We know
that the same piece of good will weigh less if it is up on a high mountain. The
law of gravity explains this phenomenon. Yet before Newton, what did people
think? We know that temperature drops on a high mountain, therefore we say, low
temperatures, for some reason, reduce the weight of goods. This is a theory. To
verify the theory, we put the same piece of good down to sea level in a
refrigerated room, measure its weight and find that it does not weigh any less.
This theory between temperature and weight is thus refuted.
Subsequently
I will explain that for every theory with explanatory power, it must be
refutable by facts but has not been so refuted. Using the reduction of
temperature to explain the reduction of weight has been refuted by facts, so
should we take this as wrong? This is an important philosophical question.
If
regardless of other circumstances, any theory refuted by facts is taken as
wrong, then all theories are wrong. This is not acceptable. Theories refuted by
facts can be remedied. Using the weight of a good on a mountain as an example,
the rationale of falling temperature has been overthrown, we can say instead
that not only is temperature lower on a mountain, but wind is also stronger.
Therefore, we carry out another experiment, putting the same piece of good in a
freezing compartment with the addition of a blowing fan to measure its weight.
Such measurement would reveal again that the temperature hypothesis is wrong.
Without
giving up, we also note that the hillside slopes. Therefore, on top of a
freezing compartment and a blowing fan, we add a tilted board and put on it the
piece of good to measure the good’s weight. Again, the temperature hypothesis
is not credible. Not discouraged, we point out that a high mountain is way
above sea level. Therefore, we spend a great deal of money building a sky-high
freezing compartment. Eventually, we duplicate the conditions of a high
mountain: with temperature freezing cold; with wind from a blowing fan; with
slope by a tilted board; with height sky-high, a piece of good really weighs
less. The temperature hypothesis is therefore confirmed. This theory is not
incorrect, yet it is an ad hoc
theory. An ad hoc theory, also a
theory but since it is too ad hoc,
has no generalized explanatory power. It is not due to a lack of content of the
theory. On the contrary, it has too much content, therefore when the content is
slightly altered, the theory is overthrown.
Any
scientific theory, even if refuted by facts, can be remedied by incorporating
more conditions. But option has to be forgone in remedying a theory. Too much
option forgone is not warranted. Option forgone will be too much if an ad hoc theory explains only a single
phenomenon but cannot be extended to theorizing other phenomena; has no
generalization function; and minimal explanatory power. Theories refuted by
facts can be remedied, and often should be remedied, yet option forgone should
not be too much. The guideline in measuring whether option forgone is too much
is based on the magnitude of explanatory power. We should not abandon a theory
when its explanatory power is not extensive – a non-extensive theory today may
be replaced by another with more extensive explanatory power tomorrow, but before
that happens, a non-extensive theory could have already been the most useful.
There is
unalterable truth in the world, yet any theory is replaceable by a better
theory. Scientific advancement is not due to a correct theory replacing an
incorrect one, but due to a theory with more extensive explanatory power
replacing a less extensive one. Advances in human thought can render what is
considered superb today replaced by another with more applications tomorrow. We
are yet to put a full stop to the capability of human thought. Science
progresses by leaps and bounds since World War II, giving us reasons to believe
that human thought may have no boundary.
If an ad hoc theory is so specific as to
explain only a single phenomenon – like the aforementioned example that
explains only the weight of an object on a high mountain – it stands at one
extreme of scientific theory which has minimal application which cannot at all
be generalized. Theories at the other extreme, however, can be outrageously
generalized so that they can never be falsified under any circumstances. They
cannot be wrong because they are devoid of content. This is what philosophy
terms tautology. An ad hoc theory has
too much content while a tautology has none. A commendable theory must lie
somewhere between an ad hoc theory
and a tautology.
The
so-called tautology refers to certain statement that cannot be falsified under
any circumstances. In a stricter sense, a tautological statement cannot be
conceived to be wrong! For instance, if I say: “A four-leg animal has four
legs.” How can this be possibly wrong? The second part of the sentence
reiterates the meaning of the first. Even if we spend loads of effort, under no
circumstances can this be conceived to be wrong. It cannot be wrong on earth, on
Mars, or anywhere within the universe. This sentence possesses powerful
generalization, but what is its content? None has it in fact! No matter how
hard we deliberate, we know this is correct, yet we do not know its content. A
tautology is empty with zero explanatory power.
In
general, a tautological statement is not as trivial as “a four-leg animal has
four legs” that can be identified at a glance. “Theories” which carry no
substance and cannot be falsified are aplenty, yet very often are not detected
even by scholarly doctors. Let me cite a few examples.
An
indispensable postulate in economics is: every behavior of every individual is
for maximizing self-interest. However, a person hurts himself if he smokes or
jumps off a building. If we say smoking or plunging from a building is because
of “maximization of self-interest”, this is a tautological statement. With all
behavior counted, using this postulate of “maximization of self-interest” to
“explain” smoking or plunging from a building cannot be falsified, since the
postulate itself generally incorporates all behavior of an individual. If the
behavior of every individual could be explained by definition and in such an
empty manner, then the entire economics would barely have any content.
Let’s quote another example. An
economist attempted to empirically test whether a private enterprise’s
production cost was the lowest possible of that enterprise. By economic
definition, in order to maximize profits, all private enterprises will do their
utmost to lower their production costs. Therefore, the hypothesis of this
economist was a tautological statement. It could never be falsified, but it
carried no content as the definition itself does not allow any behavior of
intentionally not lowering costs when presented with such an opportunity.
Friedman gave some remarkable comment on this economist’s empirical work:
“Stupid question will of course yield stupid answer!” What is a stupid
question? A question that cannot possibly have a second answer – or a question
that cannot possibly have a wrong answer – is stupid.
A tautology is not necessarily
superficial. Very often it cannot be discovered at a glance, and at times not
even by learned scholar. More than forty years ago, a Harvard University
graduate was awarded a Ph.D. in economics, with his dissertation winning an
excellence award. That dissertation was later published in a book and
vigorously trumpeted. Even more well-known was the book review by Armen
Alchian. Alchian brilliantly pointed out that the whole dissertation was a
tautology, devoid of content and could not be falsified. That book review
deeply embarrassed Harvard. Just imagine that even top-notch economics
professors in the renowned Harvard University could not discover the tautology
of a Ph.D. student, how can we underestimate the “profundity” of this kind of
logic?
I said a tautological statement
cannot be falsified, carries no content, but did not say such a statement could
not possibly be an important concept. In fact, many important scientific theories
originated from the viewpoints or concepts of tautologies. There is a
commendable feature of tautology: it can be vastly generalized. If we can
restrain or limit its scope, sometimes a falsifiable theory with content can be
devised. Its explanatory power could be so strong as to win lots of plaudits.
We can quote a few examples in
economics. It is a tautology devoid of content if the aforementioned
“maximization of self-interest” and smoking are mixed up, like by definition,
with seemingly perfect justification. But if we can insert a few constraints to
enable us to infer under what circumstances a person would smoke more, smoke
less, or quit smoking, then such a theory has content to be empirically tested.
A more distinct example, turning a
tautology into a theory with broad applications, is the quantity theory of
money in monetary theory. The starting point of this theory is obviously a
tautology: money supply (M) times the circulating velocity of money (V) equals
the price of goods (P) times the transacted quantity of goods (Q). Such an MV =
PQ equation cannot be falsified, as the former (MV) and the latter (PQ) are
merely different perspectives of the same amount. Since this equation cannot be
falsified, it becomes a definition and can thus be written as MV = PQ. Clearly
this definition does not explain anything, but since it provides a new
perspective to look at the world, it is inspiring. When appropriately
restrained, it becomes the important quantity theory of money with massive
explanatory power. Extensively learned scholars like Irving Fisher and Friedman
successfully indicated under what circumstances the circulating velocity of
money (V) would be roughly constant, then went on to specify the relationship
between money supply (M) and the price of goods (P). The quantity theory can be
amazingly applied to ever-changing situations. Ultimately, its origin was a
tautological concept.
Some people say that the Coase
theorem, vastly popular in economics for over forty years, is a tautology. I
consider the Coase theorem immensely useful since those knowledgeable can
skillfully insert constraints to generate many hypotheses capable of explaining
different phenomena. In the hands of people with varying abilities, the same
tautology could yield distinctly different clout. Those who criticize the Coase
theorem as a tautology and turn a blind eye to it have no idea of its
immensity. As to what the Coase theorem is, we will analyze in detail in Volume III.
We can draw some conclusion between
the two extremes of ad hoc theory and
tautology. An ad hoc theory has too
much content, can specifically explain a single phenomenon but its explanatory
power cannot be generalized. Yet having an ad
hoc theory is nonetheless better than having no theory at all. As well said
by Reuben Kessel: “No argument can be won with no theoretical underpinning.”
The capacity to explain a single phenomenon is better than the incapacity to
explain any phenomenon, though any commendable scientific theory must be
capable of generalization; otherwise theories could be as plentiful as
phenomena, and the world would then be a big mess.
The other extreme is: since a
tautology is too general and cannot be falsified, its content tends to be empty
and irrelevant. The explanatory power of a tautology is even weaker than that
of an ad hoc theory, yet a tautology
can be an inspiringly important concept in providing us a new perspective to
view the world. Those who believe a tautology is devoid of content and turn a
blind eye to it could have given up a treasure. Instead of abandoning a new
perspective to view the world, we should try to incorporate constraints to add
content to a tautology, hoping to turn “definition” into a theory capable of
explaining phenomena.
Greatly commendable theories capable
of explaining phenomena always lie somewhere between the two extremes of ad hoc theory and tautology. Scientific
advancement often commences from one extreme or the other and evolves
progressively toward the center.
No comments:
Post a Comment