Errors of Empiricism 4

Empiricism holds that observations are all that we have. We cannot penetrate through the observations to the hidden reality which generates these observations. Here is a picture which illustrates the empiricist view of the world:


The wild and complex reality generates signals which we observe using our five senses. The aspects of reality which we can observe are the only things that we can know about reality. The true nature of hidden reality, as it really is, independent of our observations, is unknown and can never be known to us. Influential philosopher Kant calls it the “thing-in-itself”. Noumena is the wild reality, and Phenomena is what we can perceive/observe about the reality. Quote from Encyclopedia Britannica:

Noumenon, plural Noumena, in the philosophy of Immanuel Kant, the thing-in-itself (das Ding an sich) as opposed to what Kant called the phenomenon—the thing as it appears to an observer. Though the noumenal holds the contents of the intelligible world, Kant claimed that man’s speculative reason can only know phenomena and can never penetrate to the noumenon.

Kant was enormously influential in de-railing the philosophy of science (See  Kant’s Blunder). Prior to Kant, philosophers understood science in the way we have explained in the beginning: science is about looking through the appearances in order to understand the hidden reality. However, Kant argued that this was an impossible task. All we have is appearances (phenomena), and we cannot look through them to get at the underlying hidden realities (noumena). He proposed that instead of studying the relation between appearances and reality, we should study the relationship between our thought process and the observations of the real world:


It is important to understand Kant, because his way of thinking is at the heart of Economic Modeling today. To get a deeper understanding of Kant, we provide several arguments favoring his views. Think about how a simple computer camera looks at the world. The area being looked at by the camera is represented as a square two-dimensional patch which is say 1000 x 1000 pixels. At each pixel, if the camera detects light, it puts a 1 and if it does not, it puts a 0. So, we end up with a picture of reality which is a 1000 x 1000 matrix of 1’s and 0’s. This is the OBSERVATION. Now how can we translate these observations into a picture of reality? This is the basic problem of computer vision – taking a stream of numerical inputs from the camera and translating it into a picture of reality. For example, a particular stream of 1’s and 0’s may be interpreted as a picture of a tree, by a computer vision program. As human beings, we face a similar problem. We don’t actually see the world out there. What we see is a reflection of the world within our eyes. Our minds process the image on our retina into a picture of the external reality.  Before Kant, most people thought that the image in our minds matched the external reality. What Kant said was that we have no way of knowing this. We have no way of knowing the external reality. All we can see is the image of it on our retina, and the interpretation of it in our minds. A Kantian model, which we will label a mental model later, explains how we convert streams of 0’a and 1’s into an image of reality.


For understanding the nature of models, we will need to keep these three things in mind. Reality generates observations. And our minds interpret observations as a picture of reality. Most of us think that the picture in our minds is exactly what the reality is. When I look at a tree, I do not say that my mind has interpreted an image on my retina as a tree. I say that there is a tree out there in external reality which I am seeing. However this is an over-simplified understanding. For example, when I see a mirage, I interpret the image on my retina as water, but in fact there is no water in external reality. Similarly, a fly has a compound eye, and sees the world in way which is very different from how we see it.


As opposed to Kant, traditional philosophy is concerned with the question of how the image we have formed relates to external reality (not to the bitstream of observations). Traditional philosophy would ask: which is the “correct” picture of external reality? What the fly sees or what we see? What Kant says is that there is no way to learn the answer to this question. We have no separate access to external reality apart from our observations. So instead of thinking about whether our mental pictures match true reality, we should think about how we process the stream of sensations we receive into an image (a model) of the world.  Favoring Kant, Evolutionary biologists argue that the picture that we see of the world tends to highlight those aspects which matter for our survival, and ignore or neglect those aspects which don’t. This means that the representation of reality that is captured by our senses has less to do with the true external reality, and more to do with our own survival. The point of all this is that the naïve idea that what we see is just a true picture of reality is not necessarily correct.

This idea of Kant, that we can and should abandon looking for truth – the true picture of reality – has had a powerful effect on the philosophy of science today. Especially in economics, models that we build have no relation to reality. Rather the models in use are ways of organizing our own thoughts about reality. This has led to models which are hopelessly bad. Furthermore, the IDEA that we do not need to try to match reality, has led to the impossibility of correcting bad models to make them better. All that happens is that bad models are replaced by more complex models which are even worse. To understand this better, we now discuss three types of models –

Next Post: Three Types of Models 5 

Previous Posts in this sequence:  Mistaken Methodologies of Science 1,  Models and Realities 2 Thinking about Thinking 3,

  1. Kant is right to distinguish information (images) from knowledge, but is ambiguous and subject to embodiment by Algorithm Framework Theory (AFT) which says it is innate Instructions that process information, and thereafter knowledge emerges. As the combinatorial explosions go endlessly, Kant is right again, the final truth (the noumena) would be unknown. However, due to many reasons, statics, certainties or equilibria arise somewhere in the world of “knowledge”, which are what scientists pursue and will last shortly or long, nobody definitely knows.

    Empiricism is somewhat mistaken, but where? In short, it ignores the existence of Instructions rather than the impossibility of obtaining the final truth, therefore it has to walk on only one leg, or clap with only one palm, and hence fails frequently.


  2. …nobody definitely knows the future of statics, certainties or equilibria, as the latter could sometimes be destructed by innovations. Although the final truth (the noumena) is impossible to be acquired, knowledge as computational results vary from one to another, so improvements or developments are really possible, and thus we come to a new scheme of “start + processes – eschatology”.

  3. Ken Zimmerman said:

    The whole of human existence is about uncertainty. Humans either learn to deal with uncertainty or uncertainty will deal with humans. Most social theories don’t like uncertainty, of any sort. The theories share a common failing, often fallen under by the other sciences as well. They’re founded on the ready-made notion of ‘natural objective matters of fact’ that too quickly conflated reality, unity, and indisputability. Our path lies in prying apart this now unquestioned conflation. For those of you more philosophically focused, an ontological error was made. When you look for the first, you do not get automatically the two others. And this has nothing to do with the ‘interpretive flexibility’ allowed by ‘multiple points of views’ taken on the ‘same’ thing. It is the thing itself that has been allowed to be deployed as multiple and thus allowed to be grasped through different viewpoints, before being possibly unified in some later stage depending on the abilities of the community to unify them. There are simply more agencies in the pluriverse, to use William James’s expression, than philosophers and scientists thought possible. The important ethical, scientific, and political point here is that when we shift from the world of matters of fact to the worlds of matters of concern, we can no longer be satisfied either by the indifference to reality that goes with multiple `symbolic’ representations of the `same’ nature or with the premature unification provided by ‘the natural.’ Indeed, pluralism is our ‘first principle.’ It is a rough and uncertain sea for humans to navigate. If we fail or lose our way, there is always the danger that deployment of the actors’ worlds will remain too easy because they could be taken as so many representations of what the world, in the singular, is like. In which case we would not have moved an inch and would be back at square one of social explanation-namely back to Kant’s idealism. This is also the dividing line between postmodernism, which believes its task is to add multiplicity to a world overly unified by ‘master narratives,’ and my stance here which feels multiplicity is a property of things, not of humans interpreting things.

  4. Yoshinori Shiozawa said:

    Lars Syll and Asad Zaman (in his critical posts on empiricism) pose too much importance on methodology and philosophy. I do not deny that methodology and philosophy influence the development or evolution of scientific research. But, they are rarely dominant factors. Scientists including economists are more eclectic in methods and if they find a new hopeful method they easily transfer themselves from a method to another. Characteristics of the problem they face strongly influence which method is more appropriate than another. Science evolves by internal logic of the discipline than it is influenced by methodology or philosophy.

    Analytical method is a concept that covers wide range of investigation. The actual dead-end of economics does not owe to this too wide possibility but to more concrete defects of the basic framework. Even the criticism of holy trinity of neoclassical economics (i.e. rationality, selfishness and equilibrium and not to say fallacy of composition) is too vast and ambiguous and requires more concrete examination. Much more important thing to do is to present an alternative. Accusing analytical method or empiricism does not produce this concrete alternative. Such a negative arguments can gain applause from wide range of economists, but in fact, each applauds by a different reason. There is no coherence.

    I accused these attitude (accusing methodology and philosophy) sufficiently. To escape from falling in the same negative accusation, let me suggest one point that most of readers easily acknowledge the depth of economics dead-end.

    Let us take the very basic formulation of economics: demand and supply curve. High school economics teaches us that price and quantity are determined at the crossing point of these two curves. It was Alfred Marshall who popularized this. He inserted the famous figure as a footnote (Fig. 19, Principles of Economics 1920, p.288). I contend that this framework itself is already wrong.

    Is this analytic? Is this too empiric? No! It has some flavor of analytic and empiricism. It has some flavor of synthetic and transcendent arguments. Lars Syll cannot refuse demand and supply theory (concepts of demand and supply functions and the hypothesis that prices and quantities are determined by an equilibrium point) by accusing that it is too analytical. Asad Zaman cannot refuse it claiming it is too empirical. The demand and supply theory is by no means empirical but a construction of pure imagination. It may have been partially correct in the 18th century but essentially incorrect already in the end of 19th century. The basic fallacy of modern economics lies in these concepts and theory that draws on them.

    In classical political economy time, there were already similar concepts. Ricardo fought against them but failed as a result. John Stuart Mill who believed to be a loyal successor to Ricardo started to revert to old custom. This reversion lead the economics to neoclassical revolution (or anti-revolution to Ricardian economics).

    The culmination of “demand and supply theory” is the Arrow and Debreu’s paper (1954) “Existence of an Equilibrium for a competitive economy” in Econometrica. DSGE models (standard models for New Classical and New Keynesian economics) usually assume there is only one kind of goods (sic) in the economy (situated in different time points) totally depends on Arrow and Debreu (1954). Without it, DSGE models cannot pretend to be a model of modern economy which is composed of thousands of millions of goods and services.

    So, if we beat the doctrine of demand and supply theory, whole neoclassical economics falls down. Then, is it possible to construct a new theory? If we free us from imprinted belief of demand supply theory, it is in fact possible. See for a price theory here and, for more total examination, see here.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: