Vincent J. Curtis
4 Sept 15
Concerning the bleak
verdict on validity of psychology experimental results, an extract of my paper “A Basis for
a Military Science” one of the two foundational papers of Military Science that I
wrote, is presented below. The original
paper was published in the Canadian Army Journal Vol 13.3 (2010) pp 69-85.
“At its simplest, a theory is a proposition of causation,
indeterminate in validity and significance.
S causes P. Such a proposition
can exist on paper or in the mind. The
proposition calls attention to the existence of a specific relationship between
two things or events, S and P. The
conception of a theory may be expanded to include definitions of S and P, since
it may be necessary to explain what S and P are. In addition, since S and P are related in
some way, the relationship presupposes a common subject matter in which S and P
exist.
As a theory addresses more complex issues, it comes to
include definitions, axioms, and postulates concerning a subject matter. It may come to incorporate theorems, which
are subordinate functional propositions deduced ultimately from the axioms and
postulates. It lays down characteristic
problematic propositions, or problems, to be solved. These problematic propositions are what give
significance to data obtained by observation.
For example, Ohm’s Law of electrical resistance declares a relationship
among voltage, amperage and resistance.
Mass and time are not variables of significance to Ohm’s Law, but
voltage, amperage, and resistance are.
To report the mass of the equipment in an experiment testing Ohm’s law
is to reveal that the experimenter does not understand the problem at
issue. The existence of a theory or a
conceptual analysis is prerequisite to the reliability and accuracy of
scientific observation.
In scientific research, the data obtained must be
significant with respect to the scientific proposition at issue. By significant is meant relevant to the
general proposition and able to aid in the determination of probability. A second requirement is that the data be
reliable and accurate; it must be objective.
What differentiates scientific from historical research is that
scientific research is able to test the validity of the same proposition by
different investigators at different times and places. Scientific observations are made in terms of
the concepts of the science, which is why a theory is prerequisite. It is difficult to estimate the validity of
non-quantitative information. Its
accuracy and reliability can be estimated, but only crudely, by reference to
the intellectual competence and veracity of the reporter, or by reference to
his documents and other material.
Non-quantitative knowledge, descriptions,
narratives, and characterizations have no etiological significance for they do
not provide evidence relevant to formulae or co-variation. Etiological propositions, on the other hand,
need not be quantitative, for the interdependence of variables may not be a
matter of degree but rather an all or none type of relation.
The construction of a theory is a rational process of the
human mind which transforms the material it uses. A science grows by development of its theory
and improvements of technique for gaining evidence, but a science must exist
before it can grow. Scientific
observations are made in terms of concepts of the science, not in terms of
common sense knowledge. But it is not
enough merely to propose a theory. The
Law of Non-Contradiction and the Principle of the Unity of Truth require that a
theory comport with other things that we know.
The phlogiston theory of heat and the aether theory of the transport of
light waves through the vacuum of space both fell by the wayside because they
failed to comport with knowledge we were surer of than of them.
What is valuable about science is that it provides knowledge
of determinant validity and significance.
Validity means accuracy and reliability: the truth value of the
data. Significance means susceptibility
to systematic interpretation. There are
no absolute standards for grading validity and significance.”
“A science begins with a theory or analysis of a subject
matter. A science is a body of knowledge
organized in a special way: as a compendent set of propositions. The terms of these propositions are the
variables or the concepts of the science.
A variable of a science is a term which does not refer to particular
individual things or definite aggregates of individuals. Words which refer to classes and the names of
universal characteristics are variable symbols.
In military theory, the terms platoon, battalion, division, and
commander are such variables.
There are three essential characteristics of a scientific
proposition: (1) generality, it goes beyond the evidence; (2) determinant
validity, it rests upon definite evidence; and (3) formal character of the
proposition as a relation of variables.
A fourth characteristic of a scientific proposition following the
previous three is that it is a member of a set of compendent propositions. Expressed in these propositions and in the
relational and conceptual structure of them is the theory or analysis.
”
“A science can be classified into two types, the rational
and the empirical. The difference
between them is the difference between deduction and induction. A rational science is a certain kind of
analytical exposition. A rational
science is founded upon a rational base, which consists of definitions, axioms,
and postulates. A proposition of a
rational science is proved deductively through syllogisms which begin with the
general propositions of the science. A
proved proposition of a rational science is true, not probable. A science is said to be rational when its
propositions are not only compendent but also systematically ordered. Examples of rational sciences include the
various geometries, mathematical physics, theology, politics, and ethics. The general theories of tactics and strategy
were here presented as rational sciences.
Empirical science is a body of knowledge derived by
inference from observation. Its
propositions are not systematically ordered.
A proposition of an empirical science is established as probable (not
true) by the accumulation of empirical evidence relevant to the proposition of
the science, as Ohm’s law was. The
analysis employed in empirical science never goes beyond the construction of a
set of variables, and the empirical knowledge of causation is nothing more than
knowledge of the relations which obtain in a given set of variables. Because the validity of a proposition of an
empirical science rests upon the precision of its etiology, empirical science
is either exact or it is not a science.
A pure empirical science is a straightforward etiology. More than a scheme of classification, an
empirical science is an organization of classes that are interdependent and
related by cause and effect. Empirical
science has to be exact for deviations from the empirical law of the science
are due either to inaccurate measurement or to a failure of the law. If the measurement is accurate, the law does
not hold.”
“A particular science can exist in any degree of
organization and exhibit the characteristics of both types, and it is possible
for there to be both a rational and an empirical science of the same subject
matter.
Physics and chemistry provide clear examples of what a
science is. When Sir Isaac Newton sought
to explain planetary motion, by an act of intuition he came up with the idea of
gravity. He posited the existence of
gravity as a first principle, laid down his three laws of motion as axioms, and
from that basis with the input of empirical data concerning the movement of the
planets derived his law of universal gravitation, and then derived Kepler’s
laws of planetary motion. Newton’s laws
of motion apply to any body in motion, not simply a particular set of planets
orbiting a particular star. The laws of
motion can be confirmed empirically by observations, and in fact they provide
more information than can ever be confirmed by experiment. Finally, Newton’s laws of motion are
expressible in the form of equations whose variables (force, mass, length,
time) appear in more than one equation; that is, the laws are compendent. Of his method of exposition Newton said, “I
lay down the law and derive the phenomenon from it.” Newton’s work is a classic example of the
exegesis of a rational science which becomes empirical science when empirical
data are applied to the variables of the equations.
Newton’s empirical physics happens to be quantitative, that
is the results are expressed as numbers.
But science does not have to be quantitative. Qualitative chemistry is a purely
non-quantitative branch of the empirical science of chemistry. Qualitative chemistry is concerned with the
identification of the constituent elements, compounds, and functional groups in
an unknown sample, and the immediate results of a test in qualitative chemistry
are yes or no; and ultimately this and not-this. The relationships between variables are step-functions
rather than continuous functions.
The cooperation of theoretical analysis, observation, and
inference is the essential trait of empirical scientific method. The findings of scientific research are
reported in terms of descriptive knowledge, e.g. a table of data correlating
amperage and the length of wire. The
conclusions of research, however, are never descriptive knowledge but rather
are the products of inference, e.g. the formulation of Ohm’s Law. Conclusions are general propositions
established to a certain degree of probability.
The rational sciences of Ethics, Philosophical Theology,
International Politics, and military science are non-quantitative. Hence, they do not appear to be as
‘scientific’ as Newtonian physics. They
rely upon in the input of descriptive knowledge of affairs, which can be of
unknown validity, for the inferences of the science. Nevertheless, the activities which are the
subject matters of these sciences are interpreted and understood in terms of
the concepts of their theories. To
suggest that these disciplines are ideologies or are expressions of mere
opinion is to misunderstand the nature of knowledge, inference, and the meaning
of validity and significance.
An ideology is a theory believed in to an unwarranted
degree. An ideology fails the test of: “…do
the facts lend themselves to the interpretation of the theory put upon them,
and do the inferences follow with logical necessity from its premises…is the
theory consistent with facts and within itself?” To an ideologue, the idea is more important
than facts. An ideology mimics the form
of Newton’s method of exposition, but fails to deliver the substance.
A statistical survey which purports to show correlation
between two phenomena is example of what science is not. These studies are the staple of news stories
about the medical, environmental, and psychological disciplines. What these kinds of alleged studies are is a
degraded form of raw empiricism masquerading as the science of statistics, and
they appeal to a weakness of common sense: its tendency to infer a cause-effect
relationship upon data inadequate for such an inference. This weakness comes into play when evaluating
the efficiency of a means for attaining an end, such as the efficiency of a
strategic bombing campaign to gain Victory.
The logical fallacy behind statistical surveys that seem to
relate two phenomena is called post hoc
ergo propter hoc. Correlation is not
causation. Knowledge of the existence of
characteristics of events and things does not constitute knowledge of their relationship,
and to infer a cause-effect relationship on the basis of descriptive knowledge
is a fallacy as surely as night follows day.
A science begins with a theory, and is developed by skilled observers
who know what they are looking for. I
say observers in the plural because a characteristic of scientific work is that
it can be reproduced. A statistical
survey which is not formulated to confirm a particular theory of a cause-effect
relationship between an event and a thing, and which is not taken by set of
skilled observers who know precisely what they are looking for, and which
produces results that are not exactly the same as similar work done by others,
fails the basic tests of scientific validity and significance. Even if the data of the set collected is
perfectly valid, without a theory the results are without significance. At best, a study of this kind may suggest a
more specific and in depth experiment which isolates the variables concerned
simply in order to evaluate the significance of the results of the first study.
Common sense is not science.
Common sense is easily fooled by an apparent correlation between events
and things, especially when the correlation seems strong. Common sense answers questions regarding the
adaptation of means to ends by interpreting what it observes in terms of its
experience of the world. Common sense
knowledge is often adequate for practical purposes, but common sense is all too
ready to impute the success of an endeavour to the means employed, particularly
when success is only partial. As
practical problems become more complex, common sense finds it increasing
difficult to answer questions concerning adaptation of means to ends.
The question of the efficiency of means was raised
earlier. Efficiency, a question
concerning the adaptation of means to ends, concerns the relation of one event
to another, but knowledge of the existence and characteristics of events and
things does not constitute knowledge of their relationships. If one were to ask how efficient means ‘B’ is
in producing end ‘A’, it is not sufficient to reply with a description of means
‘B’ and to state how often end ‘A’ occurred after means ‘B’ was applied. That descriptive knowledge is in itself
insufficient to establish a causal relationship between B and A. To establish that, knowledge of a scientific
character, etiological knowledge, is required.
Just as science is the source of scientific knowledge,
common sense is the source of common sense knowledge. Common sense knowledge is not the common
possession of all men, or that men who possess it possess it to the same
degree. All men are not able to
significantly interpret descriptive knowledge, and the interpretation of all
men are not of equal value. An expert in
the field may not be one who possesses scientific knowledge for there may be no
science of that field and therefore no such knowledge. However, the expert is one who possesses
rather precise knowledge of his field gained by long study and observation, and
he can interpret his knowledge wisely.
This opinion of such an expert may possess
more significance than common sense knowledge.
Where scientific knowledge relates cause and effect, common sense
knowledge relates ends and means. Means
and ends are not synonymous with cause and effect. Common sense generalizations and opinions are
never organized into compendent sets of propositions. They are members of indefinite aggregates of
generalizations, or are isolated. The
Maxims of Napoleon are an aggregation of his common sense generalizations.
Descriptive knowledge is knowledge of particular events and
things, or their existence and characteristics.
It is not of the relations of these things and events to one another as
ends and means. Descriptive knowledge
consists of narratives, descriptions, or characterizations limited in reference
to particular things or events or to definite aggregates of particular events
or things. In contrast, a proposition of
scientific knowledge never has restricted reference to particular events or
things, or definite aggregates of particular things or events. A scientific proposition is always a general
proposition.”
From the AP report:
“Among the experiments that stood up was one that found
people are equally adept at recognizing pride in faces from different
cultures.”
What, pray tell, are the covariant variables in an
experiment attempting to correlate recognition with expressions in human
faces? Degree of recognition and facial
expression? How does one quantify those
things? What is this business of
different cultures?
Psychology, sociology, criminology, and the other soft “sciences”
have a long way to go before they are real sciences.
-30-