There is a debate between realism and anti-realism. I believe the arguments brought against the other lack a certain scope regarding the enormous claims made about science. The realist philosophers in our readings represent realism with general coherent agreement. The anti-realists, on the other hand, represent a wide variety of views regarding anti-realism. The philosophers in both camps do a poor job accounting for, what seem to me, vital aspects of scientific enterprise.
When one talks of truth, or a true reality, one is not merely critiquing the sociological or psychological influences of those conducting the experiments, nor are they simply making claims about whether causation and inference are adequate tools for explaining phenomena or whether they can produce true and reliable ontological claims. It is much more than splitting hairs over observables and non-observables, true and falsity. While these are important, one needs to begin with examining the context in which this debate resides. That is, one needs to look at the embedded assumptions within the language of philosophy of science, as well as the sciences in which they refer. In addition, one needs to examine the various components that comprise the discourse: logic, context, and the principles of epistemological analysis of scientific knowledge.
A great deal of clarity could be gained if the debate focused on the aforementioned aspects. A starting point is differentiating between true relational knowledge claims and true substantive knowledge claims. This debate focuses a great deal on the nature of predicate logic and truth and falsity. But what is meant by truth? Regarding relational knowledge, truth is a result of the true conditions of predicate logic. The statement “If you are an unmarried male, then you are a bachelor” is a true statement, where M(x)->B(x). They are true because of their form and their form depends on certain immutable laws of thought.
On the other hand, substantive knowledge claims deal with the ontology of entities within the world. These are true based on experiential knowledge on the entity through acquaintance, be it through direct or indirect reference.
A vital distinction between these two types of knowledge is their relationship with context. Relational knowledge claims are contextually independent in that whatever the context, they will be true. As a result, relational knowledge is contextually indeterminate. It doesn’t matter what context a relational knowledge claim is made, the veracity does not change. Substantive knowledge, however, depends on a context for validity. The statement “Water is cold” depends on a context with determinate parameters for a truth value. In short, relational knowledge uses logical deduction to arrive at exact truth, whereas substantive knowledge uses logical induction to arrive at approximate truth.
To continue building towards clarification, we need to examine the role of language. Due to the discursive nature of philosophy, the debate within the philosophy of science seems to revolve around language, but its implications are much more than language. Language cannot be transcended. Mathematics and logic can be said to be the purest language as descriptions of truth, but these are semantically independent. That is, despite given predicates and primitive parameters, the semantic predication cannot be determined by the inference system itself. Once these have been semantically defined, the truth value can be inferred.
Regarding knowledge, language serves two primary functions of establishing knowledge, that of indexing proper names, and that of referring to descriptions. Drawing from Russell’s theory of language, there are two basic kinds of knowledge acquisition, namely acquaintance and description. Knowledge by acquaintance includes logically proper names which are referential of indexical demonstratives. Knowledge by description includes attributes that make description claims of thing in the world that can be true or false, but are not referential. That is, they are merely conceptual constructions. Within these descriptions reside indefinite descriptions, which make existence claims using indefinite articles (“A man drinking a martini…), and definite descriptions (“The man drinking the martini…), which make existence and uniqueness claims using definite articles.
What about non-existence objects? These are merely descriptions with no acquaintance. They may lead one to denote a particular in the world, but they do not ensure it. Descriptions without acquaintance may offer a adequate roadmap to their truth.
Strawson would say that referring is not asserting, but rather mentioning, something quite different from identifying. In addition, Donnellan divides definite descriptions into the referential and the attributive. Attributive descriptions denote all important, essential information about whoever or whatever. Referential descriptions contain information that is accidental and not important, but rather is instrumental in accurately picking out particulars through idexical or deictic use.
Names of objects and facts can have meaning, but only insofar as they have a context of propositions that are held together by a proper logical form. You may ask: why is this important? The answer is that logic is used to make inferences about the properties of postulated entities. Suppose we have the statement F(m, n) which stands for “m applies force to n”. In this way F is a property between m and n such that F is “applies force to”, and m and n are primitive names for Mike and Nail. So that, Mike applies force to Nail. The semantics of these variables are contextually determined and dependently attributed. They are given as substantive knowledge claims regarding the properties.
Context is necessary to define any set of data. Finding an ultimate truth is akin to finding an ultimate context. Within any given context there are phenomena that can be said to be real and true, but extrapolating beyond that context can be a fatal move for a theory.