§66. Nostalgic Empiricism

From 1919 a new Germany began to rise from the rubble of war and the ignominy of defeat. Need was felt for a more rational, sober political, aes­thetic, and philosophical culture, a neue Sachlichkeit, a new objectivity that substitutes sober, no-nonsense rationality for false fronts, functionless facades, and dangerously uncritical enthusiasm.

Rudolf Carnap, the philosophical genius of Vienna Circle logical empiri­cism, echos neue Sachlichkeit themes in the preface to his work of 1928, Der Logische Aufbau der Welt (The Logical Structure of the World). He expresses solidarity with the new orientation “in artistic movements, especially in ar­chitecture and in movements which strive for meaningful forms of personal and collective life, of education, and external organization in general. We feel all around us the same basic orientation, the same style of thinking and doing.” He introduces his work as a “call for clarity” and for “a science that is free from metaphysics,” a sober scientific philosophy for a disordered age craving rationality.¹⁴⁹

His starting point was conventionalism (commodisme), a then-new idea in the philosophy of science associated with Poincare and Duhem. Facts, observations, and discoveries were said to be relative to an ultimately conventional choice of the logical and mathematical framework in which to describe and deduce them. Scientific ontology is a matter of convenience (commodite), not insight into nature. In 1902, Poincare argued that the ax­ioms of geometry are neither synthetic a priori nor experimental.

This argument suggests an alternative to Kant’s theory of intuition and the synthetic a priori, and as “intuition” seems inconsistent with “new objec­tivity,” the alternative was welcome. To make objective sense of fleeting data requires an a priori formal structure, but the constitutive principles of struc­ture do not have to be synthetic. The powerful new methods of modern logic show that the a priori principles of objectivity are stipulated and logically analytic, with the “object” of knowledge reduced to a formal posit rather than nature’s verisimilar.

Carnap read Poincare from an early point and referred to him often in his 1920 dissertation. In notes from 1914 he summarized what he regarded as the “basic idea” of Poincare’s Science and Hypothesis: “The basic laws of math­ematics and mechanics stand over and above ‘true and false’; they are con­ventionally stipulated, because the most convenient forms of the empirical and mathematical laws can be represented by means of them, i.e., because they are the most opportune.” He noticed how experience becomes a mere formality. “Experience can serve as the foundation of the principles of me­chanics and nonetheless can never contradict them.” The elimination of sub­jectivity was already an aspiration. He cannot simply abolish experience and still be an empiricist, but he can radically de-subjectify experience, which he proposes to do by “a transition from material to structure.” Experience is significant for science only as structured by logic; unstructured experience is merely private and subjective. Experience becomes part of science only after passing through a syntactic filter that squelches the private, idiosyncratic, and willful.¹⁵¹

Russell enjoined scientific philosophers to prefer logical constructions over inference to the unobservable.

The sciences receive objective empirical content from their syntax, from the logical forms that connect evidence protocols to hypotheses. Sentences reporting the given data of sense and sentences stating abstract laws of na­ture are extensionally equivalent, the difference between them being no more than convenient terminology. “Gravity is just a linguistic abbreviation,” Carnap quipped. His Aufbau proposes a “rational reconstruction” of indi­vidual cognition, ascending from initial sensory data to objective knowledge of physical nature. The work offers a model of the “actual process of cogni­tion” reconstructed not genetically, or as it really unfolds, but rationally, as it logically could happen, the mere possibility of which confirms the objectivity of science duly formalized.¹⁵²

Experience is a problem. With Hobbes, Carnap would like to abolish it from science, but he wants to be an empiricist and cannot escape experience, since the logical analysis of meaning “pronounces the verdict of meaning­lessness on any alleged knowledge that pretends to reach above or behind experience.” It was a tradition in philosophy since Kant (though prop­erly since Ockham) that purely conceptual knowledge of nature is impos­sible and the evidence of sense indispensable, which is Carnap's dilemma. “Science wants to speak about what is objective, and whatever does not be­long to structure but to the material [i.e., perception]... is, in the final anal­ysis, subjective.” A solution appears when he says, “All objects of cognition are constituted...

Objects belong to a sequence of levels in a hierarchy of logical types, with entities on any level except the primary one defined as classes or relations of objects on a lower level. The logical construction of the world begins with elementary experiences, which are holistic momentary cross-sections of the stream of consciousness, ordered by a primitive relation of partial re­semblance. That is the empirical content of natural science in a nutshell, the experiential base, the ultimate evidence, and proof of experience's van­ishing control of statements, which makes this less an empiricism than an ironic simulacrum, a nostalgia for empiricism, an empiricism without the empiricism.

Carnap is indifferent to the epistemological argument that captivates Ayer, who transfers the incorrigibility of sense data to the statements of science. What makes Carnap’s reconstruction of scientific language imperative is the urgency of demonstrable objectivity. The sensory evidence must be reg­imented according to a logical formalism that ensures its objectivity despite a subjective genesis. Empirical evidence presupposes a prior decision on a mathematical framework, and no degree or type of experience makes the choice easier or more natural. Experience is emptied of qualities. The objects about which science is objective are not sensory, or possessed of sensory qualities at all, being individualized in purely formal terms, as layers of rela­tion without reference to qualities.

The goal of Carnap’s reconstruction is to present scientific-theoretical descriptions as pure relation descriptions. Science does not say what iron is, what electricity is, those being metaphysical questions about essence, while science deals only with structural properties.

Mathematician Max Newman challenged the purely structural concep­tion of science. He was responding to Russell, who in The Analysis of Matter (1927), advocated a structural-description idea of science along Carnap’s lines. Newman’s argument was that no purely relational theory, divested of qualitative grounding in concrete instances, can be a physical theory of nature. A pure structural description can be realized by any items what­ever (dominos, marshmallows), provided we have enough of them, with no way to tell whether a theory in physics was about atoms, say, and not marshmallows. Emile Meyerson made a comparable point when he argued that transforming a calculation into a physical statement requires interpreta­tion; until we introduce an intuition, a sensuous quality, there is nothing that the formalism is talking about. Since anything the least concrete must come from somewhere other than a pure structure description, to restrict science to such descriptions means that it has nothing to say about nature.¹⁵⁵

A science of pure relations would also be a science without quantities, be­cause a science without qualities is a science (if this is still the right word) without measurements, which presuppose qualitative difference. Pure quantity in a mathematical formula is physically meaningless apart from stipulations regarding how to observe and measure it, and that is not possible with a purely quantitative universe of discourse. Nothing could be observed; only topological and metric structures would exist. Even a simple meas­urement of length requires an observable difference between endpoints.

More suave than Carnap yet to the same purpose, Bas van Fraassen disavows the association of empiricism and experience. “Empiricism today should not be saddled with the confusions that surrounded the notion of ex­perience,” above all, the deplorable assumption that experience is the ulti­mate source of scientific evidence. But as a critic observes, he ignores the more important work of experience, which is our principal resource for improving opinions. “Experience makes a robust rational contribution to empirical thought, both in exposing defects in our view as well as in guiding its improvement” For a theory to say something of the observable also makes a large draw on experience, even without the assumption that it is the ul­timate source of evidence. Experience has a more important role in exper­imentation and empirical science, one that logical empiricists incorrigibly overlook.¹⁵⁷