Towards an Objective Theory of Rationality

4. An Objectivist Epistemology

4.1 Preliminaries

Book cover: The Aim and Structure of Physical Theory by Pierre Duhem

A number of objectivist epistemologies have been proposed, and they have failed for various reasons. Contemporary epistemologies, such as conventionalism, pragmatism and falsificationism, in their attempts to overcome the problems of earlier foundationalist epistemologies, have led to a revision of some epistemological terms such as 'truth' and 'knowledge'. Such revisions are not undesirable in themselves, so long as it is understood that what is being offered is a revision of common-sense notions. If such a revision leads to greater clarity and a deeper understanding of the subject, this is well and good. However, some recent epistemologies introduce clearly subjective elements into rational appraisal, for example, considerations of simplicity (Poincaré's and Duhem's conventionalism) and psychological utility (James' pragmatism).

What I want to do here is to outline an objectivist epistemology that does justice to our commonly understood notions of 'truth', 'knowledge', 'rational', and so on, without sacrificing a thorough-going objectivism. I will develop seven criteria for assessing evidence-statements offered in support of a theory. The criteria can be divided into two types. The first type specifies the required logical relationships between the evidence-statement and the theory under evaluation. These criteria are relatively uncontroversial as they appear in other theories of rationality. I will call criteria of this type 'criteria of dependence'. The second type of criteria is what I refer to as 'criteria of independence'. They specify how the evidence-statement needs to be independent of the theory under evaluation and perceptual bias. It is this second set of criteria that marks off this theory of rationality from other more subjective schema.

How can we contrast an objectivist epistemology with its subjectivist rivals? I think we do that by considering what it means to be objective. For an entity to be objective, be it a person, social group, report or belief, it must of necessity be impartial, impersonal, disinterested and detached. Now contrast these requirements for objectivity with what it means to be subjective. To be subjective is of necessity to be prejudiced, biased or aligned. In a nutshell, to be objective is to be in a relation of independence from biasing factors.

Before discussing the criteria in some detail, I will deal with some preliminaries clarifying the meanings of terms and the nature of truth-bearers. I begin by giving an account of 'truth'. This term, as with most others found in natural languages, is vague. I think, though, that for our present purposes, its meaning may be adequately formalized as: 'what is actually the case'. I shall take the primitive truth-bearers to be interpreted sentences or propositions. It does not matter here which it is, but for convenience I shall assume it is interpreted sentences. So, 'x is true' means 'x states what is actually the case', where 'x' specifies an interpreted sentence.

This interpretation of 'true' is liberal enough to include logical truths, such as 'It is raining now or it is not raining now', in the class of true sentences. It is also broad enough to include subjunctive conditionals and statements about the past and future as contingent. I cannot expand on this anymore here, but must rely on the reader's intuitive understanding of the concept to some extent. Needless to say, I think that coherence and pragmatist theories of 'truth' are explicating a notion that has little bearing on what we commonly understand by this term.

Next, I indicate the epistemic nature of truth-bearers in an objectivist epistemology. An objectivist epistemology stipulates that the epistemic status of a truth-bearer is a function of the properties of the truth-bearer itself and is independent of its relationship with subjective states, such as beliefs, preferences and attitudes. Alternatively, subjectivist and relativist epistemologies render the epistemic status of truth-bearers as the function of beliefs, preferences, attitudes, or some other subjective state of an individual or group of individuals. In the objectivist epistemology proposed here, then, whether a sentence, when interpreted, is true or false is dependent on the properties of the sentence, that is, whether it states what is actually the case. In contrast, its epistemic status is independent of its relations with subjective states. So, a sentence is true or false independently of whether some individual or group of individuals believe the sentence, or whether the sentence evokes praiseworthy thoughts or feelings of contentment, and so on.

I said that the primitive truth-bearers are sentences. I said this because we ordinarily also speak of beliefs as being true or false, and I think that we can give an account of this way of speaking by using the account of the epistemic status of sentences given above. A person's 'belief' can always be expanded into 'belief that x is true', where 'x' specifies an interpreted sentence. For example, 'Mary believes that New Delhi is the capital city of India' can be analysed as 'Mary believes that "New Delhi is the capital city of India" is true'. And Mary's belief is a true belief only if 'New Delhi is the capital city of India' is a true sentence. So, beliefs are also truth-bearers, but in a derivative sense.

I introduced the notion of 'true belief' in order to lead into the development of a theory of 'rational belief'. Not all contingent sentences of a language are indubitably true or false.[6] (This may also be the case for analytic sentences.)[7] An adequate objectivist epistemology, then, must specify acceptable criteria for rational belief in those cases in which the set of evidence statements do not lead indubitably to conclusive verification or falsification of a belief. This usage of the term 'rational', in which a belief may be correctly judged as 'rational' although further investigation may demonstrate its falsity, is given in ordinary language. We may therefore characterize the meaning of 'rational belief', sufficient for our purposes here, as 'epistemically justified belief', where 'epistemically justified' is further specified as 'epistemically justified relative to the acceptable set of available evidence-statements E'.

On this ordinary language account, for a person to assert genuinely that 'x is true' is to express a belief that 'x', where 'x' specifies an interpreted sentence. Furthermore, it is 'rational' for a person to assert genuinely that 'x is true', that is, to express a belief that 'x', iff 'x is true' is epistemically justified relative to the set of acceptable evidence-statements E available at the time. It may be 'rational' to believe that 'x' and to assert genuinely that 'x is true' even though 'x' may be false. This is in accord with our intuitive understanding of the meaning of 'rational'.

Book cover: Philosophy of Science: A Very Short Introduction by Samir Okasha

This leads us to the meaning of the term 'knowledge'. I shall accept here the conventional analysis of the meaning of 'knowledge' as 'justified true belief'. On this account, then, a person 'knows' that 'x' iff that person's belief that 'x' is a rational belief and 'x' is true. One important consequence of this analysis is that it is logically impossible for a person to 'know x' and 'x' be false. We can also conclude, then, that the set of our present rational beliefs constitute our sum of knowledge, for they are rational and true.

It may be asked: On what basis can we be assured that our current 'rational' beliefs are in fact true? My answer is that it is rational to assert them to be true since they are epistemically justified relative to the evidence. This usage is also in accord with our intuitive understanding of these terms, except in so far as that we may want to say that not all of our present rational beliefs constitute knowledge. We may want to say that a rational belief constitutes knowledge only if the evidence in support of the belief exceeds a minimum specified strength. We can only begin to solve this problem by first agreeing to a set of criteria for rationally appraising beliefs. I shall now attempt to formulate such a set.

The formulation of such principles is no easy matter, for what we appraise are beliefs about sentences on a number of different theoretical levels. We appraise observation sentences, such as 'You sped through that "Stop" sign and hit my car.' On the next level, we judge causal hypotheses, such as 'Smoking has a tendency to cause lung cancer.' More general hypotheses that explain a process in terms of some underlying microstructure or overlying macrostructure are also appraised. The operator gene theory of cancer and the atomic theory of chemical affinity are examples of the former, while Marxist analysis of revolutions in terms of social institutions and Einstein's explanation of the motions of celestial bodies as being a function of the topology of space are examples of the latter. In appraising these high level theories, it is clear that we are not judging isolated sentences, but a complex set of interconnected sentences, the members of which change with the development of the theory.

On an even higher level, we judge entire belief-systems that include a cosmology, ontology and epistemology in one package deal. In these high-level systems, these elements are interrelated to form one systematic whole and to replace one element necessarily involves modifying the others. Vedanta Hinduism, traditional Aboriginal religion and the Newtonian dualist–mechanistic empiricist world view are just such belief-systems.

This introduces a further complication. The details of an objectivist theory of rationality are a function of the ontology-cosmology that we choose to adopt. This is partly for the simple reason that our theories of observation will depend upon the way we think the world is structured and the way we think we are built. Conversely, details of our ontology-cosmology will be a function of how we think we rationally appraise theories.

The solution to these difficulties can be found if there are certain fundamental objective criteria for the appraisal of ontology-cosmologies independent of such theories. I think that there are such criteria that follow from a general objectivist epistemology and, furthermore, the application of such criteria justifies a realist framework within which it is possible to further develop an objectivist theory of rationality. In the remainder of this essay, I shall explicate such fundamental criteria, briefly argue that the application of such criteria favours a realist framework and then go on to develop further a theory of rationality within such a framework.


  1. [6] I will give reasons below for thinking that no contingent sentences can be considered indubitably true or false.
  2. [7] Here I use 'analytic' in the broad sense of 'analytically true' or 'analytically false'. Because of the fluidity of the analytic–synthetic distinction in natural languages, my comments here only apply to formalised, or roughly formalised, systems of sentences in which the meanings of sentences is conventionally fixed.

Copyright © 2016

You will be interested in

Share This

  • twitter
  • facebook
  • linkedin
  • googleplus
  • gmail
  • delicious
  • reddit
  • digg
  • newsvine
  • posterous
  • friendfeed
  • googlebookmarks
  • yahoobookmarks
  • yahoobuzz
  • orkut
  • stumbleupon
  • diigo
  • mixx
  • technorati
  • netvibes
  • myspace
  • slashdot
  • blogger
  • tumblr
  • email
Short URL:


PDF Download Towards an Objective Theory of Rationality

Download this essay

PDF Download Imre Lakatos and a Theory of Rationality

Download complete essay on Imre Lakatos and a Theory of Rationality