Logical Empiricism

by Maurice Cornforth

Maurice Cornforth, who lives in London, first studied philosophy at London and then at Cambridge University. He is the author of Science and Idealism: An Examination of Pure Empiricism and Modern Logic (1947), and of several articles and papers on philosophy, and is now at work on a continuation of Science and Idealism, dealing with pragmatism and recent theories of semantics and semiotic, and also on a larger book on logic and the theory of knowledge.

How Logicians Announced a Great New Advance in Philosophy

In 1914 Bertrand Russell, having completed the Principia Mathematica, announced that a new way of thinking had "gradually crept into philosophy through the critical scrutiny of mathematics." This type of philosophy, he said, "represents, I believe, the same kind of advance as was introduced into physics by Galileo: the substitution of piecemeal, detailed and verifiable results for large untested generalities recommended only by a certain appeal to the imagination." [1]

The key to this new philosophy was its method, the method of Logical Analysis. Philosophy, said Russell, should not attempt to compete with natural science in working out a theory of the universe, or theories about particular parts of the universe. It was a mistake to suppose "that a‑priori reasoning could reveal otherwise undiscoverable secrets about the universe." [2] On the contrary, facts and generalizations about the world, positive knowledge, must be acquired empirically, partly through ordinary perception, partly by the more refined technique of natural science. The task of philosophy was then to subject the propositions established through ordinary perception and by science to a logical analysis. Such an analysis could not establish any new truths. But by making clear the meaning of truths already known it would remove difficulties and confusions and impart a new clarity to our knowledge.

This idea that the task of philosophy is rather to make an analysis of the meaning or reference of empirical knowledge than to establish transcendental truths by a priori reasoning was not in itself anything very new. Already Mach's Analysis of Sensations, published in 1885, had set out to make clear the meaning of all scientific propositions by showing how every science was in truth concerned with the order and arrangement of sensations or sense data and with nothing else. And, indeed, the same thing had been put forward more than a century earlier in Berkeley's Principles of Human Knowledge, published in 1710.

Russell, therefore, was only restating something which had been common ground amongst empiricists for over two centuries. What he claimed as specifically new was the discovery of a logical technique for restating and refining the traditional method of empiricist philosophy.

Thus in his latest work, A History of Western Philosophy, Russell says: "Modern analytic empiricism . . . differs from that of Locke, Berkeley and Hume by its incorporation of mathematics and its development of a powerful logical technique. It is thus able, in regard to certain problems, to achieve definite answers, which have the quality of science rather than of philosophy. It has the advantage," he continues, "as compared with the philosophies of the system‑builders, of being able to tackle its problems one at a time, instead of having to invent a block theory of the whole universe. Its methods, in this respect, resemble those of science." [3]

The "powerful logical technique" to which Russell refers is the technique of modern mathematical logic. He had in mind that the methods which had proved successful in mathematical logic and in mathematical analysis could be applied in philosophy.

Mathematicians of the nineteenth century had been able to overcome difficulties in the differential calculus by the definition of infinitesimal quantities in terms of the limit of a series of real numbers. Russell himself believed he had shown in Principia Mathematica that the whole of mathematics could be derived from logic by means of Frege's definition of cardinal numbers as "classes of classes similar to a given class." If, then, the nature of numbers, the nature of infinitesimals, and so on had been clarified by a technique of analytic definition, could not the philosophical problems of the nature of the world be likewise solved by a technique of logical analysis, analyzing and defining terms whose meaning was in doubt by an exact logical technique? This would then replace the dubious a‑priori theorizing of the builders of philosophical systems.

Science and the Crisis of Philosophical Systems

The use of the method of logical analysis was a reflection of the general state of crisis into which the whole activity of philosophical system building had fallen as a result of the rapid development of natural science.

Already in the nineteenth century the futility of philosophical system building was becoming evident. In terms of eighteenth century science, philosophy could still seem to such thinkers as Leibnitz to have a separate function in interpreting the world, which science by itself could not fulfill. But as Frederick Engels was to point out, the great unifying discoveries of nineteenth century science made such systems once for all unnecessary. The last "great" philosophical system was that of Hegel. More and more has it become obvious that system building and the controversy of systems is barren of results. The systems can no longer satisfy the demands which men put upon them. The tasks of philosophy must be different from the tasks formulated by the builders of systems.

It is worth noting that this issue was already quite squarely put by Marxists some time before it was formulated by anybody else.

"Modern materialism," wrote Engels in Anti‑Dühring (1878), "no longer needs any philosophy standing above the other sciences. As soon as each separate science is required to get clarity as to its position in the great totality of things and of our knowledge of things, a special science dealing with this totality is superfluous. What still independently survives of all former philosophy is the science of thought and its laws—formal logic and dialectics. Everything else is merged in the positive science of nature and history." [4]

The new formulation which dialectical materialism provided of the tasks of philosophy, and the contribution it made had little influence upon the content of academic philosophy. The official philosophers, the professionals, continued to build their systems. The past hundred years, indeed, have seen the production of many systems, the latest being that of A. N. Whitehead. Sometimes the system builders have tried to use parts of science to bolster up their fantasies—for instance, Bergson, or Whitehead himself. Others, like the English and American "absolute idealists" of fifty years ago, cheerfully ignored science altogether, saying it was concerned merely with "appearances."

But the continuance of this type of philosophy could not but provoke a reaction. It had been challenged by dialectical materialism, and it was again challenged by the representatives of philosophical empiricism. Thus we find that, historically, Russell's announcement that philosophy should concern itself with logical analysis and not with the attempt to "reveal otherwise undiscoverable secrets" by "a‑priori reasoning" was flung down as a challenge against the contemporary English "absolute idealists"—Green, Bosanquet, Bradley, and others. He realized clearly enough that their system building could not stand up to scientific criticism, that it was based on confusions and logical fallacies, and that it was wholly useless and undesirable from the point of view of extending scientific knowledge and the application of scientific methods to theoretical problems.

For Russell and those who thought on similar lines, logical analysis, which should in their opinion become the sole concern of philosophy, did not aim at building up its own system of the world, but at interpreting and clarifying the results of science.

Still, such an interpretation of science needed to have some definite content. The theories of formal logic, to which Russell himself made such a great contribution, provided only the bare logical form in which propositions should be stated. But an interpretation that made clear the propositions  of  science must not only make clear their logical form; it must make clear exactly what sort of things they are about—their content.

It is a noteworthy fact that in seeking the ideas in terms of which the content of science could be clarified Russell and his fellow work looked backwards, to the results of previous empirical philosophy.

In his three chief works of philosophical analysis—Our Knowledge of the External World, The Analysis of Matter, and The Analysis of Mind—Russell proceeded on the classical empiricist principle, which he formulated as follows: "It can be laid down, quite generally that, in so far as physics or common sense is verifiable, it must be capable of interpretation in terms of actual sense‑data alone." [5]

For Russell, mathematical analysis had already made clear the nature of number—what we are talking about when we talk about numbers—in terms of the abstract logical categories of individual things and their properties and relations, as developed in the logical theory of classes. In the same way, he thought, philosophical analysis would make clear the nature of the mind and its relationship to the body, the constitution of the material universe and the existence of the external world—what we are talking about when we talk of mind, body, matter, and so on—in terms of the categories previously employed by philosophical empiricism; namely, impressions, sensations, sense‑data.

But this type of interpretation, interpreting all scientific statements in terms of sense‑data and so analyzing the objective material world out of existence—"the extrusion of permanent things'' was the phrase Russell used [6] —produced a philosophy in essence no different from the sensationism of Mach or the subjective idealism of Berkeley. The "powerful logical technique" was, in fact, nothing but a technique for saying in a new and rather more difficult language what had often been said before.

Thus the renunciation of system building and the path of a Galilean advance in philosophy led, in the hands of the first exponents of the method of logical analysis, to nothing better than a restatement of the old philosophy of subjective idealism. This was duly carried to its logical conclusion in a system of complete solipsism with the publication in 1921 of Wittgenstein's Tractatus Logico‑Philosophicus.

Logical Empiricism and the Program of Analysis of Language

How to preserve the method of logical analysis without being confronted by the paradoxes of subjectivism and solipsism to which Russell's and Wittgenstein's treatment had led? How to formulate an empirical philosophy which would not be bound by the traditional empiricist ruling that the known world must be constructed out of sense data? These were the problems that faced all those philosophers who followed in the footsteps of Russell. The development of the contemporary school known as logical positivism, or logical or logistic empiricism (the examination of which is the business of the present essay) has been conditioned by precisely these problems.

What was the starting point of this development?

A study of Russell's basic works of philosophical analysis shows that his whole method aimed at the production of some theoretical “construction" of the world. Thus, in Our Knowledge of the External World Russell speaks of the task of "discovering what sort of world can be constructed" [7] exclusively out of elements concerning whose existence there can be no possible manner of doubt, i.e., sense‑data. He explains that people ordinarily suppose that the world is constructed out of "permanent things." Yet considerable doubt attaches to the existence of such permanent things. Logical analysis, therefore, will "construct the world" out of elements to whose existence no doubt attaches, and these elements are of the nature of sense‑data.

For Russell, then, logical analysis was a way of revealing the essential nature of the world, i.e., that the world is constructed out of sense‑data, and not out of permanent material things, or spiritual monads or other dubious kinds of entities to which traditional a‑priori philosophy had made reference.

In accordance with this, the analytic philosopher J. T. Wisdom wrote: "The philosopher asks, What is the Self? What is the State? What is Time? . . . The philosopher is asking for a certain kind of definition of the Self, of the State." [8] This definition, provided by the method of logical analysis, will answer the question as to the real ultimate essential nature of things—of the Self, or the State, or Time, as the case may be.

Now it is easy to see that such a conception of the aims of the analytic method can well be criticized from the point of view of a strict empiricism. And it is this criticism which the logical positivists, or logical empiricists, provided. They pointed out that to use a logical‑analytic method for "constructing a world," or discovering the ultimate essential nature of things, was merely to repeat in a new way the old errors of the philosophical systems. They claimed that those who professed to be empiricists and to use a logical method should definitely renounce all attempts at metaphysical construction. Their contention was that, while Russell and his associates had begun the criticism of traditional metaphysics, they had not carried it through with nearly sufficient rigor. A still greater transformation of philosophy was required in order to make it truly empirical, scientific, and free from metaphysics.

An important step in this new onslaught against metaphysics was made in the enunciation of the "principle of verifiability" by Moritz Schlick, the founder of the so‑called Vienna Circle of empirical philosophers. This principle said that "the meaning of a proposition is its method of verification." A proposition for which no method of empirical verification exists is, then, meaningless. A great many of the propositions of traditional philosophy are in this way to be criticized as nonsensical. And it is asserted that no proposition contains any other element of meaning over and above what is given in its method of empirical verification.

What happens when this principle of verifiability is applied to Russell's analysis of the external world in terms of sense‑data? Russell said that the world must be constructed out of sense‑data and not out of permanent things. Yet what is the method of verification of this statement? In actual fact, the empiricists now contended, there is no method of verification for the statement that the world is composed of sense‑data, if this statement is intended to contradict the statement that the world is composed of permanent things. You can say whichever you like, it makes no essential difference. One mode of expression may be better for one purpose, the other for other purposes. They are merely two different forms of words, two alternative languages for expressing the same facts, and not rival statements of the nature of reality. [9]

The conclusion was drawn that philosophers have been misled by language. Russell himself, in formulating the logical‑analytic method, had already remarked upon this. But he was now criticized on the ground that, instead of carrying forward a thorough logical critique of language, he had allowed himself to be diverted into what were essentially metaphysical speculations, whose meaninglessness was revealed in the light of the principle of verifiability. The task of philosophy, however could not be to construct the world, or discover the real nature of things, by extrascientific methods—whether methods of a‑priori system building or methods of logical analysis. The task was represented as a quite different one—namely, "critique of language; for the past methods of formulating the problems of philosophy had always rested "on the misunderstanding of the logic of our language." [10]

The new conception of the tasks and methods of philosophy which emerged from the principle of verifiability was most clearly and systematically stated by Schlick's associate, Rudolph Carnap.

"A philosophical, i.e., a logical, investigation must be an analysis of language," Carnap wrote. [11]

"The questions dealt with in any theoretical field,” he wrote further, “. . . can be roughly divided into object‑questions and logical‑questions . . .  By object‑questions are to be understood those which have to do with the objects of the domain under consideration, such as inquiries regarding their properties and relations. The logical questions, on the other hand, do not refer directly to the objects, but to sentences, terms, theories, and so on, which themselves refer to objects." [12]

Object‑questions are only to be answered by the empirical methods of natural science. They cannot be answered by extrascientific philosophical means. Philosophy, as distinct from science, must be exclusively concerned with logical questions, i.e. with questions about the logic of the language in which we express our statements about objects. Philosophers must not inquire into the nature of objects, must not try to analyze the nature of Reality; they must leave that to science. They must confine their inquiries to the study of the logic of language. And this study will do the job which they really want done—it will enable them to reach agreed solutions of philosophical problems, and in particular it will throw a flood of light upon the problems of the philosophy of science.

The school of logical empiricism, then, which arose out of the logical‑analytic method of Russell and out of the criticism of this method, may be said to be characterized by two guiding principles:

(1) The repudiation of all "metaphysics," i.e. all doctrines about the world as a whole, the ultimate nature of Reality, and so on, not capable of empirical verification; together with the acceptance of the empirical view that all knowledge is derived from experience and must be verified in experience.

(2) The formulation of the task of philosophy as being the logical analysis of language, and specifically the logical analysis of the language of science.

This means that critically the aim is to destroy "metaphysics." Positively, it is to make clear the significance of science; for the analysis of scientific language will show what science is doing, together with the connections of the sciences in one single universal science.

Despite the fact that rather widely varying views on different questions have been put forward in the give and take of discussion between members of the school—by such writers as Carnap, Neurath, Frank, Tarski, Reichenbach, C. W. Morris, and Popper—the above represents the principles of method and programmatic aim which they all have in common.

In the light of these common principles, the logical empiricists, in the period preceding the recent war, set about organizing international congresses of scientific philosophy and launched the project of an International Encyclopedia of Unified Science. Their aim was to establish a scientific activity of philosophizing on an international scale, which would be comparable with the natural sciences and mathematics in the steady building up of accepted and certifiable theoretical results.

Having now briefly examined the historical background of logical empiricism and its general program, I propose in what follows to discuss the actual content of its critical rejection of metaphysics and of its attempt at an analysis of science, and then its general tendency as a philosophy and whether it serves the interests of the advance of scientific knowledge and of social progress.

Metaphysics and the Theory of Knowledge

In his Philosophy and Logical Syntax Carnap says that the function of logical analysis is to analyze all assertions in order to make clear the sense of each and the connections between them. The principal task in this connection is to find out the "method of verification." The assertions of traditional metaphysics are found to admit of no method of verification. Since to be able to find no method of verification is the same as to be able to find no meaning, the metaphysical theories are therefore meaningless, nonsensical. They appear to have a sense only because they arouse emotions. Carnap distinguishes an "expressive" or emotion‑arousing use of words from the "representative" use, and says: "Metaphysical propositions are . . . like laughing, lyrics and music, expressive." [13]

The criticism of metaphysics shows in particular, Carnap continues, that all general assertions about the "Reality" of so‑and‑so are metaphysical and senseless. For example: Is Matter real? Is Mind real? Is there really an External World? All such questions, and the theories purporting to answer them, are senseless.

All this makes it very clear that in Carnap's opinion the traditional problems of philosophy must be ruled out as senseless and as "pseudo‑problems." Natural science, on the other hand, being systematic verifiable theory, alone will stand the test.

But with regard to science, philosophy must not ask such questions as: What kind of objects does science deal with? Does science refer to objective external realities, or is it on the contrary a technique for predicting the order of sensations? For, in Carnap's words, "the questions of the kinds of facts and objects referred to by the various languages [of science] are revealed as pseudo‑questions." [14] Philosophy deals with the language of science—with its syntactical and semantical rules—and by so doing may hope to clear up the kind of difficulties that receive expression in such "pseudo‑questions."

By such arguments as these it is definitely claimed that empiricism has rid itself of the taint of subjectivism. For the view that "only sense‑data exist," that the world is to be "constructed out of sense‑data," and so on, is rejected as "metaphysical. But the view that "external material things exist independent of their being perceived—the “realist" view, which is a part of any materialist philosophy—is like­wise rejected for precisely the same reason. In this connection, Carnap writes:

"Sometimes the views of the Vienna Circle have been mistaken for a denial of the Reality of the physical world. But we make no such denial. It is true we reject the thesis of the Reality of the physical world: but we do not reject it as false but as having no sense, and its idealistic antithesis is subject to exactly the same rejection. We neither assert nor deny these theses, we reject the whole question." [15]

And again: "We make no assertions as to whether the given is real and the physical world appearance, or vice versa; for logical analysis shows that such assertions belong to the class of unverifiable pseudo‑statements." [16]

Is it the case that by this downright "rejection" of all "metaphysics" the method of logical analysis has really succeeded in separating itself from subjectivist trends of philosophy? The view which asserts the existence of the material world independently of its being perceived or thought about, which is generally counterposed to subjective idealism, is not denied, but is rejected. Subjective idealism, too, is rejected, but not denied. The position is a delicate one for the logicians who have thus distinguished rejection from denial. Just what it is that they claim to be rejecting, and where such rejection leads, is worth inquiring into.

"The material world exists"—"The material world does not exist." "The material world contains processes which are reflected in various ways in our perceptions and thought but exist independently of such reflection"—"What we call material things are nothing but complexes of sense‑data." Here we have statements which have often been debated by philosophers and continue to be debated, but the controversy between which, we are now told, is without meaning. What is at stake in this controversy? What is at stake is our whole point of view with regard to human knowledge.

Both the upholders and the detractors of the existence of the material world, together with those who say the question is nonsense, may be at one in recognizing that all knowledge is based on experience. This recognition is the common denominator of a wide variety of views—materialists, subjective idealists, and logical empiricists, all share it. The differences arise over the question of the source of our experience itself.

The materialists who affirm the existence of the material world are saying that our perceptions have their source in the interaction between ourselves and our material environment; that this interaction is such that our perceptions reflect or mirror the external reality well enough to serve generally as a trustworthy guide in practice; and that, by the use of scientific methods and logical modes of thinking, our thoughts and theories can likewise be made fairly adequately to reflect various aspects of the material world in which we live.

On the other hand, those who deny this are saying that our sense‑experience is ultimate, absolute, an ultimate given datum, into the origin and significance of which there is no sure means of inquiring, but which can at most be a subject of speculation. Thus Berkeley, denying the source of sensation in the external material world, said it must be called forth in our minds by the will of God; Wittgenstein, writing two hundred years after Berkeley, did not mention God (be­cause he thought that would be "nonsense"), but spoke instead about "the mystical." True, these same philosophers recognize the right of science to speak about the way our sensory apparatus is stimulated by impulses originating from the external environment. But, according to them, these very statements of science are themselves to be interpreted as having no other subject matter than sense‑data. For Mach, for instance, the brain, like everything else, was nothing but  "a complex" of sense‑data, so that to say to say that sensations are aroused by impulses affecting the sensory centers of the brain was only to establish a relationship between two complexes of sense‑data, and was not to establish the source of experience itself.

These were the kind of considerations which Lenin evidently had in mind when, explaining and defending materialism, he wrote: "All knowledge comes from experience, from sensation, from perception. That is true. But the question arises, does objective reality belong to perception, i.e., is it the source of perception? If you answer yes, you are a materialist. If you answer no . . . you deny the objective content of experience, the objective truth of your experimental knowledge." [17] "Materialism and idealism differ in their respective answers to the question of the source of our knowledge and of the relation of knowledge . . . to the physical world." [18]

Whether in this connection you "reject" the source of our experience in the material world or whether you "deny" it—whether you deny that our experience has a material basis or reject it as meaningless­—does not make a great deal of difference. It is possible to argue about "the meaning of meaning." But at all events the question of the reality of the material world as the source of experience has had enough mean­ing to divide philosophers into two rival camps according to the way they have answered it, and it still has that meaning. When logical empiricists "reject the thesis of the Reality of the physical world" they thereby reject the objective reference of science, the objectivity of knowledge, the material basis of experience. In this way they place themselves in the same camp as the older empiricists, the subjective idealists, although they protest that they reject subjective idealism as well.

This fact is exemplified in the argument developed by Carnap and others to the effect that it is possible to use either "a physical object language" or "a sense‑datum language" at will: the difference between saying that a thing is "a complex of atoms" and saying that it is "a complex of sense‑data" is merely a difference in choice of language. If that is so, then the possibility of ever being able to express the dependence of our "sense‑data" on the circumstances of our material existence is a‑priori ruled out, and every essential point has been conceded to the side of subjective idealism.

Materialists—unlike some of the representatives of what is often called philosophical "realism"—justify their "thesis of the reality of the physical world" and of the material basis of experience, not by general "metaphysical" arguments a priori, but by appealing to experience itself. The materialist considers that this "thesis" is "verified," not indeed by any particular crucial experiment, but by the whole development of science and technique. And the "thesis" has its basis in the entire structure of a materialist theory of knowledge, which regards knowledge as a social product whose foundations lie in the interactions of men with one another and with their material environment.

From this point of view the essential weakness of the "antimetaphysical" theory of logical empiricism is to be found precisely in its theory of knowledge—or, rather, in its lack of a theory of knowledge. A philosophical formulation of the standpoint of science, embodying the justification of a scientific view of the world and the rejection of a‑priori philosophical theorizing, must have as its basis scientific theory of knowledge. No system of formal logic or abstract "principle of verifiability" will serve as a substitute. The very principles of the "logical analysis of language" remain suspended in the air so long as they have no basis in a concrete examination of the actual material social foundations of the thought process which is expressed in language. The "principle of verifiability" itself remains a mere phrase so long as the principles of the actual concrete process of building up knowledge and verifying it are ignored.

Thus the rejection of metaphysics, announced by logical empiricism, is undertaken on a most inadequate basis. This school of philosophy claims to be able to distinguish those propositions which have sense from those which have not, and to single out and reject the propositions of  metaphysics as belonging to the latter category. But its apparatus for carrying out this task is defective. It does not proceed from any scientific analysis of the actual process of human knowledge. And in the absence of such an analysis, the upshot of its "antimetaphysical" program is to reject the objectivity of scientific knowledge.

In this respect logical empiricism remains in line with the whole development of empirical philosophy since Berkeley. But Berkeley’s open rejection of knowledge of the material world, and the "construction" by Mach or Russell of an alternative "world" out of sense‑data, are now replaced by the "logical" explanation that the whole question of the sources and objects of knowledge is to be rejected.

But you neither solve nor get rid of a crucial question by "rejecting" it.

The Logical Analysis of Science

I said above that logical empiricism had a double aim—critically to destroy metaphysics, positively to make clear the significance of science. How does it proceed with the positive aim?

In the first place, we may consider the results of a simple application of Schlick's principle of verifiability. According to this, the sense or meaning of all propositions of science is given by the "method of verification." The verification of propositions of astronomy, for instance, is given by looking through telescopes, so that the meaning of these propositions is to be elucidated in terms of the sense‑data of astronomers. Again, the verification of propositions of atomic physics is given by observations off electrical apparatus, so that the meaning of these propositions is to be elucidated in terms of what we see when we operate such apparatus, i.e., in terms of such sense‑data as flashes on screens, and pointer readings.

In other words, the scientific theories of astronomy and physics, and similarly all other scientific theories, are to be understood as short‑hand expressions of what we may expect to observe, or what sense‑data we may expect to experience, under various specifiable conditions. That exhausts their meaning. To say, for instance, that the theories of astronomy reflect the stellar universe existing independent of all experience, or that the theories of atomic physics reflect the physical world existing indep endently of all experience, is to be regarded as senseless metaphysics.

It is not difficult to see that this subjectivist interpretation of science is far from satisfactory. There are parts of science, for instance, which deal with the past history of the earth, with geological periods before the appearance of man and before the existence of any experience. What is the meaning of these particular propositions of science? According to the present interpretation they must be understood simply as predictions of future sense‑data—and yet we can recognize that this cannot exhaust the meaning of such propositions. Another difficulty arises in connection with psychological propositions about other people's mental states. Yet another difficulty arises in connection with psychophysical propositions about the dependence of mental upon physiological phenomena.

These difficulties need not be elaborated here, because they are obvious and have often been mentioned by contemporary philosophers of all schools. The logical empiricists themselves soon became aware of such difficulties, and therefore tried to revise the purely subjectivist interpretation of science which was the first result of the application of the principle of verifiability.

In his books Philosophy and Logical Syntax, The Unity of Science, and The Logical Syntax of Language, Carnap developed the view that previous empiricist studies in the analysis of science had been vitiated by the attempt to explain the meaning of scientific propositions. This led to a subjectivist metaphysics which he had to confess was just as bad, in its way, as a "realist" metaphysics.

But on the contrary, he explained, "the logic of science is nothing else than the logical syntax of the language of science." [19] And: "By the logical syntax of a language, we mean the formal theory of the linguistic rules of that language—the systematic statement of the formal rules which govern it, together with the development of the consequences which follow from these rules." [20]

Thus philosophy will not inquire what science means, what scientific propositions are about, since in the light of antimetaphysical criticism "the questions of the kinds of facts and objects referred to by the various languages [of science] are revealed as pseudo‑questions." In brief: "The philosophy of a science is the syntactical analysis of the language of that science." [21] By sticking, then, strictly 'to `syntactical analysis" the difficulties of subjectivist interpretation are to be avoided. Such difficulties just do not arise. It will soon be found, however, that other difficulties arise instead.

Proceeding to the "logical syntax of the language of science," Carnap wrote: "Science is a system of statements based on direct experience and controlled by experimental verification. . . . Verification is based on protocol statements." Protocol statements are "statements needing no justification and serving as the foundations for all the remaining statements of science. [22]

Instead, therefore, of  trying to interpret the meaning of scientific statements, the task of the logical analysis of science is now to be regarded as being to show how science, the whole system of scientific statements, is derived from protocol statements according to certain formal rules. The protocol statements serve the function of reports on the basic observational data of science and on the observed results of experiments designed to verify scientific theories.

Where, then, does the syntactical analysis lead? It will be found to lead to results no less paradoxical than those of 'the previous subjectivist analysis. These results were summed up by Neurath as follows:

"Sentences are to be compared with sentences, not with 'experiences,' not with a 'world,' nor with anything else. All these senseless duplications belong to a more or less refined metaphysics, and are therefore to be rejected. Every new sentence is confronted with the totality of sentences which are present and which have been brought into agreement. Then a sentence is called correct if it can be brought into the system. Whatever we cannot systematise is rejected as incorrect. Instead of rejecting the new sentences we can also, wherever we find it generally difficult to make a decision, alter the whole system of sentences until the new sentence can be included . . . . In the present theory we always remain within the realm of speech‑thinking." [23]

Neurath here correctly pointed out that, if you adhere to a strictly syntactical analysis, then the only account that can be given of science is that it is a changeable system of sentences which have to be continually "brought into agreement," and the correctness of any scientific theory depends on how it can be fitted into the existing system of scientific sentences, and on nothing else. Even with regard to the basic "protocol," it cannot be said that the protocol is certified because it expresses the ultimate experiential data on which the rest of science is based: this, too, would be a non‑syntactical statement, another “senseless duplication" belonging to "a more or less refined metaphysics.'' 

The analysis began by saying that science is "based on direct experience" and "controlled by experimental verification." But no account of such basis and control can be given in terms of a strictly syntactical analysis—these are non‑syntactical, non‑formal concepts.

Thus if the subjectivist analysis based on a simple application of the principle of verifiability led to a severing of the connection between science and the objective world by making science deal purely with predictions of future sense‑data, the new syntactical analysis severs the connection even more completely by ruling out any sort of consideration of what science deals with and making it appear simply as a system of sentences, the correctness of each of which depends solely. on whether it can be "brought into the system."

But such an analysis can give no account at all of what science is doing, why it is doing it, what is its value, or why we should prefer it to mysticism or metaphysics. In other words, it is altogether wide of the mark as an analysis of science. Its chief feature is that it continues to "reject" the idea that science reflects the real world and gives us objective truth.

This syntactical analysis is not, however, the last word of logical empiricism. As this type of philosophy is led from one impasse into another it continually produces some new device wherewith to try to extricate itself.

In the first volume of his new Studies in Semantics Carnap announces that "the field of theoretical philosophy is no longer restricted to syntax but is regarded as comprehending the whole analysis of language, including syntax and semantics and perhaps also prag­matics.” [24]

This reflects the widening out of the logical study of language to include not only logical syntax, but also "semantics" and "pragmatics," already announced in the second number of the International Encyclopedia of Unified Science. [25] According to this improved analysis, the functioning of a linguistic sign—a particular case of what is called "semiosis," or the functioning of signs in general—involves the sign in a threefold relationship; namely, to other signs in the same language, to what it designates, and to the people who use and interpret the sign. Logical syntax studies signs in relation to other signs; semantics studies signs in relation to what they designate; pragmatics studies signs in relation to the people using and interpreting them. The whole logical study of language, which is a part of "semiotic," includes all three special studies. The logical analysis of science, then, is incomplete if it confines itself to the syntactical aspects of science: it must include the semantic aspect, "and perhaps also pragmatics."

Thus logical empiricism starts the analysis all over again, inspired with better hopes of success.

The essential feature of logical syntax is that it studies the formal rules employed in the construction of a given set of statements, i.e., the rules of formation, which determine which statements have sense or are permitted within the system, and the rules of transformation, which determine which statements are formally compatible with which other statements, which are formally incompatible, and how statements can be formally deduced one from another. Semantics, on the other hand, is concerned with the rules determining the truth or falsity of statements, i.e., with the determination of truth conditions for statements.

If we consider any system of statements, such as a science, then the assignment of semantic rules for determining the conditions of the truth and falsity of statements belonging to that system will constitute an "interpretation" of the system. The logical analysis of science is, then, concerned, not only with the syntax of the language of science, but also with the semantic rules employed, i.e., with the mode of determining the truth or falsity of statements belonging to the science, or with the interpretation of the science.

In his contribution to the International Encyclopedia, Carnap indicates the way in which a more complete logical analysis, which includes semantics, can apply to the science of physics. He distinguishes between the "elementary" and the "more abstract" terms employed in physics.

The elementary terms are those which refer to what is directly observed—such as the pointer readings, flashes on screens, tracks in Wilson chambers, and so on. He points out that "singular sentences with elementary terms can be directly tested." In other words, the system of the science includes semantic rules for determining the truth or falsity of all such sentences. The sentence, "The pointer‑reading is n," is true if and only if the pointer gives the reading n.

On the other hand, the theorems of the science expressed in the "more abstract" terms—for example, about light waves, and electrons—can only be indirectly tested. Such indirect test consists in deriving from them singular sentences with elementary terms in accordance with the syntactical rules of the science, and then utilizing the semantic rules to find out whether such singular sentences are in fact true. For example, we test whether an electron has been emitted by observ­ing the track in the Wilson chamber; we test whether an alpha‑particle has been emitted by observing the flash on a screen, and so on.

Therefore, Carnap concludes, in the science of physics we need give no interpretation of any of the more abstract terms or theorems. It is enough if the science includes syntactical rules for connecting the more abstract theorems with elementary singular statements, and only the latter need be interpreted. [26]

Much the same conclusion is presented in Hans Reichenbach's recent Philosophical Foundations of Quantum Mechanics. There he distinguishes between "phenomena" and "interphenomena"—a distinction which amounts to the same thing as Carnap's elementary and more abstract terms. Statements about phenomena may be determined as true or false, but statements about interphenomena will have the truth‑value of "indeterminate." In other words, we need rules for connecting the two sets of statements in the system of the science; but an interpretation as true or false need only be given to the statements about phenomena. Quantum mechanics includes rules for connecting statements about electrons with statements about tracks in Wilson chambers and lines on photographic plates; but the former need be regarded as neither true nor false, truth is to be ascribed only to the latter types of statements.

Here, then, we have an analysis of science of an apparently most elaborate kind, presented with a wealth of logical terminology. What does it all amount to? It is interesting to compare it with the much simpler account of the work of science—or, at least, of physical science—which was given in 1927 in the Gifford Lectures of the late Sir Arthur Eddington, and which earned some notoriety at the time.

Eddington said: "The whole subject‑matter of exact science consists of pointer readings and similar indications." [27] As to what the pointers signified, he went on, this was "inscrutable." What difference is there between this simple statement and the logical theory that truth or falsity can only be assigned to "singular sentences with elementary terms" and that the "more abstract" statements of science can be given no interpretation? There is no difference.

Eddington also said: "Something unknown is doing we don't know what—that is what our theory amounts to.” [28] What difference is there between this simple statement and the logical theory that all statements about "interphenomena" have the truth‑value indeterminate? Once again, there is no difference.

Thus having fled from a subjectivist analysis of science into the realms of pure logical syntax, and having then sought the aid of semantics, the logical empiricists find themselves to all intents and purposes again saddled with a subjectivist analysis. They are once again interpreting science in the classical empiricist way in terms of predic­tions of observations. They are once again saying that scientific theories are to be understood as shorthand expressions for what we may expect to observe, or what sense‑data we may expect to experience, under various specifiable conditions. This semantic account of the objects the theories refer to, plus the syntactical account of the formal connection between the elementary and more abstract statements of science, plus the pragmatic account of how scientific statements arouse definite expectations on the part of the people who interpret them, exhausts the meaning of scientific theory.

Limitations of an Analysis of Language

In what does the inadequacy of this whole analysis of science consist? It consists in its inability to show the source of scientific knowledge as knowledge of the objective world, to show the grounds for confidence in science as knowledge and as guide to action, or to show how we can best organize scientific research in order that it may serve the ends of human progress.

Whence does this inadequacy arise?

It arises from the narrow preoccupation of the logical empiricists with questions of the logical analysis of language.

According to C. W. Morris in the International Encyclopedia of Unified Science, the study of science falls entirely under the study of the language of science, because this includes not only the formal or syntactical structure of scientific theory, but also its relations to the objects designated (semantics) and to the people using it (pragmatics). [29] But in fact, far from the study of the language of science comprehending the whole study of science, it leaves out the most essential features, which nevertheless have to be reckoned with if the work of science is to be properly appreciated.

For an all‑round, rather than an abstract and one‑sided, understanding of science it is necessary to take into account such facts about science as the following:

Science is a social activity; and the theories of science, set forth in the statements in scientific language, are products of this social activity.

The history of science shows that the problems tackled by science are ultimately connected—though the connection may be far from direct or simple—with the problems set by the development of social production and production technique, and by social organization.

In the case of the physical, and no less the biological, sciences, there is a close relationship between theory and technology. Theories are devised in the light of an experimental technique. Theories about the constitution and laws of various physical or biological systems are founded on the use of a technique for effecting controlled changes in those systems and for recording and measuring the changes produced. In this way the more science finds out about the constitution and laws of motion of material systems, the better the technique that can be devised for finding out more; and the better the technique, the further can scientific theory advance. Advance in theory leads to advance in technology, advance in technology leads to advance in theory, and neither can advance without the other.

It is just these essential considerations about science that are neglected and glossed over in the abstract logical study of the language of science. That such is the case is most glaringly shown in the further contribution of Leonard Bloomfield to the International Encyclopedia.

There an analysis is made of what is called a typical "act of science." Such an "act" is said to consist of the following steps: observation; report on observations; statement of hypotheses; calculation; prediction; testing by further observations. It is then pointed out that all but the first and last steps are "acts of speech.” [30]

In the light of such an analysis of the work of science, it is very natural to suppose that we should simply accept the initial and end observations as "given"—as hard facts, so to speak—and that the "study of science" should fall, as Morris suggested, entirely under the study of the language of science. And it is also natural to come to the positivistic conclusion that the objects designated by the language of science consist of the observational data‑pointer readings and the like in the case of the physical sciences; and that the pragmatic function of science is to organize expectations of the occurrence of such observational data.

The whole inadequacy of the analysis is contained in the lack of analysis of what is involved in "observation."

If we consider, for example, what is involved in the observation of a pointer reading, then it becomes very clear that the "act of science" is far from beginning with the observation. For in order to get a pointer reading (or a flash on a screen, a black line on a photographic plate, or a track in a Wilson chamber) there is first necessary the technical process of devising and constructing scientific instruments. This is an act of social technique, which consists in the production of a physical system whereby the scientist‑technician will be able to control the occurrence of certain processes and record or measure them.

The basic observational data for scientific theory are not "given" but produced.

The observation and the record of the observation, which serve as starting point for a further development of scientific theory, are themselves produced by the application of a technique; and this technique has its foundation in already established theoretical knowledge, in the light of which it seeks to win further knowledge of and control over physical processes.

The real "act of science" has its basis, then, not in a mere "observation," but in an activity of social technique. The subject matter of scientific theory is by no means the observations—"pointer readings and similar indications," as Eddington said—but the objective material processes to which the technique relates, and which are registered, recorded, measured by means of the observations obtained by the use of the technique. Similarly the theories of science are tested in the further application of technique, and in the success or otherwise of new techniques invented in the light of the theories.

The observations of science are obtained in such a way that they throw light on the actual constitution and laws of physical systems. And scientific theories, if properly understood and used, serve the ends of increasing our all‑round understanding of ourselves and the universe, of increasing our power to use natural processes for our own ends and our ability to organize our own social affairs. In this way, it may be added, the test of scientific theories is by no means confined to a laboratory test, but is effected in the whole application of science in social life.

The analysis of science, then, ought to show it as arising out of social life, as both engendered by and engendering social techniques. Such an analysis of science is materialistic. Traditional empiricism regarded science simply as a theory or set of theories whose aim was to sum up the regularities discerned in past observations and predict regularities in future observations. Logical empiricism only restates this in terms of "an analysis of scientific language." Materialism, on the other hand, sees science as a social activity arising in the course of the development of production and of production technique. It sees the observational data of science as obtained in the course of men's struggle to understand and master natural and social forces. It sees scientific theory, therefore, not as a summary and prediction of observations, but as the best account which we are able to give of the world we live in and of our own place in it. [31]

It is clear enough that all scientific results are stated in language, and that, as scientific activity becomes specialized, so a specialized use of language is developed in the statement of scientific theory. The logical empiricists are undoubtedly right in stressing the importance of the logical study of this language. It becomes important not only for the philosophical understanding of science and its significance for us today, but also for the development of science itself, since theoretical difficulties arise which can be shown to be associated precisely with the use of the instrument of language.

But the study of the language of science does not include the whole study of science. On the contrary, the study of language ought itself to proceed on the basis of a study of the social technique of scientific knowledge, which uses language as an instrument.

To propose a merely abstract study of language is to rule out in advance the possibility of showing how the scientific statements whose language it is proposed to analyze express objective knowledge, or are anything more than a means of referring to observational data and organizing expectations of future observational data. And such is, indeed, the outcome of logical empiricism.

Those who wish to preserve and develop what is positive and useful in the logical study of language, but to overcome the shackles of the traditional empiricist standpoint, must seek to do so on the foundations of a materialist theory of knowledge. On the solid basis of the materialist theory of knowledge, achieved in dialectical materialism but still awaiting a fuller development, philosophers may hope to show the sources of our knowledge of the objective world. And when they study the language of science from this point of view, they will study it as the means of expressing such knowledge.


It remains to summarize the principal conclusions that emerge from the criticism of logical empiricist philosophy from the materialist point of view.

Logical empiricism has the central theoretical weakness that it seeks to criticize the traditional philosophical systems, and to analyze and assess the meaning of science, with the help of nothing but an abstract logical theory of language, and ignoring the necessity of a scientific study of the actual process of knowledge. For this reason it is led continually to misrepresent the nature of our knowledge as knowledge of the objective world.

In essentials logical empiricism is nothing but a continuation of the tradition of the subjectivist theories of the older empiricism, which it restates in new forms and with a new terminology. In relation to science, the analysis of which is the central aim of logical empiricism, this tradition consists in ignoring the fact that science enlarges our knowledge of nature and of humanity and our practical control over natural and social forces; instead it represents science as a set of theories summarizing and ordering what is given in observation.

That this whole mode of "analysis" of science is a misrepresentation is evident if a concrete and historical consideration of the actual process of building scientific knowledge is substituted for the abstract preoccupation with the analysis of language. But it is also worth considering some of the broader consequences of this misrepresentation, because this helps to place logical empiricism in its correct perspective in the picture of contemporary ideologies.

The consideration of the character of science as knowledge is connected with the consideration of the part which the development of science plays or can play in modern society. The application and planned development of science can mean abundant food, shelter, health, rest, culture, and happiness for every human being. This demands a social reorganization, which in turn depends on a scientific understanding of the laws of social development and of the conflicting forces operating in contemporary society.

A philosophy which correctly recognizes how the development of modern science is a development of our knowledge of nature and mankind cannot but recognize and stress the above facts, and such a philosophy is an indispensable possession of those who are taking part in the struggle for progress. On the other hand, the misrepresentation of the character of science which is made by logical empiricism is such as to obscure the whole social function of science and to confuse the issues of the fight to realize the progressive potentialities of scientific knowledge. In relation to the task of extending scientific knowledge and securing its development and utilization in the service of human progress, the program of "logical analysis" bears a barren and scholastic character. In this respect it is certainly no "progressive philosophy"—a title which is often claimed for it. Quite the reverse. By rejecting the objectivity of scientific knowledge it obscures the significance of science as a weapon of enlightenment and progress.

he fact should always be borne in mind in discussing philosophy that, though the philosophers may lecture in universities and engage there in learned discussions which are quite incomprehensible to the majority of mankind, nevertheless these universities are factories of ideas which go forth into the world and play a part quite independently of the intentions of their creators. The ideas propagated in the discussions and writings of logical empiricists certainly do not help people to form a scientific picture of the world and man's place and prospects in it, or to understand how science can be utilized in the service of the common man. But this very scholastic negativity, this very failure to present a scientific philosophy comprehensible to the common man, means that the way is left open for the deception of the people by supernatural, idealistic, and antiscientific illusions. If the scientific philosophers put it out that science is concerned with nothing but pointer readings, the common man will agree with Eddington that in that case science does not tell us very much. It is in vain for philosophers to pretend that they are energetically combating idealism and irrationalism and supernaturalism as meaningless metaphysics, when they have nothing to put in their place, and when they just as energetically combat the scientific materialism which is the only practical alternative.

C. W. Morris in his latest work, in which he claims "to lay the basis for a comprehensive and fruitful science of signs," [32] succeeds in carrying linguistic, or rather "semiotic" studies to a new point, where the door is deliberately opened wide for any form of extrascientific or antiscientific ideology. The new "science" reduces the significance of all signs, from the buzzer which summons experimental dogs to their dinner to the utterances of scientists and philosophers, to their function as organizers of behavior, and the whole process of knowledge consists in building up effective sign systems. A place is readily found for religion as "prescriptive‑incitive discourse" which "lays down the pattern of behaviour which is to be made dominant in the total orientation of the personality." As for philosophy, it is "discourse dominated by the systemic use of signs in its greatest comprehensiveness." [33] The worst of it is that C. W. Morris takes no account of the fact that powerful interests antagonistic to progress are in the habit of loudly engaging in "prescriptive‑incitive discourse" and "systemic use of signs" for their own ends and in order to deceive and hoodwink people. To expose such activities, to show up their deceptiveness and falsity, and to arm the common people with a scientific view of the world, is an essential task of a progressive philosophy. Such a task, the philosophy we have been considering is utterly unable to fulfil.

Logical empiricism set out to abolish "metaphysics" and to establish the foundations of a new type of scientific philosophy. Yet it succeeds only in continuing in new forms the traditional empiricist philosophy which rejected the material basis of experience and the objective reference of scientific knowledge. And the very latest outcome of the new "science of signs" is to reinstate speculative system‑building—not to mention religion—as a "systemic use of signs."

Logical empiricism has failed to make a scientific approach to science itself. We would not deny the positive value of some of the specialized work which has been done in connection with the study of language and signs, nor the technical advance which modern formal and mathematical logic has made over the traditional logic. Here there is no doubt a basis for fruitful discussion between materialists and empiricists. But in estimating the philosophy as a whole, materialists can come to no other conclusion than that it misrepresents the whole character of scientific knowledge.


1 Russell, Our Knowledge of the External World (1914), p. 4. [—> main text]

2 Ibid., p. 5. [—> main text]

3 Russell, A History of Western European Philosophy (1945), p. 862. [—> main text]

4 Engels, Anti‑Dühring, p. 32. [—> main text]

5 Russell, Our Knowledge of the External World, p. 81. [—> main text]

6 Ibid., p. 107. [—> main text]

7 Ibid., p. 72. [—> main text]

8 J. T. Wisdom, "Ostentation," in Psyche, Vol. XIII. [—> main text]

9 This is argued at length by A. J. Ayer, Foundations of Empirical Knowledge (1940), chap. 1. [—> main text]

10 Wittgenstein, Tractatus Logico‑Philosophicus, Preface. [—> main text]

11 Carnap, The Unity of Science (London, 1934). [—> main text]

12  Carnap, The Logical Syntax of Language (1937), p. 277. [—> main text]

13 Carnap, Philosophy and Logical Syntax (1935), p. 29. [—> main text]

14 Carnap, The Unity of Science. [—> main text]

15 Carnap, Philosophy and Logical Syntax, p. 20. [—> main text]

16 Carnap, The Unity of Science. [—> main text]

17 Lenin, Selected Works, Vol. XI, p. 19. [—> main text]

18 Ibid., p. 316. [—> main text]

19 Carnap, The Logical Syntax of Language, p. xiii. [—> main text]

20 Ibid., p. 7. [—> main text]

21 Carnap, Philosophy and Logical Syntax, p. 88. [—> main text]

22 Carnap, The Unity of Science. [—> main text]

23 Neurath, "Sociology in Physicalism," Erkenntnis, Vol. II, p. 603. [—> main text]

24 Carnap, Studies in Semantics, Vol. I, "Introduction to Semantics" (1942), § 39. [—> main text]

25 C. W. Morris, Foundations of the Theory of Signs. [—> main text]

26 Carnap, Foundations of Logic and Mathematics, §§24, 25. [—> main text]

27 Eddington, The Nature of the Physical World, p. 252. [—> main text]

28 Ibid., p. 291. [—> main text]

29 C. W. Morris, Foundations of the Theory of Signs, §1. [—> main text]

30 Bloomfield, Linguistic Aspects of Science. [—> main text]

31 A very interesting approximation towards a materialist standpoint is made in Hans Reichenbach's Experience and Prediction, in the views he there develops concerning probability. Unfortunately there is not space here to discuss Reichenbach's analysis. [—> main text]

32 Morris, Signs, Language and Behavior (1946), p.  v. [—> main text]

33 Ibid., pp. 146, 234. [—> main text]

SOURCE: Cornforth, Maurice. "Logical Empiricism," in: Philosophy for the Future: The Quest of Modern Materialism, edited by Roy Wood Sellars, V.J. McGill, Marvin Farber (New York: The Macmillan Company, 1949), pp. 495-521.

Science versus Idealism by Maurice Cornforth

Maurice Cornforth on William Blake vs. the Fetishism of Language

Materialism and the Dialectical Method by Maurice Cornforth

Marxism and the Linguistic Philosophy by Maurice Cornforth

Science and Evaluation by Maurice Cornforth

Partisanship and Objectivity in Theoretical Work by Maurice Cornforth

Maurice Cornforth on Partisanship and Objectivity by Ralph Dumain

Vienna Circle, Karl Popper, Frankfurt School, Marxism, McCarthyism & American Philosophy: Selected Bibliography

Positivism vs Life Philosophy (Lebensphilosophie) Study Guide

Marx and Marxism Web Guide

Offsite links:

Partisanship and Objectivity in Theoretical Work by Maurice Cornforth
(also on The Autodidact Project web site

"Maurice Cornforth's Contribution to Marxist Metaethics" by Renzo Llorente

On Reappraising Maurice Cornforth” by Edwin A. Roberts

Origins of the Private Language Argument by Jan Dejnozka

Maurice Cornforth - Wikipedia, the free encyclopedia (minimal information)

Home Page | Site Map | What's New | Coming Attractions | Book News
Bibliography | Mini-Bibliographies | Study Guides
My Writings
| Other Authors' Texts | Philosophical Quotations
Images & Sounds | External Links

CONTACT Ralph Dumain

Uploaded 3 December 2005
Links added 30 Dec 2008 & 4 July 2009

Site ©1999-2009 Ralph Dumain