Logical positivism (logical empiricism, neo-positivism) originated in Austria
and Germany in the 1920s. Inspired by late nineteenth and early twentieth
century revolutions in logic, mathematics, and mathematical physics, it
aimed to create a similarly revolutionary scientific philosophy purged of
the endless controversies of traditional metaphysics. Its most important
representatives were members of the Vienna Circle that gathered around Moritz
Schlick at the University of Vienna (including Rudolf Carnap, Herbert Feigl,
Kurt Gödel, Hans Hahn, Karl Menger, Otto Neurath, and Friedrich Waismann)
and the Society for Empirical Philosophy that gathered around Hans Reichenbach
at the University of Berlin (including Walter Dubislav, Kurt Grelling, and
Carl Hempel). Although not officially members of either group, the Austrian
philosophers Ludwig Wittgenstein and Karl Popper were, at least for a time,
closely associated with logical positivism. The logical positivist movement
reached its apogee in Europe in the years 1928-1934, but the Nazi seizure
of power in 1933 marked the effective end of this phase. Thereafter, however,
many of its most important representatives emigrated to the United States.
Here logical positivism found a receptive audience among such pragmatically,
empirically, and logically minded American philosophers as Charles Morris,
Ernest Nagel, and W. V. Quine. Thus transplanted to the English speaking
world of "analytic" philosophy it exerted a tremendous influence--particularly
in philosophy of science and the application of logical and mathematical
techniques to philosophical problems more generally. This influence began
to wane around 1960, with the rise of a pragmatic form of naturalism due
to Quine and an historical-sociological approach to the philosophy of science
due mainly to Thomas Kuhn. Both of these later trends, however, developed
in explicit reaction to the philosophy of logical positivism and thereby
attest to its enduring significance.
HISTORICAL BACKGROUND
Kant and the neo-Kantians. Kant, in the positivists' eyes, had made a lasting
contribution to scientific philosophy--particularly in his rejection of
the possibility of super-sensible, metaphysical knowledge and his reorientation
of theoretical philosophy around the two questions "How is pure mathematics
possible?" and "How is pure natural science possible?" In
answering these question Kant developed his famous defense of synthetic
a priori knowledge--knowledge independent of sensible experience yet nonetheless
substantively applicable to the empirical world. For Kant, the mathematical
physics of Newton brilliantly exemplified such synthetic a priori knowledge,
through its reliance on Euclidean geometry and fundamental laws of motion
such as the law of intertia. Kant's theory of a priori faculties of the
mind--the faculty of pure intuition or sensibility and the faculty of pure
understanding--was then intended to explain the origin of synthetic a priori
knowledge and thus make philosophically comprehensible the possibility of
Newtonian mathematical physics.
After the intervening dominance of post-Kantian idealism, a number of German
speaking philosophers renewed the call for a scientific, epistemological,
and non-metaphysical form of philosophy. But these neo-Kantian philosophers
also had to face an important new challenge to the Kantian synthetic a priori:
the nineteenth century development of non-Euclidean geometry by Gauss,
Bolyai, Lobachevsky, Riemann, and Klein. Although some neo-Kantians attempted
to defend the uniqueness and a priority of Euclidean geometry nonetheless,
others--especially those of the Marburg School such as Paul Natorp and Ernst
Cassirer--aimed to generalize the synthetic a priori beyond its particular
embodiment in classical, Euclidean-Newtonian mathematical physics. This
latter tendency was similar in important respects to ideas the logical positivists
were to elaborate.
Nineteenth century predecessors: Helmholtz, Mach, Poincaré. These
three thinkers, through their efforts to comprehend the radical changes
sweeping through nineteenth century science, initiated a new style of scientific
philosophy later taken up and systematized by the logical positivists.
The changes in question included the rise of non-Euclidean geometry, the
formulation of the conservation of energy and general thermodynamics, and
the beginnings of scientific physiology and psychology. Hermann von Helmholtz
made fundamental contributions to all three areas. He based geometry on
the postulate of "free mobility" of rigid bodies, and, since
all classical geometries of constant curvature--negative, positive, or zero
(Euclidean)--satisfy this postulate, he opposed the Kantian commitment to
the a priority of geometry: whether space is Euclidean or non-Euclidean
is an empirical question about the actual behavior of rigid bodies. In
physiology, Helmholtz articulated a general principle of psycho-physical
correlation whereby our sensations correspond to--but are in no way pictures
or images of--processes in the external physical world. These processes
consist, in the end, of microscopic atoms interacting via central forces,
and, on this basis, Helmholtz developed his famous interpretation of the
conservation of energy.
Ernst Mach and Henri Poincaré can be seen as reacting, in diverse
ways, to Helmholtz. Mach attacked especially atomism and the idea of a
psycho-physical correlation between two incommensurable realms, and he
advanced a program for the unity of science based on immediately perceptible
"elements" or "sensations." The task of science consists
solely in seeking correlations among such elements (as in phenomenological
thermodynamics), and all dualistic and atomistic tendencies are to be purged
as metaphysical via historico-critical analysis. This Machian empiricism
exerted a decisive influence on the logical positivists. Poincaré,
on the other hand, influenced the positivists primarily through his philosophy
of geometry. He agreed with Helmholtz's emphasis on the free mobility of
rigid bodies but disagreed with Helmholtz's empiricism. According to Poincaré,
the idea of a rigid body is an idealization that cannot be straightforwardly
instantiated in the physical world. By freely choosing one of the three
classical geometries as, so to speak, a definition of rigidity, we then
first make it possible to carry out empirical investigations with real physical
bodies. Physical geometry is thus neither synthetic a priori nor empirical:
it is "conventional."
RELATIVISTIC PHYSICS
Albert Einstein's special (1905) and general (1915) theories of relativity
entered this volatile intellectual situation as a revelation. And the relativistic
revolution in physics directly stimulated Schlick, Reichenbach, and Carnap
to initiate a parallel revolution in scientific philosophy. All three thinkers
agreed that relativity--especially through the general relativistic description
of gravitation via a (four dimensional) geometry of variable curvature--definitively
refutes the Kantian idea that Euclidean geometry is synthetic a priori.
Moreover, relativity arises from critical reflection on the empirical significance
of spatio-temporal concepts in physics (in particular, the concept of simultaneity
and the concept of motion) and thus demonstrates the fruitfulness of Mach's
basic point of view. At the same time, however, through its use of sophisticated
abstract mathematics, relativity also illustrates the limitations of Machian
empiricism (according to which even mathematical concepts have an empirical
origin). All three thinkers therefore attempted to formulate an intermediate
position that would do justice to both Machian empiricism and the continued
importance of a priori, mathematical elements in physics. Poincaré's
concept of convention came to play a central role.
Schlick, Reichenbach, and Carnap first pursued rather different paths.
Whereas Schlick emphasized from the outset that the Kantian synthetic a
priori has no place at all in the new relativistic context, Reichenbach
and Carnap initially attempted to salvage important aspects of Kantianism.
Reichenbach began by distinguishing the idea of necessary and unrevisable
truth from the idea of necessary presupposition of a given scientific conceptualization
of nature. For Reichenbach, relativity refuted the former but embodied
the latter. Kant was right that the necessary presuppositions of Newtonian
physics included Euclidean geometry and the laws of motion. In moving to
relativistic physics, however, these are replaced by fundamentally new presuppositions.
We thus end up with a relativized version of the Kantian a priori (as constituting
the presuppositions of a particular theory). Carnap, by contrast, began
by distinguishing metrical from topological features of physical space.
The latter are indeed synthetic a priori as Kant thought (they even depend
on a kind of pure intuition), but the former--as general relativity has
shown--essentially involve the behavior of empirically given bodies. We
thus end up with a weakening of the Kantian a priori (from metrical to topological
features).
These early attempts to salvage aspects of the synthetic a priori did not
survive, however. For Schlick's view that relativity is simply incompatible
with Kant eventually won the day. Although the distinction between Poincaré's
conventionalism and Helmholtzian empiricism was not entirely clear (and
Reichenbach, in particular, preferred to associate his later viewpoint with
Helmholtz rather than Poincaré), both Reichenbach and Carnap soon
came to replace the Kantian notion of the a priori with Poincaré's
concept of convention. Yet this form of conventionalism (unlike Poincaré's)
was forged in the crucible of a revolutionary new physics and thus demonstrated
the vitality and relevance of a new philosophy.
LOGIC AND THE FOUNDATIONS OF MATHEMATICS
Whereas the positivists appealed to Poincaré's concept of convention
(as realized, so they thought, in relativistic physics) to give a new answer
to Kant's question concerning the possibility of pure natural science, they
appealed to modern developments in logic and the foundations of mathematics
to give a new answer to Kant's question concerning the possibility of pure
mathematics. There were in fact two distinguishable sets of developments
here. The formal point of view, typified by David Hilbert's logically rigorous
axiomatization of geometry, freed geometry from any reference at all to
intuitively spatial forms and instead portrayed its subject matter as consisting
of any things whatever that satisfy the relevant axioms. Geometry is rigorously
and a priori true, not because it reflects the structure of an intuitively
given space, but rather because it "implicitly defines" its subject
matter via purely logical--but otherwise entirely undetermined--formulas.
Mathematical truth, on this view, is identified with logical consistency.
The "logicism" of Gottlob Frege and Bertrand Russell, by contrast,
aimed to construct particular mathematical disciplines (especially arithmetic)
within an all-embracing system of logic. On this view mathematical disciplines
(like arithmetic) indeed have a definite subject matter about which they
express truths: namely, the subject matter of logic itself (propositions,
classes, etc.). As thus purely logical, however, such pure mathematical
disciplines express merely analytic truths and hence are in no way synthetic
a priori.
Hilbert's formal point of view was pursued especially by Schlick, who in
a sense made the notion of implicit definition, together with the associated
distinction between undetermined form and determinate (given) content, the
centerpiece of his philosophy. The logicist point of view, by contrast,
was pursued especially by Carnap, who studied with Frege and then was decisively
influenced by Russell. Indeed, Carnap was inspired by Russell's conception
of "logic as the essence of philosophy" to reconceive philosophy
itself on the model of the logicist construction of arithmetic. He began,
in Der logische Aufbau der Welt (1928), by developing a "rational reconstruction"
of empirical knowledge--an epistemology--within the logical framework of
Russell's and Whitehead's Principia Mathematica. By defining or "constituting"
all concepts of empirical science within this logic from a basis of subjective
"elementary experiences," Carnap's reconstruction was to show,
among other things, that the dichotomy between empirical truth and analytic/definitional
truth is indeed exhaustive.
Yet the logic of Principia Mathematica was afflicted with serious technical
difficulties: the need for special existential axioms such as the axioms
of infinity and choice. Partly in response to such difficulties, Ludwig
Wittgenstein asserted in his Tractatus Logico-Philosophicus that logic has
no subject matter after all: the propositions of logic are entirely tautological
or empty of content. Carnap eagerly embraced this idea, but he also attempted
to adapt it to the new, post-Principia technical situation--which involved
the articulation of the "intuitionist" or "constructivist"
point of view by L. E. J. Brouwer and the development of meta-mathematics
by Hilbert and Kurt Gödel. In Logische Syntax der Sprache (1934) Carnap
formulated his mature theory of formal languages and put forward his famous
"Principle of Tolerance"--according to which logic has no business
at all looking for true or "correct" principles. The task of
logic is rather to investigate the structure of any and all formal languages--"the
boundless ocean of unlimited possibilities"--so as to map out and explore
their infinitely diverse logical structures. Indeed, the construction and
logical investigation of such formal languages became, for Carnap, the new
task of philosophy. The concept of analyticity thereby took on an even
more important role. For this concept characterizes logical as opposed
to empirical investigation and thus now expresses the distinctive character
of philosophy itself.
THE VIENNA CIRCLE
Otto Neurath, Hans Hahn, and the physicist Philipp Frank initiated a discussion
group in Vienna, beginning in 1907, in which they considered a combination
of Machian empiricism with Poincarè's new insights into the conventional
character of physical geometry. Deeply impressed by Schlick's work on relativity
theory, they arranged (apparently with Einstein's help) to bring Schlick
to the University of Vienna in 1922 to take over the Chair in Philosophy
of the Inductive Sciences previously held by Mach. What we now know as
the Vienna Circle quickly took shape. Reichenbach, who had become acquanted
with Carnap through their common interest in relativity, introduced him
to Schlick in 1924. In 1925 Carnap lectured to the Circle in Vienna on
his new "constitutional theory of experience" and became assistant
professor under Schlick in 1926. The Circle then engaged in intensive discussions
of Carnap's epistemology and Wittgenstein's Tractatus. Wittgenstein's view
that all propositions are truth-functions of "elementary propositions"
was combined with Carnap's constitution of scientific concepts from a basis
of "elementary experiences" so as to create a new, logically rigorous
form of empiricism according to which all meaningful--scientific--propositions
are reducible to propositions about immediately given experience. And this
was articulated as the "official" philosophy of the Vienna Circle
in the famous manifesto Wissenschaftliche Weltauffassung in 1929.
Neurath was the driving force in thus turning the Vienna Circle into a public
philosophical movement. Trained in economics and the social sciences, Neurath
was extremely active politically as a scientific neo-Marxist. In particular,
he took the community of natural scientists as the model for a rationally
organized human society, and, on this basis, he advocated a reorganization
of both intellectual and social life from which all non-rational, "metaphysical"
elements would be definitively purged. In this sense, Neurath saw the philosophical
work of the Vienna Circle as a reflection of the wider movement for a neue
Sachlichkeit then current in Weimar culture--as typified, for example, by
the Dessau Bauhaus. As in the wider culture, this movement stood in philosophy
for a rejection of individualism in favor of the cooperative, piecemeal,
and "technological" approach to problems exemplified in the sciences,
and it was therefore particularly hostile to what was perceived as a return
to the metaphysical system-building of post-Kantian idealism by influential
German philosophers such as Martin Heidegger. Carnap was especially sympathetic
to Neurath's broader philosophical-political vision and clearly expresses
this vision in the Preface to the Aufbau. Schlick, by contrast, preferred
a more individualistic model of philosophy and resisted the idea of a "movement."
This divergence between a "left wing" and a "right wing"
of the Circle emerged in the sphere of epistemology in a debate over "protocol-sentences"
in the years 1930-34. At issue was the status of the basic propositions
or protocols in which the results of scientific observation are recorded.
It had initially appeared, in Carnap's constitutional system of the Aufbau,
that such propositions must express private, subjective sense-experience.
For Neurath, however, this view was inconsistent with the publicity and
intersubjectivity required by science. He therefore advocated a more naturalistic
conception of protocols as sentences accepted by the scientific community
as recording the results of observation at a given time. These sentences
must thus be expressible within the public and "physicalistic"
language of unified science and hence, like all other sentences, are in
principle revisable. Schlick was deeply shocked by Neurath's view--which
he took to represent an abandonment of empiricism in favor of the coherence
theory of truth. Carnap attempted, in typical fashion, to mediate the dispute:
at issue was simply a choice between two different languages in which to
formulate or rationally reconstruct the results of unified science. Although
Neurath's thoroughly intersubjective "physicalistic" language
(where, as Karl Popper emphasized especially, every sentence is revisable)
was clearly preferable on pragmatic grounds, Carnap held that this choice--like
every other choice of formal language--is in the end conventional. Empiricism,
in Carnap's hands, is itself framed by conventional and hence non-empirical
choices.
EMIGRATION, INFLUENCE, AFTERMATH
The rise of the Nazi regime set off a wholesale migration of logical positivists
to the English speaking world. Carnap, who had become professor at Prague
in 1931, moved in 1936 to the University of Chicago. Reichenbach, who had
fled to Istanbul in 1933, moved in 1938 to the University of California
at Los Angeles. (After Reichenbach's death in 1953 Carnap took over his
position at UCLA beginning in 1954.) Neurath, after leaving Vienna for
The Hague in 1934, fled for England in 1940--where he worked in Oxford until
his death in 1945. Friedrich Waismann fled for England as well, where he
lectured at Oxford from 1939. Philip Frank emigrated to the United States
(also from Prague) in 1938 and settled at Harvard in 1939. Karl Menger
took up a position at Notre Dame in 1937, and Kurt Gödel became a member
of the Institute for Advanced Study at Princeton in 1940. Herbert Feigl
went first to the University of Iowa in 1933 and then to the University
of Minnesota in 1940, where he founded the influential Minnesota Center
for the Philosophy of Science in 1953. Carl Hempel joined Carnap at the
University of Chicago in 1939 and, after teaching at Queens College and
Yale, settled at Princeton in 1955. (Schlick was murdered by a student
at the University of Vienna in 1936.)
The growth of philosophy of science in the United States was decisively
shaped by the work of Carnap, Reichenbach, and Hempel. Reichenbach influenced
especially the development of philosophy of physics through his work on
geometry, relativity, and the direction of time. Hempel published extraordinarily
influential papers on the logical analysis of explanation and confirmation
and thereby furthered the ideal of scientific philosophy first articulated
by Carnap. Carnap himself continued the construction of formal languages
in which such concepts as testability, modality, and probability could be
rationally reconstructed or "explicated" and thus contributed
further to the same ideal. Indeed, Carnap's explication of concepts through
the construction of formal languages influenced the English speaking world
of analytic philosophy far beyond the borders of philosophy of science.
Developments in formal semantics and philosophy of language, in particular,
rested on Carnap's initial work on modality.
The Carnapian ideal of explication is based on a sharp distinction between
logical and empirical investigation, analytic and synthetic truth. In his
Logical Syntax of Language Carnap had attempted a general explication of
the concept of analyticity itself--a general formal method for distinguishing,
within the context of any given formal language, the analytic from the synthetic
sentences of that language. After accepting Alfred Tarski's semantical
conception of truth in 1935, however, Carnap abandoned the approach of Logical
Syntax and frankly admitted that (although explications for various particular
languages could still be constructed) he now had no generally applicable
explication of the concept of analyticity. After studying with Carnap in
the early 1930s, W. V. Quine then exploited this situation to attack the
concept of analyticity as such and, on this basis, to attack the Carnapian
ideal of logical explication as well. Philosophy, for Quine, is itself
a kind of empirical science--a branch of human psychology or "naturalized
epistemology." Moreover, at the same time that Quine was articulating
this new philosophical vision, Thomas Kuhn published The Structure of Scientific
Revolutions in the International Encyclopedia of Unified Science edited
by Carnap and Charles Morris. Whereas Carnap had relegated the (conventional)
choice of scientific language to the limbo of pragmatics, Kuhn concentrated
on those factors--especially social factors--which, in a scientific revolution,
determine precisely this kind of choice. These ideas, in harmony with Quine's
more general naturalistic point of view, then led to historical and sociological
approaches to the study of science and thus, in the end, to the decline
of logical analyses of scientific language in the Carnapian style.
Bibliography
For surveys of the development of logical positivism in its historical context
see A. Coffa, The Semantic Tradition from Kant to Carnap (Cambridge, 1991);
P. Frank, Modern Science and Its Philosophy (Cambridge, Mass., 1949); V.
Kraft, Der Wiener Kreis (Wien, 1950), trans. as The Vienna Circle (New York,
1953); R. Haller, Neopositivismus (Darmstadt, 1993). A delightful short
introduction is M. Geier, Der Wiener Kreis (Reinbeck bei Hamburg, 1992).
See also R. Carnap, "Intellectual Autobiography," in P. Schilpp,
ed., The Philosophy of Rudolf Carnap (Lasalle, 1963).
For the major writings of the logical positivists see M. Schlick, Gesammelte
Aufsätze (Wien, 1938), Allgemeine Erkenntnislehre (Berlin, 19181, 19252)--trans.
as General Theory of Knowledge (Lasalle, 1974); R. Carnap, Der logische
Aufbau der Welt (Berlin, 1928)--trans. as The Logical Structure of the World
(Berkeley, 1967), Logische Syntax der Sprache (Wien, 1934)--trans. as The
Logical Syntax of Language (London, 1937), Introduction to Semantics (Cambridge,
Mass., 1942), Meaning and Necessity (Chicago, 19471, 19562), Logical Foundations
of Probability (Chicago, 1950); O. Neurath, Gesammelte philosophische und
methodologische Schriften (Wien, 1981); R. Reichenbach, Gesammelte Werke
(Braunschweig/Wiesbaden, 1977-), Philosophie der Raum-Zeit-Lehre (Berlin,
1928)--trans. as The Philosophy of Space and Time (New York, 1958), Experience
and Prediction (Chicago, 1938), The Rise of Scientific Philosophy (Berkeley,
1951); C. Hempel, Aspects of Scientific Explanation (New York, 1965). See
also L. Wittgenstein, Tractatus Logico-Philosophicus (London, 1922) and
K. Popper, Logik der Forschung (Wien, 1935)--trans. as The Logic of Scientific
Discovery (London, 1959). Translations of many of the most important writings
of the logical positivists appear in The Vienna Circle Collection, ed. R.
Cohen. See, in particular, M. Schlick, Philosophical Papers (Dordrecht,
1979); O. Neurath, Empiricism and Sociology (Dordrecht, 1973); H. Reichenbach,
Selected Writings (1978); H. Hahn, Empiricism, Logic, Mathematics (Dordrecht,
1980); K. Menger, Selected Papers (Dordrecht, 1979). A very useful short
collection is A. Ayer, ed., Logical Positivism (Glencoe, Ill., 1959). Many
original manuscripts and letters can be found in the Archive for Scientific
Philosophy at the University of Pittsburgh.
For the development of non-Euclidean geometry and philosophy of geometry
after Kant see R. Torretti, Philosophy of Geometry from Riemann to Poincaré
(Dordrecht, 1978). For neo-Kantian reactions to late nineteenth and early
twentieth century developments see P. Natorp, Die logische Grundlagen der
exakten Wissenschaften (Leipzig, 1910); E. Cassirer, Substanzbegriff und
Funktionsbegriff (Berlin, 1910) and Zur Einsteinschen Relativitätstheorie
(Berlin, 1921)--both bound together as Substance and Function & Einstein's
Theory of Relativity (New York, 1953). For Helmholtz see Philosophische
Vorträge und Aufsätze, ed. H. Hörz and S. Wollgast (Berlin,
1971), Schriften zur Erkenntnistheorie, ed. P. Hertz and M. Schlick (Berlin,
1921)--trans. as Epistemological Writings (Dordrecht, 1977), Selected Writings
of Hermann von Helmholtz, ed. R. Kahl. For Mach see Die Mechanik in ihrer
Entwicklung historisch-kritisch dargestellt (Prague, 1883)--trans. (from
the 9th German edition) as The Science of Mechanics (Lasalle, 1960), Beiträge
zur Analyse der Empfindungen (Jena, 1886), 5th enlarged ed., Die Analyse
der Empfindungen (Jena, 1906)--trans. as The Analysis of Sensations (New
York, 1959). For Poincaré see La Science et l'hypothese (Paris,
1902), La Valeur de science (Paris, 1905), Science et méthode (Paris,
1908)--all three bound together in The Foundations of Science (Lancaster,
Pa., 1913).
For the positivists' reaction to the theory of relativity see M. Schlick,
Raum und Zeit in der gegenwärtigen Physik (Berlin, 1917)--trans. as
Space and Time in Contemporary Physics (Oxfrord, 1920); H. Reichenbach,
Relativitätstheorie und Erkenntnis Apriori (Berlin, 1920)--trans. as
The Theory of Relativity and A Priori Knowledge (Berkeley, 1965), Axiomatik
der relativistischen Raum-Zeit-Lehre (Braunschweig, 1924)--trans. as Axiomatization
of the Theory of Relativity (Berkeley, 1969), R. Carnap, Der Raum. Ein Beitrag
zur Wissenschaftslehre (Berlin, 1922). The principal developments in logic
and the foundations of mathematics to which the positivists were reacting
are collected in J. van Heijenoort, ed., From Frege to Gödel: A Source
Book in Mathematical Logic, 1879-1931 (Cambridge, Mass., 1967). See also
P. Benacareff and H. Putnam, eds., Philosophy of Mathematics: Selected Readings
(Englewood Cliffs, 1964). For Wittgenstein's influence on the Vienna Circle
see F. Waismann, Wittgenstein und der Wiener Kreis (Frankfurt, 1967)--trans.
as Wittgenstein and the Vienna Circle (Oxford, 1979). For the protocol-sentence
debate, with special emphasis on Neurath's role, see T. Uebel, Overcoming
Logical Positivism from Within (Amsterdam-Atlanta, 1992). For the positivists'
emigration see H. Feigl, "The Wiener Kreis in America," in D.
Fleming and B. Bailyn, eds., The Intellectual Migration: Europe and America,
1930-1960 (Cambridge, Mass., 1969). A. Ayer, Language Truth and Logic (London,
19361, 19462), was extraordinarily influential in popularizing the positivist
movement within the English speaking world. For the debate between Carnap
and Quine on analyticity see R. Creath, ed., Dear Carnap, Dear Van: The
Quine-Carnap Correspondence and Related Work (Berkeley, 1990).
MICHAEL FRIEDMAN
10