| |
Feyerabend is critical of the views of Hempel and
Nagel, and he takes a fundamentally different view,
fundamental because Feyerabend advances his
pragmatic theory of observation in opposition to the
Positivist naturalistic view of observation.
This point of departure places Feyerabend in
the same company as Einstein, Popper and Hanson, all
of whom reject the Positivist separation of theory
and observation.
On the Positivist view observation statements
are the products of natural processes that supply
the observation language with its meanings.
Feyerabend on the other hand affirms an
artifactual theory of meaning, when in
"Explanation, Reduction, and Empiricism"
he bases his pragmatic theory of observation on the
distinction between nature and convention.
In his view this distinction implies,
contrary to the Positivist view that the
observational status of a statement must be
separated from its meaning.
Thus Feyerabend says that an observation
sentence is distinguished from other sentences of a
theory not by its meaning content but by the cause
of its production, by which he means that its
production conforms to certain behavioral patterns.
His pragmatic theory of observation gives
Feyerabend an alternative to any reductionist thesis
such as Nagel's.
He maintains that when a transition is made
from one theory to another theory of wider scope,
which Nagel calls the secondary and primary sciences
respectively, what actually happens is something
much more radical than the incorporation of an
unchanged theory into the context of the primary
theory, unchanged with respect to the meanings of
the secondary theory's main descriptive terms as
well as to the meanings of the terms of its
observation language.
What happens is not a reduction, but is the
complete replacement of the ontology and perhaps the
formalism of the secondary science by the ontology
and the formalism of the primary science, and a
corresponding change in the meanings of the
descriptive elements of the formalism of the
secondary theory, where these elements of the
formalism of the secondary theory are still used.
Feyerabend states that contrary to the
Positivist reductionist thesis, the replacement
affects not only the theoretical terms of the
secondary science, but also at least some of the
observational terms occurring in its test
statements. He opposes the Positivist thesis that a comprehensive theory
merely orders facts, and maintains that a general
theory has a deeper influence on thinking.
This deeper influence is the semantical
influence of the context of the primary theory on
the empirical statements and vocabulary of the
secondary theory. The consequence of the distinction between nature and
convention that separates observability and meaning,
is what Feyerabend calls the "contextual theory
of meaning".
This theory of meaning description implies a
wholistic approach in his view, because he says that
the contextual determination of meaning is not
confined to a single scientific theory or even to a
single language.
Thus the unit of language involved in the
test of a specific theory is not just the theory
taken together with its own consequences, but rather
is a whole class of mutually incompatible and
factually adequate theories.
This class is the context by which meanings
are to be made clear. Feyerabend's rejection of the
Positivist naturalistic causal theory of meaning and
his proposal of his conventionalist contextual
theory of meaning, lead him to attack two basic
assumptions that he finds in Nagel's theory of
reduction and explanation.
These assumptions are (1) deducibility and
(2) meaning invariance.
Meaning variance is one of the reasons that
deducibility is impossible, but in addition to
meaning variance, there are purely quantitative
reasons why deducibility is impossible.
In his treatment Nagel gave the reduction of
Galileo's physics to Newton's physics as an example
of a homogeneous reduction, one in which there is no
meaning change resulting from the reduction.
But Feyerabend says that there is a
quantitative deviation between the Galilean and the
Newtonian physics, an inconsistency due to the fact
that one and the same set of observational data is
compatible with very different and mutually
inconsistent theories.
This inconsistency that makes deduction
logically impossible, has two reasons. Firstly universal theories always make claims about phenomena
that are beyond those that have actually been
observed or that might be available at any
particular time; it is this characteristic that
makes them universal.
Secondly the truth of any observation
statement, such as a statement reporting a
measurement reading, can be asserted only within a
certain margin of error. The first reason allows for theories that differ in domains
where experimental results are not yet available.
The second reason allows for such differences
even in those domains where observations have been
made, provided that the differences are restricted
to the margin of error in the observations.
The principal reason that deducibility is
impossible in explanation and reduction of general
theories is the inconsistency produced by the
meaning variance, the semantical change resulting
from the change of context.
To illustrate this Feyerabend considers the
purported reduction of Aristotle's theory of motion
to Newton's theory.
In this case Newton's theory offers the same
quantitative measurements as Aristotle's, and there
is no quantitative inconsistency.
The reduction is achieved in the apparently
simple manner of equating the concept of impetus in
the Aristotelian theory with the concept of momentum
in Newton's theory. On Newton's approach the meanings of the descriptive terms in
the impetus theory are fixed by the procedures and
assumptions of the theory.
But Feyerabend maintains that the concept of
impetus as fixed by the usage established in the
Aristotelian theory of motion cannot be defined in a
reasonable way in Newton's theory, because the usage
involves laws that are inconsistent with Newtonian
physics. Thus
contrary to Nagel, the concept of impetus is not
explicable in terms of the theoretical primitives of
the primary science in a reduction, even if equating
impetus with momentum is proposed as a physical
hypothesis instead of an analytical one. Such a
physical hypothesis merely says that wherever
momentum is present, then impetus will also be
present, and the measurements will be the same in
both cases.
Feyerabend also finds meaning variance in the
purported reduction of phenomenological
thermodynamics to the kinematic theory of gases, the
heterogeneous reduction case considered in detail by
Nagel. He describes Nagel's view as the claim that
the terms in the statements which have been derived
from the kinetic theory with the help of correlating
hypotheses will have the same meanings that they
originally had within the phenomenological theory.
And he states that Nagel repeatedly
emphasizes that these meanings are each fixed by its
own procedures, that is by the procedures of the
phenomenological theory, whether or not the theory
has been or will be reduced to some other
discipline. Thus
the term "temperature" as fixed by the
established usages of phenomenological
thermodynamics, as Nagel says, is such that its
application to concrete situations entails the
strict nonstatistical law. Feyerabend states that the kinematic theory does not offer
such a concept.
There does not exist any dynamical concept in
the phenomenological law, while on the statistical
account fluctuations between two levels of
temperature is allowed.
He therefore says that the thermodynamic
concept and the kinetic statistical concept of
temperature are incommensurable, and that
replacement rather than incorporation or derivation
characterizes the transition from a less general
theory to a more general one.
Feyerabend notes that both he and Nagel say
that incorporation into the context of the
statistical theory changes the meanings of the main
descriptive terms of the phenomenological theory,
but he adds that this is double talk by Nagel,
because the law that has been reduced is no longer
the same law. He
says Nagel's view of change of meanings is somehow
supposed to leave untouched the meanings of the main
descriptive terms of the discipline to be reduced.
There is a sense in which Nagel's view
involves double talk. This double talk is not an inconsistency in Nagel's thesis,
but rather is a logical consequence of his
semantical thesis, the view that the terms in
science are equivocal and have multiple meanings.
But Feyerabend prefers to reject any such
equivocation that would permit semantical continuity
through the reduction.
Instead he prefers to retain the univocity in
the terms at any point in time, and to affirm a
change from one meaning of a univocal term to
another new one, even at the expense of a semantical
continuity in the empirical explications. Consideration of the nature of this semantical
discontinuity introduces the roles of inconsistency
and especially incommensurability.
The
Sapir-Whorf Hypothesis
In his "Explanation, Reduction, and
Empiricism" Feyerabend describes two ways in
which theories can be related to each other such
that meaning variance may occur.
Those two ways are inconsistency and
incommensurability.
Given two historically successive theories
denoted T
and T' respectively, the theory T
will differ from the theory T',
either (1) if T is inconsistent with T'
in the domain of deduced empirical laws where T and T' overlap, or (2)
if the set of empirical laws that follow from theory
T' will be incommensurable with those following from T.
When the relation is inconsistency, the two
theories are commensurable, which is to say
semantically comparable.
Feyerabend references Popper saying that the
new and superior theory T'
implies laws that are different from and superior to
those implied by theory T.
In this case the laws deduced from theory T'
correct and replace those deduced from T,
as occurred in the case of Newton's theory
correcting and replacing Kepler's and Galileo's
laws. When
theories T and T' are
incommensurable, however, they do not have any
comparable observational consequences.
It is not even possible to say that the
empirical laws that are deduced from one are
superior or inferior to those that are deduced from
the other. This semantic incommensurability is admitted by Feyerabend's
pragmatic theory of observation.
On this theory of meaning nature does not
determine the content of thought and therefore does
not guarantee consistency or even comparability of
meaning. Instead
the content of thought is a human artifact not
unlike any work of art, and there may result
differences between people's thinking that are so
fundamentally different that they may admit no basis
for comparison or common denominator; they may be
incommensurable.
In his "On the 'Meaning' of Scientific
Terms" (1965), reprinted in Realism,
Rationalism, and Scientific Method, Feyerabend
describes a theory and its predecessor as
incommensurable, if prior to the time the theory is
proposed, there exists no more general concept
having an extension that includes the extensions of
the concepts of the two theories. He considers Einstein's relativity theory to be
incommensurable with Newtonian celestial mechanics,
because prior to Einstein the Reimann metric did not
include time, and he says that this change in the
transition to Einstein's theory was drastic enough
to exclude common elements between the two theories.
He also considers quantum theory to be
incommensurable with classical physics, because
prior to its advent the conservation laws were not
applied to virtual states.
Later Feyerabend further elaborates on his
concept of semantic incommensurability by drawing
upon the Sapir-Whorf hypothesis and specifically
upon Whorf's thesis of linguistic relativity.
Both Kuhn and Feyerabend briefly reference
Whorf in their works published in the 1960's, and
Feyerabend's elaboration of his thesis of semantic
incommensurability is to be found in his Against
Method published in 1975.
But before turning to this 1975 work, a
summary of the Sapir-Whorf hypothesis is in order.
Benjamin Lee Whorf (1897-1941) was a cultural
anthropologist and linguist by avocation, who
received a BA degree in chemical engineering in
1918, and spent his career with an insurance company
eventually becoming Assistant Secretary, an officer
of the corporation.
He became interested in linguistics in 1924
and was almost completely self-educated in
linguistics except for some nondegree courses that
he took from Edward Sapir, a cultural anthropologist
and linguist at Yale University.
Sapir encouraged Whorf to study the language
of the Hopi American Indians, and he financed
Whorf's field studies.
These studies occasioned Whorf's formulation
of the Sapir-Whorf hypothesis, the thesis of
linguistic relativity for which Whorf is now best
known. Whorf wrote many articles, but few of those
that he submitted to academic journals were accepted
and published in his lifetime in spite of the
intrinsic merit of the papers.
A posthumous anthology of his writings titled
Language,
Thought and Reality was published in 1956 (ed.
Carroll, MIT Press). It may be said that there is an
earlier and a later, expression of Whorf's thesis.
The earlier statement made in the 1930's is
his thesis of cryptotypes or covert categories,
while the more mature statement is the explicit
statement of linguistic relativity made in
"Science and Linguistics" in 1940.
Whorf exemplifies the idea of the cryptotype
with grammatical categories for gender.
Gender may be manifested either by overt or
by covert indicators.
They are overtly manifested by morphemes,
which are formal markers that occur in such
languages as Latin or German.
They are covertly manifested in English by
what Whorf calls their "reactance", their
association with definite linguistic configurations
such as lexical selection, word order that is also
class order, or in general by some kind of
patterning. More
precisely, overt categories are those having a
formal mark that is present in every sentence
containing a member of the category, while covert
categories are all others, even those that are
marked nonphonetically but only in certain types of
sentences. And
he defines his idea of reactance as a special type
of rapport, an idea that is roughly equivalent to
the general idea of structure in language.
Rapport is the linkage between the elements
of language that enables these elements to have
semantical effect, and it is governed by what Whorf
calls "an invisible central exchange". This central exchange of linkage bonds is what gives rise to
the covert categories, or cryptotypes that are
submerged, subtle and elusive meanings corresponding
to no actual word, yet which have a functionally
important role in the grammar of a language.
Words of a covert category are not
distinguished by a formal mark but rather by a
semantical class, by an idea that gives the
grammatical class its unity, which is manifested by
common reactance. Semantically the covert category is what Whorf calls a
deep persuasion of a principle behind some
phenomenon, like the ideas of inanimation,
substance, force, or causation.
The thesis that language structure controls
thought, which Whorf sets forth in his theory of
covert categories, is central to his theory of
linguistic relativity. He locates his development of linguistic relativity in the
history of cultural anthropology in the lineage of
Franz Boas and Edward Sapir.
Boas had shown that a language could be
analyzed sui
generis, that is, without forcing upon the
language the categories of the classical tradition.
Then in 1921 in his book Language
Sapir inaugurated the linguistic approach to
thinking, demonstrating the importance of
linguistics to cultural anthropology.
According to Whorf comparative linguistics
now reveals that the background linguistic system,
the grammar of each language, is not merely a
sentence-producing instrument for voicing ideas but
rather is the shaper of ideas.
And this is the essence of his thesis of
linguistic relativity.
The human mind cuts up nature, organizes it
into concepts, and ascribes significance, because
men are parties to an agreement that holds
throughout the speech community, and that is
codified in their language.
Not all observers are led by the same
physical evidence to the same picture of the
universe, unless their linguistic backgrounds are
similar or in some way can be calibrated.
For Whorf's term "calibrated" one
is tempted to substitute Feyerabend's term "commensurated",
except that Feyerabend does not believe that
semantically incommensurable theories can ever be
commensurated.
Whorf further elaborates on his linguistic
relativity thesis in his "Language, Mind and
Reality" (1942).
In the context of a discussion of the Mantric
Art of India he distinguishes two great levels: the
realm or level of meaning or lexication, and the
higher and controlling level of patterning of
sentence structure that guides words which occur at
the lexical level and that is more important than
words. Lexication,
the partitioning of the whole manifold of experience
and the assigning of the parts to words, makes the
parts stand out in artificial and semifictitious
isolation. This
process of lexication is controlled by the
patterning function of sentence structure and thus
by the organizing at a higher level, where the
combinatory scheme occurs.
These patterns are not individual sentences,
but rather are schemes of sentences and designs of
sentence structure.
The patterns are manifested by using the
mathematical or grammatical formulas into which
words, values or quantities may be substituted. Each
language does this partitioning and patterning in
its own way, and each has its own characteristic
form principles that make consciousness a mere
puppet, whose linguistic maneuverings are held in
unsensed and unbreakable bonds of pattern.
These passages suggest similarities between
Whorf's view and Feyerabend's contextual theory of
meaning, save for the fact that Feyerabend does not
restrict the term "meaning" to a lexical
function.
As it happens, Whorf explicitly states in
several of his later articles that his thesis of
linguistic relativity applies to empirical science.
He views it as applicable not only because
science including mathematics consists of language,
but also because an awareness of the effect of
language on the foundations of thought will
facilitate what he describes as science's next great
march into the unknown.
He expresses regret that philosophers and
mathematicians do not even have apprenticeship
training in linguistics, and he states the opinion
that further development in logic will proceed with
the investigation of the structures of diverse
languages. Like
later philosophers, Whorf views the various
specialized sciences as different languages, because
he finds that there exist communication problems
among the researchers in the different specialties,
just as there are such problems among the speakers
of different natural languages.
He maintains that these communication
problems do not simply breed confusion about details
that the expert translator could resolve.
The problems are much more perplexing, since
the language of science is a sublanguage, which
incorporates certain points of view and certain
patterned resistances to widely divergent points of
view. These resistances not only isolate artificially the
particular sciences from one another, but they also
operate to restrain the scientific spirit from
taking the next great step in its development, a
step which entails viewpoints unprecedented in
science and involving a complete severance from
tradition. This
great episode will unify the diverse sciences, and
will be based on the discovery of the aspect of
language consisting of patterned relations.
The approach to reality through mathematics
as used in science today is merely one special case
of this. Whorf
proposed that there is a premonition in language of
an unknown and vaster world, which is quite
different from the world as it is currently
understood through the structure of the
Indo-European languages, which insist on
substantives. The
apparent necessity of substances is purely a result
of the Ayrian grammar. The logic of Aristotle is provincial, because it is based on
the ideology of substantives, while modern physics
with its emphasis on fields casts doubt on this
ideology. Whorf
prognosticates the emergence of a new type of
language for science that is even more universal
than that presently used, because it will be a
transcendental logic of relations of pure
patternment
Whorf was more prescient than he probably
knew. If
there is a language of pure patternment, it is the
mathematical statement of the modern quantum theory,
which does not translate unambiguously into the
substantive language of ordinary discourse.
Even the practice of scientific realism does
not resolve the issue of whether the electron’s
wave and particle aspects are instantiated as two
aspects of one and the same entity, as the
Copenhagen advocates maintain, or whether they are
instantiated as two separate entities, as Bohm
maintains, because mathematics does not contain
substantive syntactical categories.
The individual in mathematics is the
measurement instance and not the substantive entity.
Thus Hanson’s observation that the
mathematical expressions of the wave mechanics and
the matrix mechanics can be transformed into one
another does not support his thesis that such
transformability implies the correctness of the
Copenhagen interpretation.
And Bohm is correct in maintaining that the
wave-particle issue occurs in what he calls the
“informal” language and not in the formalism.
It is ironic that Feyerabend did not exploit
Whorf’s insights during the years that Feyerabend
was supporting Bohm’s hidden-variables
interpretation in opposition to Hanson’ defense of
the Copenhagen duality thesis.
|
|
