Criteria of Theoreticity: Bridging Statement and Non-Statement Viewby Gerhard Schurz



Logic / Philosophy



Jacques Ellul, Ivan Illich

Statement By

H. Daniel Lang



Criteria of Theoreticity: Bridging Statement and Non-Statement View

Gerhard Schurz

Received: 31 August 2012 / Accepted: 27 September 2013  Springer Science+Business Media Dordrecht 2013

Abstract In this paper I reconstruct and compare criteria of theoreticity that have been developed by Carnap, Sneed and proponents of the Munich school of structuralist philosophy of science. For this purpose I develop a unified framework in which one can transform model-theoretic theory representations into linguistic ones, and vice versa. This bridges the gap between statement and non-statement view and allows a precise comparison of linguistic and model-theoretic criteria of theoreticity. In the final part I suggest a system of improved definitions of theoreticity and pre-theoreticity. 1 Introduction: The Problem of Theoreticity

Beginning with Carnap (1936/1937), logical empiricists realized that scientific theories contain non-logical concepts such as ‘‘electron’’ or ‘‘gravitational force’’, which are not definable by observation concepts. These concepts were called theoretical concepts (or terms), as opposed to empirical concepts, whose meaning is either directly given by observation, or is explicitly definable by observation concepts (more on this notion at the end of sec. 2).

In many writings, logical empiricists (e.g. Carnap 1956) used the notion of ‘‘theoretical’’ in the wide sense of ‘‘non-empirical’’. I state this in a first definition:

Definition 1 (Theoreticity i.w.s.) A non-logical concept or term u (of some interpreted language l) is theoretical in the wide sense (i.w.s.) iff it is not empirical.

To avoid misunderstandings, let me first point out that certain often heard critiques of the linguistic approach of the logical empiricists are wrong. Since

G. Schurz (&)

University of Du¨sseldorf, Du¨sseldorf, Germany e-mail: 123


DOI 10.1007/s10670-013-9581-x

Carnap (1947), if not before, a linguistic framework l was understood as a language together with a semantic interpretation of l, typically in the form of a class of semantic models. So the critique that the linguistic account is purely ‘‘syntactical’’ is off the mark. In what follows we will use the terms ‘‘concept’’ and ‘‘term’’ interchangeably, since each of our terms goes along with a corresponding concept. Moreover, since Carnap (1936/1937) logical empiricists did no longer make the assumption that every non-logical term of a language is either primitive or explicitly defined (as asserted in Stegmu¨ller 1976, 11).

According to def. 1, theoretical concepts i.w.s. coincide with non-empirical concepts. Based on an assumed notion of empiricity, Carnap divided the non-logical concepts of the language into empirical and theoretical concepts i.w.s. This view has been called the dual-level conception of scientific language (for a nice exposition see Tuomela 1973). It has been criticized that according to this view, being theoretical or empirical is regarded as a purely linguistic property of concepts (cf.

Stegmu¨ller 1976, 23ff; van Fraassen 1980, 4; French 2008). Again, this critique is unfounded. For Carnap, the notion of ‘‘empiricity’’ was not a linguistic but an epistemological and psychological notion, and his division of concepts was based on epistemological and psychological grounds. Of course, one may doubt that a clear borderline between empirical and non-empirical concepts exists; but this is a different question which is addressed below.

Nevertheless def. 1 of theoreticity i.w.s. was unsatisfying for logical empiricists.

Their original program was to show that all (non-logical) scientific concepts are reducible to empirical concepts, or in other words, that all empirically undefinable concepts are superfluous metaphysics. Now they found such concepts in the heart of science, in theories of physics. So the important task was to find a characterization of theoretical concepts which distinguished them not only from empirical concepts, but assigned to them a positive function for empirical science, and thus distinguished them from empty metaphysics. This is the problem of theoreticity.

In a famous paper which has been called ‘‘Putnam’s challenge’’, Putnam (1962) argued that logical empiricists had only given a negative characterization of theoretical terms (e.g. being ‘‘undefinable’’ or ‘‘merely partially definable’’), but not a positive characterization. Putnam and other scientific realists (e.g. Maxwell 1962) suggested capturing the positive function by means of a realistic metaphysics, according to which it is the function of theoretical terms to refer to unobservable entities. However, this characterization is unsatisfying insofar as it leaves it unclear how a ‘‘reference to something unobservable’’ is epistemically accessible at all.

What philosophers of science need is an explication of theoreticity which is not based on free-floating metaphysical reference assumptions, but on the role which theoretical concepts play in a given theory. 2 Theoreticity According to Carnap

It is not true that the logical empiricists had nothing to say about the role of theoretical concepts. The story began with a discovery of Carnap in (1936/1937, 44). He discovered that dispositional concepts such as ‘‘x is soluble in water’’ are

G. Schurz 123 not definable by observation concepts, if one uses only truth-functional logical operations. Carnap recognized that a dispositional concept is partially definable, by a so-called bilateral reduction sentence, in short a BR. He goes on to observe that also most concepts of scientific theories—so called theoretical concepts—are only partially definable, as follows:

Definition 2 A bilateral ‘reduction’ sentence (BR) has the following form: For all objects or systems x: If x is exposed to measurement conditions M, then: x has disposition or theoretical property u iff x exhibits measurement result R.