Paul Krugman still wrong. The mainstream model of an economy based on general equilibrium, rational utility maximization, and money neutrality is one of a possible world that doesn't exist and can't exist in a monetary production economy.
There is nothing wrong with constructing models of possible worlds, and, in fact, all models exist in possibility space, not real space. But it is wrong to claim or imply that such models of possible worlds apply to the real world when there is evidence that they do not. This is what science is about, and it is the difference between doing science and doing mathematics.
Conventional economic theory is largely mathematical with an impeccable logic pedigree from axioms, but it not scientific in that it lacks an empirical warrant. The models are attractive but vacant.
Lars P. Syll’s Blog
Polanyi and Keynes on the idea of ‘self-adjusting’ markets
Lars P. Syll | Professor, Malmo University
Lars P. Syll’s Blog
Polanyi and Keynes on the idea of ‘self-adjusting’ markets
Lars P. Syll | Professor, Malmo University
7 comments:
"There is nothing wrong with constructing models of possible worlds, and, in fact, all models exist in possibility space, not real space. But it is wrong to claim or imply that such models of possible worlds apply to the real world when there is evidence that they do not."
Tom, is there a bit of a contradiction in your words here? It seems you're saying that all models are not real and any model that isn't real can't apply to the real world. What then would be the purpose of ever creating models if we can never apply them to the real world?
From that, what do you think of the notion that: all models are wrong but some are useful? If that's the framework we forced to deal with since all models are wrong, if usefulness is the main criteria, then does "real"ness even matter?
"Tom, is there a bit of a contradiction in your words here?"
Feature (not a bug) of the dialectic method... Tom is Jesuit (Plantonist) trained (PhD Philosophy Georgetown) ie QUALIFIED in the dialectical method... they are Platonists ie dialectical... "thesis + anti-thesis = synthesis"
This is the way that methodology works... its anti science... ie its in contrast to science... one vs. the other...
Its an inequality... btw not judging here.. ie gotta be a purpose for both methods imo...
Here this is FIGURATIVE LANGUAGE from the Greek Scriptures in the context of agriculture that the Lord may have thought the uneducated Israelites MIGHT be able to understand (but ofc we can see they didnt...):
"Matthew 7:15-20 New King James Version (NKJV)
You Will Know Them by Their Fruits
15 “Beware of false prophets, who come to you in sheep’s clothing, but inwardly they are ravenous wolves. 16 You will know them by their fruits. Do men gather grapes from thornbushes or figs from thistles? 17 Even so, every good tree bears good fruit, but a bad tree bears bad fruit. 18 A good tree cannot bear bad fruit, nor can a bad tree bear good fruit. 19 Every tree that does not bear good fruit is cut down and thrown into the fire. 20 Therefore by their fruits you will know them."
Where Tom sez here: "largely mathematical with an impeccable logic pedigree from axioms, but it not scientific in that it lacks an empirical warrant"
Tom is putting that from Jesus to His disciples (active learners) in a Philosophical (not figurative "agricultural") context... leaving out the figurative language...
Tom is just TELLING you directly ie he's using the didactic method here.... ie TELLING YOU exactly how it is... ie suck it up cupcake...
See here in figurative language: "9 Every tree that does not bear good fruit is cut down and thrown into the fire."
Paul sez here via NON figurative language ie didactic method: 25 Wherefore, putting off the false, let each be speaking the truth with his associate, " Eph 4:25
Einstein comes close but not quite here: "insanity is doing the same thing over and expecting a different result"...
If something is not predictive (turns out to be false) then you have to discard it and go back to the drawing board sorry that's just the way it works suck it up cupcake morons... get tough and dig back in ....
Responding to SDB
I don't like the notion that "all models are wrong." For the most part, most models are not exact replicas and don't purport to be be. The fact that models are simplifications doesn't imply that they are "wrong." The purpose behind most modeling is to address what is important to some issue and disregard the trivial. The art of modeling is wielding Ockham's razor with respect to a particular purpose. There is nothing "wrong" with modeling astronomy with the earth as the center. But it's a whole lot more economical to put the sun at the center and use Newton's laws rather than Ptolemaic epicycles. Some models are more useful than others.
It's not that models are "wrong." What's wrong is the claims made about them, e.g, claims that go beyond the scope of the model.
Perhaps I was not clear enough here about possible versus actual. It is key, so I'll go deeper into it. This is a brief summary of issues in philosophy of science that have long been debated.
All models, whether formal or informal, conceptual or mathematical, are logical, in the sense of articulating possible structure and function. Static models are essentially structural, whereas dynamic models add function.
Models are logical artifacts used to describe possible "worlds." "World" here means system. Different possible worlds are different systems of entities, properties and relationships. Logical systems are represented symbolically, conceptually suing words and mathematically using variables, constants and operations, etc. Mathematical models also based on definitions and rules that are conceptual.
For example, novels are models of possible worlds, along with action that takes place therein. They are not meant to describe the real world, although historical models correspond to some degree with the facts of history, at least to the degree these are known. Often the boundary between fact and fiction is somewhat obscure. A model sets forth a possibility. That possibility can only be known to be actual by checking it against the real world through observation.
It is impossible to determine the degree a model may represent reality from the model alone. Some models don't claim to be representational of reality, but some do. Those that do can only be shown to represent reality by providing evidence that is external to the model, e.g. empirical data, showing that a possibility is an actuality.
Model set forth possibilities. The degree to which the possibility that a model articulates is representative of actuality is confirmed by comparing the model to the real world, that is, to "the facts," "evidence," "data."
A model can be internally consistent or inconsistent, which can be determined from the construction of the model based on the rules of logic and math. But the internal consistency says nothing about the real world.
Similarly, axioms are assumptions. Axioms and postulates are stipulations that are assumed as starting points. They can only be known to be true on the basis of evidence. If the assumptions are true and the logical correct then the conclusion can be known to be true by deductive logic. But axioms are stipulated in modeling, so the model needs to be put to the test. Deduction allows for hypothesis generation. If the results are negative, then the model (assumptions) need to be adjusted since the possibility the model describes is shown not to correspond to reality as claimed.
For example, a hypothesis is generated from a model that predicts a certain result in possibility space as an event occurring at certain coordinates. But on mapping the possibility space against real space, no such event is observed at those coordinates. Fail. If the logic is valid, then something must be off with the assumptions.
continued
continuation
The first step is checking for consistency, which is logical. Then, if the model is claimed to be representative, deduce a theorem from axioms and then formulate a hypothesis and design an experimental protocol for testing it. This is what science normally does.
If a model purports to represent the world but fails the test it is not very useful for purpose. But the only way to tell is to check the model against evidence that is external to the model through testing either the assumptions or hypotheses empirically. This is called "methodological naturalism" in science.
For example, a description is a conceptual model of a possible state of affairs. The description is true if and only if the description modeling a possible state of affairs in the world corresponds to an actual fact in the world. No amount of understanding or inspection of the description can guarantee the truth of that statement. Determining the truth of a description as a possibility requires evidence by checking putative claims against facts — actuality.
Of course, a description can also fail as a model if it violates the rules of logic. Then it is not false but a nonsense. At the level of theory is this inconsistency, failure to give sufficient meaning to terms, and the like that can be determined from the model alone.
Most models in science are not simple descriptions but general descriptions. Such models describe a "possible space" that can fit many facts and events. There are two ways this can happen.
The first is a general model that claims to fit the facts/events precisely. The second is a general model based on probabilities.
In classical physics, models are general descriptions that fit all the facts precisely, so any test can call the model into question. Of course, the simple form of an equation applicable in a vacuum may have to be adjusted for relevant factors like friction.
In other sciences — life sciences, social sciences and psychology, the general descriptions are generally based on probabilities, so it is more difficult and challenging to get a definitive answer similar to physics and precise results as in engineering. For example, in economics many models assume cet. par. when things are always changing and they use static modeling when the real world is dynamic and complex. Moreover, data may be lacking or less than ideal.
So it may be difficult to impossible to check the model against world with any degree of precision. Much of the precision in such situations is only wrt to special cases, while the really interesting issues are other special cases that may be either impossible for a model to handle or simply excluded from the scope of the model. This greatly limits model usefulness as a general description.
Models are called into question by false hypotheses. If hypotheses cannot be generated from the model that enable comparison of the model with reality, the model is of limited usefulness. This is a problem in modeling other than physics because the subject matter is elusive, on one hand, and other other, ethical issues prevent carrying out experiments. So possibility can only be reduced to probability.
end
I'll try to save you some Jesuit/Platonist oxygen here Tom:
"Wherefore, putting off the false" Eph 4:25
5 words to your Platonist "word salad" (pardon the figurative language...)
Post a Comment