![Free-eBooks.net](/resources/img/logo-nfe.png)
![All New Design](/resources/img/allnew.png)
and regression can be used to relate different aspects (variables) in the data
set. The strength of the relationships that are recognized in the data set can
the tested using t-scores, F-ratios, Chi Square and other methods. At the end
of the day, none of these methods can prove causation: they can only show
correlation. The concept of causation depends on a theory (or hypothesis)
about the relationship between the variables. Statistical methods allow a test
of the hypothesis or theory. The hypothesis cannot be proven it can only be
disproven and the hypothesis rejected. Statistics can be used as evidence to
support or reject a perception of causation.
47
3.3 Foundations of “Science”
CHANCE
If the world of events were truly a machine subject to the law of causes,
events would be deterministic. Bronowski argues that the recognition of the
law of chance is central to the method of science. It adds “statistical law” to
the concept of “causal law.” (Bronowski,. p 82) Causal law states that event B
is caused by event A and therefore, event B will follow event A 100 time out of
100 occurrences of event A. Statistical law is based on the notion that event B
will “probably” follow event A. The process is described as one where:
“We look for a trend or systematic difference. But the line of this trend
will itself be blurred by the unsteady hand of chance or random
fluctuation. We cannot get rid of this random scrawl. But we can from it
determine a measure of random variation, and use that to draw round
the trend an area of uncertainty. If the area is small enough by
standards which are agreed between us, then the trend is established,
and we know the limits within which it is likely to lie. (Bronowski, p
92)
The concept of probability provides the method by which observations of an
extraordinarily complex world can be interpreted. It gives us information and
knowledge that may not be “true” but is useful.
In a complex world, there may be many reasons for a lack of certainty in
causes. There may be other hidden or unrecognized forces that influence the
relationship between event A and event B. If event A results in event B 90%
of the time we may believe that A “causes” B. If the occurrence of A results in
event B 30% of the time, other “causes” of B may be more important.
Probability is a key idea in the understanding of causes. Statistics provides the
means to state that with 95% confidence (or some other percentage) event A
is correlated with event B.
USEFULNESS AND “TRUTH”
Knowledge held at any time may be “true” or “not true.” Knowledge that is
true may or may not be useful. Knowledge may be useful whether it is “true”
48
3.3 Foundations of “Science”
or not. Before the Copernican Revolution, a common belief was that the Earth
was a stationary center of the universe. This was the Ptolemaic system
attributed to Claudius Ptolemy [127-151 AD], a Greek mathematician and
astronomer who lived in Egypt. In this system, the sun, stars, planets and
moon circled the Earth in repeated patterns. Complex models were
constructed to explain and predict the paths of the objects. These models
worked with reasonable accuracy and were useful to plan for seasons, planting
of crops, and to prepare for floods. The models were useful, but “wrong.” New
information obtained through observation and measurement showed there
were simpler explanations for the paths of the celestial bodies. The Copernican
or heliocentric view gained dominance. Galileo [1564-1642] verified the
Copernican system with a new technology (the telescope). Johann Kepler
[1571-1630] improved on Galileo’s findings and calculated equations to
explain the elliptical orbits of the planets about the sun. As we accept “new
knowledge” about the cosmos and subatomic matter, we replace old truths
with new truths.
49
3.4 Explanation, Prediction and Storytelling
3.4 EXPLANATION, PREDICTION AND STORYTELLING
Explanation and prediction are two of major objectives of science. These
two goals are not symmetrical: it is possible to explain an event or
phenomenon without being able to predict the probability of its occurrence: at
the same time, it is possible to predict an event without being able to explain
its nature or causes. Mark Blaug identifies two problems that arise from the
“Symmetry Thesis.” First is the problem: “the history of science contains a
number of theories which appear to explain natural phenomena, without
however predicting them even in a statistical sense.” (Blaug, 1986, p 274)
Darwin’s theory of evolution is cited as an example.
Second is the problem: “…science, and particularly social science, abound
in rules-of-thumb that yield highly accurate predictions about both natural and
social events despite the fact that we may have absolutely no idea why these
rules-of-thumb work as well as they do. (Ibid.)
Whether explaining or predicting, science places value on precision and
rigor of the process. However, one should avoid using the same criteria to
evaluate scientific models with different objectives. It is also necessary to
avoid attempts at precision and rigor that are not possible. Thomas Mayer
cautions economists (the warning applies to all disciplines):
“...we should draw a much sharper distinction that is usually done
between two types of economic theory. One, formalist theory is abstract
theory that is concerned with high-level generalization and looks
towards axiomization. The other, empirical science theory focuses on
explaining past observations and predicting future ones. While both are
perfectly legitimate, applying the criteria appropriate to one to evaluate
the other generates confusion and misunderstanding. (Mayer’s
book)...is a plea for a more modest economics that recognizes the
inherent difficulty of making precise and indubitable statements about
the actual world, accepts that there is a trade-off between rigor and
relevance. I certainly agree that one should be as rigorous as one can
be: I just oppose trying to be as rigorous as one can not be.” (Mayer, p
50
3.4 Explanation, Prediction and Storytelling
7)
An emphasis on rigor and precision may result in attempts to develop
theories or models that are esoteric and of little interest to anyone other than
the scientist-author.
In addition to explanation and prediction, science and the stories of science
also create, shape and transmit individual and social values. Often this is an
unintended effect rather than a conscious objective. The study of the evolution
of methods in a discipline, such as economics, will hopefully create a greater
awareness of this role and a greater understanding of one of the important
effects.
3.5 LOGIC
Several processes can be used in the discovery, creation and justification
of knowledge. Instinct, intuition, abduction, deduction, induction and authority
are examples of sources of knowledge. Appeals to authority as a justification
for acceptance of knowledge is common but is not a reliable source. Instinct,
intuition and introspection were once of great importance, but are not often
seen as credible as “science” when seeking justifications for “knowledge” in
Western, industrial societies. Research in the cognitive sciences and behavioral
economics has recently been investigating intuition as a means of decision-
making. Daniel Kahneman (a psychologist) received the Nobel in economics
for work in cognitive processes and intuition in economic decisions. However,
most discussions of methods in science place primary emphasis on inductive
and deductive processes.
51
3.5.1 Deductive Reasoning
3.5.1 DEDUCTIVE REASONING
Aristotle (384-322BCE) is usually credited with formalizing syllogistic or
deductive reasoning. Deductive reasoning is a process that starts with a set of
premises (or a priori truths) or general principles and through rules of logic,
“deduces” a conclusion about a specific case. There are usually two premises:
a major premise and a minor premise. If the general principle or major
premise were that all the water in the lake was safe to drink, then deductive
reasoning would conclude that a specific glass of water from the lake (the
minor premise is the water is from the lake) is safe to drink. The internal logic
could be correct but if either of the premises were false, correct deductive
logic would not yield true conclusions.
3.5.2 INDUCTIVE REASONING
Francis Bacon (1561-1626) is credited with formalizing inductive
reasoning. J.E. Creighton argues that Bacon’s N
ovum O
rganum was to replace
Aristotle as the preeminent guide to the process of acquiring knowledge.
“Bacon did for inductive logic what Aristotle did for the theory of the
syllogism. It is of course, incorrect to say, as has sometimes been said,
that Bacon invented the inductive method of reasoning. … What Bacon
endeavored to do was to analyze the inductive procedure, and to show
what conditions must be fulfilled in order that truth may be reached in
this way.” (Bacon, pps vii-viii)
Inductive reasoning is the process of inferring information from empirical
observations. If several glasses of water were taken from a lake and each
glass of water was shown to be safe to drink, it might be “inferred” that the
water in the lake is safe to drink. Because all the water in the lake was not
(and possibly could not) be tested there is some probability that all the water
in the lake is not safe to drink. Empiricism is rooted in the inductive process
52
3.5.2 Inductive Reasoning
and is based on empirical observations. Statistical inference is an application
of the inductive method.
While inductive methods are useful, there are pitfalls to avoid. Observations
might be incomplete or the interpretation of the observation(s) could be
incorrect. The selection of which phenomena to observe and the sequencing of
the “facts” can alter the conclusions reached. The application of inference and
inductive methods requires judgment and caution in the interpretation of data.
3.5.3 ABDUCTIVE REASONING
Abduction is a creative process from which hypotheses arise. Abduction is
similar to induction. The differences are that abduction is less formal process
that consists of a combination of intuition, experience, observation, deductive
reasoning and generates hypotheses which could be wrong. Abduction is the
insight that occurs with less conscious formal reasoning than either induction
or deduction.
It is the purpose of inductive and deductive reasoning to test the
hypotheses that emerge from the process of abduction.
3.6 EPISTEMOLOGY AND ECONOMIC METHODOLOGY
Epistemology is the study of the nature and limits of knowing. Economists
are confronted with an ocean of facts and data that are reputed to support a
plethora of theories and laws that purport to be the “truth” about economic
behavior. Any discipline, whether it is economics, physics, biology or …,
advances because someone questions the received wisdom: both extensions
of ideas and new ideas that are created as reactions against result from
questions about the received wisdom. If a scientist, economist or practitioner
of any discipline has the “truth,” their only task is to make sure others accept
53
3.6 Epistemology and Economic Methodology
that “truth.” A bit of humility about what one thinks they know is not a bad
thing. A quick survey of some of the basic ideas in epistemology provides an
enlightened humility.
3.6.1 A TAXONOMY OF KNOWLEDGE
Joel Mokyr classifies knowledge as propositional and prescriptive
knowledge. Mokyr, an economic historian, relates the problem of human
knowledge to economic growth and the economic problem. Propositional
knowledge is “.. knowledge (that is to say beliefs) about natural phenomena
and regularities.” (Mokyr, p 4) Prescriptive knowledge is instructional or
knowledge about techniques about how to do something. (ibid)
3.6.1.1 PROPOSITIONAL KNOWLEDGE
In Mokyr’s taxonomy, propositional knowledge (Ω) can take two forms.
He describes these as (1) “the observation, classification, measurement, and
cataloging of natural phenomena.” And (2) “the establishment of regularities,
principles and ‘natural laws’ that govern these phenomena and allow us to
make sense of them. ” Mokyr’s characterization of propositional knowledge is:
“Science, as John Ziman has emphasized, is the quintessential form of
public knowledge, but propositional knowledge is much more: the
practical informal knowledge about nature such as the properties of
materials, heat, motion, plants and animals: and intuitive grasp of basic
mechanics (including the six ‘basic machines of classical antiquity: the
lever, pulley, screw, balance, wedge and wheel): regularities of the
ocean currents and the weather: and folk wisdom in the ‘apple-a-day-
keeps-the-doctor-away’ tradition. Geography is very much a part of it:
knowing were things are is logically prior to the instructions of how to
go from here to there.” (Mokyr, p 5)
He argues that for the economic historian what matters is the collective
knowledge of what society, as a whole, knows (the union of all statements of
such knowledge). Confidence and consensus about knowledge as well as
54
3.6.1 A Taxonomy of Knowledge
access to and transmittal of that knowledge is of great importance to how
propositional knowledge is used. Mokyr characterizes the development of new
propositional knowledge as “discovery, the unearthing of a fact of natural law
that existed all along but was unknown to anyone in society. ” (Mokyr,, p 10)
3.6.1.2 PRESCRIPTIVE KNOWLEDGE
Prescriptive knowledge (λ) is the knowledge about how to do something:
it is technique or instructional knowledge. This prescriptive knowledge is
defined as “sets of executable instructions or recipes for how to manipulate
nature. ” (Mokyr,, p 10) The addition to this prescriptive knowledge is called an
“invention.” Prescriptive knowledge is not right or wrong it is successful or
unsuccessful. Mokyr argues that the industrial revolution and the associated
economic growth began when prescriptive knowledge came to be based on
proportional knowledge. Individuals can learn to do things without knowing
why they work. Once you know why techniques (prescriptive knowledge)
work, (propositional knowledge), it is easier to invent improvements to old
techniques and develop of new ones.
3.6.1.3 AN EXAMPLE
Knowledge about baking includes an understanding of the effects of
altitude, leavening, moisture, temperature, gluten and a host of other
phenomena on cakes. This knowledge is propositional knowledge. A cake can
be baked by someone in San Francisco with a recipe (prescriptive knowledge)
and no knowledge about the effects of altitude on cakes. The recipe will work
as long as person doesn’t try to bake a cake in Santa Fe, NM (elevation 7200
feet). To modify the recipe so it will work at the new elevation requires
propositional knowledge. The development of new recipes (λ) requires some
proportional knowledge (Ω).
55
3.6.2 Brief Survey of Epistemology
3.6.2 BRIEF SURVEY OF EPISTEMOLOGY
Karl Popper [1902-1994] is the primary architect of falsification as a
method of science. In his T
he L ogic of Scientific Discovery , 1934, he outlines
the basic approach taken in what is called the scientific method. He proposes
that scientific knowledge grows through a process of making hypotheses abut
the nature of problems and the falsification or testing of those hypotheses.
Popper argues that it is the duty of every scientist to try to disprove or reject
his or her hypotheses. If a hypothesis cannot be rejected by empirical
evidence, it may be retained as “probably true.” All knowledge then is
probabilistic: it has not yet been falsified. The process is subject to what
statisticians call Type I and II (or alpha and beta) errors. Type II errors occur
when a false hypothesis is accepted as “true.” When a “true” hypothesis is
rejected as false a Type I error has occurred.
Thomas Kuhn [ The Structure of Scientific Revolutions, 2cd ed, 1962,1970]
offers another explanation for the evolution and change of scientific thought in
the “hard sciences.” His explanation is often applied to economics and social
sciences. Kuhn used the concept of “paradigms” and paradigm shifts to explain
the process. The term, paradigm, is often used and abused in discussions.
Kuhn’s approach is essentially a “truth by consensus” which is contained in
the paradigm. This paradigm (and its associated “truth by consensus”) is
practiced until there are “anomalies” or problems that the existing paradigm
cannot explain. Then an alternative paradigm with greater explanatory powers
replaces it. He argues that a science operates within a paradigm. This
paradigm is characterized by,
• the “community structure of science’
• or the “disciplinary matrix” which consists of symbolic generalizations
(deployed without question),
• shared commitments to a set of beliefs and a set of values.
56
3.6.2 Brief Survey of Epistemology
The members of the science use this paradigm to resolve anomalies. When
an anomaly of major significance or a large number of anomalies cannot be
explained, the paradigm must be questioned and a new paradigm for that
science developed. In this manner “science progresses.”
Imre Lakatos’ method is expressed in his book, Proofs and Refutations,
[Cambridge University Press: Cambridge, 1976]. Lakatos’ approach, while in
the tradition of one of his teachers, Karl Popper, is critical of both Popper and
Kuhn. He advocated a more sophisticated form of falsification of “groups of
theories” and combined it with “scientific research programmes (SRP’s)” which
were more specific than paradigms. Lakatos’ SRP consists of two elements,
the “hard core, protective belt” and the “positive heuristic.” (Pheby, John,
Methodology and Economics: a Critical Introduction, M.E.Sharpe, 1988,, p 56)
The hard core is constructed of “basic axioms and hypotheses” that are
accepted without question and is used as a defense mechanism. The positive
heuristic is the body of theories and problems that drive the research
programmes. (Pheby, p 56)
Kuhn’s approach can be contrasted with that of Karl Popper and Imre
Lakatos. Popper saw the advancement of knowledge as the result of the
falsification of testable hypotheses. Those hypotheses that were not disproved
were accepted as “probably true.” Lakatos took the middle ground. Rather
than falsifying a hypothesis or the whole paradigm, he felt that the process
was based on “scientific research programs.” A school of economic thought
may represent a paradigm (in a Kuhnian sense) or a scientific research
program (in a Lakatian sense).
A more extreme view is expressed in Paul Feyerabend’s book, Against
Method (Verso: London, 1988, originally published by New Left Books, 1975).
He advocates an approach to science that has been called “theoretical
57
3.6.2 Brief Survey of Epistemology
anarchism.” Feyerabend argues that the “success of science cannot be used as
an argument for treating yet unsolved problems in a standardized way” and
scientific achievements can “be judged only after the event.” (Feyerabend, p
2) Feyerabend’s approach to the methodology of science is radically different
because of his objectives. He claims his purpose is “humanitarian not
intellectual” in that he wants “to support people not advance knowledge.” He
is “against ideologies that use the name of science for cultural murder.”
(Feyerabend, p 4]) While he does not disavow the title of “theoretical
anarchist,” he does provide insights into the evolution of science and
knowledge. Feyerabend summarizes some of his insights:
“Neither science nor rationality are universal measures of excellence.
They are particular traditions, unaware of their historical grounding.”
(Feyerabend, p 231)
“Yet it is possible to evaluate standards of rationality and to improve
them. The principles of improvement are neither above tradition nor
beyond change and it is impossible to nail them down.” (Feyerabend,
p 248)
“Science is a tradition among many and a provider of truth only for
those who have made the appropriate cultural choice.” (Feyerabend, p
256)
“The entities postulated by science are not found, and they do not
constitute an ‘objective’ stage for all cultures and all of history. They are
shaped by special groups, cultures, civilizations: and they are shaped
from a material which depending on its treatment, provides us with