This article suggests a definition for the term “philosophy”. It is a summary of a book, “The Cat That is Not There”, published by Magnes Publishing House (Hebrew University Press), 2010. The title comes from a dictum attributed to William James: “A philosopher is a blind man searching in a dark room for a black cat that isn’t there.”  

Why define “philosophy”? 

It is clear why philosophers are interested in the definition of “philosophy”, but why should a layman care? Here is one reason: philosophy is the only field in which you can find problems 2500 years old that are still open, and that despite tremendous efforts no tangible progress has been made towards their solution. You may find this as motivation to try your own luck against the problems. But a more reasonable approach is to try to understand what in the nature of philosophy makes the existence of such problems possible. 

People regard philosophy with a mixture of awe and suspicion. Awe because its problems look deep, suspicion because no concrete insights emerge from philosophical discussions. But nobody, including philosophers, is sure what precisely philosophy is. What is its subject matter? And is it the topic that makes a discussion philosophical, or the way the topic is studied? Philosophy is concerned with human thinking, but human thinking is part of the world – why should its study be different from that of any other subject? The object of a philosophical discussion is always a fata morgana, that disappears when you get closer. If it becomes tangible, it no longer belongs to the realm of philosophy. There is undoubtedly something unique about philosophy, setting it apart from all other branches of knowledge. 

Beyond all this, the definition of “philosophy” is interesting because it bears on the philosophical problems themselves. At least, the definition given in this article does.  

Peculiarities

Why is the philosophical discussion meaningful, if it destroys everything great, interesting and important? Because what we destroy is nothing but a tower of cards. (Wittgenstein) 

The touchstone of any definition of philosophy should be the ability to explain its many peculiarities. For example, the fact mentioned above, that two and a half millennia of research have not brought any progress on the main problems. As Wittgenstein put it, “Today’s philosophers are not any nearer to understanding reality than Plato. Isn’t it amazing how far Plato advanced?” In other fields robust edifices of knowledge are constructed, one solid layer upon another. Nothing of the sort exists in philosophy. As Wittgenstein’s remark cited above testifies, every construction is accompanied by just as much destruction. Nothing is agreed upon, and the general spirit is that of constant debate. “There is undoubtedly confusion, absurdity and puzzlement in philosophy” (Peter Strawson). More than in any other field, philosophical study usually relates to the sayings of previous researchers rather than to the object of study. All these puzzling characteristics must have a common origin, and more likely than not, one that can be sharply defined.  
 

What kind of problem is “what is philosophy”? 

Let me start by expropriating the problem of “what is philosophy” from the possession of philosophers. It is not a philosophical problem at all. The last statement may sound circular, because it depends on the definition of “philosophy”, but one property of philosophy that is agreed upon by all is that it is not empirical. A problem answerable by observation cannot be philosophical. However, the definition of “philosophy” (like all other definitions) is empirical. Finding it means identifying the conceptual structure that people recognize as “philosophy”. This should be done by scrutinizing philosophical writings, to find their common underlying structure.  

The ease and confidence with which people recognize philosophical problems testify to the sharpness of this structure. This is not to mean that it is easy to discover: the fact that a mechanism (in this case, that of recognizing philosophical discussions) operates well in our minds does not mean we necessarily know how it operates. In this respect, mental mechanisms are not different from physical ones: having a well functioning digestive system does not mean its owner knows how it works.  
 

“Clarification of concepts” 

The most common definition of philosophy is that its job is to clarify concepts. According to this definition a philosopher is somebody who stops life’s race for a moment, to examine his concepts or values. There are two sides to this venture: generality and order. The philosopher asks about values higher than people usually consider, and tries to put order in his conceptual system. For example, an educator may stop his daily work and ask “What is education? Is it aimed at molding the personality of the student? Do we wish to impose our views, or just exert some gentle guidance?”  

There are quite a few problems with this definition. 

1. Concepts are part of the world. It is not clear why putting order in them should be different from putting order in a room.
2. Clarifying concepts of another person, or one’s own past concepts, or the concepts of some society disjoint from ours, does not induce the impression of being philosophical. Clarifying the ancient Greeks’ concept of causality is not philosophical. “How does John use the term ‘education’? When teaching his students, does he try to mold their personalities? How liberal is our present educational system?” – all this is descriptive, and hence non-philosophical.
3. The philosopher will claim that the meaning of the question was changed by the above transformation. His questions were not merely descriptive, but normative – what is causality indeed? What should our view of education be? Should we attempt to mold our students’ personalities? But then this is a matter of decision, namely choice of values, and like description, a decision, too, cannot be philosophical.
4. There are philosophical problems, actually the most famous ones, that do not fit this definition. The Mind-Body problem and that of Determinism-Free Will do not ask for clarification of concepts. They pose paradoxes. Hillary Putnam once said that all philosophy starts from a sensation of paradoxality. Philosophical problems often point at something disturbing that needs rectification, not only clarification.
5. The main objection is that clarifying a concept is the duty of the person developing it. A concept is in constant interplay with reality, being adapted to fit the outside world, and thereby also be clarified. And this interplay takes place in the brain of the field expert.

 
Something else is happening here, beyond clarification. There is some secret, about the relationship between observing the concepts and using them.  

Collingwood’s definition 

One philosopher talked about this secret, and his definition is, in my opinion, the most illuminating of all existing definitions for “philosophy”. This was R.G. Collingwood, an Oxford historian and philosopher of history, who lived between 1889 and 1943. He wrote a book on the definition of philosophy, “An Essay on Philosophical Method”. In his last book, “The Idea of History”, published posthumously, he distilled from it the following definition:  

Philosophy is thought of the second degree, thought about thought. […] This is not to say that it is the science of the mind, or psychology. Psychology is thought of the first degree; it treats mind in just the same way in which biology treats life. […] Philosophy is never concerned with thought by itself; it is always concerned with its relation to its object, and is therefore concerned with the object just as much as with the thought. (the emphasis is mine, R.A) 

And here is a key concept – non-separation: 

[…] Philosophy cannot separate the study of knowing from the study of what is known.  

A simple example: for the psychologist the question “what is truth” means how does the concept of truth operate in our minds. When asked in philosophy, the question, means also at the same time what the concept should be, namely what truth really is. So, the philosopher looks at the person who uses the concept of truth, and at the same time regards himself as being in the position of that person.  

It may be helpful to have in mind the following diagram:  

 *

The top arrow designates the conceptual relation studied. What Collingwood says is in psychological inspection the inspector and the inspected are distinct. The inspection is done as if through a glass partition. In philosophy the inspector and the inspected are one and the same person. The inspector also studies the world, just as does the inspected, meaning that he puts himself in the latter’s position.  

This paper is an elaboration of this observation. It tries to show the ubiquity of the non-separation structure in philosophy. And it takes the observation one step further, in a claim that this structure is not really possible, and that the assumption of non-separation is infeasible.  
 

A definition 

So, here is my definition: 

Philosophy studies human thinking, while assuming that the conceptual system studied is identical with the one used for the study.  

In other words, philosophy arises when there is no separation between observer and observed. But I need to be more precise: it is only assumed that the two are not separate. Because, as we shall see, non-separation is not possible in reality. Separation is unavoidable. When a person observes a part of himself, he necessarily creates in himself a separate authority, observing that part.  

Pure and impure philosophical problems  

The definition will be easier to support if made more specific. The more specific claim is that there are two types of philosophical problems: those that are generated purely by the non-separation assumption, and those which only get their philosophical flavor from it. Problems of the first type do not carry any content relating to reality. It is as if somebody defines a number as “itself plus 1”, assumes that his definition is valid, and then cries “help! A paradox! A number that is equal to itself plus 1!” Such problems disappear once separation is enforced. Problems of the second type are in fact empirical problems about reality, but the non-separation assumption lurking behind them prevents us from seeing this. Once separation is enforced in such problems, they become ordinary problems about reality.  

Problems of the first type I shall call “pure”. These tend to be the “big” problems: skepticism, the Mind-Body problem, Determinism-Free Will.  They have the character of paradoxes, which is no surprise: faulty assumptions are bound to lead to contradictions. Each of these problems contributes a philosophical component to the other problems in its field.   
 

The effect of separation 

How can one prove that a problem involves non-separation between observer and observed? There is a very simple way: applying separation. Namely, asking the problem about somebody else. It is an empirical fact that that this causes the problem or its philosophical flavor to disappear.  

For example, “what is justice” becomes after separation “how does the concept of justice operate in John’s brain”, or “what are the justice values in our society”.  “Is the world part of my dream” becomes “is the world part of John’s dream”. The feeling that this changes the nature of the questions needs explanation, but apart from this the effect of separation is clear: it either causes the problem to disappear, implying that the problem is purely philosophical (this is for example what happens in the question of the dream), or strips it of its philosophical flavor (like in the problem about justice).  

Lewis Carroll applies separation – skepticism 

I shall first show the non-separation assumption in the three “big” problems mentioned above. Of these three, the non-separation is most explicit in skepticism. The recurring metaphor in skeptical problems is that of the jail: we are captives of our perceptions, our language and our modes of thinking. We cannot know anything about the world as it really is, since it reaches our knowledge only through mediators. The most extreme form of this doubt is the fear that there is nothing at all out there, it is all part of one’s dream.  

As usual, the first step should be applying separation, namely asking the problem on somebody else. Luckily, somebody already did it for us, in an amusing way. In “Through the looking glass”, the red king is asleep on the ground, and Twiddledee explains to Alice that she is nothing but part of his dream. When he wakes up, Twiddledee explains, she is going to disappear – puff – in the air. Skeptical arguments, it turns out, disappear in thin air once formulated on another person. Whether the entire world is only in the red king’s mind is an inane question. The problem is totally dependent for its existence on the assumption of identity between observer and observed.  

But this means that the problem is inane also when asked about oneself. Observing yourself cannot be different in principle from the observation of another person.

As Socrates says in “Carmides”, a thought cannot be a thought of itself. It must be disjoint from its object. Looking at myself I must construct within myself an observing entity. And then the answer to the question “is there indeed a flower vase on the table, or is it part of my dream?” is “I checked. There is indeed a flower vase on the table. I am not dreaming”. Since in this case the answer is given by another part, not the one under doubt, there is no self reliance in this answer. It is another person that checks the first person’s perception.  

The skeptical stance accuses us of circular reasoning. Guilt of circularity plays a central role in epistemology (the theory of knowledge) as also in the philosophy of values. “Epistemology is nothing but an exercise in skepticism”, as Alfred Ayer put it. It is the fear of self reliance that is responsible for the feeling of the ground disappearing beneath one’s feet, so typical of philosophy. But it turns out that the accuser is the one responsible for the fault: he himself built a circular structure of concepts.  Having done so, he blames those of us who feel standing on solid ground of blindness. The philosophical skeptic is a person cutting the branch on which he sits (or more precisely, believes he does so) and then complaining of being hung in the air.  

Determinism-Free Will, the idle argument and Newcomb’s paradox 

In skepticism the circular structure is transparent. In the Determinism-Free Will problem it is much less so. Before unfolding the way this problem is generated by non-separation, let me just say the principle behind it: it is based on the impossibility of fully understanding the causality governing one’s own actions. The observation of one’s own decision process leads to delicate circularity problems.  

The fact that the problem of Determinism-Free Will vanishes once separation is applied is widely agreed upon. When looking at another person, there is no feeling of contrast between the fact that his actions are free and their being subject to physical laws. The causal chain leading from past event to his future choice is not different, in principle, from any other causal chain. Indeed, many philosophers looked at the problem from this angle, and denied the existence of a contradiction. Hume, Mill and others claimed that there is no contradiction, because the opposite of freedom isn’t being determined by laws, but coercion. In the 1920-s, Moritz Schlick wrote that the fact that the problem survived this critique is one of the greatest scandals of philosophy. But the argument of Hume and Mill is dependent on being said from the point of view of an onlooker. The Australian philosopher Campbell, for example, complains that the Hume-Mill claim is no solution: it applies only to the actions of another person, not oneself. When looking at myself, I feel the generator of the action, meaning that the action could not be determined by laws, hence the feeling of contradiction.  

Things look different from the first person viewpoint. Here there seems to be a real problem, whose concise formulation is this:  

Determinism creates an equivalence between past and future events. But this is impossible, because we can decide about the future, while we cannot decide about past events.  

This argument cannot take off when formulated from an onlooker’s point of view. Simply because looking at another person, we do not think of determining events, but only in terms of causal links. Here is an example clarifying this point, as well as explaining the name the argument gained: “the idle argument”. Suppose that I have an exam tomorrow. Assume, furthermore, that an extraordinarily wise person, with unlimited powers of prediction, predicted the grade I will get. He even wrote down his prediction on a piece of paper, which of course he hides from me, since otherwise I could act in such a way as to give him the lie, making the prediction task circular and therefore impossible. Since the action of writing took place in the past, I cannot change the content of the note. Thus I cannot also change the grade I will get tomorrow. I can just as well sit idly, and not study. Of course, the outcome will be that tomorrow I will find in the note, as well as on the grades billboard, the number “0”. Note that from an onlooker’s point of view, there is no problem: he can be aware of the causal link. The punch line “therefore it is wiser not to prepare for the exam” cannot be said about another person. A decision is only on my own actions (and if I give advice to somebody, I become part of his system).  

In 1960 the physicist William Newcomb took this argument and gave it an attire of a paradox, in which it gained enormous popularity. Here is one formulation: suppose that you are standing near a well, with a $100 bill in your hand. You are pondering whether to throw the bill down the well or not. The wise person from the last paragraph predicted your choice, and he told you that if he predicted that you would throw the bill he put $1000 in your bank account half an hour ago. If he predicted that you would not throw, he did nothing. The paradox is now in the existence of two arguments, both appearing to be completely valid, while being contradictory. One says that your present action will not change what happened half an hour ago. What you have in your bank account will not change, so throwing the bill is simply losing $100. The other argument is that if you throw the bill the wise man had predicted that, and you will be you will be $900 richer.  

The point of divergence of the two arguments is the possibility of determining the past. The first argument plainly claims that determining the past is impossible. The second argument does an equally simple thing: it points out a way of doing precisely that. The story of the wise man provides a link between my future action and a past event, and using this link I can determine the past event.  

Note, first, that this is nothing but the idle argument, reformulated. The starting point of both is an equivalence between a future event A and a past event B. Newcomb’s argument is “Since I can determine A, I can also determine B”. The idle paradox says “I cannot determine B hence I cannot determine A”.– the same thing, formulated using negation.  

As already mentioned, once the problem is formulated about somebody else it vanishes. There is no feeling of contradiction in believing that somebody else’s decision has been predicted. The decisive step, “hence I can do that and that so as to determine the past event”, just cannot be taken in this case, since you cannot decide for the other person. From aside, you don’t feel that the decision determines the past event. There is a causal link, but no determination. The feeling of determining the outcome of the event completely depends on being in the position of the decider.  

This in itself is clear indication of a circularity phenomenon underlying the paradox. As to the nature of this circularity – at the risk of doing injustice to the argument, I will present only its final step (one has to read the book for the full chain of ideas). But to delineate it properly, let me break it into two parts, each receiving a heading of its own.  

What do we “determine” 

In the second argument (for throwing the bill) the decider is using a causal link between his future action and the past event. The crucial observation is that this link must go through the decision process. Namely, the wise man has to know the motives leading to the decision. This is so because of the way we perceive “decision”. A causal link between the action and a past event, that does not go through the deliberation, generates the feeling of coercion. There are many such causal links, determined by external factors: when I deliberate whether to go to a movie tonight or not, I have to know something about the location of the movie theater, the possibilities of transportation, the list of movies shown (all determined by past events). But all these I perceive as not given to my choice. I feel that I “determine” only that part of the event that is exclusively linked with my deliberation process.  

Deciding on your decision process

But if this is the case, then the second argument (that which recommends throwing the bill) says this: “I would like my motives to be such and such, so that the wise man predicted so and so”. So, the second argument is an attempt to decide about one’s motives. It is the decision process that is to be decided on. The person is trying to determine what his motives should be – a manifestly circular task. So, the second argument is wrong because it hides circular reasoning. Since the idle argument is nothing but Newcomb’s paradox in reverse, this solves also the idle argument.  

Summary: determinism from outside, free will from inside 

Summing up, the feeling of incongruity between determinism and free will stems from adopting two positions at the same time – those of the decider and of the onlooker. From the onlooker’s point of view, there is no problem in believing that things are determined. The decider cannot see that. Understanding in full the causal links going through his decision process, or using these causal links to “determine” things, leads to problems of circularity. In the book, by the way, two more such circularity problems are pointed out.   

The Mind-Body problem 

The Mind-Body problem, like that of Determinism-Free Will, presents a paradox. It is in the apparent contradiction inherent in the interaction between entirely different realms: the physical and the mental. My pain has no mass or speed, yet it causes my arm to raise – an event in physical reality. In the reverse direction, a needle prick, which is a physical event, causes the mental event of a feeling of pain. How can two entities playing in disjoint playgrounds interact?  

Here, too, many philosophers agree that applying separation causes the problem to disappear. Looking through a glass partition at a community of ants, the concepts we construct about their minds are formed from their physical behavior, and so the problem does not arise at all. The same is true also with regard to human beings. The concept of another person’s mind is constructed from physical events just as the concept of a university (and here I am using a simile of Gilbert Ryle) is constructed from components such as the students, the buildings, the curriculums.  An onlooker knows about another person’s wish to raise his hand by the words he says (and words are part of the physical world) and by his actions. If the onlooker is a physiologist, he can also observe messages shot through the neurons, both towards the hand and towards the part of the person’s brain responsible for the verbal expression of the wish.  

This claim has been proposed by many philosophers, but the philosophical community is unimpressed. The reason is that this is true only from aside. In self inspection, things are different. Looking at myself, I cannot accept that my wish to raise my hand is merely neurons shooting command in my brain. I also know that I want to raise my hand. I don’t have to observe my reactions in order to know that I love somebody. I also know it, first hand. To know that I understood an idea, I don’t have to put myself to a test, or hear my words. The essence of the Mind-Body problem, as many philosophers agree, is in the private access people have to their minds. Only the owner of a feeling or thought can really know what he feels or thinks. And this knowledge is not about his reactions, but about something lacking physical attributes. 

But this means that the person does not create in himself a separate observing authority. If he did, that authority would have observed the person’s mental events just as it would another person’s mind, namely the observing authority would learn about the person’s mind by his reactions. If this were the case, the Mind-Body problem wouldn’t have arisen. So, underlying the Mind-Body problem there is a special assumption, that a person’s knowledge of his mental events is very different from his knowledge of all other phenomena in the world: it is direct. A person does not need to hear the words “I see light” or watch his pupils contract in order to know that he sees light. He knows it directly from the sensation of light. I will call this the “direct knowledge assumption”.  

The source of all western philosophy

      All western philosophy circles around the Mind-Body problem (Karl Popper) 

“Direct knowledge” means non-separation between observer and observed. The same sensation leads to the knowledge “there is light” and to the knowledge “I know that I see light”. First order and second order are identified. But much more that that is true: the direct knowledge assumption is the epitome of non-separation, the one inducing the feeling of non-separation in all other fields. The “cogito ergo sum” of Descartes is nothing but its statement: “I think therefore I am” sounds more sensible than “The cat sits on the mat therefore the cat is” only because of the feeling that your knowledge of your own mind is very different from your knowledge of the rest of the world. So, the Cartesian view of the mind starts there.  

It is also from this assumption that skepticism is born. When my friend reports that he sees a vase of flowers and I doubt the truth of his claim, all I have to do is compare his report with the world. This is possible, since I am comparing two events of the same category – the report is part of the world, as is the vase. When I ask the same question about myself, I am comparing two things from different categories. You do not have to hear your report about the vase, you know it directly. And you cannot compare two phenomena belonging to different realms.  

This explains Popper’s dictum, cited above. 
 

The is-ought dilemma 

Finding the non-separation structure in the “big” problems is strong evidence for its ubiquity in philosophy. However, most philosophical problems are of different nature – they do carry some real content. Their philosophical flavor is induced by two factors: the Hume-an is-ought dilemma, and a fear of circularity. These are anything but separate: the is-ought dilemma is in fact a special case of the fear of circularity.  

For example, the question “what is causality” does not disappear upon separation. “How does the concept of causality operate in John’s mind” is not a void problem, and “how do today’s scientists use the notion of causality” is a non-trivial question.

At the end of the day, all meaningful things that can be said refer to these versions of the problem. But the philosopher refuses to see this, because there is doubt lurking behind: how do we know? What guarantees that the present perception of the notion of causality is correct? Description is not enough. The real question is what should the concept of causality be. We don’t want the “is”, we want the “ought”. 

But the “ought” task should be given to the scientist. Forming the concept of causality is part of doing science. So, the philosopher is actually trying to play the role of the scientist. But he will also refuse to see himself as merely forming the concept: he wants justification. Thus he is trying to do both – form the concept and describe it. He is trying to play the role of the observer and the participant in the game (in this case, the scientific game) at the same time.  
 

An exercise in skepticism

      Epistemology is nothing but an exercise in skepticism (Alfred Ayer) 

The is-ought confusion is the main contributor to the philosophical ingredient in the theories of values (though, as we saw in the last paragraph, also in the philosophy of science). In epistemology, the main contributor is skepticism, or in another terminology, fear of circularity. A descriptive answer to a question hits against the doubt – are we not assuming the requested? And, in another formulation – is it in the world or in me?  

Take for example the problem practically opening Wittgenstein’s Tractatus: “does the world consist of facts?” Facts are pieces of information. So, from aside the question is “does John separate his perception of the world into chunks which he calls ‘facts’?” (the answer is probably “no”, not all information is given in chunks. But this is immaterial for our question here. We are asking why does the question sound philosophical). When asked about myself, I start doubting – is the partition into facts in my mind, or in the world? 

Is the non-separation structure feasible?

Think, if you please, if there is sight which is not a sight of something in the world, but it is its own sight? (Socrates, in “Carmides” by Plato) 

I will be happy if the reader has gone with me this far, and agrees with the above definition of philosophy. But I would like to make one step further. I want to claim (as I did here and there already throughout the paper) that the non-separation structure does not really exist. It is impossible. When you look at a part of yourself, you necessarily erect in yourself a new entity which looks at that part. The assumption that the same part serves in two roles at the same time is like assuming that a radio broadcaster can broadcast his very present action of broadcasting. One way of realizing the impossibility of non-separation is trying to program a computer to look at another part of itself. It is necessarily a separate part of the program that will look at the first part. There can be no difference of principle between looking at another person and at yourself.  

As already remarked, faulty premises lead to faulty consequences. It is thus no wonder that philosophy is permeated with paradoxality.  
 

A necessary Copernican change of view  

I have tried to show that the starting point of all philosophy is in the Cartesian view of the mind: that a man knows his mind in a unique way, having direct access to his mental events that nobody else has. Realizing that this view is faulty requires a change of viewpoint of the sort that Freud named “Copernican revolution”. Freud ascribed to himself such a revolution, in the discovery that man does not necessarily know about his personality more than others do, and that he does not know his inner motives and wishes more than an outside observer does. To this there should be added another step, relating not only to content but also to the way in which one learns about one’s mind. The way a person knows about his mental events is in principle the same as that by which he knows about the rest of the world.  

Man has many illusions about his place in the world: that he is the crown of creation, that the world was created to cater for his needs, that he was created in God’s image, that his life has special meaning. Philosophy adds yet another illusion: that human thought is a special phenomenon, and more importantly that a man knows his thoughts in a special way. Like all anthropocentric illusions, this leads to problematic conclusions. It is time that we renounce also this assumption.