20th WCP: The Reference of Theoretical Terms

A popular explanation for the success of scientific theories is made by presupposing scientific realism. The realist's thesis maintains that typically the theories of "mature" science are approximately true and that observational terms and theoretical terms do actually refer, i.e. they denote entities.

Therefore, it is part and parcel of the realistic claim that it is "reference" which explains theory "success". But if we or the realist are not able to clarify what "reference" is and a fortiori cannot specify the reference to theoretical "objects", we, and especially the realist, become entangled in a vicious circle, without any further independent criterion — apart from the success of the theory — which shows that the term is indeed a referring term.

It is therefore clearly necessary to clarify the notion of reference.

Needless to say, "reference" is a relational term; but immediately a problem arises here namely, that we not only habitually imagine the relata, but we are also convinced that a relation can only be a relation between entities in a strict (Quinean) sense.

There are various kinds of references resp. relations described as referential. This applies, e.g., to the referential relation which is usually called "intentionality". If we consider the traditional distinction between reference and meaning and analyze the meaning of a term we always find at least one referential component (intentional or intended object). This referential mapping process is not an unimportant aspect of linguistic reference, but in the present context we are considering the kind of relation called "denotation" or "extension".

The notion of meaning and the concept of reference are nonsubstantial constructions of interpretation; yet nevertheless I would like to argue in favor of a reference-theoretical approach.

Theoretical terms have often caused diverse problems for theoretical scientists, depending on their particular "metaphysical" background. At first sight, an elegant solution appears to be given by Rom Harré's approach (1986): Harré understands the reference to theoretical as well as to empirical entities as a "material practice". He offers two explications for "material practice":

Observers (materially embodied [Harré, 102]) are involved in a physical/physicalistic relation to a material object; experimentators produce parts of reality on a theoretical basis.

Regarding the production of reference it turns out to be indifferent whether the object is an object in the world or just "in" a theory, for it is trivial that no one is able to think of objects without conceptual mediation. However, the distinction between physical and theoretical entities is theory-dependent itself, as is also the idea that a speaker might refer to something and how he would do so.

Secondly, Harré advocates a so called referential realism, i.e. he tries to develop his concept of realism further than traditional truth realism (1) and metaphysical realism. Within the range of his scientific program he distinguishes three kinds of scientific theories dependent on three different levels of being and their respective associated cognitive or operational-experimental practices: type 1 theories contain descriptions and definitions of observable phenomena, they reach therefore beyond the level R1 of the directly observable. In type 2 theories "objects" are represented or postulated, which are — in Harré's opinion — not presently (not yet actually),but in principle, observable; while the entities postulated in type 3 theories (covariances, symmetries, hence mathematical features) are principally beyond all observation (Harré 70).

Harrés conception aims at a transformation of type 2 theories to type1 theories. It is feasible to consider the method he suggests for this explicitly within the scope of this paper. Let me only indicate here that Harré (referring to L. Roberts) distinguishes two kinds of reference production: the so-called DC- and IP-references. ("DC" stands for the demonstrative pronoun and complement: it could be called deixis as well, in connection with a general term.) DC-references achieve the selection of an object out of an extension set, determined by the general terminus, by pointing at, referring to something by similar procedures. They differ from the IP-references in that in the DC-associations the general termini indicate perceivables. The IP-references are so-called after the "I" for "indefinite pronouns" and "P" for the "individuing predicate". In a way, the IP-references are also dependent on perceptions, the "I" refers to a that; it's true that this perception is not explained by some other perception, but the predicate level is defined by mere theoretical constructs.

The type 2 theories deal mainly with IP-references (e.g., "Whatever is the cause of this phenomenon is an X") and Harré's "aim" in science is to transform (or reduce) IP-references to DC-references. Therefore, he introduces the beautiful metaphor of referent hunting with which — in my opinion — he unnecessarily imputes that one can speak about "reference" only when there is a referent in quite a specific sense — and, according to Harré, that would be an observable existence. Of course, Harré doesn't assume that one could somehow see or taste neutrinos. Obviously, these are merely indirectly perceivable through specific technically produced effects or phenomena. Secondhand reference, however, is reference-theoretically "indecent", since it is not direct reference. There is, in principle, no direct acquaintance with the "object" and, therefore, it is difficult — to say the least — to find a distinction (in this relation) between Harré's "reference chase" and Carnap's correspondence rules, except, perhaps, by pointing to the fact that retransformating R2-entities to R1-entities would in the long run hinder the progress of cognition in the case where the re- or transformation attempts are made indiscriminatey, firstly, in all theories of various disciplines and, secondly, independently of the respective intended function of the theory (explanation, prediction, "production instruction"). When reference is seen functionally — and I think this is the only way to view it — we can attribute a functional status to the "objects" of reference. That means that certain "objects" of reference "exist" only within the reference act itself, and are therefore only dependent on the respective intended function.

For a long time it has been generally accepted that the distinction between observational and theoretical terms is, if not questionable in principle, nevertheless highly dependent on aims, formulations of the question, conventions, practices of speech and not least on levels of technology, especially of measurement technology.

Our entire world access is conceptually defined (dependent on our cognition equipment), is so to speak interpretation-infused. Thus, Pegasus, bald present French kings, unicorns and wooden irons, nouns originating from adjectives like redness and blueness, phlogiston, neutrinos and their mass, Snow White, as well as the objects of direct perception themselves are — metaphorically spoken — in the same boat. Nevertheless, it is necessary and possible to make distinctions. There remains particularly the (practically) necessary distinction between observable and theoretical terms, as well as the determination of a demarcation criterion between theoretical and other non-observable entities.

The thesis states that possible objects of reference are generated on different levels of interpretation, (2) become reference objects by speaking acts, which means that on the one hand a succeeding reference procedure presupposes its objects, and on the other, that this "object" is not given independently (from interpretation), but necessarily drafted in the form of an interpretation result, formed and structured in processes of explanation, which partly (IS1+2) cannot take place in any other way, i.e. independently from interpretation "grasps", which are at least partly influenced by a given (already found) speech practice (IS3) and partly quite consciously and purposefully constructed for certain aims.

Not only the objects of terms which are traditionally not understood as being of a referring type, but also theoretical entities should be understood as higher level interpretative constructs, and thus as something rather linguistically introduced, that "exists" only by dependence on a certain language. On the other hand, the referent is, partially because of linguistically independent conceptions, types of action, practices, etc., definitely and correctly to be understood as not being of just a linguistic type in the narrow sense.

In summary: Theoretical terms are insofar referring terms as they, occuring in hypotheses or laws, are used to explain reality (i.e. real events and objects), make predictions possible and to develop technological procedures. Their objects of reference ("referents") are given only under a description: more specifically, they are significantly (selected and selectable) indepently of a description. This is however — following Frege and Russell — , a serious obstacle for at least direct reference. Nota bene, the distinction between direct and nondirect reference is difficult to specify: Referring terms are characterized by objects being imputed (hypostatized) or projected outside of the mind; and this makes reference basically indirect. There may be reasons for distinguishing reference to perception-based objects as a direct reference. But theoretical terms can also be understood as directly referring — therefore you need only to refer to the definition that a term is only referring in the case that the truth — and realization conditions are understood as being singular; and this is what they are — as is well-known — when there is an x, that makes a sentence exactly true, if x fulfils the predicate F ( ): Whether this is a theoretical, empirical or even mystical predicate, is unimportant in this context.