
which science provides. However, the common perception of classical mechanics
as a program for eliminating purposive descriptions of natural phenomena
is a consequence that such descriptions were the result of importing subjective
categories into the description. They provide terms for an "objective description"
because they refer to what is directly observable to any normal
experiencing subject and can be communicated by unambiguous measurements
expressed in the language of mathematics.
Classically, it is possible to hold that unambiguous descriptions which
these terms made possible referred to a reality behind the experienced
phenomena. To a classical realist, the "objectivity" of mechanical descriptions
implied much more than that they were unambiguously communicable. Objectivity
could be grounded in a reality behind the phenomena because the terms defining
the state of the system could be consistently held to refer to the properties
of an independent reality. If one assumed some sort of causal link between
the mental and the physical, it was possible to consider the properties
of the observed phenomena as the mechanical causal effects of properties
possessed by an independent reality interacting with the perceiving subject's
Subjectivity and the Description of Experience
Bohr's complementarity was vitally concerned with how to reconcile the
empiricist conviction that, ultimately, individual subjective experiences
are the foundation of all knowledge of nature with the professed goal of
objectively describing those experiences. Classical physics had an answer,
but it was based on a presupposition that the quantum postulate required
discarding. Therefore, a reanalysis of the concept of objectivity and a
radical redefinition of this criterion of scientific knowledge form the
philosophical heart of complementarity.
Bohr was very much aware that the requirement of objectivity imposed
on the scientific description of nature was in paradoxical contrast to
the subjective status on which such scientific description was based. He
writes: "Yet occasionally just this 'objectivity' of physical observations
becomes particularly suited to emphasize the subjective character of all
experience."  Here Bohr clearly regards a "physical observation" as
an objectively given datum. However, to count as an observation, the experienced
phenomena of that interaction must be described unambiguously. The basis
for this observation is the experimenter's experience of the phenomena.
Since that experience is an everyday experience, it must be described by
the usual terms for describing everyday objects. In this way, "objectivity"
of "physical observations" emphasizes the "subjective character of all
Unfortunately, this use of "subjectivity" by Bohr was reckless, as he
came to realize from the persistent attempts of critics to interpret complementarity
subjectively. Thus, in later years, every essay contains a disclaimer of
any subjective intentions, for example:
...the decisive point is that in neither case [i.e.,
in neither quantum physics nor in relativity] does the approximate widening
of our conceptual framework imply any appeal to an observing subject, which
would hinder unambiguous communication of experience. 
Nevertheless, the classical tendency to regard experience as subjective
leads to reading complementarity as endorsing a subjective foundations
for science. Bohr was very aware of this temptation:
In view of the influence of the mechanical conception of nature
on philosophical thinking, it is understandable that one has sometimes
seen in the notion of complementarity a reference to the subjective observer,
incompatible with the objectivity of scientific description. Of course
in every field of experience we must retain a sharp distinction between
the observer and the content of the observations, but we must realize
that the discovery of the quantum of action has thrown new light on the
very foundation of the description of nature and revealed hitherto un-noticed
presuppositions to the rational use of the concepts on which the communication
of experience rests. In quantum physics, as we have seen, an account of
the functioning of the measuring instruments is indispensable to the definition
of the phenomena and we must, so to speak, distinguish between subject
and object in such a way that each single case secures the unambiguous
application of the elementary physical concepts used in that description.
Bohr believed this shifting of the distinction between subject and object
also caused problems in psychology and biology, as noted in part 6 of this
review. In the essay from which the above paragraph was taken he continues
by pointing out:
While in the mechanical conception of nature the subject -
object distinction was fixed, room is provided for a wider description
through the recognition that the consequent use of our concepts requires
different placings of such a separation. 
These comments indicate that Bohr believed the task of securing objectivity
of empirical knowledge is presumed to refer to the description of
experience, or in other words, what happens after the subject has
the experience rather than how experience originates. Bohr avoids grounding
objectivity in either the objects contribution or the subjects contribution
to the formation of experience. His view was that objectivity is
secured by understanding the proper use of descriptive concepts and so
it is a matter of how we describe an experienced phenomena rather
than how such experience originates. Whether consciousness is regarded
as reducible to a physical process or as a mental presentation of ideas
is irrelevant to complementarity. What concerned Bohr was not the
interplay of distinct ontological orders, but the necessity of describing
experienced phenomena unambiguously.
Observation, for Bohr, did not involve any mysterious mental/physical
interactions. In order to describe an observation in physics, both systems
which interact to produce the observation must be capable of being described
as purely physical systems. Bohr is emphatic in his claim that there is
no need for new observational language in quantum physics. Since
classical descriptive concepts have unambiguous empirical meaning, quantum
experimental language presents no need to be described in any fashion other
than that of classical physics.
When speaking of "observing" an atomic system, if Bohr's use of "observing"
is confused with his use of "experiencing" it may be philosophically misleading.
An observer "experiences" the phenomena that confirms the theory. Some
of these phenomena may be described as observations of atomic systems.
Classical descriptive terms have unambiguous reference only for describing
phenomenal objects, and since "atomic system" cannot refer to a phenomenal
object, classical descriptive terms cannot be used to describe atomic systems.
Therefore, if any description of the atomic system is at all possible,
such a description cannot be communicated unambiguously.
Bohr's complementarity dramatically alters the classical description
of nature. "Observer" and "object" become categories of description
having precise meaning only in the context of a particular description
of an experiential event.
The Ideal of Objectivity
Epistemology addresses the fundamental question of how to formulate
a description of experience to meet the criterion of objectivity. Classically,
a scientific description was held to be "objective" because it was believed
that it determined properties possessed by the object as it exists independently
apart from any observational determination of those properties. The methodology
of science was built around the goal of developing the means for determining
such properties and at the same time eliminating any element that arises
from the role of the observing system in acquiring that empirical knowledge
necessary to confirm scientific theory.
Bohr's argument was that this understanding of objectivity of a scientific
description of nature must be revised. For Bohr, the objectivity of scientific
description is grounded in the use of concepts used for describing the
experience. These concepts secure that objectivity not by playing any role
in the origin or formation of experience, but by guaranteeing that communications
expressed in terms of such concepts are unambiguous. Bohr writes:
Every scientist, however, is constantly confronted with the
problem of objective description of experience, by which we mean unambiguous
communication. Our basic tool is, of course, plain language. ...we shall
not be concerned here with the origin of such language, but with its scope
in scientific communication, and especially with the problem of how objectivity
may be retained during the growth of experience beyond the event of daily
The main point to realize is that all knowledge presents itself within
a conceptual framework adapted to account for previous experience and that
any such frame may prove too narrow to comprehend new experiences... 
As Bohr understood natural science, the objectivity of a description cannot
be secured by separating the "objective" properties possessed by an independent
reality from the "subjective" properties which exist only in relation to
an experiencing subject. In complementarity, the definition of "objectivity"
requires the scientists goal to be development of a conceptual scheme or
framework for describing the phenomena in a way which can be communicated
Once the quantum postulate is accepted, it is necessary to realize that
any description of a phenomena is a description of a physical interaction
in which the distinction between observed object and observing system is
made for the purpose of unambiguously describing the interaction as an
observation of a specific phenomenal object. To secure unambiguous communication
of the result of this observation, it is necessary that the description
includes a precise specification of the whole physical situation in which
the observation occurs.
Bohr's epistemological lesson teaches that on the presuppositions of
the classical framework, descriptions expressed by those classical mechanical
concepts can be considered unambiguous. But the exploration of the atomic
system dictates a need to adopt a new presupposition about these phenomena
to that expressed by the quantum postulate. This change then renders descriptions
expressed in classical terms ambiguous, consequently a new, more
general framework is needed to restore unambiguity.
Because it is necessary to describe the observation as a causal process,
which must be done if that observation is to be interpreted as an observation
of an object system, it is also necessary to apply classical dynamical
conservation principles. And to do this, there must be a theoretical representation
of the system isolated from the interaction. But in quantum formalism it
is impossible to derive a classical mechanical state of the system that
would enable us to "picture" the object apart from the observation. Therefore
it must be recognized that the theoretical representation of the atomic
system is an abstraction, not a picture of a concrete object.
The generalization of the classical mechanical framework which Bohr
advocated does much more than eliminate the ambiguities that result from
quantum formalism. Ultimately, in proposing to revise the basis for objective
scientific descriptions, complementarity alters the very concept of physical
reality. In Bohr's complementarity, the distinction between the subject
that experiences and the phenomena that is experienced is made within
the description of the phenomena as an observational interaction, whereas
in classical mechanics the distinction is made between the experiencer
and the experienced. Bohr writes:
The notion of complementarity does in no way involve a departure
from our notion as detached observers of nature, but must be regarded
as the logical expression of our situation as regard objective description
in this field of experience. The recognition that the interaction between
the measuring tools and the physical system under investigation constitutes
an integral part of the quantum phenomena has not only revealed an unsuspected
limitation of the mechanical conception of nature, as characterized
by the attribution of separate properties in physical systems, but
has forced us, in the ordering of experience, to pay proper attention to
the conditions of observation. 
Bohr's claim was challenged by his student and friendly critic, Wolfgang
Pauli. Upon reading Bohr's first draft of the essay from which the above
quote is taken, Pauli replied to Bohr in his style which was overladen
with sarcasm and irony, for which he become infamous for:
Under your great influence it was indeed getting more and more
difficult for me to find something on which I have a different opinion
than you. To a certain extent I am therefore glad, that eventually I found
something: the definition and the use of the expression "detached observer".
...According to my own point of view, the degree of this "detachment" is
gradually lessened in our theoretical explanation of nature and I am expecting
further steps in this direction. ...it seems quite appropriate to call
the conceptual description of nature in classical physics, which Einstein
so emphatically wishes to retain, "the ideal of the detached observer".
To put it drastically the observer has according to this ideal to disappear
entirely in a discrete manner as a hidden spectator, never as actor, nature
being left alone in a predetermined course of events, independent of the
way in which the phenomena are observed. "Like the moon has a definite
position" Einstein said to me last winter, "whether or not we look at the
moon, the same must also hold for atomic objects, as there are no sharp
distinctions between these and macroscopic objects. Observation cannot
an element of reality like a position, there must be something contained
in the complete description which corresponds to the possibility
of observing a position, already before the observation has been made."
...I considered the impredictable [sic] change of the state by a
single observation in spite of the objective character of the result of
every observation and notwithstanding the statistical laws for the frequencies
of repeated observation under equal conditions - to be an abandonment
of the idea of the isolation (detachment) of the observer from the course
of physical events outside himself. 
It is apparent that, even though Pauli worked with Bohr on his Como paper
and always regarded himself as an advocate of complementarity, both he
and Einstein confused the distinction made in the description of phenomena
as an interaction between the observing system and the atomic object with
which it interacts with the distinction between the subject that experiences
the whole phenomena and the phenomena the subject experiences. Bohr did
not realize the phrase "detached observer" confused the issue by
tending to identify the experiencing subject, the "observer" in classical
physics, with the "observing system", which, according to Bohr's view,
is treated in the description of the phenomena as a physical system partially
causing the phenomena given to the subject. Bohr replied to Pauli like
As always, you touch on a very central point. A phrase like
"detached observer" ... used in connection with the phrase: "objective
description" ... had to me a very definite meaning. In all unambiguous
account it is indeed a primary demand that the separation between the observing
subject and the objective content of communication [i.e.,
the phenomena to be described] is clearly defined and agreed upon. ...
this condition is indispensable in all scientific knowledge. ... It appears
that what we have really learned in physics is how to eliminate subjective
elements in the account of experience, and it is rather this recognition
which in turn offers guidance as regards objective description in other
fields of science. To my mind this situation is well described by the phrase
"detached observer". 
Bohr insisted that the description of a phenomena must include a description
of the whole interaction between observing system and observed object,
all of which forms part of the whole phenomena to be described by an adequate
theory. Pauli was certainly misunderstanding Bohr's intention in the phrase
"detached observer". While there need not be a separation between observer
and observed on a physical level, and indeed there cannot be if there is
to be an observation at all, by requiring a description of a "detached
observer", objectivity demands that subjective elements in the accounting
of experience are eliminated. Thus, the "observer", in this sense, is detached
from the observation.
When Bohr refused to define how experience originates, he turned his
back on the whole representational tradition of epistemology. Although
there is a superficial resemblance between complementarity and Kantian
epistemology, the apparent similarities are created by equating such terms
as "experience" and "objectivity", which change in meaning in the shift
from Kantian epistemology to complementarity.
A true Kantian would never argue, as Bohr does, that a physical description
would demand a change in the proper use of concepts which give experience
its form. If Bohr were a Kantian, he would argue, as he does, that the
classical concepts are indispensable, but then his claim that the quantum
postulate demonstrates that the classical framework is no longer tenable
would be a complete non sequitur. Bohr's rejection of the Kantian
outlook follows from the fact that as he understood Kant, the critical
philosophy was designed to show that classical concepts were independent
of the content of experience and thus could not be shown to be inadequate
by any empirical discovery.