Chapter 4
Demarcation and the philosophy of science
I remarked earlier that there has been an outbreak of new `sciences'
in the past century, many of which share few characteristics with the
old natural sciences, and that this is in part attributable to
scientism. We now need to consider in more depth the extent to which
it is justifiable to criticize this development. We need to address
whether there is a clear enough definition or understanding of what
natural science is to justify distinguishing it from non-science.
4.1 The History and Philosophy of Science discipline
With roots extending back to the early nineteenth century, but
blossoming into an identifiable independent discipline in the mid
twentieth, a whole field of academic study has grown up, known as the
History and Philosophy of
Science
.
Its practitioners study the
particular questions surrounding how science works, what scientists
have done and said, and what the influence of science is upon the
wider culture. The field encompasses historians who specialize in
scientific topics or personalities, philosophers who try to understand
and explain the logic, methods and epistemology of science, and
sociologists who study the social dynamics of science and scientists.
To give an impression of the relative size of the field, the largest
academic Department of History and Philosophy of Science (HPS) in the
United Kingdom, at Cambridge University
,
has 5 professors, while the
Department of Physics (the Cavendish) at Cambridge50 has 37. So the professional discipline of HPS, which at Cambridge is taught
as part of Natural Sciences, is rather modest in size. Many
universities don't have departments but have HPS scholars whose
professional homes are in other
more traditional departments. Actually, a lot
of the published analysis of science, particularly
what is written for the general public, is authored not by HPS
specialists but by working scientists themselves, and by journalists.
This is a situation not unlike, for example, political history, where
the books and articles are written by three main groups:
historians, the politicians themselves, and journalists. These
roles are not mutually exclusive, of course. Politicians can sometimes
be historians and journalists, and vice versa; but its is worth being
aware of the distinctive emphases of the various perspectives. I write
obviously as a working scientist, but in the present context with the
hope that what I write is accessible and interesting to a broad
audience. I also intend to account fairly for what the academic HPS
specialists have discovered, although I don't pretend to their
professional focus. I have had the unalloyed pleasure of
personal acquaintance with some distinguished historians and sociologists of
science, and the more ambivalent experience of being the object of study of
others. My first move is now to turn the tables on these folks for a
moment and draw some conclusions from simple observations of what they
do.
Recall that I have insisted on using the word science not as it was
classically derived, referring to any systematic knowledge, but as
referring to the study of nature
:
natural science, what was commonly
called natural philosophy
prior to the nineteenth century.
You might have lingering doubts about the appropriateness of this
insistence. And there might be experts in what I have called
non-scientific fields who are still smarting at the designation, who
feel in it an implied slight. If so, then I want to draw authority
from the History and Philosophy of science for my position. Not so
much in what HPS scholars say is science or non-science, but
what they actually demonstrate by their actions is the primary subject
matter of their field. If one takes any sample of the writings of a
spectrum of HPS academics or of their journals, it becomes immediately
obvious that what they mean by science is in fact natural
science. Here and there one finds a few studies of borderline fields
like psychology, for example, for which considerable hesitation is
justified in calling it a natural science; there is also a certain amount
of reflexive study of HPS concepts and development itself; but by far
the overwhelming bulk of what HPS specialists study is what the common
person would call science: physics, chemistry, biology, geology,
astronomy, physiology, and so on. I notice that from time to time
philosophy of science scholars refer, often with some detectable
awkwardness, to `social science'
.
They might, apparently somewhat
reticently, grant to social science the title. But even if that lip
service is rendered, it is almost vanishingly rare that they would
actually study the field of sociology as scientific knowledge, or
instance sociological discoveries as examples of science at
work. So this is my first conclusion from observation of the HPS
field: it supports by its actions the predominant usage of `science'
to mean natural science, and not just knowledge of any systematic type.
4.2 Scientism, sociology and socialism
This state of affairs would undoubtedly surprise the figures of the
early nineteenth century who offered analyses of scientific methods
that they sought to apply not only to nature but to society. In doing
so they addressed the philosophy of science, but also helped to found
the discipline of sociology
, and through their more
indirect influence were instrumental in the birth of various political
movements and theories, such as Marxism. Of these figures, Auguste
Comte
(1798-1857), the founder of what became
known as Positivism
, is undoubtedly the most
notable.
In a study of some of these sociological figures by
Nobel-prize-winning economist F. A. Hayek,
The Counter-Revolution of Science (1952), the English word
scientism
acquired (from French antecedents) the
derogatory implication that is widespread today. At the beginning of
the first part, entitled "Scientism and the study of society", Hayek
delineates scientism thus
During the first half of the nineteenth century a new attitude made
its appearance. The term science came more and more to be confined to
the physical and biological disciplines which at the same time began
to claim for themselves a special rigorousness and certainty which
distinguished them from all others. Their success was such that they
soon came to exercise an extraordinary fascination on those working
in other fields, who rapidly began to imitate their teaching and
vocabulary. Thus the tyranny commenced which the methods and technique
of the Sciences in the narrow sense of the term have ever since
exercised over the other subjects. These became increasingly concerned
to vindicate their equal status by showing that their methods were the
same as those of their brilliantly successful sisters rather than by
adapting their methods more and more to their own particular
problems. And, although in the hundred and twenty years or so, during
which this ambition to imitate Science in its methods rather than its
spirit has now dominated social studies, it has contributed scarcely
anything to our understanding of social phenomena, not only does it
continue to confuse and discredit the work of the social disciplines,
but demands for further attempts in this direction are still presented
to us as the latest revolutionary innovations which, if adopted, will
secure rapid undreamed of progress.51
I repeat. This is a definition and
assessment
of scientism in sociology not by
me or by a scientist but by an economist, a world leading practitioner
of one of the most distinguished social disciplines. Hayek's work
provides a valuable overview and insight into the scientisms of the
early nineteenth century from which, however, we shall have space to
mine only a few nuggets.52
Auguste Comte's apprenticeship, his early
thinking and writing, was under the tutelage of Henri
Saint-Simon
. Saint-Simon, an
adventurer and failed entrepreneur who at the age of 38
turned to intellectual pursuits,
was himself the architect and advocate of a scientistic approach to
the analysis of society that was to regard "our social relations
as physiological phenomena".
His intellectual ambitions were adorned, from the beginning, with bizarre
religious overtones. In the "Cult of Newton"
he proposed, apparently in all seriousness,
that a council consisting of three mathematicians, three physicists,
three chemists, three physiologists, three authors, three painters and
three musicians, elected by the whole of mankind, should become the
representatives of God on earth.53
Saint-Simon's ability to attract to his coterie dynamic
young thinkers, particularly from the
École Polytechnique, opposite which he lived,
provided him with an intellectual and literary depth that he lacked in
himself. That symbiosis was handicapped by many acolytes' eventual
estrangement because of Saint-Simon's unsupportive stance toward
liberty. Nevertheless, after Comte, a nineteen-year-old, expelled
polytechnicien, became his secretary in 1817, much of subsequent
Saint-Simonian doctrine bears the strong stamp of Comte; and Comte's
most influential
System of Positive Policy (1824) first
appeared as part of Saint-Simon's Catéchism des Industriels.
Comte's positive system, in this first incarnation, looks forward to the
disappearance of the liberty of conscience from society. Just as such
liberty has no place in astronomy, physics, chemistry, etc., it ought
not to do so in society. Its disappearance will
happen when politics has been elevated to the rank of a natural
science: the new science of Social Physics
.
This expectation is in accordance with Comte's most famous dictum, the
Law of Three Stages.
"Each branch of knowledge is necessarily
obliged to pass through three different theoretical states: the
Theological or fictitious state; the Metaphysical or abstract state;
lastly the Scientific or positive state", which of course is what
constitutes true knowledge.
Comte's classic work began in 1826 as a series of lectures and
extended to six volumes appearing between 1830 and 1842 as the
Cours de philosophie positive
. His task is
defined by the Law of Three Stages. Physics, chemistry and biology
have passed through the preliminary stages and thus become positive
knowledge. Comte is the one who is going to conduct this passage for
the knowledge of society. Without delving in detail into what
this transition was supposed to require, let the following
characterization and critique from Hayek's analysis of Comte's
scientism stand as a pithy summary of the program.
... since the habits of thought which man had acquired in
interpreting the actions of his own kind had long held up the study of
external nature, and the latter had only made real progress in
proportion as it got rid of this human habit, the way to progress in
the study of man must be the same: we must cease to consider man
anthropomorphically and must treat him as if we knew about him as
little as we know about external nature. Although Comte does not say
so in so many words, he comes very near doing so, and therefore one
cannot help wondering how he could have failed to see the paradoxical
nature of this conclusion.54
The sociological part of the Cours expanded to a size exceeding
the sum of the other three volumes which Comte devoted to the natural
sciences. The most crucial part of the sociological analysis,
according to Hayek is
... the attempt to prove the basic contention, which Comte, as a
young man of twenty-six, had expressed in a letter to a friend when
he promised to show that "there were laws governing the development
of the human race as definite as those determining the fall of a
stone." History was to be made a science, and the essence of all
science is that it should be capable of prediction. The dynamic part
of sociology was therefore to become a philosophy of history, as it
is commonly but somewhat misleadingly called, or a theory of history
as it would be more correctly described
.
The idea which was to inspire so much of the thought of the
second half of the nineteenth century, was to write "abstract
history," "history without the names of men or even people." The
new science was to provide a theoretical scheme, an abstract order
in which the major changes of human civilization must necessarily
follow from each other.
As Hayek, writing during the Second World War, was increasingly aware,
this program's eventual chilling product would be the rationale for
the twentieth century's most despotic totalitarian regimes. It makes
little difference whether the subsequent development set out to be
principled, left-wing, academic, and intellectual, or whether it was
employed by manipulative right-wing demagogues to add verisimilitude
to their populist mythologies. The scientistic dehumanization of
history and the portrayal of society as as subject to inevitable laws
of dynamics became the suitably malleable base material from which
could be fashioned political rationalizations bearing the spurious
honorific of `scientific'. From today's perspective of a post-Soviet
and postmodern world, we can be thankful for the much reduced
persuasiveness of this scientistic sociology. But while scientisms are
still alive and well, as they undoubtedly are, it is salutary to
recall the consequences, however indirect, of some of the earliest
identifiable examples.
I have pointed earlier to the religious character of scientism, its
status as a comprehensive world-view, based on prior metaphysical
commitments, and as a guiding framework for understanding all of
existence and motivating moral decisions. A major objection of
secularist advocates to this characterization rests upon the absence
in scientism of the clerical hierarchic authority and public rituals
that characterize most theistic religions. Whether such features are
universal in religion is debatable. And whether scientism today
possesses them is also open to interpretation when one thoughtfully
considers the ways by which scientific orthodoxy is adduced and
enforced, or the ways that academic ceremonies, for example, mimic
those of organized religion. But regardless of the current state of
scientism, a remarkable feature of the early nineteenth century
scientisms was the attempt to embody them in explicit new religions,
complete with all the trappings of traditional faiths.
The last of Henri Saint-Simon's works published in his lifetime was
entitled New Christianity
and appealed
to a divine core of truth focused on "the improvement of the moral
and physical existence of the poorest class". It reflected his view,
articulated a few days before he died, that "The Catholic System was
in contradiction with the system of sciences and of modern industry;
and therefore, its fall was inevitable. It took place, and this fall
is the signal for a new belief which is going to fill with its
enthusiasm the void which criticism has left in the souls of
men".55
His followers, led by Barthélemy-Prosper Enfantin,
set out to found
the organized religion of this new belief. Their efforts began with a
short-lived journal, moved into what might be considered a `house
church' hosted by Hippolyte Carnot (brother of Sadi, the discoverer of
the Carnot cycle), and progressed to public lectures on the
Doctrine de Saint-Simon, and thence to a "Family over which
Enfantin and Bazard presided as the two Supreme Fathers - new popes
with a college of apostles ... "56,
services, public confession of sins, itinerant preachers, and the
founding of local centers throughout the country. It ended in a move
to a monastic community complete with menial labor and vows of
celibacy.
Auguste Comte
quickly disassociated himself in
the 1820s with the Saint-Simonian religion. But in his own later
writings, the Systeme de politique positive (1851-54), and the
Catéchism positiviste (1852) he is concerned to move beyond
intellectual analysis to "moral regeneration"
which is emotional, imaginative, and subjective. To do so led him to
found his own new Religion of Humanity
complete with four-fold
hierarchic priesthood, and a high-priest, who was of course Comte
himself. Its public worship was to be practiced through commemorative
acts in celebration of our beloved dead. Comte's obsessive
prescriptive detail, which led many even of his supporters to allege
mental instability, governs, for example, the number and length of
daily private prayers, his nine personal sacraments, eighty-one annual
festivals, the saints, the icons to be used in Positive churches, and
the stipulation that they should all face towards the source of their
enlightenment: Paris.
The Religion of Humanity gained a significant following even
outside France. Although in England the main London church never had
more than 137 contributors, and there were never more than six or
seven English congregations outside London, it did appeal to some
influential figures. John Stuart Mill
for example, whose System
of Logic (1843) brought Comte's sociology to the attention of the
English public, gave support which, though qualified by his distaste
towards the cultic details he regarded as "ridiculous", included
financial support. And novelist George Eliot
and her partner Henry
Lewes attended Positivist church services, gave financially, and
personally promoted Positivism.57
4.3 Scientific method and demarcation
For much of the twentieth century philosophers
generally, and philosophers of science particularly, sought mightily
for methodological descriptions or definitions of science. These
efforts can be thought of as being of two types. The more ambitious is
to try to identify and explain the methods that science uses to obtain
its knowledge. One such method is Induction
. Francis
Bacon's analysis of Induction is a large part of his claim to be the
father of modern science. But in today's terminology, his ideas were
an early example of the philosophy of science. Indeed, since Bacon
practiced little or no science himself, it is better to consider him
the father of the philosophy of science, than of science itself.
Subsequently Induction has been examined logically and historically,
in the hopes of being able to articulate a more or less well defined
procedure, or set of procedures, that science uses, or should use, to
ensure that it discovers reliable descriptions of the
world. Nineteenth century philosopher John Stuart Mill, whose `Mill's
Canons'
are still considered an important encapsulation of inductive
methods, considered Induction the foundation of all knowledge,
including the knowledge of logic itself. His view was that the rules
of logic are themselves discovered by Induction. In the twentieth
century, the logical strengths and weaknesses of Induction, its
presuppositions, and the ways that it has in fact been used by
scientists have been extensively studied. Other logical and
epistemological schemes have also been proposed and subjected to
similar critical evaluation, as we shall shortly discuss.
The second type of definition of science is less ambitious. It takes
as its objective not a comprehensive description of science's methods,
but more modestly a criterion or set of criteria that distinguish
science from non-science. This would be a solution to what is called
the problem of demarcation
:
demarcating the boundaries between science
and non-science. Obviously if one were successful in the more
ambitious program of comprehensively describing science's methods,
that would provide a solution to the problem of demarcation. Yet one
can imagine a solution to demarcation that does not incorporate a
solution of the more ambitious methodological challenge.
To give away the punch line, the current opinion in philosophical
circles is that both of these programs has failed. Before discussing
more fully this apparent failure however, I want to point out that it
gives rise to a paradox. It is this. Despite having concluded that
there is no satisfactory working definition of what science is, the
History and Philosophy of Science has not collapsed and vanished as an
academic field. It is tempting to describe colloquially the apparent
failure of demarcation
by concluding
"HPS does not know what it is talking about". The conclusion I
actually draw from the paradox that HPS still seems to be alive and
well is different. It is that in fact HPS does know what it
talking about, but that it is unable to say what it is talking
about. In other words, there actually are some intuitive ways by which
science is identified, as evidenced by the pretty clear boundaries of
the topics that HPS does actually study, but that, at least thus far,
these intuitions have not been sufficiently clear-cut or logically
explicable to survive the rigors of HPS criticism.
What brought a new perspective to the philosophy of science in the
late nineteenth century is generally thought to be the modern
understanding of logic. George Boole
is best known today as the
progenitor of Boolean Algebra, the symbolic treatment of logic, in his
An Investigation of the Laws of Thought on Which are Founded the
Mathematical Theories of Logic and Probabilities (1854). Boolean
Algebra is an embodiment of the rules of logic, abstracted (from the
syllogism, for example) into an algebra, in which statements about sets
are represented by symbols and are subject to combining operations
such as Or, And, Not, etc. Actually, in Boole's own analysis, the
rules of this algebra are not quite in their modern form, which was
developed by others in the succeeding 30 years. The rules are somewhat
comparable to, but subtly different from, the high-school algebra of
numbers we all know. Boole's legacy was not widely known except to
mathematicians and logicians until the 1930s, when it was realized
(notably by Claude Shannon
) that these laws are precisely what is
needed to analyze the operation of combinations of switches and
relays. Today, because of the importance in electronics of digital
circuits, which fundamentally consist of electronic switches, no
engineering undergraduate education can be considered complete without
at least a passing acquaintance with Boolean Algebra. But, to return
to our topic, in the late nineteenth century the development of
symbolic formalization was carried further, and with even higher
ambitions, by Friedrich Frege
. He sought to demonstrate that
arithmetic is a branch of logic in his Basic Laws of Arithmetic
(1893). This path was pushed to what was hoped to be a logical
conclusion by Bertrand Russell
and Alfred North Whitehead
, in their
Principia Mathematica (1910-13), whose aim was to derive the whole of
mathematics from well-defined axioms and symbolic logic.
The undoubted clarification of logic itself through these symbolic
mathematical developments, and their apparent58 success in establishing a firm logical
foundation for mathematics, was of great significance for the
philosophy of science. It removed much of the mystery from classical
logic, for anyone who cared to master the mathematical methods; and it
offered a rigorous representational model, an example or paradigm,
which it appeared might be extensible to the whole of
science.59
Logical Positivism
was so called mostly because of its
ambition to fill this role. It asserts that there are only two types
of statement that are meaningful. First, Analytic statements,
including logic and mathematics, are true, in a sense by
definition. Second, Empirical statements must be verified
empirically; they include, and are best represented by, the statements
of natural science. Any other types of statement, especially
statements that cannot in principle be experimentally verified, are
nonsense - not just false, but meaningless. Thus Logical Positivism
not only reinforced the ambitions of the old Positivism to turn all of
knowledge into positive science, but it also proffered a definition of
what would qualify as science, in effect a demarcation, and ruled out
definitively the cognitive value of any undertaking that failed to
conform to its standard. If Logical Positivism were right, then every
discipline had to be reinvented in its terms. As
Rudolph Carnap, the
best known of the Vienna Circle
that founded Logical Positivism, put
it "To pursue philosophy can only be to clarify the concepts and
sentences of science by logical analysis. The instrument for this is
the new logic."60
Passing over the elementary, but none the less telling, critique that
Logical Positivism is itself neither Analytic nor Empirical, and
therefore by its own lights is nonsense, its most serious difficulty
is perhaps the whole notion of verification. It rapidly became clear
that what counts as verification could not be adequately articulated,
and that in any case, acknowledged theories of natural science
frequently did not seem to qualify.
Falsificationism
(sometimes called Deductivism) is associated
primarily with
Karl Popper, whose Logik der Forschung (1934)
followed closely on the heels of the logical positivists, although its
English translation, Logic of Scientific
Discovery61, did not appear till
1959. Popper's position is critical of the positivists, but still
seeks a description of science's methods in terms of logic. Popper's
approach is based on the elementary observation that a universal
proposition such as "all swans are white" can not logically be
proved, no matter how many white swans we see, but can be logically
disproved by the observation of a single black swan. He
concluded that verification of universal propositions through
processes of induction and confirmation is not what science
does. Instead, the way science works is by systematic attempts to
falsify supposed universal laws. When a law successfully
survives many such attempts, when it passes the most stringent of
potentially falsifying tests, it is regarded as having thereby gained
strong corroboration
. Although Popper does not advocate a
thorough-going scientism like the logical positivists that dismisses
unfalsifiable statements as ipso facto meaningless; and he thus avoids
their worst excesses of self-contradiction; he does promote
falsifiability as a demarcation criterion, as the test of whether a
statement is or is not scientific. We shall see in a moment that
replacing verification with falsification, while it appeals to a
clearer deductive logic, has some serious problems of its own.
Logical empiricism
was the term Carl Hempel (1905-97)
preferred to describe his position, though he was part of the
positivist Vienna circle, and unlike Popper accepted Induction as a
process of verification. He was predominantly again a logician, and
an advocate of the analysis and identification of science with
logic. He is most famous for his
Raven Paradox62, which runs
approximately like this.
The
statement "All ravens are black" is logically equivalent to "All
non-black things are non-ravens". The observation of a black raven is
inductive evidence in support of "All ravens are black", and hence
of its equivalent, "All non-black things are non-ravens". The
observation of a non-black thing that is a non-raven, for example a
white shoe, is inductive evidence in support of "All non-black things
are non-ravens", and hence of its equivalent "All ravens are
black". But it appears unreasonable to regard the observation of a
white shoe as evidence that all ravens are black.
This paradox has proven extremely fruitful in the analyses of
Induction that it has provoked. It would be inappropriate here to try
to list or evaluate all the different types of resolution that have
been proposed. However, they all have in common that they invoke
something more than the operations of pure logic on the isolated
case. Either they quantify the degree of support by appealing to
presumptions about the wider context, or else, seemingly more
drastically, they abandon Boolean two-value logic, and appeal to
other more generalized `logics'. Let's illustrate with a couple of
examples of the appeal to external information. One can argue (in a
way that can be formulated quantitatively using
Bayes' theorem in
probability) that the weight of support given by the white shoe is far
less than that of the black raven because we know that ravens are a
tiny fraction of all the things there are, whereas the number of
non-black things is not. Or one can argue that it depends on one's
knowledge of the external structure of the universe whether in fact a
black raven observation supports the universal proposition all ravens
are black. Suppose we happen to know that we are standing in one of
two aviaries of which one aviary possesses just one raven (and it
is black) out of ten thousand birds, while the other aviary has one
hundred ravens, just one of which is white, out of ten thousand birds,
and we see a black raven among the thirty other birds in our field of
view. The best guess is then not that "all ravens are black" in our
little aviary universe, but actually that, because we see a raven at
all, we are probably in the one-hundred-raven aviary, in which not all
are black. The observation of a black raven in this situation is
evidence against the proposition that all ravens are
black. Although this example may seem contrived, it illustrates that
the evidentiary value of an observation, even whether it is for or
against a universal law, depends upon the structure of the `universe',
and our prior knowledge of it. The reason why we don't consider the
white shoe evidence that all ravens are black is that we intuitively
incorporate into our thinking our background knowledge of the
universe, which is not present in the purely logical analysis of the
statements.
Hempel's objective was to acknowledge and solve a serious challenge to
describing Induction, conceived as a purely logical process, as the
foundation of science's method. The predominant view today is that what he
succeeded in doing was to clarify the seriousness of the challenge to
such an extent that the whole logical positivist/empiricist program
became untenable.
Under-determination
is a label
that gathers together several critiques of simplistic views of the
relationship between theory and empirical data. At the beginning of
the twentieth century,
Pierre Duhem had made the important point
"... the physicist can never subject an isolated hypothesis to
experimental test, but only a whole group of hypotheses; when the
experiment is in disagreement with his predictions, what he learns is
that at least one of the hypotheses constituting this group is
unacceptable and ought to be modified; but the experiment does not
designate which one should be changed."63
There is no such thing as a
simple empirical test of a theory; "hypotheses are tested in
bundles". A theory is always tested in conjunction with lots of other
auxiliary hypotheses
which are necessary to render the results of the
theory into experimental form. Acts of observation are not
independent of theory, they are
`theory laden', at the very least
because even the techniques of observation rely upon theory. Therefore
there is no such thing as the unambiguous experimental falsification
of a theoretical hypothesis. When a result is observed which
contradicts deductions from a hypothesis, either the hypothesis is
false, or something in the rest of science used to deduce the
falsified result is erroneous, or the observation itself is
mistaken. Duhem believed that this meant there could be no such thing
as a Crucial Experiment
,
because theories are under-determined by the
experimental data. We have to choose not between two competing
hypotheses but among an infinite number of possible hypotheses. Duhem
illustrates this with reference to the experiment of Foucault on the
speed of light in water that helped to distinguish between the wave
and particle theories of light, saying "Light may be a swarm of
projectiles, or it may be a vibratory motion whose waves are
propagated in a medium; is it forbidden to be anything else at all?"
He believed not.
In his celebrated critique entitled The Two Dogmas of
Empiricism64,
W. V. Quine, in 1951, argued
that neither Logical Positivism's distinction between analytic and
empirical statements, nor the supposition that individual empirical
statements can be reduced to immediate experience were
supportable. Instead "... our statements about the external world
face the tribunal of sense experience not individually but only as a
corporate body." This echoes Duhem's assertion, but extends it to all
of knowledge, including logic itself: "The totality of our so-called
knowledge or beliefs, from the most casual matters of geography and
history to the profoundest laws of atomic physics or even of pure
mathematics and logic, is a man-made fabric which impinges on
experience only along the edges. Or, to change the figure, total
science is like a field of force whose boundary conditions are
experience. A conflict with experience at the periphery occasions
readjustments in the interior of the field."
Sociology of Science
in the form of
Thomas Kuhn's seminal The Structure of Scientific
Revolutions65, which first appeared in 1962,
gave the tottering edifice of logicisms the final shove that brought
their ambitions to ruins. However, these particular philosophies of
science could hardly be considered more than collateral damage from
the bombshell he dropped on the whole enterprise of systematizing
science's methods. Kuhn's perspective was based on an observational
sociology of science. He treated theorizing about logical structures
of scientific warrant as a fruitless form of armchair philosophy, and
set out instead to document and analyze what scientists actually
do. While both the details and the conclusions of his work have been
extensively criticized, many of the scientific traits he identified
rang true in the experience of scientists, and he gained such
credibility that no subsequent HPS work could ignore it. [Scientists
themselves can ignore it, and many have done so, just as they
have in large numbers ignored the whole of the History and
Philosophy of Science discipline; but that is another tale.] Kuhn
observed that, in contradiction to Popper, scientists don't in general
try to falsify their theories; they defend them. What's more, theories
aren't immediately considered defunct if a single example of a
falsifying observation arises. Instead, they are shored up by
adjustment of auxiliary hypotheses
.
This much was not really revolutionary, but Kuhn's more major point
concerned those rare but exciting moments of history when a
transformation occurs of our conceptions of the universe, or more
modestly at least, of a scientific discipline. Such revolutions
replace, for Kuhn, the picture of slow, steady, or stumbling,
self-correction, which is the more traditional image of science. He
regards `normal' science
as being the quiescent period of gradual
self-consistent development of the reigning
paradigm. It leads
eventually to an accumulation of explanatory strains arising from the
accommodation of perplexing experimental results or theoretical
dissatisfactions. Those strains continue to grow, but the paradigm is
not overthrown until it can be replaced wholesale by a new one that
relieves the strain by a more satisfying explanation of the results
that puzzled the prior one. And the process begins over. The formation
of paradigms, and the replacement of one paradigm by the other is, by
contrast with normal science, a discontinuous and logically
inexplicable revolution. In the end, science seems to be little more
than opinion, expert opinion granted, but still just opinion. There
is, in Kuhn's words "no standard higher than the consent of the
relevant community": a situation that has been colorfully
characterized as scientific mob rule.
So much so, that Paul Feyerabend
seemed to
many to represent an extrapolation of the same intellectual path when
he argued in Against Method66 (1975)
that there is no scientific method; that science is, and should be,
anarchic; and that theories are successful because of the
no-holds-barred rhetorical tactics of their advocates, not because of
any particular correspondence with reality, or conformity with
accepted epistemological schemes.
Feyerabend is completely explicit that his stand is motivated to a
large degree by a rejection of what he dubs the chauvinism of science,
"... let me repeat that for me the chauvinism of science is a much
greater problem ... Scientists are not content with running their own
playpens in accordance with what they regard as the rules of
scientific method, they want to universalize these rules, they want
them to become part of society at large ...". It is a charge of
scientism.
Later sociological studies of science have gone to extreme in
developing this theme, arguing that there is nothing special about
science, that its knowledge is no more certain than any other type of
knowledge, that its knowledge is culturally determined, and that
purely cultural analyses of science (e.g. feminist, or third-world
perspectives) might lead to sciences with different technical
content. We'll explore these postmodern viewpoints later, but I
remark here that scientists (and I mean natural scientists) generally
regard the notion that there might be a feminist chemistry (for
example), having different technical content from our current
chemistry, as utterly ludicrous.
Research Programs is the terminology
advocated in place of paradigms by
Imre
Lakatos, Feyerabend's contemporary and sparring partner in the
metaphorical bout of rationalist (Lakatos in the blue corner) vs
irrationalist (Feyerabend in the red corner). A Research Program
"... is a series of successive theories with their associated
auxiliary assumptions, instrumental procedures, etc." What
distinguishes the worthy from the unworthy Research Programs is
whether or not they are `progressive'
.
All right, but what is progressive? How is that to be measured or
assessed? Several plausible measures of progressiveness are offered by
the advocates of this viewpoint. Independent testability
arising from the ability to test auxiliary hypotheses independently,
"in different bundles"; Unification, meaning the
successful application of a small family of problem-solving strategies
to a broad class of cases; and
Fecundity, designating the
property of theories that open up new and profitable lines of
investigation.67
The vagueness of these `progressiveness' measures is not just my brevity
here. It is the outward evidence that the attempt to
define science in terms of mathematically precise logic has been abandoned.
4.4 Demarcation disputes
One might think these increasingly abstruse debates about the status
and identity of science an area of philosophical specialization
with little practical significance. However, these matters of
demarcation have been brought very much into the American public eye
in recent years by the role they play in battles about high-school
biology. The 1925 show trial of
John Scopes for teaching evolution in
Dayton, Tennessee, is surrounded by sufficient contemporary and
subsequent commentary - and drama - to fill several bookshelves. We
shall not expand it further here and will postpone discussion of
evolution itself to a later section. But it is appropriate to comment
on the influence of the ongoing legal skirmishes upon demarcation.
Although the guilty Scopes verdict was overturned on the absurd
technicality that the $100 fine exceeded the judge's authority,
anti-evolution law remained on the statutes
of Tennessee and several other US states. Text-books worked cautiously
around these laws until in 1963
a Pennsylvania court ruled that Bible reading and the Lord's Prayer
were not appropriate for public schools but that neutrality between
"those who believe in no religion [and] those who do believe" must be
observed. Then in 1968 the Arkansas Supreme Court struck down its
anti-evolution statute, and was followed in this action by Mississippi
two years later.68 Thereafter
repeated attempts by activists to introduce creationism into the
curriculum have repeatedly been overturned by the courts based on the
opinion that it advanced particular religious preferences and
entangled the state with religion, in contravention of the First
Amendment clause "Congress shall make no law respecting an
establishment of religion, or prohibiting the free exercise
thereof;". The logic by which it follows from this clause that
religion is to be effectively banished from public school curricula is
hard to discern, but that is not our present topic. The strategy
adopted by
`Creation Science' activists increasingly,
in the face of
these reverses, was to portray creationism as science and to
argue that, as such, it should be taught alongside evolution. A
statute worded along these lines in Arkansas was struck down in 1982,
after testimony from a host of expert witnesses. Judge Overton's
decision was widely reprinted, even in the journal of the American
Association for the Advancement of Science (AAAS), Science.
Louisiana's similar statute arrived by a tortuous legal route at the US
Supreme Court on 10 December 1987. Its defenders argued that it was
not religious but scientific. Seven of the nine justices were
unconvinced. Thereafter, the Intelligent Design
(ID)
movement69 went to even
further efforts to ensure that their ideas were free from religious
taint. A number of local school-board initiatives to include it in the
curriculum, or to mandate disclaimers weakening evolution, gained some
temporary advantage, leading to the celebrated case in
Dover,
Pennsylvania, 2005. Judge Jones ruling on that case, identifies ID as
"a religious alternative masquerading as a scientific theory"
and castigates the school board
for precipitating a pointless trial.
The key observation for us here, is that these high-profile legal decisions
hinge on the question of whether certain opinions and teachings are or
are not science. That this has become the deciding question is a
remarkable sign of the dominance of scientism in our
culture. Scientism leads to acrimonious arguments about whether
opinions are or are not science because the scientistic ethos gives
special status to science that it does not give to non-scientific
disciplines. The result is that the demarcation of what is or is not
science becomes not merely an academic philosophical discussion, but a
vital legal matter that decides practical questions of deep importance and
emotional significance in the minds of most of the American public.
It is impossible that even academic discussion should escape the
influence of this practical significance. In the first place, the
various legal verdicts have undergone strong scrutiny by commentators
sympathetic to the losing side (the creationists). There is plenty of
scope for the criticism that the verdicts are based on an
over-simplified understanding of what constitutes science. It is
in the nature of legal argument to require sufficient simplification
and crispness as to justify the decision.
In the process many of the
subtleties are inevitably lost. The critics are naturally led to
emphasize the difficulty of demarcation,70
because the broader and the
fuzzier are science's boundaries, the easier it is to argue that
creationism is inside them. One might have expected the opposite
influence upon those who support the verdicts: a preference to regard
the demarcation as easier or more definite. However, there does not
seem to have been a significant countervailing initiative among
philosophers of science to clarify demarcation; although one can
perhaps detect a trend away from the more extreme relativism of the
sociological critiques of science. Scientists themselves are generally
of the opinion that they know science without difficulty when they see
it.
4.5 A more modest characterization
Since the program to define science by uncovering its logical methods
has thus far failed; since every attempt to identify a process by
which one could even establish what is science and what is not science
is judged to have been found wanting; since, in fact, it has become a
celebrated cause that the problem of demarcation is unsolved; and
since this question has become a high-profile legal matter: I am in
dangerous waters. I am asserting that there are two identifiable
characteristics of science, reproducibility and Clarity.
Am I therefore claiming to have solved the demarcation problem?
No, I am not claiming to have solved the demarcation problem. What I
am observing is that, despite the difficulties that undoubtedly
exist in specific demarcation, there are in fact identifiable
characteristics of science. These characteristics don't provide
algorithms either for the practice or the identification of science,
but they are nevertheless truly part of science. They are not
exhaustive. They are not all the characteristics of science;
many others might be identified to differing degrees in different
branches of science; but reproducibility and Clarity are vitally
important ones. Moreover, and this is the most important point, these
relatively non-controversial characteristics on their own constitute a
substantial restriction of the scope of science. They don't seek to be
definitive enough to resolve all disputes about vexed questions of
identifying and demarcating science. They certainly don't resolve all
the logical or methodological puzzles about how science actually
works. But they are important qualitative elements that are, in actual
fact, part of the way that science operates. In short, I am not
setting out to provide a comprehensive solution of demarcation, but I
am claiming to be able to identify some characteristics of what any
solution must look like. Modest answers to parts of problems are
sometimes what one must settle for.
Another factor that I want to plead in mitigation of my apparent
hubris is that the difficulty of demarcation is substantially
amplified by scientism
.
What I mean is this. First, philosophically,
demarcation between science and non-science in the context of
scientism, is equivalent to the demarcation between sense and
nonsense, rationality and irrationality, knowledge and
superstition. If one accepts scientism, the belief that science is all
the knowledge there is, then that demarcation of meaningfulness is
what one is trying to solve. Many philosophical arguments are about
that demarcation. The demarcation of meaningfulness, of all knowledge,
is beyond my interests here. I have a much narrower intent which is to
identify (some of) those principles which characterize natural science
- regarded as a subset, not the totality, of all valid rational
knowledge. I think that this problem is soluble, at least in part, and
that one should not discount the identifiable characteristics of
natural science just because of failures of a wider program. Second,
politically, since scientism has embroiled the problem of demarcation
in high-profile legal questions that raise emotions on both sides, the
difficulty of demarcation is made significantly greater. People see
that practical decisions, about which they care very much, hinge on
questions of scientific demarcation. No wonder there is great
sensitivity about this issue from all quarters. But my whole aim here
is to repudiate the scientism that leads to the enhancement of these
difficulties. If, as I am saying, science is not all the
knowledge there is, then the weight that demarcation has to bear is
reduced to a scope that is both more manageable and less sensitive.
My ultimate defence is that I am content if necessary to regard the
characteristics of repeatability and Clarity as partial
definitions of what I mean by science. In doing so, I might be
ruling out some studies that claim to be, or perhaps even widely are
regarded as, science. If that happens, then I say, so be it. This is
not simply a semantic stratagem on my part. The characteristics I
have identified are possessed by those sciences that are responsible
for the high epistemological prestige and compelling explanations of
the natural world that we attribute to science. It is those natural
sciences which are being implied almost always when science's
authority is being cited. So if I have by definition ruled out some
disciplines that have a reasonable claim to being scientific in the
sense of systematic, I have ruled out only those that are irrelevant
to the most significant debates. I recognize that this will sound
chauvinistic coming from a physical scientist. But I insist once more
that any perception of chauvinism in this position arises from the
self-same scientistic viewpoint I am at pains to deny. I find it
utterly inconsistent, for example, that toward the end of his splendid
introductory book What is this thing called science?,
philosopher of science,
Alan Chalmers
criticizes philosophers who emphasize experiment,
observing that their consequent failure to provide a "universal
account of science" ... "could conceivably be avoided by identifying
science with experimental science, but this would hardly serve to
appease those who wish to call themselves political scientists or
Christian scientists, for example."71
Yet Chalmer's whole book has been
about natural science. I didn't notice a single one of his many
instructive illustrative exemplars of science being about
sociological, political, or religious topics. If political
scientists are offended, so what? Why should these folks have to be appeased?
Why should one suddenly regard exclusion of political science as a
disqualifying fault in an identification of science's topics,
characteristics, or methods?
Finally, and this seeks immediately to soften any harshness of my
previous impatient sentences, I need to be crystal clear that I have
no intention to discount or disparage academic disciplines that I
regard as not being science. That
political science
is not a science in the way
I mean it, does not change its scholarly or practical value. I do not
subscribe to scientism. I believe there is deep meaning, truth,
relevance, and insight in non-scientific studies pursued with
intelligence and rigor. But their merits have to be really their own,
not the reflected glow of a terminological anachronism.
In summary, then, the discipline of History and Philosophy of Science
does not have simple answers to the questions, what is or is not
science? or what methods does science use? Science cannot be, or at
least has not so far been, formalized into a purely logical
structure. But HPS, like science itself, nevertheless appears to have
intuition about what science is, which is definite enough to know what
it is talking about. Natural science is what HPS studies. Although
strict demarcation is fraught with peril, a peril greatly enhanced in
recent debates by a scientism that artificially inflates the stakes,
there are identifiable characteristics of science. Attempts to
turn other disciplines, especially social disciplines, into explicit
positive science, after the manner of the natural sciences, have a
long history - of failure.