The problem in any kind of existence begins from a very old distinction between appearance and reality. Appearances are obviously how things seem to us in our perception although not everything that we perceive does really also exist. How things seem to us is a property of our perceptual apparatus—senses, mind, brain, etc. Reality, on the other hand, is supposed to be independent of this perceptual apparatus. Therefore, how do we know that what appears to us is also real? Could it be that we are hallucinating or dreaming and what appears to us does not in fact exist?
The Problem of Illusion
The existence of hallucination and dreams is undeniable and we know about them being hallucinations or dreams only when we come out of them. The skeptic therefore argues: Could it be that many things that we presently consider real—chairs, tables, houses, and planets—could also be dreams and we will realize that only when we “wake up”?
Of course, this problem is not unique to perception, but applies to any kind of thought, belief, cognition, or intention. For instance, I might think that the sky is purple, or believe that the Earth is flat, or intend the ownership of someone else’s property. In each such instance, I have a sensual, mental, intellectual, or intentional state which exists in my perception but not in reality.
How do we know that our mental state is false? How do we know that we are mistaken? Philosophers have debated this problem for millennia, and various solutions to this problem have been offered. Let’s take a closer look at the most promising of these solutions, and their associated problems.
The Method of Induction
One viewpoint—called Induction—supposes that the problem of hallucination or dreams can be solved by repeated observation. If repeated observations produce the same perception, then the repeated confirmation should eliminate the possibility of doubt. For instance, we don’t have the same dream every day, and if we were to be dreaming then repetition will solve the problem. With such confirmation, we can be sure that we are actually not dreaming or hallucinating. Scientific experiments must, for instance, be repeated before their results can be considered real. If something has been repeatedly confirmed, then we would be sure that the observation in fact reveals to us the nature of reality.
While this approach to knowledge works in many practical situations, it also fails for equally many phenomena. Take an ordinary visual illusion, such as the illusion of size, for example. The tracks of a train appear to converge at a distance and the planet Jupiter appears much smaller to us. No matter how many times we see it, and no matter how many different people see it, it always appears the same. Clearly, in such situations, repeated observation does not solve the problem of illusion because everyone experiences the same illusion every time, and repetition doesn’t confirm or deny the truth of the experience. Thus, while Induction works in many cases, it also fails in many cases.
Knowledge by Coherence
Another view—called Coherentism—says that truth lies not in the perception of individual facts, but in the coherence between multiple facts. For instance, we can fly towards Jupiter or walk towards the converged rail track and it will appear much larger to us. The apparent convergence of the tracks is therefore an illusion because it does not sit well with the other facts about it obtained subsequently. Knowledge, in this view, arises not from individual observations, but from the coherence between them. As we collect more and more facts, we suppose that coherence amongst the facts begins to develop. Some facts are mutually consistent while others are not. We must now consider the mutually consistent facts to be true, because they are confirmed by the existence of the other facts.
Coherentism, however, gives rise to a new problem that to empirically determine if two facts are consistent, we must simultaneously observe them, although as facts accumulate, it becomes harder and harder to observe them simultaneously. You might argue that we don’t have to observe the facts simultaneously if we just record them in some way (such as on paper) to compare them later on. However, there is a fundamental difference between observations and recordings—the observations are based on sensation, while the comparison of recordings involves the cognition of meanings.
The Problem of Representation
For instance, if we measure the size of Jupiter at different time instances, and record these observations before comparison, the measurement of size only requires sensation but the recording and the comparison between recordings requires us to use a language of encoding the observation. If we were using alphabets, numerals, or other techniques to record, the observation of those symbols has to be interpreted to denote the facts uncovered in the sensation. How material objects become representations of meanings is highly problematic in science. We suppose that material objects are things-in-themselves, and not things about other things. How can they symbolize meanings?
Of course, this problem is not limited to the paper recording of observations. Even when instruments are used to perform measurements, there is an element of symbolism involved. Pointer movements or detector clicks are facts in their own right, but they are also representations of the facts outside the instrument. If we only treated the pointer movement and detector click as a fact (like the naked eye observation of a planet), then we could not use the pointer movement to infer that the mass of an object is 5 kilograms. We could only say that the pointer moved. When instruments are used in measurements, the instrument transforms the facts about the world into a symbol of the facts. Now we don’t just say that the pointer moved 300. We also say that the mass is 5 kilograms. In the naked eye observation the facts are not interpreted, but in the instrument observations the facts are interpreted. We treat the instrument and the world differently: we interpret the instrument, not the world.
This fundamental difference between the instrument and the world is crucial to the success of Coherentism because we cannot observe everything at once. To even compare the facts, we must record them in matter. However, when meanings can be recorded in matter, then we have broadened the nature of matter in a very fundamental way—matter is not just things-in-themselves, but also things about other things. We now have a choice: we can exclude meters and meter recordings from the world, or we must change our view of matter to include meters and meter recordings.
Both induction and coherence thus fail in some fundamental ways. Repeated observation is not a guarantee of truth, and coherence amongst facts requires the consideration of symbols. Of course, symbolism is not a logical problem like induction, but it necessitates a shift in our thinking.
The Nature of Rational Knowledge
Of course, we haven’t yet eliminated alternative methods to knowledge. Critiques of empiricism are not new, and the alternative to empiricism is called reason or Rationalism. This viewpoint argues that the collection of individual facts is not sufficient because we must somehow connect these facts into a theory. The world is not piece-meal, although our observations are piece-meal. To create reality from our observations we have to postulate something ‘behind’ the perceptions. But how do we know what lies ‘behind’ the perceptions? We have no empirical access to this unifying reality.
The scientist postulates a role for intuition in science by which we can come up with ideas about the nature of reality which would become axioms in a scientific theory. While we cannot see what lies behind our perception, we can postulate some reality and check its consequences. This view is called Foundationalism and it claims that to know if perception is true, we must simply assume some premises and check their truth against the observations. For instance, we can assume that the world is electrons and protons governed by a mathematical theory and compute the predictions of the theory. If the predictions match the observations, the premises are ratified and can be considered true.
However, now we hit a new problem called the underdetermination of theories by experiments—a given observation is consistent with many premises. For instance, my observation of Jupiter being a small planet is consistent with the planet indeed being small and less far away or being much larger and much farther away. How do we know which of these is true? The only way to find out is to generate more predictions of the theory and validate them against facts. We now have the same problem of Coherence that we saw before: if we have to collect many facts, then we must record them. In Coherence, a symbol was required to denote only facts about the world. In Foundationalism, the symbol must denote the ideas in the theory in addition to facts about the world. These two representations are different from the physical properties of the symbol themselves!
Ideas vs. Things
If Coherence poses a problem of reference and meaning in observation, Foundationalism complicates this problem even further. For instance, the symbol p can denote the idea of momentum, or the momentum of a specific particle. In the theory, p only denotes an idea. When some numbers are plugged in, a value of p is computed, and this same symbol now denotes an object. If the ideas are in our mind and the facts are in the world, then there are symbols which point both to our mind and the facts, and these pointers are quite different from the physical properties of the symbols themselves.
We now have a different problem: not only are some objects symbols of ideas or facts, but we also don’t a priori know what they are symbols of. Sometimes, a symbol may point to an idea, and at other times it might point to some facts. How do we know when the symbol points to either ideas or facts? Lest you think this is just a notional issue in the use of science, it actually lies at the root of many paradoxes in mathematics, including the famous Gödel’s Incompleteness theorem which arises because the same symbol is alternately interpreted as a meaning and a name. As a name, a number points to an object (a fact), and as a meaning, it denotes its properties. You can call someone Mr. Barber, but this doesn’t make him a barber. Similarly, you can arbitrarily change the name of an object, but that doesn’t change the object’s properties. However, since we cannot know when a symbol points to ideas or facts, we also don’t know how to distinguish them in principle. And this inability leads to logical contradictions.
This is a fundamental problem in Foundationalism because it indicates that there can never be a consistent and complete theory of nature. We can build consistent theories of nature that exclude many parts of nature—e.g., meters and meter readings—and in this case the theory will be incomplete. If, however, we add meters and meter readings into nature, then the theory will be inconsistent.
The Incompleteness of Scientific Theories
The problem of incompleteness has now appeared in all areas of science. We find that theories are increasingly confirmed to a greater and greater extent—and therefore their truth is less doubtful with each passing year—although this confirmation only pertains to certain aspects of the observation. Theories predict some aspects of the phenomena very accurately, but do not predict the other aspects at all. This is clearly unsatisfactory because we don’t suppose that there are different realities that cause different parts of a phenomenon; we suppose that there is one reality that causes the entire phenomena. If the theory only explains parts of this phenomenon, then the theory is not real.
This would still be a minor issue if the incompleteness was tied to a particular theory; we might say that only that theory is incomplete. However, as it turns out, every fundamental theory in every fundamental area of science—mathematics, physics, computing, and biology—is incomplete. All these problems of incompleteness are closely connected: they arise when science deals with object collections rather than with individual objects. In mathematics, this is connected to the nature of numbers, because numbers are properties of collections. In computing, this is connected to program semantics, because the meaning of a program is tied to the program as a whole. In physics, this is connected to the inability to predict matter distribution since the same total matter or energy in a collection can be distributed in many ways. In biology it is connected to the question of structure and function, because these are only properties of the collection. In every area of science, the scientific premises fail to deal with collections.
The Search for a New Foundation
The problem essentially reduces to the need for a new foundation in which collections—rather than objects—are the most fundamental entities. These collections cannot be epiphenomena of their parts; rather, the parts must be the epiphenomena of the collections. In current foundations, objects are real and collections are byproducts. These collections are denoted by concepts—chairs, tables, houses, and planets—but these concepts are not fundamental and hence they have no role in scientific theories. If, therefore, collections were more fundamental than objects, then concepts would be more real than the objects they collect. The objects would represent refined concepts while the collection would denote an abstract concept. The more abstract the concept, the more real it would be and the most real thing would also be the most abstract. In this foundation, chairs, tables, houses, and planets would be more real than atoms and molecules, although there must be entities even more abstract and real. We cannot therefore assert a single type of reality. We can only assert a tiered and graded world which begins in the most real and abstract, but gradually becomes ever more unreal and contingent.
In other words, the things we consider the most real in current science—i.e. sub-atomic particles, atoms, molecules, etc.—are relatively the most unreal in the new foundation. By contrast, the things that we consider unreal in current science—i.e. chairs, tables, houses, minds, societies, cultures, living beings, etc.—are more real in the new foundation. Unreality, of course, does not imply non-existence. Even the unreal exists. However, the unreal is an epiphenomenon. Reality is therefore not connected to existence; it is rather connected to what is fundamental or foundational in science. The epiphenomena exist as well; however, they are not fundamental, and therefore they are relatively unreal.
We might say that the world that we see, taste, touch, smell and hear is an epiphenomenon of the world that we don’t currently perceive, although we might be able to conceive it. Since the real is more abstract, the things that we see are not as real as the things that we don’t see. In fact, what we don’t see—because they are more abstract than the things that we can see—is more real. To see what we don’t see, there is obviously a need to develop new kinds of perceptions. These perceptions will begin to perceive in the world, what we can only—at present—conceive. For instance, now we would begin to perceive the existence of tables, chairs, houses, living beings, societies, cultures, etc. These things would not be epiphenomena of their parts; rather the parts would be epiphenomena of the wholes. I explore this theory in more depth in Moral Materialism.
The Problem of Hallucination Revisited
The above approach to science tells us how mistakes, misperceptions, hallucinations etc. can be distinguished from truth. The distinguishing criterion is that we must look in the deeper recesses of the observer, which also represent ever more abstract ideas. If, for instance, we see a visual illusion, before we know it to be the truth we must suspect our perceptual apparatus—the senses, in this case. The senses are more abstract than the objects they perceive: for example, objects can appear yellow or red but the property of color cannot be attributed to those objects; it can only be attributed to the eyes. In so far as red and yellow are refinements of the idea of color, color is logically prior to red and yellow. If something red appears to be yellow (or vice versa) then there is something wrong with the way we have conceived color differently from how it should normally be conceived. This change in our notion of color is in effect like a pair of colored goggles that distort the notion of color to tint the world differently.
Philosophers have been looking in the world to find methods by which the errors in perception, thinking, intention, and belief can be solved, when in fact the problem begins in how our perceptual apparatus is contaminated. If our perception or thinking is distorted, every method in epistemology to eliminate that error must fail in one way or another. Why would we not look at the observer’s mechanism of perception to understand what causes illusions, hallucinations, or misperceptions? I conducted such an attempt in the book Sāńkhya and Science, bringing in insights from Indian philosophy.