p We have already given a critical analysis of phenomenalist epistemology which presents acquisition of knowledge as combining of subjective "sensory data”. Another, more sophisticated variety of philosophical subjectivism has much greater currency in present-day Western works on the philosophy of science. Until recently, the view prevailed amongst West European and American specialists in the logic and methodology of science that only cognition at the pre-scientific level (perception and knowledge recorded in terms of everyday language) may deal with actually existing objects. From this standpoint, scientific theoretical knowledge is different in character: it merely records in a special schematic form the regularly recurring dependences existing between the objects of prescientific experience. Of course, account is taken of the fact that acquisition of scientific knowledge implies employment of artificially created objects, in particular, apparatus, measurement instruments, etc.

p The actual existence of the latter is by no means rejected. Moreover, it is believed that natural and artificial objects are equally objects of cognition. To be more precise, cognition is thought to be concerned with establishing definite relations between various combinations of sensuously perceived natural and artificial objects, for that is exactly what the process of measurement consists in, and scientific cognition is in this case limited to performing various measurement operations. In terms of this conception, for instance, the object studied by the microphysicist is not the processes in which electrons, positrons, and other objects inaccessible to the senses are involved but the behaviour of the corresponding devices: oscillations of their indicators, appearance of light spots on displays, etc. In other words, it is assumed that apparatus and measurement devices do not at all mediate the cognitive relation 156 to the objects that are not given to the subject in pre- scientific experience but appear themselves as the objects of knowledge (the adherents of this view insist that naturally created things become objects of scientific cognition only in their relation to the apparatus and measurement instruments). Only that is regarded as real which can be directly observable. Everything else, including objects that are specified at the theoretical level only, are regarded as certain subjective fictions which, although playing a certain role in the cognitive process, do not by themselves have real objects as referents. As we have seen, these are the characteristic arguments of the operationalist doctrine.

p This trend of thinking is based on the opposition of knowledge as a record of the directly observable to knowledge resulting from a whole ensemble of assumptions, suppositions, and arguments. Indeed, to make judgements about the behaviour of microobjects from instrument readings, one must be acquainted both with the theory of the domain of reality under study and with the theory describing the work of the device itself, enabling us to correlate the instrument readings with the corresponding characteristics of the phenomena studied. The adherents of this conception believe that the objects for which concepts are introduced in these complicated arguments and assumptions cannot be real in the same degree as the artificial and natural objects of our everyday experience-stones, trees, tables, chairs, machines, apparatus, etc. But it is easy to show a lack of logic in this argument.

p In ordinary life we have to use all kinds of mediator objects all the time for the simplest observations—- spectacles, the magnifying glass, or, say, simple window glass. In the same way, the surgeon uses the probe in examining a wound. In all these cases man studies those objects the relation to which is mediated by artificially constructed devices rather than the mediators themselves. To be consistent, one must also recognise that even in cases of elementary observation the subject’s relation to the object is mediated by the environment filling the space between the two. The singling out of a real object implies in all instances a reliance (usually unconscious) on a number of assumptions concerning the behaviour of the mediator object. But that is not all. As we have tried to show in the previous chapter, even those man-made objects which do not function directly as mediators in observation (labour implements, the objects of everyday life, etc.), are actually instrumental in the social mediation of perception, for it is in the objects of the "artificial environment" that the historically accumulated experiences of object- 157 transforming and cognitive activity, are objectified, and in particular, the Standards and norms of perception are reified. The knowledge of the real object can only be singled out of the varied sensory information with the help of such norms and standards. Any perception is, as we have endeavoured to show, a complex mediating activity implying assumptions, hypotheses, schematisation, etc.

p Supporters of this variety of subjectivism do not doubt the actual existence of the objects of ordinary pre- scientific experience. Moreover, in their view, those infinitesimal or extra-large objects that can only be studied with the aid of special scientific apparatus (microscopes and telescopes), also really exist. But in this case interpretation of the results of observation requires conscious use of a number of branches of theoretical physics covering, in particular, the propagation of light waves in outer space, in the Earth’s atmosphere, a system of lenses, the eye, etc. Does it not mean that knowledge acquired through of a number of assumptions, suppositions, and theoretical reasoning, can also relate to real objects? Why must we then negate the existence of actual referents, e.g., of the objects of modern microphysics? The adherents of this view reply that observable and non-observable objects must be distinguished. The knowledge of observable objects, they believe, relates to actual referents although it may imply certain assumptions, hypotheses, and arguments. As for non-observable objects, their existence is fictitious.

p Indeed, not all objects, magnitudes, and parameters with which a certain scientific theory is concerned, are actually observable. Let us ask ourselves this question, however: does that mean that a certain object, now non-observable and studied at the given moment on the theoretical level only, will never become observable at all? Apparently not. For example, although the theory of molecular structure of matter was originally merely a theoretical hypothesis and there was no way of observing molecules in direct experience, the molecules of many substances can now be observed through electronic microscopes.

p The justice of this is recognised by the scholars holding the view here criticised. But they point to the essential difference between molecules and such subatomic objects as the electron. The knowledge of many objects that are studied purely theoretically does not cancel the possibility of their eventual fixation in experience by some instruments of experimental inquiry. But there are theoretical objects (the electron included) which cannot in principle be observed. Only that is real which can be observed actually or potentially. Objects that are non-observable 158 in principle do not exist as real objects—that is the conclusion drawn by the adherents of the system of views considered here.

p The objects with which modern microphysics deals are indeed regarded as non-observable in principle. But what does observability or non-observability of objects and their characteristics that are studied in scientific theory mean?

p Of course, only those objects can be observable which are in some way or other included in the process of acquiring sensory information. However, we have endeavoured to show in the previous chapter that already at the level of ordinary pre-scientific perception, the knowledge of the characteristics of observed phenomena is not identical to the information received through sensory channels, being determined by specific referential meanings. It is these referential meanings, object-hypotheses, the standards of perception rather than sensory information by itself, that determine what precisely is observed or perceived. In scientific theoretical thinking, theory rather than sensory information by itself determines which of the objects, magnitudes, or parameters studied in the theory can be actually or potentially observable. Theory has to take into account such circumstances, accidental relative to the objects studied, as the size of man’s body and the specific traits of his perceptual system. The fact that men as physical bodies belong to the class of macro-objects and that man as the subject of perception can therefore use only macro-objects as apparatus proves to be essential for microphysics.

p These circumstances determine the possibility of including certain objects in the very process of acquiring sensory information, i. e., in the act of experiential observation. However, only within the framework of a definite scientific theory can it be established what specific objects studied by science may or may not be included in the process of observation and for what reasons, what the meaningful characteristics of objects of both kinds are, and what precisely is observed. The properties of the subject’s perceptual system are also considered in terms of the given theory. In any case, the observability or non-observability of the given objects of scientific knowledge depends in principle on definite characteristics of these objects, assumed or established in the theory, and does not directly coincide with their existence or non-existence. The objects that are not in principle observable by man can actually exist. (That means that if observations were carried out by an intelligent being strongly differing from man in its natural properties, e.g., if it were comparable in size with 159 micro-objects, it might record in an experimental way .many of the objects that cannot in principle be observed by man. On the other hand, a radical revision of a given scientific theory and a different choice of the basic assumptions will inevitably affect the notions of the observability or non-observability as a matter of principle.)

p Grover Maxwell, a modern American specialist in the philosophy of science, considers the following purely hypothetical case as an illustration of the thesis of the possibility of actual existence of objects unobservable in principle.

p Suppose, he argues, that new types of micro-objects are discovered by science that are at present unknown and which interact with electrons under certain circumstances in such a way that the interaction does not disturb their eigenstate. Suppose also that a drug is discovered which alters the human perceptual apparatus—perhaps even activates latent capacities so that a new sense modality emerges. Finally, suppose that with our altered perceptual system we are able to perceive (not necessarily visually) by means of the newly discovered type of micro-objects in a manner roughly analogous to visual perception in which, as is well known, photons participate. Under certain additional conditions which we shall not characterise here, we might be able to "observe directly" the position and other characteristics of some electrons. It would follow of course, that quantum theory would have to be altered in some respects, since the newly discovered type of microobjects does not conform to all its principles. At the same time the revision of the theory does not in this case provide any grounds for concluding that the electrons observed are not the same objects that were regarded as nonobservable in principle from the standpoint of old theoretical notions. No one will doubt the reality of the electrons observed. But if these are the same objects that were not observed earlier, it is obvious that we had no right to doubt their actual existence before that either. However improbable the hypothetical case considered here might seem, it does not involve any logical or conceptual absurdity, concludes Maxwell.^^47^^

p In a conversation with Werner Heisenberg Einstein said once: "From the principled positions it is absolutely incorrect to desire that a theory should be founded on observable magnitudes only. For in reality it is all precisely the other way round. Only theory decides what one can observe... Your assertion that you introduce only observable magnitudes is actually an assumption about a property of the theory on which you are working.”^^48^^

p Thus the experimenter does not observe absolutely the same objects that are the objects of perception at the prescientific level. The scientist records in experience (one may even say "sees directly”) objects, processes and situations which are not ordinarily perceived at all: changes in electric voltage, a drop in strength of current in the circuit, etc. The main point here is not, of course, a change in the sensitivity of the perceptual system but the emergence of new referential meanings determined by the accepted scientific theory.

p In this respect, observation aided by theory is in principle similar to ordinary perception: in both cases the referential content of what is observed is determined by a system of object-hypotheses and not by sensory information itself. It would be a mistake, however, to slur over the differences between the two processes, as Kuhn is inclined to do, for instance. The American scientist correctly stated the extremely important fact that theoretical concepts do not serve simply for interpreting the results of ordinary perception obtained regardless of their utilisation but are included in the act of scientific observation itself determining its nature and results. Yet Kuhn is hardly justified in going still further and insisting that scientific observation is of the same subjective and direct nature as ordinary perception, that in both cases there is no conscious interpretation or extended subjective reflexion.^^49^^ This notion of Kuhn is closely linked with the main idea developed in his book—the view that successive replacement of scientific paradigms is similar to changes in the structure of the perceptive field resulting from a "switch in visual gestalt”. But the ability to "see directly”, through the medium of apparatus readings, the objective processes indicated by the devices assumes an extensive education in which the behaviour of the devices is consciously correlated with the behaviour of the object studied. Even when this education is completed and the scientist sees directly, as it were, those objective processes which are for him meaningfully defined by a system of theoretical concepts, fixation of observed objects also assumes the functioning of ordinary pre-scientific perception: in order to observe the strength of current in a circuit from ampermeter readings, the subject must be able to perceive the ampermeter itself and the motion of its needle as objects of ordinary experience. Thus the “ givenness” of the objects of scientific research in experience includes observation of two objects simultaneously: of the object of everyday experience and of that thing whose referential meaning is grasped by the subject in terms of 161 the concepts of some theory (both of these things exist objectively and actually, although on different levels of reality, so to speak). When a person becomes a scientist, he does not cease to be the subject of ordinary pre-scientific experience and of practical activity associated with it. For this reason, the system of referential meanings which serve to maintain this activity, being included in the mechanism of ordinary perception, cannot in principle be replaced by the referential meanings defined at the level of scientific cognition (though Paul Feyerabend suggests the opposite). The higher levels of cognitive activity do not cancel the functioning of the mechanisms of ordinary perception but are in a specific manner superimposed on these mechanisms incorporating them in more complex syntheses. It would therefore be wrong to insist, for instance, that the referential meanings perceived in language study are generated by language, although mastering it signifies a new stage in the interpretation of perceptions: in actual fact they are basically formed already at the pre-linguistic level of cognition, in the course of practical object-oriented activity, although language does introduce something new in them. Observation in scientific cognition does not exclude the functioning of the mechanisms of ordinary perception. The astronomer observing the Sun as a cosmic body at a certain distance from the Earth and subject in its movements to theoretically formulated laws, cannot at the same time get rid of the impression, shown to be illusory by science, that the Sun moves relative to the immobile Earth.

p The development of science eliminates the illusions of pre-scientific cognition. But a scientific theoretical picture of reality does not at all imply a negation of the objective reality of those objects (as well as of their aspects and relations) with which man deals at the pre-scientific level, and neither does it negate the truth (relative truth, of course) of many assertions of the so-called common sense. This applies not only to such objects of ordinary experience as tables, trees, stones, etc., but also to properties of these objects which are commonly referred to in philosophy as secondary: colour, smell, etc. It would be inconsistent to assert that only electromagnetic waves of definite length and not colours and smells exist objectively and really, and at the same time to recognise the objective reality of the objects of ordinary pre-scientific experience, that is exactly the view held by those who divide the perceived qualities of objects into primary and secondary. Physical theories do not include the concepts of secondary qualities but, more than that, they do not include the concepts of 162 the objects of ordinary pre-scientific experience. If we were to regard as real only those objects to which physical theories directly refer, we should conclude that in actual fact only definite combinations of atoms and molecules rather than trees, rocks, and tables exist in reality.

p In actual fact, cognition at different levels deals with real objects and real characteristics of these objects. However, objective reality itself is multidimensional, it has many levels, and different objects may belong to different levels of reality. Ordinary macro-objects and the secondary qualities inherent in them exist at that level of objective reality to which ordinary pre-scientific experiences belong. Scientific cognition, physics in particular, penetrates into a deeper level of objective reality, whose existence does not cancel the reality of the objects of ordinary experience.

p A system of theoretical concepts reflects the characteristics of actually existing objects, including those that are actually or essentially non-observable. The meaning of these concepts is thus not reducible to an ensemble of the laboratory operations of measurement, as operationalists believe. On the contrary, the measurement itself only becomes possible when we know what to measure, that is, when the general characteristics of the objects measured are theoretically specified. It is exactly scientific theory that makes it possible to select from the entire diversity of experience those facts and dependences between them the investigation of which will permit the scientist to single out the essential characteristics of the objects under study. Measurements that are performed outside the context of a well-developed theoretical system formulating the essential dependences between objects, including non- observable ones, turn out to be absolutely meaningless, as a rule. And it is not just the fact that measurement results are subsequently theoretically interpreted that is important here. Well-developed theoretical conceptions are a necessary premise of meaningful measurements themselves, for only the former indicate the object and the mode of measurement itself. The measurements performed outside of a correlation with the essential dependences of a definite type of objects do not express, strictly speaking, an act of cognition, just as acquisition of information from the environment uncorrelated with objects is not yet cognition.

p “We often hear,” writes Kuhn, "that they [the laws expressing quantitative dependences] are found by examining measurements undertaken for their own sake and without theoretical commitment. But history offers no 163 support for so excessively Baconian a method. Boyle’s experiments were not conceivable (and if conceived would have received another interpretation or none at all) until air was recognized as an elastic fluid to which all the elaborate concepts of hydrostatics could be applied. Coulomb’s success depended upon his constructing special apparatus to measure the force between point charges. (Those who had previously measured electrical forces using ordinary pan balances, etc., had found no consistent or simple regularity at all.) But that design, in turn, depended upon the previous recognition that every particle of electric fluid acts upon every other at a distance. It was for the force between such particles—the only force which might safely be assumed a simple function of distancethat Coulomb was looking. Joule’s experiments could also be used to illustrate how quantitative laws emerge through paradigm articulation. In fact, so general and close is the relation between qualitative paradigm and quantitative law that, since Galileo, such laws have often been correctly guessed with the aid of a paradigm years before apparatus could be designed for their experimental determination.”^^50^^

p Using as an example the revolution in chemistry carried out by Dalton, Kuhn shows that one and the same operation applied to nature through different paradigms may indicate quite different aspects of the patterns of nature. Moreover, an old measurement operation in a new role may produce other experimental results.^^5^^ *

p Einstein’s analysis of the procedures for measuring time (we may recall here that it was this analysis that was the starting point of Bridgman’s formulation of the doctrine of operationalism) is far from being a mere description of a "directly given" operation, implying in actual fact a number of theoretical premises. Simultaneity can only be defined if we postulate that the velocity of light in vacuum is the same in all directions and invariant relative to the motions of source and receiver. This, postulate is logically prior in the special theory of relativity to any experimental measurement of the velocity of light, because it is used in the very definition of the time scale at distant points.^^52^^

p Even the simplest prescriptions for measurement operations used in science usually follow from theoretical considerations. True, instructions for laboratory operations may be formulated in such a way that their theoretical foundations will be camouflaged, but that does not mean at all that these foundations do not actually exist.

p Thus scientific theories determining the meaning and character of experimental procedures contain as often as 164 not knowledge of such objects and parameters which are not observed and are not measured directly. Let us now note that there are also scientific theories, the most fundamental ones, actually, (usually referred to as substantive in the literature on the methodology of science) which are not applied directly to interpretation of observation data at all but are only correlated with the empirical world in combination with other theories (the latter are commonly referred to as “observational” or “interpretative”) and on condition that a number of additional assumptions are made. Generally speaking, the question of experimental application of a fundamental scientific theory (and, in this connection, of its experimental verification) proves to be far from simple, and usually the search for methods of experimental application of such a theory requires considerable efforts for further elaboration of the theory itself and the construction of a number of additional theories, hypotheses, etc..

p An isolated theory is never directly linked up with an experiment: an act of such association implies the use of a whole hierarchy of theories and hypotheses including those from other domains of knowledge, the theory of experimental devices, a number of hypotheses linking up the non-observable with the observable, certain idealisation assumptions, etc. The experimental data themselves are formulated in terms of a definite (“interpretative”, or “observational”) theory.

p We shall not consider in detail the important problem of the interrelation ef the empirical and theoretical components in scientific knowledge. In the context of the problems of immediate interest to us, it is important to stress the relative independence of knowledge recorded in theory from various (potentially infinite) ways of empirical or experimental application of that knowledge. Of course, theory must be experimentally tested, that is to say, it must be linked up through a whole chain of mediations with experimental results. The question as to how this verification is carried out is fairly complicated and is now intensively studied in the literature on the methodology of science. At the same time the meaning, the content of theoretical knowledge, is not directly determined by the modes of its association with experimental data.

p In the Western philosophy of science, the conception of logical positivism has until recently prevailed; it regarded a scientific theory as an uninterpreted formal calculus given meaningful, empirical interpretation in terms of the so-called correspondence rules connecting the terms of 165 the theoretical language with those used in the sentences of observation allegedly recording only direct sensory data. The rules of correspondence must, from this standpoint, be necessarily included in the structure of the scientific theory itself, for only these rules transform an uninterpreted formal calculus, which refers to nothing, into a theory that is, knowledge about a definite class of objects.53 in reality, however, the meaning or content of a theory is not determined directly by its empirical applications, for meaning is specified "from above" rather than "from below”, through a model interpretation of theoretical assertions. The number of potential modes of empirical application of the given theory is in actual fact infinite (and all possible applications cannot be foreseen in advance), whereas the number of "correspondence rules" used in the given theory is strictly limited, according to logical positivism. And finally, the most important point is this. Assertions linking the given theory with experimental results do not simply correlate theoretical and non-theoretical, "purely observational" terms but themselves belong to an auxiliary theory (auxiliary relative to the given one). Thus these statements (it would be imprecise to call them "correspondence rules”, in any positivist sense) are not included in the structure of the given theory itself, which is relatively independent from various possible modes of its empirical application. Knowledge recorded in a theory reflects the essential dependences between real objects, far from being a set of prescriptions for carrying out laboratory operations; neither is it restricted to listing the ways of direct practical transformation of objects.

p Here we would like to draw a far-reaching analogy between theoretical knowledge and knowledge recorded in ordinary perception. One may recall that the referential meaning of the perceptual image is amodal, that is, relatively independent from the type of modality (visual, auditory, tactile, etc.) in which perception is implemented. This is apparently a general property of the cognitive relation—relative independence of knowledge from the modes of its correlation with actual sensory information (the independence is indeed relative, for knowledge is impossible in the absence of any mode of such correlation).

Thus cognition may be a reflection of real objects both in ordinary perception and in scientific thinking— at both the empirical and theoretical levels of the latter.