Introduction

Western Philosophy in the Past Century, Analytic vs. Continental:

For nearly a century, philosophy in the Western world has divided itself rather sharply into two hostile camps.  Unfortunately, it is not as easy to characterize the difference between the modern camps as it was to distinguish the rationalists from the empiricists.  It is very difficult to specify any difference of doctrine between the two camps, as both are quite diverse in that respect, and it is hard to find any doctrine which doesn’t have defenders on both sides.  There are stylistic differences; it probably wouldn’t be that hard to program a computer to distinguish which side a given work came from on the basis of word-counts of particular jargon terms common within one group or the other.  But the differences in jargon do not, for the most part, seem to connect to any deeper differences.  And yet, despite the shallowness of most of the differences (a cynic might wonder if it was because of them), understanding and communication between the two sides is in general far less than seemed to exist between the classical rationalists and empiricists.

As the name suggests, continental philosophy is particularly associated with the European continent, especially France and Germany.  For a long time analytic philosophy has been concentrated in English-speaking countries, and so it is sometimes also called Anglo-American philosophy (continental philosophers are far more likely to call it this than are analytic philosophers).  But while the geographical divide may contribute to the persistence of the disagreement, it doesn’t explain what the disagreement is.  Further, the geographical divide is far from total.  There are philosophy departments in the English-speaking world which are primarily continental in orientation, and philosophy departments in continental Europe which have analytic leanings.  More strikingly, while analytic philosophers have largely ignored continental philosophy, continental philosophers have had considerable influence in other academic fields in the English-speaking world, particularly literary fields.

While I can’t explain the current division very well, as it seems to have stopped making sense (an observation which has been made frequently, for some time, by people on both sides; this observation has done little to repair the division), I do think I can say a lot about how it originated.  I’ll start quite far back, and contrast the modern division with the division we’ve discussed in this course.

Classical empiricism was most popular in Britain, while the classical rationalists were concentrated on the continent, but there is no smooth evolution from classical empiricists and classical rationalists to the current state of things (not to mention that while many analytic philosophers would admit to empiricist sympathies, most continental philosophers would vigorously deny having much in common with rationalists).  Rather, the correlations with geography seemed to grow weaker over time.  Especially by the late 19th century; various forms of idealism with clear rationalist ancestry were becoming trendy in Britain, alongside a continuing tradition of empiricism, and in continental Europe empiricist tendencies (usually calling themselves “positivist”) were increasingly prominent rivals to the idealist views there.

The late 19th century also saw one of the crucial developments in the history of analytic philosophy, which corresponds to one of the few substantial differences between the analytic and continental schools (and also between modern analytic philosophy and traditional empiricism).  The German philosopher Gottlob Frege introduced a new system of symbolic logic, one which represented an extraordinary improvement over any previous system of logic.  He toiled in obscurity for decades (late 19th century Germany seems to stand out for failing to recognize philosophical genius, considering how little recognition Nietzsche also received in his own time), but in the early 20th century, his ideas finally began to get some attention.  Bertrand Russell in England discovered Frege’s ideas, and presented the new symbolic logic to English-speaking philosophers; the Austrian philosopher Ludwig Wittgenstein ironically encountered these ideas through Russell and would then bring them to a much larger audience in Austria and Germany.  But Frege had one direct influence of note as well; Rudolf Carnap took Frege's courses in logic and the philosophy of mathematics and philosophy of language.

Simultaneously, developments in the foundations of physics at the beginning of the 20th century produced strong echoes in philosophy.  Since the time of Newton, physics had been taken as the paradigm of settled knowledge; Kant thought that geometry had to be synthetic a priori because it revealed the shape of Euclidean space.  The fall of classical physics led to a tremendous amount of soul-searching (primarily among the Kantians, but that would include a large segment of German philosophers).  It also seems to have provoked not a little soul-searching among some of the scientists.

These two trends came together in the 1920s, alongside other influences, to produce the Logical Positivist movement, centered around the group now known as the Vienna Circle, which consisted primarily of philosophically-inclined scientists.  The leader of the Vienna Circle, Moritz Schlick, earned his Ph.D. in physics, studying under Max Planck.  Rudolf Carnap, considered by many the most talented of the group, was close to a Ph.D. in physics himself when WWI broke out, though when he returned to his studies after the war he wrote a dissertation on the philosophy of space and time, influenced by both Einstein’s physics and Kant’s ideas.  The circle also included mathematicians like Hans Hahn and social scientists like the economist and sociologist Otto Neurath, as well as a few with more orthodox philosophy backgrounds.

Even among the members of the Vienna Circle, there was considerable diversity of views, though the situation isn’t nearly as bad as trying to describe what analytic philosophy involves these days.  The members of the Vienna Circle were united in combining a strong sympathy toward traditional empiricist ideas, a deep interest in modern science, and, most distinctively, a belief that the methods of the new logic could be helpful in dealing with philosophical problems.

The Logical Positivist philosophy was connected to the more general Modernist movement, and while the Logical Positivists would have denied any direct connection between science and ethics, they did, as a group, incline toward pacifism and democratic and socialist ideals, and tended to think that there was some kind of harmony between those ideals and the modern scientific worldview; conversely, they seemed to think that various kinds of anti-scientific idealism and neo-rationalist obscurantism (as they saw it) were connected to the sort of authoritarianism and nationalism which produced the first world war.  The political leanings of the Vienna Circle were also shown by some of their activities outside of philosophy; many of them were, for example, involved in an adult education project in Vienna, motivated by their egalitarian ideas about helping people outside of the traditional elites.

The rise of the Nazis was an obvious threat to pretty much everyone in the movement.  Some of them were Jews, some of them were communists, and all of them had lots of Jewish or communist friends and leftist political affiliations of varying kinds.  As a result, nearly all of them found ways to leave Europe in the 1930s, with some of them ending up in Britain, but most of them coming to America.  There seems to have been much less migration of philosophers not associated with Logical Positivism, for whatever reason (there were of course some who found Nazi rule congenial, such as Heidegger, but this explanation can’t be applied to a broad range of cases.  Very few, if any, of the members of the still vast Kantian tradition joined the Nazi party;[1] nonetheless, they mostly did not leave Germany.  Perhaps they were less politically active than the Logical Positivists had been, and so found it easier to just keep their heads down and avoid attention when the Nazis came to power).

In Britain, in the mean time, the idealism which had been popular in the late 19th century and some way into the 20th had considerably declined in popularity after WWI (partly, no doubt, due to a sentiment similar to that in Germany which connected idealism to nationalism and warmongering, as well as in Britain an additional sentiment that associated idealism with Germany).  In America, the end of the 19th century and the beginning of the 20th century had seen the appearance of a distinctive home-grown philosophical movement, the pragmatism of Pierce, James, and Dewey.  Pragmatism was strongly empiricist in inclination, and actually shared with Logical Positivism some of the fascination with logic; Pierce was a sophisticated logician, and in the 1930s C. I. Lewis was keeping this tradition of logically sophisticated pragmatism very much alive in America.

Thus, the Logical Positivist exiles found their new homes full of apparent philosophical allies.  This migration seems to be the primary cause of the analytic/continental split.  The departure of the Logical Positivists from the European continent skewed philosophy there sharply away from sympathy toward empiricism, paying attention to formal logic, and paying attention to science.  Conversely, the Logical Positivists were to have great influence in America.  While the Logical Positivists themselves had less influence in Britain, empiricism remained strong there and Wittgenstein also helped popularize some similar notions.  WWII also disrupted communication between continental Europe and Britain and the Americas.  After the war, the philosophers on either side of this divide apparently felt that they had little to talk about, and so mostly didn’t talk, a situation which has continued, despite many other changes, since.

Notably, one subsequent development which entirely failed to produce any reconciliation was the collapse of the Logical Positivist movement in America, which happened somewhere around the 1950s.  There are many stories as to what happened; one school of thought which was long popular but which has become a minority view is that W. V. O. Quine and some others refuted some central Logical Positivist mistakes (as in, it’s no longer widely believed that the refutations worked; it’s not just that people no longer think those refutations were the real reason for the collapse).  I will briefly give my view.

In order to survive as foreigners in America during the McCarthy era, the leading Logical Positivists greatly toned down any display of their leftist political sympathies.  They came to focus much more narrowly on philosophy of science, and also sought to de-politicize their approach to science as much as possible, in order to avoid drawing any attention to their own political views.  They were in fact so successful at this that a lot of more recent writing associates the Logical Positivists with conservative political views (people being apolitical is generally good for the status quo, of course, so advocating apolitical views is sometimes a covert way of serving the conservative cause).  This seems to have also produced a new generation of philosophers in America who thought what the Logical Positivists did was rather dull and irrelevant, since the Logical Positivists themselves were trying to hide any interesting applications of it.  A cynic might suggest that since logic and science are hard, any reason for thinking maybe you didn’t need to understand those to do philosophy would be welcome among lazier philosophers.  Thus, arguments that the Logical Positivists were on the wrong track, and philosophy shouldn’t be done that way, were widely accepted, even though again most philosophers these days will tell you the arguments that won the day weren’t all that good.

The shift was not complete; even after the fall of Logical Positivism, one might try to distinguish analytic from continental philosophy on the basis that some analytic philosophers took logic and science very seriously, while hardly any continental philosophers did.  Conversely, one area in which the Logical Positivists were definitely not leaders was history; it would also not be a completely misleading generalization to say that analytic philosophers have continued that tradition somewhat, and so that continental philosophy is more focused on history than analytic philosophy.  But even these are just tendencies.  There are definitely continental philosophers who are respectful of science, and analytic philosophers who aren’t, and analytic philosophers who are careful historians of philosophy, and continental philosophers who aren’t.  I don’t know of any continental philosophers who take formal logic very seriously, but that seems to be a minority tendency in analytic philosophy as well these days (unfortunately, if you ask me, but that’s the situation).

Verification

The Logical Positivist movement is probably most infamous for having advocated something called the “verification principle”, the doctrine that any claim which cannot be verified or falsified by some possible experience is meaningless.  Carnap is commonly taken to be presenting some form of such a principle in “Overcoming Metaphysics.”  His view is that sentences acquire their sense from their logical connections with protocol sentences.  Protocol sentences are simply statements about observation, so Carnap takes their cognitive meaning to be exhausted by the observations they describe.  Actual observation determines whether a protocol sentence is true or not, and the logical connections between the protocol sentences and other statements enable us to use them to deduce the truth or falsity of those other statements; thus, it is the machinery for determining the truth or falsity of statements which determines their meaning.  So this does seem to be a verification principle (and he describes his criteria for sentences having sense in terms of verification on occasion).

Prospects for describing the full logic of most ordinary language expressions are dim, and prospects for fully specifying the protocol sentences connected to any claim of remotely interesting strength more so.  Thus, the usefulness of any verification principle like this as a test is highly questionable; requiring that it be satisfied perfectly would likely leave us with just about everything anyone ever says being nonsense.  A more relaxed version would surely be needed.  But we can’t specify exactly how relaxed it is permitted to be; drawing that line precisely would require us again to know exactly how each statement is connected with protocol sentences, so we could somehow count up the connections (or something) and see if we were close enough.  Since the whole problem is that our knowledge isn’t that detailed, it seems that the best we can say is that some statements seem to be pretty well connected to experience, and some not so much.  The verification principle may still say that we should view the latter with more suspicion, but this hardly seems to be the sharp response to metaphysics Carnap promises.

I think the verification principle in fact is not the sharp implement Carnap thinks he has, and not merely because it can’t be.  The way Carnap actually approaches the effort to expose a piece of metaphysics does not consist in insisting that there is no experience which could verify or falsify it.  Rather, Carnap uses his idea that the components of sense are logical relationships and protocol sentences to analyze the sense of particular claims.  In the cases he treats at greatest length (the Heidegger material), he doesn’t even get to the stage of looking for protocol sentences; “the nothing itself nots” is logically defective in such a way that it isn’t logically related to anything at all, and so it is unnecessary to ask whether any of the things it is logically related to are protocol sentences which could be used to verify it.

Furthermore, Carnap explicitly considers the possibility that his own analysis of the logic of Heidegger’s claims (on which they seem to lack any sense at all) could be a misinterpretation; his response to this possibility is to examine Heidegger closely, to see how Heidegger says his words should be interpreted.  He seems to think it is important that Heidegger’s own statements explicitly rule out any effort to find a sensible alternative interpretation.  Carnap only levels the accusation of nonsense when the possibility of sensible interpretations has thus been fairly clearly ruled out, and where he has a diagnosis of what, other than sense, is probably going on.  To take another example, unlike some of the lesser positivists, Carnap did not classify Freud’s theories as nonsense, despite the difficulty in devising empirical tests.  Devising empirical tests can be difficult in legitimate science (a fact which Carnap’s background in science may have helped him see more clearly), and psychologists since Carnap’s time have in fact made progress in devising such tests.  The more recent research hasn't always gone so well for Freud, but that’s not the issue; Freud's theories must have been meaningful for it to be possible for recent empirical results to have shown them to be wrong, which seems to have happened in some cases.  In any event, those Carnap denounced as metaphysicians were those who did not make any effort to work out tests or clarify meaning, instead insisting as a matter of principle that this was impossible.

Similar points may be made about his other examples.  In the criticism of Descartes’ use of the “I think,” Carnap claims that existence is something that applies to predicates, and not to objects.  In Frege’s logic there is a very sharp distinction between predicates and objects, and there are some very deep logical motivations for keeping such distinctions sharp (the “theory of types” Carnap mentions at one point is needed to prevent a certain kind of contradiction from arising).  That “existence” isn’t a predicate that can be ascribed to objects was already Kant’s complaint against anything like Descartes’ ontological argument, and Frege’s logic may seem to provide means to make Kant’s point even more sharply.

However, while the existential quantifier (the backwards “E” thing) definitely applies to predicates and never to names of objects, it’s also an artificial logical innovation.  This introduces a difficulty; Carnap didn’t think his logic revealed what natural languages meant all along.  Also, in his later writings he says explicitly that you can construct a logic any way you want, as long as you explain the rules (and this is largely implied in the present reading as well).  Can’t someone invent an existence predicate that applies to objects?

The answer is that certainly one can.  Unfortunately, the application of Carnap’s objections to metaphysics is never as straightforward as some of its enthusiasts have thought.  But Descartes didn’t think he was using some special sense of “existence” in either presenting the “cogito” argument or in developing his ontological argument.  He doesn’t seem to have thought there were multiple senses of existence.  Further, he seemed to think that obvious features of the ordinary sense of existence were responsible for the validity of his arguments.  If he was using a special sense of existence, it would have been necessary for him to show that his special sense had the features his arguments require, and that the conclusions had the significance he attributed to them.  I don’t wish to go into the present state of discussions of “existence” in logic in detail, but I can say that prospects for devising anything coherent which serves Descartes’ purposes do not appear bright.

Are Value Statements Meaningless?

In later writings, Carnap began to explicitly distinguish between what he called “cognitive meaning” and other kinds of meaning sentences could have.  “Cognitive meaning” refers to whatever it is which can make a sentence true or false, so it corresponds to the “sense” he talks about in “Overcoming Metaphysics.”  He proposes the name “optative meaning” for the meaning of value judgments, claims about what’s good or right or the like.

Carnap talks about metaphysics more than about ordinary value judgments in “Overcoming Metaphysics,” but he says enough about value judgments to make it clear that he thinks they don’t have cognitive meaning; that, say, “stealing is wrong” does not count as a true statement (or a false one).  Presumably, he thought that there were no candidates for protocol sentences to ground the cognitive meaning of such statements.  Protocol sentences must be acceptable to all (and regarded as observational by all).  Carnap, like many, seems to have thought that people disagree about values to such an extent that there could be no value protocol sentences on a par with protocol sentences about objects.

It should be noted that this is not completely uncontroversial.  Some people claim that there are some (near enough) cultural universals to ground the meanings of ethical terms.  Carnap apparently disagreed.  Nonetheless, it is probably significant that Carnap chose not to waste much time going after explicitly ethical philosophy in “Overcoming Metaphysics.”  Precisely because the idea that moral judgments aren’t a kind of factual judgments is widespread, that they should be dealt with differently and may have something to do with individual preferences, stating something in the form of a moral judgment, as “x is right” or “x is good” can be seen as less misleading than the usual metaphysical practice.  Even though these forms look like declarative sentences, they not only don’t mean the same as any declarative sentence (according to Carnap), they are also frequently not taken to mean the same as any declarative sentence.  In other words, the form of a treatise on ethics may be at least a less unsuitable form for the expression of values than the form of a treatise on metaphysics, since it is at least less misleading.

Overcoming Metaphysics

Carnap believed that many claims of philosophical truth are really disguised value-judgments.  Nietzsche also argued for this view; perhaps it is for that reason that Carnap mentions Nietzsche relatively favorably at the end (it is also probably not a coincidence that the first word of Carnap’s title was one of Nietzsche’s favorite words).  I would further maintain, though Carnap does not state this explicitly in his paper, that one of the reasons he found this procedure so suspect was that he saw it as immunizing the value-judgments from criticism.  Since they’re not really the sort of thing that can be true or false, the prospects for refuting them are non-existent.  The same is true of the prospects for proving them, of course, but anybody who is sympathetic to the values in question is likely to apply lax standards of proof, and accept the terrible arguments given by the metaphysicians in their favor.

Carnap thought that value judgments should be treated entirely differently from that; he thought they concerned what we wanted, and how to negotiate between ourselves how to get there.  Taking the judgments to be truths eliminates the need for such discussion and negotiation, and cuts short any examination of whether we really want to buy into those values, whether we might not be sacrificing other things of equal or greater importance.  Again, this seems to fit with Carnap’s favorable attitude toward Nietzsche, since Nietzsche seemed to view the value issue in this way.

Obviously, the values people will be most inclined to want to shelter are those which are most dubious, and to which they are most emotionally attached; Carnap seems to have thought of metaphysics as not only containing values but being especially likely to contain reactionary values which he desperately wanted brought out into the open and criticized.

If I’m right in my interpretation of Carnap, it may be a point in favor of the accuracy of Carnap’s analysis that his primary target in this essay joined the Nazi party a couple of years later.  Then again, maybe with the benefit of hindsight I’m giving Carnap too much credit; perhaps he merely got lucky in choosing such a perfect target.


Overcoming Metaphysics through Logical Analysis of Language

by Rudolf Carnap (translated by Aaron Boyden)

  1. Introduction.
  2. The Meaning of a Word
  3. Metaphysical Words without Meaning
  4. The Sense of a Sentence
  5. Metaphysical Pseudo-Sentences
  6. Senselessness of all Metaphysics
  7. Metaphysics as an Expression of Values

1.  Introduction

From the Greek skeptics to the 19th century empiricists, there have been many enemies of metaphysics.  They objected in quite diverse ways.  Some said that the doctrines of metaphysics were false, because they contradicted empirical knowledge.  Others said that they were merely uncertain, because their questionings exceeded the boundaries of human knowledge.  Many anti-metaphysicians explained that preoccupation with metaphysical questions is sterile; whether one could answer them or not, it is unnecessary to concern oneself with them; let us occupy ourselves entirely with practical business, which confronts an active person every day!

Through the development of modern logic, it has become possible to give a newer and sharper answer to the question of the validity and authority of metaphysics.  The research into “applied logic,” or “epistemology,” which sets itself this task, through logical analysis of the intelligible content of scientific sentences and thus the clarification of the meaning of the words (“concepts”) which occur in those sentences, leads to a positive and a negative result.  The positive result applies to the realm of empirical science; the particular concepts of the various branches of science are made clearer; their logical-formal and epistemological connections are revealed.  For the realm of metaphysics (including any science or philosophy of values or norms) the logical analysis leads to a negative result, that the ostensible propositions of this realm are complete nonsense.  Thus a radical overcoming of metaphysics is achieved, which was not yet possible for the earlier anti-metaphysicians.  Indeed, related thoughts are found in some previous reflections, for example in those of nominalist inclination; but it is only now that the decisive development is possible.  The progress of logic in recent decades has provided the tools for making a sharp case.

When we say that the so-called propositions of metaphysics are nonsense, this word is meant in its strongest sense.  In a weaker sense, it is sometimes customary to call a statement or a question nonsense when presenting it would be entirely unproductive (e. g. the question:  “what is the average total body weight of every person in Vienna whose phone number ends with '3?'”); or also a statement that is quite obviously false (e. g. “in 1910, Vienna had 3 residents”), or one which is not merely empirically but logically false, that is contradictory (e. g. “persons A and B are each one year older than the other”).  Statements such as these still make sense, even if they are false or pointless; it is only propositions which make sense which can in any way be classified as (theoretically) productive or unproductive, true or false.  In contrast, to be nonsense in the strong sense is for an arrangement of words, from a particular language, to not constitute any statement at all.  Sometimes such an arrangement of words appears at first glance to be a statement; in these cases we call it a pseudo-sentence.  Thus, our thesis is that the apparent statements of metaphysics are revealed by logical analysis to be pseudo-sentences.

A language consists of vocabulary and syntax; that is of a supply of words that have meanings, and also rules for constructing sentences; these rules describe how the different kinds of sentences can be constructed from the words.  Accordingly, there are two kinds of pseudo-sentences:  either they contain words, which are merely mistakenly believed to have a meaning, or the words they contain have meaning, but have been put together contrary to syntax so that they make no sense.  We will soon show examples of pseudo-sentences of both kinds in metaphysics.  Later we will give our grounds for saying that metaphysics as a whole consists of such pseudo-sentences.

2.  The Meaning of a Word

If a word (in a specific language) has a meaning, it is customary to say that it picks out a “concept”; if it merely appears that a word has a meaning when really it has none, we speak of a “pseudo-concept”.  How can we explain the origins of such words?  Is each word not introduced into the language in order to pick out something specific, so that from its first application it has a specific meaning?  How could traditional languages contain meaningless words?  Originally, every word (apart from rare exceptions, of which we will later give an example) had a meaning.  In the course of historical development, some words change their meaning entirely.  And so, sometimes it also happens that a word loses its old meaning without acquiring a new meaning.  Thus, we get a pseudo-concept.

In what does the meaning of a word consist?  What conditions must a word satisfy in order for it to have a meaning?  (Whether the conditions are explicitly described, as for the words and symbols of modern science, or whether they are implicit, as with most of the words of traditional languages, does not matter for our purposes).  First, the syntax of the word must be given, that is the way in which it appears in the simplest sentence form in which it can occur; we call this sentence form its elementary sentence.  The elementary sentence form for the word “stone” is, for example, “x is a stone”; in sentences of this form are placed for “x” designations from the category of things, for example “this diamond” or “this apple.”  Second, for the elementary sentence S containing the word, answers must be given to the questions which can be formulated in these ways:

  1. From which sentences is S deducible, and which sentences are deducible from S?
  2. Under which conditions is S true, and under which is it false?
  3. How is S to be verified?
  4. What is the sense of S?

(1) is the correct formulation; the formulation (2) follows the manner of speaking of logic, (3) that of epistemology, and (4) that of philosophy (phenomenology).  That what the philosophers mean by (4) is captured by (2) was noted by Wittgenstein:  the sense of a sentence lies in the criterion for its truth.  ((1) is the “meta-logic” formulation; in a forthcoming work[2] an explicit presentation of meta-logic as a theory of syntax and sense, that is deducibility relations, will be given).

For many words, and certainly for the overwhelming majority of scientific words, it is possible to give their meanings by tracing back through other words (“constitution,” definition).  For example:  “'Arthropods' are animals with segmented bodies and limbs, and exoskeletons of chitin.”  This answers the aforementioned question for the elementary sentence form for “arthropod,” namely the sentence form “x is an arthropod.”  It is clear that a sentence of this form can be deduced from premises of the form “x is an animal,” “x has a segmented body,” “x has segmented limbs,” and “x has an exoskeleton of chitin,” and that conversely these sentences could also be deduced from it.   Through these specifications of deducibility (in other words:  through the criteria of truth, the method of verification, the sense) for the elementary sentence for “arthropod,” the meaning of the word “arthropod” is secured.  In this way, every word in the language can be traced back through other words, ultimately to the words in the so-called “observation sentences” or “protocol sentences.”  It is through this derivation that the world acquires its meaning.

We can leave entirely aside the question of the content and form of the basic sentences (protocol sentences), which has as yet received no final answer.  It is customary in epistemology to say that the basic sentences refer to “the given.”  However, there is no agreement concerning the question of what should count as given.  Some maintain the opinion that the statements about the given concern simple qualities of sensation and feeling (for example “warm,” “blue,” “joy,” and the like); others incline toward the view that the basic sentences concern complete experiences and similarity relations between them; a further view suggests that even the basic sentences already talk about things.  Regardless of the differences between these conceptions, it remains certain that a sequence of words only has a sense when its deducibility from protocol sentences can be specified; and similarly, a word only has meaning when the sentences in which it occurs can be traced back to protocol sentences.

Once the meanings of words have been specified by these criteria (in other words:  through the deducibility relations of their elementary sentences, through their truth conditions, through their methods of verification), one cannot, subsequent to the establishment of these criteria, choose what one wishes to “mean by” the word.  One cannot give less than the criteria, as they are required to give a word a sharp meaning, but one equally cannot give more, as the criteria fully specify everything.  The criteria implicitly contain the meaning, and all that remains is to set this out explicitly.

Let us say, for example, that someone proposes the new word “teavy,” and tells us that there are things which are teavy, and things which are not teavy.  In order to locate the meaning of this word, we would ask for the criteria; how can it be determined in a concrete case whether a specific thing is teavy or not?  Now let us suppose that he refuses to answer this; he tells us that there are no empirical tests for teavyness.  In this case, we would not consider the introduction of the word permissible.  If he continued to employ the word, insisting that there are teavy and non-teavy things, only it is beyond the weak, limited understanding of human beings, this eternal mystery of which things are teavy and which are not, we would recognize this as empty chatter.  Perhaps this person would assure us, though, that he really does mean something by “teavy.”  From this would follow only the psychological fact that he associates various images and feelings with the world.  The word would not acquire a meaning from this.  If no criteria for the new word are specified, the sentences in which it occurs say nothing.  They are merely pseudo-sentences.

Secondly, consider the case where the criteria for a new word, say, “toovy,” are specified, such that the sentence “this thing is toovy” is true if and only if the thing is quadrangular.  (For our example it is unimportant whether the criteria are explicitly given, or whether we determine them through observing the affirmative and negative uses of the word).  Here we would say that the word “toovy” has the same meaning as the world “quadrangular.”  And we would recognize it as illicit if those who employed the world told us that what they “meant by” the word something other than “quadrangular;” that even though every quadrangular thing is also toovy, this only results because quadrangularity is the infallible sign of toovyness, which is itself a mystery, an indescribable quality.  We would respond that, once the criteria have been specified, it has been specified that “toovy” means “quadrangular,” and that one no longer has the freedom to mean this or that by this word.

The results of our analysis will be briefly summarized.  Let “a” be some word, and “S(a)” the elementary sentence in which it occurs.  The necessary and sufficient conditions for “a” to have a meaning can then be given in each of the following formulations, which ultimately say the same thing:

  1. The empirical criteria for “a” are known.
  2. It has been specified which protocol sentences “S(a)” can be deduced from.
  3. The truth conditions for “S(a)” have been specified.
  4. The method of verification for “S(a)” is known.[3]

3.  Metaphysical Words without Meaning

For many words of metaphysics it will be shown that they do not satisfy the conditions that have been given, and so that they lack meaning.

Take as an example the metaphysical expression “principle” (taken as a principle of being, not as a presupposition or axiom).  Various metaphysicians give answers to the question, what is the (highest) “principle of the world” (or of “things,” “being,” “existence”), for example:  water, number, form, flux, life, spirit, idea, the unconscious, action, the good, and so forth.  In order to discover the meaning of the word “principle” in these metaphysical questions, we must ask the metaphysicians under which conditions a sentence of the form “x is the principle of y” would be true, and under which it would be false.  In other words, we ask for the criteria or for the definition of the world “principle.”  The metaphysician answers roughly as follows:  “x is the principle of y” says “y emerges from x,” “the being of y rests upon the being of x,” “y arises through x,” or something of the sort.  But these words are ambiguous and indefinite.  They often have a clear meaning; for example we say of a thing or an event y that it emerges from x when we observe that after things or events of type x, usually or always things of type y follow (causation in the sense of lawful connection).  But the metaphysician tells us that he does not intend such an empirically specifiable relationship; otherwise, the metaphysical theses would be simple experimental claims of the same kind as those of physics.  The word “emergence” cannot here have the meaning of a temporal and causal sequence that it usually has.  However, no criteria for another meaning are given.  Thus, the apparent “metaphysical” meaning of the word, in contrast with any empirical meaning it might have, does not exist.  If we consider the original meaning of the word “principium” (and the corresponding Greek word “άρχή”), we notice the same development.  The original meaning of “beginning” is removed; it is no longer to mean the temporally prior but rather the prior in some other, special, metaphysical aspect.  The criteria for this “metaphysical aspect” are, however, never given.  In both cases the word loses its earlier meaning without being given a new one; the word remains as an empty husk.  From its earlier meaningful period it is associated with various representations; they attach themselves to new representations and feelings through the circumstances in which the word is used.  But the word does not acquire a meaning from this, and it remains meaningless so long as no one can give a way of verification for it.

Another example is the word “god.”  Apart from the variations in the use of this word in particular circumstances, we must distinguish the linguistic usage in three distinct cases or historical periods, although they overlap temporally and regionally.  In the mythological use the word has a clear meaning.  This word (and parallel words in other languages) refers to various material beings which rule on Olympus, in the sky, or in the underworld, and which possess power, wisdom, goodness and happiness in greater or lesser degrees.  Alternately, the word can signify spiritual beings, which do not have anthropomorphic bodies, but which in various ways influence things and events in the empirical world and so are empirically accessible.  In the metaphysical use, on the other hand, “god” signifies something super-empirical.  The meaning of anything material, or of anything spiritual that is immanent in the physical, is completely removed.  Since no new meaning is given, it becomes meaningless.  To be sure, it usually appears as if the word “god” is given a meaning in the metaphysical use.  But the definitions which are given soon reveal themselves to be pseudo-definitions.  They lead to logically impermissible constructions (which will be explained later) or they lead to back to other metaphysical words (such as “ultimate ground,” “the absolute,” “the unconditioned,” “the independent,” “the self-created,” and the like), but in no case to the truth conditions of elementary sentences.  For this word not even the first requirement of logic is satisfied, namely the requirement of providing its syntax, that is the form of its occurrence in elementary sentences.  The elementary sentence must have the form “x is a god,” but the metaphysician either rejects this form entirely, or at least when he gives it he fails to give the syntactical category of the variable x.  (Categories are, for example:  bodies, properties of bodies, relations between bodies, numbers, etc.).

Between the mythological and the metaphysical uses stands the theological use of the word “god.”  Here there is no unique meaning, but rather one shifts back and forth between the approaches already given.  Some theologians have a clear empirical (and so in our terminology “mythological”) concept of god.  In this case there are no pseudo-sentences; but the disadvantage for the theologian is that under this meaning the claims of theology are empirical and hence subject to the judgment of empirical science.  Other theologians employ the metaphysical usage.  In still other cases the usage is unclear, either they employ sometimes this, sometimes that usage, or sometimes, because they are not clear, they can be interpreted in either way.

Just as with the examples of “principle” and “god,” so are most of the other specifically metaphysical terms without meaning, for example:  “idea,” “the absolute,” “the unconditioned,” “the eternal,” “the being of existence,” “non-being,” “things in themselves,” “absolute spirit,” “objective spirit,” “essence,” “being-in-itself,” “being-in-and-for-itself,” “emanation,” “manifestation,” “articulation,” “the self,” “the other,” and so forth.  With these expressions it is no different as with the word “teavy” in the earlier fictitious example.  The metaphysician tells us, that he can't give us empirical conditions; when he insists that there is nonetheless something which he “means by” his words, we know that this only shows that some images and feelings are associated with the word, which do not give it a meaning.  The metaphysical apparent sentences, in which such words occur, have no sense, and say nothing; they are pseudo-sentences.  We will discuss later how to explain their historical appearance.


4.  The Sense of a Sentence

We have so far considered those pseudo-sentences in which there are meaningless words.  There is a second kind of pseudo-sentence.  They contain words with meaning, but the words are assembled in such a way as to have no sense.  The syntax of a language determines which sequences of words are permissible and which are not permissible.  The grammatical syntax of natural languages, however, does not entirely prohibit senseless sequences of words.  Take as an example the following two arrangements of words:

  1. “Caesar is and,”
  2. “Caesar is prime.”

The arrangement of words (1) is constructed contrary to syntax; the syntax requires that the third word not be a connective, but rather a predicate, either a noun (with an article) or an adjective.  For example, “Caesar is a general” is constructed correctly syntactically; it is a sequence of words with sense, a genuine sentence.  In the same way, the arrangement of words (2) is syntactically correct, as it has the same grammatical form as the example just given.  Nonetheless, (2) is nonsense.  “Prime” is a property of numbers; it cannot be affirmed or denied of a person.  Since (2) appears to be a statement, but is not one, says nothing, expresses neither a true nor a false proposition, we call this sequence of words a “pseudo-sentence.”  That the grammatical syntax does not prohibit this easily leads one to the impression that one is dealing with a statement, albeit a false one.  However, “a is prime” is false if and only if a is divisible by a natural number other than a or 1; this cannot be asserted if “Caesar” is substituted for “a.”  This example is chosen so that the nonsense is easy to reveal; for most metaphysical so-called statements it is not so easy to recognize that they are pseudo-sentences.  That in everyday language it is possible to generate a senseless sequence of words without violating the rules of grammar is one of the reasons that the grammatical syntax is unsatisfactory from a logical point of view.  If the grammatical syntax precisely expressed the logical syntax, no pseudo-sentences could be generated.  Were the grammatical syntax to distinguish not only the parts of speech, the noun, the adjective, the verb, the conjunction, etc., but within each of these types made the further distinctions required by logic, these pseudo-sentences could not be composed.  If nouns were subdivided into more types, according to whether the properties of bodies, or of numbers, or the like could be applied to them, then the words “general” and “prime” would belong to different word types, and (2) would be just as linguistically unacceptable as (1).  In a correctly constructed language, all senseless sequences of words would thus be like our example (1).  They would already be ruled out automatically by the rules of grammar; so that in order to avoid nonsense one would not need to pay attention to the meanings of the individual words.  Instead it would suffice to notice the types of the words (their “syntactic category,” for example:  thing, property of things, relation among things, number, property of numbers, relation among numbers, and so forth).  If our thesis that the statements of metaphysics are pseudo-sentences is correct, then metaphysics could not be expressed in a logically correctly constructed language.  This is why the great philosophical project of the modern logicians, of constructing a logical syntax, is so important.

5.  Metaphysical Pseudo-Sentences

We will now present examples of metaphysical pseudo-sentences, in order to demonstrate how they violate logical syntax, even though they are permitted by traditional grammatical syntax.  We select some sentences from the metaphysical doctrine which has in recent times had the strongest influence in Germany.[4]

“Being only will be investigated, and besides – nothing; being alone and further – nothing; solely being and beyond it – nothing.  How is it with this nothing?...  Is there only nothing because there is the not, that is negation?  Or is it the reverse?  Is there negation and the not only because there is the nothing?...  We affirm:  The nothing is more fundamental than the not and the negation....  Where do we seek the nothing?  How do we find the nothing?...  We know the nothing....  Dread reveals the nothing....  Why and for what we dread, there is 'really' – nothing.  In the deed:  the nothing itself – as such – was there....  How is it with the nothing?...  The nothing itself nots.

In order to show that the possibility of generating pseudo-sentences depends on logical shortcomings of the language, we present a schema below.  The sentences under I are grammatically and logically unobjectionable; they have sense.  The sentences under II are grammatical analogs of those under I (with the exception of B 3).  The sentence form II A (as question and answer) does not conform to the requirements which a logically correct language would impose.  It nonetheless has sense, as it may be translated into in a correct language; thus sentence III A shows the sense that II A has.  The inconvenience of sentence form II A is shown by the way we can, with grammatically unobjectionable operations, move to the nonsense sentence forms II B, which were taken from the quote above.  These sentence forms cannot be generated in the correct language in column III at all.  Still, their senselessness is not obvious at first sight, as one is easily deceived by the analogy with the sentences in I B, which have sense.  The error demonstrated is because in our language, unlike a logically correct language, there can be grammatically identical form between sense and nonsense sequences of words.  To every sentence in words a formalization in symbolic logic is attached; these formulas allow the inconvenient analogy between I A and II A and the way they allow the senseless constructions in II B to be especially clearly recognized.

I.  Sensible sentences of the customary language.

II.  Emergence of nonsense from sense in the customary language.

III.  Logically correct language.

A.  What is outside?

                                         O(?)

     Rain is outside.

                                         O(r)

A.  What is outside?

                                            O(?)

     Nothing is outside.

                                            O(n)

A.  There is nothing (exist nothing, is lacking) something that is outside.

                            ~(x)O(x)

B.  How is it with this rain?  (That is:  what is the rain doing?  Or: what can be said about this rain?)

                                          ?(r)

B.  “How is it with this nothing?”

                                             ?(n)

B.  None of these forms could be constructed.

1.  We know the rain.

                                          K(r)

1.  “We seek the nothing.”  

     “We find the nothing.”    K(n)

     “We know the nothing.”  

2.  The rain rains.

                                          R(r)

2.  “The nothing nots.”

                                              N(n)

3.  “There is only the nothing, because...”

                                              E(n)

A closer examination of the pseudo-sentences under II B shows some recognizable differences.  The generation of the sentence (1) results simply from the mistake that the word “nothing” is used as the name of an object, because the customary language uses this form to express a negative existential sentence (see II A).  A correct language does not use a special name for this purpose, but rather a recognizable logical form of the sentence (see III A).  In sentence II B 2 there is something new, namely the generation of the meaningless word “to not;” the sentence is thus nonsense twice over.  We previously explained that the meaningless words of metaphysics normally emerged because some meaningful word, through its metaphorical employment in metaphysics, came to lose its meaning.  Here we encounter a rare case where a new word is introduced, which already from the very beginning has no meaning.  Sentence II B 3 is also rejected for two reasons.  It shares with the previous sentences the mistake of treating “nothing” as a referring noun.  Besides that there is also a contradiction.  If it were permissible to use “nothing” as a name or a referring expression, the definition would deny the existence of this object, but in sentence 3 there is a reference to it.  Thus, if the sentence were not senseless, it would be contradictory and so absurd.

Upon seeing the huge logical errors which can be found in II B, one may speculate that in the quoted essay perhaps the word “nothing” has been given an entirely different meaning than formerly.  This conjecture is reinforced, when we read further that dread reveals the nothing, that in dread the nothing itself is there as such.  Here the word “nothing” appears to have a definite emotional constitution, perhaps of a religious type, or at least something, which is the basis of the feeling, which is to be signified.  Were this the case, then the sentence II B would not contain the logical mistakes attributed to it.  However, the beginning of the quoted passage shows that this interpretation is not possible.  From the presentation of “only” and “and besides nothing” it is clearly established that the word “nothing” has the customary meaning of the logical particle which is used to express negative existential sentences.  This introduction of the word “nothing” is followed immediately by the central question of the essay:  “how is it with this nothing?”

Any doubts that we might have misinterpreted are eliminated when we see that the author of the essay is completely clear in asserting that his questions and sentences conflict with logic.  “Question and answer in regard to the nothing are similarly in themselves paradoxical....  The commonly accepted so-called basic rules of thinking, the principle of non-contradiction, the universal ‘logic,’ strikes these questions down.”  So much the worse for logic!  We must topple its authority:  “when the power of the understanding is thus broken in the field of questions about the nothing and being, then that decides the destiny of the authority of 'logic' within philosophy as well.  The idea of 'logic' dissolves itself in the vortex of fundamental questions.”  Could serious science accept the vortex of illogical questions?  The answer to this is also given:  “the alleged seriousness and superiority of science become absurd, when it does not take the nothing seriously.”  So we find good support for our thesis; a metaphysician himself draws the conclusion that his questions and answers are not compatible with logic and the scientific way of thinking.

The difference between our thesis and that of the earlier anti-metaphysicians is now clear.  Metaphysics is not for us “mere fantasy” or “myth.”  The sentences of a myth do not contradict logic, but only experience; they are thoroughly sensible, since they’re false.  Metaphysics is not a “superstitious belief” one can only believe true or false sentences, but not senseless sequences of words.  Neither can metaphysical sentences be considered “working hypotheses;” since for a hypothesis the derivable consequences for empirical sentences are required, and that is precisely what the pseudo-sentences lack.

With respect to the so-called limitations on human capacities for knowledge, the following objection is often made in defense of metaphysics:  the metaphysical sentences cannot be verified by humans or any finite beings; but they can perhaps be considered speculations concerning what a being with higher or ultimate capacities for knowledge would answer, and as such speculations they would remain sensible.  Against this objection we would respond as follows.  When the meaning of a word is not given, or the sequence of words is not syntactically assembled, no question has been set.  (Think of the pseudo-questions:  “Is the table teavy?”, “Is the number seven holy?”, “Are the even or the odd numbers darker?”.)  Where there is no question, not even an omniscient being can answer.  The objector would perhaps say:  as the sighted can communicate a new kind of knowledge to the blind, so could a higher being perhaps communicate metaphysical knowledge, for example whether the apparent world is the guise of a spirit.  Here we must consider what “new knowledge” amounts to.  We can certainly imagine that we could encounter animals which could report a new sense to us.  If these beings proved Fermat’s last theorem for us or discovered a new scientific instrument or revealed a heretofore unknown natural law, then new knowledge would have been reported to us by their help.  This is because we could test such things, just as the blind can understand and test the whole of physics (and thus all of the statements of the sighted).  If, instead, these beings said something to us that we could not verify, then we also couldn’t understand it; for us there would thus be no communication, but merely noises uttered without sense, though perhaps also with associated images.  Through another being, whether it knows less or more or everything, our knowledge can only be quantitatively augmented; no knowledge of an entirely new type could be acquired.  What is uncertain for us, can become more certain with help; but what is incomprehensible, senseless, cannot become sensible through the help of another, no matter how much they know.  Thus, not even a god or a devil can help us to metaphysics.

6. Senselessness of all Metaphysics

All of the examples of metaphysical sentences which we have analyzed have come from just one essay.  However, the conclusions we have drawn apply as well, sometimes in verbally exactly the same ways, to other metaphysical systems.  That essay was right to quote Hegel approvingly (“the pure being and the pure nothing amount to the same”).  The metaphysics of Hegel has exactly the same logical character as we have found in the modern metaphysics.  The same also applies to the majority of metaphysical systems, though the types of linguistic expressions and so the types of logical mistakes vary to greater or lesser degrees from the examples given.

It is unnecessary to provide further examples of metaphysical statements from different systems for analysis.  Instead, only the main types of errors will be indicated.

Perhaps the majority of the logical mistakes which occur in the pseudo-sentences follow from logical defects in the employment of the word “to be” in our language (and the corresponding words in most languages, or at least most European languages).  The first mistake is the ambiguity in “to be”; it is used both as a copula for attaching a predicate (“I am hungry”), and also as indicating existence (“I am”).  This error is compounded because the metaphysicians are often unclear about this ambiguity.  The second error lies in the form of the verb in the second meaning, existence.  The verbal form feigns a predicate where there is none.  It has long been known that existence is not a predicate (see Kant’s refutation of the ontological argument for the existence of God).  But only the modern logic fully establishes this point:  it places an indication of existence in a syntactical form such that it cannot be applied as a predicate to anything standing for an object, but only to a predicate (see, for example, sentence III A in the table earlier).  Most metaphysicians since ancient times have allowed themselves to be led into stating pseudo-sentences by the verbal and predicative form of “to be,” for example, “I am,” “God is.”

We find an example of this mistake in the “cogito, ergo sum” of Descartes.  We will ignore the objections which have been made against the premise of the forgoing reasoning – whether the sentence “I think” is an adequate expression of the intended meaning or whether it perhaps contains a hypostasis – and will instead look at the two sentences entirely from the standpoint of formal logical.  Two logical errors are noted.  The first is in the conclusion, “I am.”  The verb “to be” is certainly intended here to convey the sense of existence, since a copula cannot be employed without a predicate; the “I am” of Descartes is always understood in this way.  But then this sentence violates the aforementioned logical rule, that existence can only be asserted in relation to a predicate, never in relation to a name (subject, proper name).  An existence sentence doesn’t have the form “a exists” (as here:  “I am” amounts to “I exist”), rather “there exists something of such and such a type.”  The second error lies in the move from “I think” to “I exist.”  If an existence claim is to be deduced from the sentence “P(a)” (“the a has the property P”), this can only assert existence in connection to the predicate P, not in connection with the subject a.  From “I am European” what follows is not “I exist” but rather “something is European.”  From “I think” what follows is not “I exist” but rather “something thinks.”

That our language expresses existence with a verb (“to be” or “to exist”), is in itself not a logical mistake, but instead just inconvenient, dangerous.  Through the verbal form one is easily led to the mistaken conception that existence is a predicate, and then to such logical associations and eventually senseless expressions as we have already described.  It is for this reason that such forms as “the existent” and “the non-existent” have played such a large role in metaphysics.  In a logically correct language such forms cannot be constructed.  It appears that in Latin and German, perhaps through the precedent of the Greeks, the forms “ens” and “seiend” were introduced for the purposes of metaphysics; thus was the logic of the language made worse, when it was believed that a lack was being filled.

Another very commonly encountered violation of the logical syntax is the so-called “category mistake.”  The previously mentioned mistakes resulted from employing an expression with a non-predicative meaning as a predicate.  Here, a predicate is employed as a predicate, but as a predicate of the wrong “category;” this is a violation of the rules of the so-called “theory of types.”  An artificial example is the sentence discussed earlier:  “Caesar is prime.”  Personal names and number words belong to distinct logical categories, and so also predicates for persons (such as “general”) and predicates for numbers (“prime”).  The category mistake is, unlike the previously discussed employment of the verb “to be,” not exclusively found in metaphysics, but rather occurs often in everyday language.  There it does not usually lead to nonsense; the category ambiguities in words are easily removed.

Example:  1.  “This table is larger than that one.”  2.  “The height of this table is larger than the height of that table.”  Here the word “larger” expresses in (1) a relationship between objects, while in (2) it is employed for a relationship between numbers, and thus two different syntactical categories.  The error is here not substantial; for example, it could be eliminated by writing “larger1” and “larger2”; “larger1” is then defined in terms of “larger2” by declaring sentence form (1) synonymous with sentence form (2) (and some other similar forms).

Because category confusions in everyday language produce no harm, it is customary to completely ignore them.  But while this is extremely convenient for common uses of language, in metaphysics it has unfortunate consequences.  Here there are category mistakes, following the custom of everyday language which allows them, which, unlike the everyday language examples, cannot be translated into logically correct form.  Pseudo-sentences of this type can be found often in, for example, Hegel, and also in Heidegger, who in copying many of the idiosyncrasies of Hegel’s style also retains many of his logic defects.  (For example, predicates which should be applied to objects of a particular type are instead applied to predicates applying to such objects, or to “being” or to “existence” or to a relation between such objects).

Although we have found that the many metaphysical sentences are senseless, the question arises whether there is a core of sense in metaphysics which would remain if we threw out the nonsense.

One could conclude from our discussion so far that metaphysics is in great danger of falling into nonsense, and so that one must take care to mind such dangers when engaged in metaphysics.  However, in fact the situation is that metaphysical sentences cannot make sense.  This follows from the task which metaphysics sets itself; it wishes to find and establish knowledge which is completely outside empirical science.

We explained earlier that the sense of a sentence is found in the method of verifying it.  A sentence only says what can be verified about it.  Thus, whatever a sentence says, it can only concern empirical matters.  Anything that lies in principle beyond experience can’t be said, or thought, or asked.

The (sensible) sentences can be broken down into the following categories:  first, there are sentences, which because of their logical form alone must be true (“tautologies” as Wittgenstein calls them; they correspond roughly to Kant’s “analytic judgments”); they say nothing about reality.  The formulas of logic and mathematics are of this type; they are not themselves descriptions of reality, but only rules for transforming such descriptions.  Second, there are the negations of such sentences (“contradictions”); they are refutable, and so, again because of their form, false.  For all the remaining sentences the difference between truth and falsity is determined by the protocol sentences; they are (true or false) empirical sentences and belong to the domain of empirical science.  If one constructs a sentence which does not belong to one of these categories, it is automatically senseless.  Since metaphysics seeks to make statements which are neither analytic, nor accessible to empirical science, it must employ words for which no criteria are given and which are thus meaningless, or it must assemble meaningful words in such a way that neither an analytic (or contradictory), nor an empirical sentence results.  In both cases it must produce pseudo-sentences.

Logical analysis thus issues the judgment that any alleged knowledge which is claimed to go above or behind experience is nonsense.  This judgment applies to all speculative metaphysics, to all alleged knowledge from pure thought or pure intuition, which thinks it can dispense with experience.  But the judgment applies as well to any metaphysics which begins with experiences, but which through special inferences would know what lies outside or behind experience (such as, for example, the neo-vitalist theses that there are “entelechies” in living things which are not physical; the question concerning “real causal connections” over and above the existence of lawful regular successions; the talk of “things in themselves”).  The judgment also applies to all normative philosophy, to all ethics or aesthetics as normative disciplines.  The objective acceptability of a norm or value cannot (even with the concepts of normative philosophy) be empirically verified or deduced from empirical sentences; it can thus not be expressed at all (by any sensible sentences).  To put it another way:  either one gives empirical criteria for “good” and “beautiful” and all the other predicates of normative sciences, or one does not do so.  A sentence with such a predicate would in the first case be an empirical judgment, but not an evaluative judgment; in the second case it would be a pseudo-sentence; a sentence which expresses a value judgment cannot be constructed.

Finally, the judgment of nonsense applies also to those metaphysical movements which are inaccurately called epistemological, namely realism (insofar as more is intended than the empirical fact that the past displays a certain regularity which makes the application of inductive methods possible) and its opposite, subjective idealism, solipsism, phenomenalism, and positivism (in its older sense).

But what remains for philosophy, when all the sentences that say anything are of an empirical nature and so belong to the sciences?  What remain are not sentences, or theories, or systems, but rather a method, namely the method of logical analysis.  We have shown the negative application of this method in the preceding; it was used to expose meaningless words and senseless pseudo-sentences.  In its positive employment it clarifies the sensible concepts and sentences which are the logical principles of the sciences and mathematics.  The negative application of the method is important and necessary in the present historical situation.  However, while it can’t be examined here, the positive application is more productive, and will continue to be useful in the future.  This kind of logical analysis of fundamental presuppositions is what we call “scientific philosophy,” in contrast to metaphysics; the majority of the contributors to this journal[5] work on this project.

The question of the logical character of the sentences in which logical analysis is presented, for example the sentences of this and other logical essays, can here be answered only tentatively; they are partly empirical and partly analytic.  These sentences about sentences and parts of sentences correspond partly to what we call pure meta-logic (for example, “a sequence, which is composed of an existential quantifier and a name for an object is not a sentence”), partly to descriptive meta-logic (for example “the sequence of words on such and such page of such and such book is nonsense”).  This meta-logic will be treated on another occasion; at that time it will also be shown that the meta-logic which is used to speak about the sentences of a language can itself be formulated in that language.

  1. Metaphysics as an Expression of Values

When we say that the sentences of metaphysics are complete nonsense, that they say nothing, there will be those, even among those sympathetic to our viewpoint, who will have a feeling of discomfort.  Could it really be true that so many men of different times and places, including some of the greatest minds, expended so much effort, indeed fervor, in the pursuit of metaphysics, when in reality this consists of nothing but senseless strings of words?  And how can we understand the enormous impact of these works on readers and hearers right down to the present day, when they contain not even folly, but in fact nothing at all?  These thoughts are justified, as metaphysics does indeed have a content; only it is not theoretical content.  The (pseudo-)sentences of metaphysics do not serve to represent states of affairs, whether those which obtain (then they would be true sentences) nor those which don’t obtain (then they would be at least false sentences); they serve to express values.

Perhaps we may assume that metaphysics developed out of Myth.  The child is angry at the “evil table” which hurt him; primitives try to appease the demon who brings earthquakes or give thanks to the god who brings the rain for their crops.  Here natural phenomena are personified, giving a quasi-poetic expression to the way humans feel about their relationship to the world.  The heritage of mythology is found on the one hand in poetry, which by various means brings the relevance of the myths to life to the forefront, and on the other hand in theology, in which the myths evolve into a system.  What, then, is the historical role of metaphysics?  Perhaps as a substitute for theology upon the appearance of systematic, conceptual thought.  The (alleged) supernatural sources of theological knowledge are replaced by natural, although (allegedly) super-empirical sources of knowledge.  On closer examination, we can find beneath the various changing approaches the same motives which were seen in mythology.  We find that metaphysics also springs from the desire to express an attitude toward life, toward the circumstances in which humans live, the feelings and ambitions directed toward the world, toward the community, toward their projects, toward vicissitudes of fate.  This attitude toward life is present, usually unconsciously, in everything humans do and say; it even appears in their faces and the way they walk.  Many have the desire to express these attitudes in another, more concentrated and uniquely significant way.  Should such people have artistic inclinations, they find the possibility for such expression in the creation of works of art.  The way in which the style and form of artistic works reveals values and attitudes toward life has already been clearly described by various sources (for example, by Dilthey and his students).  (The expression “world-view” is often employed for this; we avoid it because of the way it is ambiguous between values and theories, which our analysis seeks to distinguish).  For our purposes what is important is that art is an adequate, metaphysics is an inadequate means of expressing such values and attitudes.  In and of themselves there naturally need be no objections to any forms of expression which might be chosen.  However, metaphysics presents itself in a form in which it pretends to be something which it is not.  This form is that of a system of sentences, with (apparent) relations of justification between them, that is the form of a theory.  Thus, it promises a theoretical content, which we have seen it does not deliver.  Not just the readers, but also the metaphysicians themselves are caught up in the deception that the metaphysical sentences say something, that they describe the way things are.  The metaphysician believes that he travels a path, in which there are ways of going true or false.  In reality, he has said nothing, but has only expressed something, like an artist.  That the metaphysicians deceive themselves cannot be concluded from their use of language as a form of expression and their use of declarative sentences; the lyric poets do the same, but without engaging in the same self-deception.  But the metaphysicians give arguments and try to prove their claims, and they write polemics against metaphysicians of other views, and try to refute their treatises.  The lyric poets on the other hand do not seek in their poems to refute the sentences in the poems of other poets; because they know that they are in the realm of art and not theory.

Perhaps music is the purest medium of expression for the attitude toward life, since it is most sharply free from descriptions of any objects.  The harmonious attitude, that a metaphysician might wish to express with a monistic system, is more clearly expressed in Mozart’s music.  And if the metaphysician expresses a dualistic-heroic attitude in a dualistic system, perhaps he only does this because he lacks the talent of Beethoven, to express this in an adequate medium.  Metaphysicians are musicians without musical talent.  Instead they have a strong drive to work in a theoretical medium, through connections between concepts and ideas.  Instead of satisfying the theoretical drive in science and the desire for expression of attitudes in art, the metaphysician confuses the two and produces something which is absolutely not knowledge and which is somewhat unsuitable to the expression of values.

Our contention that metaphysics is a substitute, although an unsuitable one, for art, is also shown by the fact that the metaphysician in whom perhaps the strongest artistic inclinations existed, namely Nietzsche, was least inclined to the mistakes resulting from these confusions.  The greater part of his works have primarily empirical content; they present, for example, historical analysis of artistic phenomena or historical and psychological analysis of morality.  In the work in which he most strongly expresses what others express through metaphysics and ethics, namely in “Zarathustra,” he does not employ the misleading theoretical form, but rather openly the form of art, of poetry.


[1] A recent paper argued that no Kantian ethicists did, but I haven’t closely examined the paper’s methodology for identifying the Kantian ethicists in 1930s Germany.

[2] [Presumably Carnap refers here to his The Logical Syntax of Language, considered by many his most important work.]

[3] The logical and epistemological conceptions underlying our position can only be briefly given here; see also:

        Wittgenstein, Tractatus Logico-Philosophicus, 1922.

        Carnap, Der Logische Aufbau der Welt, 1928.

        Waismann, Logik, Sprache, Philosophie.  (In preparation.) [Never actually published.  Waismann was in close communication with Wittgenstein at this time, and wanted to put together some more understandable version of Wittgenstein's philosophy.  Wittgenstein was never happy with Waismann's efforts, and Waismann eventually gave up.]

[4] The following quote (italics in original) is taken from:  Martin Heidegger, “Was ist Metaphysik?”  1929.  We could as well have taken passages from another of the recent or traditional metaphysicians, though these illustrate our thesis especially well.

[5] [This paper originally appeared in Erkenntnis, the journal of the Logical Positivists in the early 1930s.  Erkenntnis was shut down during the Nazi era.  The modern analytic philosophy journal Erkenntnis was named in honor of the old Logical Positivist journal, but is not a direct continuation of it.]