Towards an Objective Theory of Rationality

6. A Realist Theory of Knowledge

6.3 Criteria for Accepting Observation Statements

Observation statements play a key role in building and testing our knowledge of the world and of ourselves. They often function as evidence for our explanatory theories. Criterion 1 in my theory of knowledge already stipulates that a statement is acceptable as evidence for a theory only if it satisfies the condition that it is true. What are the epistemological implications for my theory of rationality given the theory-ladenness of observation statements on the one hand and the propensity for perceptual bias from subjective factors on the other?

Given the above considerations, I can now add two specific conditions for the acceptance of an observation statement as evidence for a theory. They are as follows.

The conditions for an observation statement to be evidence for a theory are:

Criterion 6:
the statement is interpreted in terms of accepted physical and observational theories;
(dependence on accepted presuppositions)
Criterion 7:
the statement is not the result of distorting subjective influences, such as prior beliefs and experiences, expectations, needs, values and personality.
(requirement of independence from bias)
Book cover: Fads and Fallacies in the Name of Science by Martin Gardner

These two new criteria serve to fill out the picture of my theory of knowledge within a realist framework. They further explicate my Criterion 1; the condition that a statement tendered as evidential support for a theory be true. Criterion 6 ensures that the observation statement does not presuppose an unsupported observational theory or other such auxiliary hypothesis. Criterion 7, on the other hand, serves to eliminate perceptual biases from the observer's psychological state or viewing conditions.

How do these two criteria operate in practice? Two examples illustrating the application of Criterion 6 follow. The first example concerns an observer who reports that the earth did not move as he jumped in the air. The observation report is proposed in support of the Ptolemaic-Aristotelian cosmology with a fixed earth at the centre of the cosmos. The application of Criterion 6 disallows this observation report as the report presupposes the false theory of dynamics that entails that a person will not move unless acted upon by a force.

In the second example, an observer attends a public 'psychic' demonstration after which he reports that the 'psychic' on show had supernaturally bent his car key. Applying Criterion 6, this observation report is rejected as evidence for psychic abilities because it presupposes the false theory that untrained observers are able to follow the hand movements of people versed in the skills of misdirection.

The purpose of Criterion 6 is not to prevent the acceptance of a new physical or observational theory. Our current theories are always open to revision in the light of further evidence. However, if an observation statement relies for its veracity on a new, untested theory (that is, on an unverified auxiliary hypothesis), then the observation statement in question cannot be accepted until the untested auxiliary hypothesis on which it relies has undergone and passed independent tests. In this sense, Criterion 6 formalises a degree of epistemic conservatism. We do not discard our current established theories on the basis of an anomalous observation statement until the rival theoretical assumptions grounding the observation statement are themselves tested and verified.

What follows next are two examples of the application of Criterion 7; the requirement of independence from subjective influences. For the first example, consider an observational report of a U.F.O. sighting from the night before. Subsequent investigation of the report reveals that the observer had an avid prior interest in U.F.Os. The application of Criterion 7 methodologically excludes the acceptance of the observation statement as it cannot be ruled out that the observer was influenced in his perception by his prior belief in the existence of extra-terrestrial visitors.

In the second example, a report in a widely circulated newspaper of a sighting of the Yeti is followed by a large number of similar sightings. An analysis of the timing of the reports shows that there were no reported sightings between the time of the original sighting and its publication in the newspaper. Applying Criterion 7 to this case again rules out the acceptance of the later observation reports as it is possible that these reports are the result of the increased expectation by the local inhabitants of seeing the Yeti generated by the newspaper report itself.

It is important to note that additional evidence may come to light that increases the veracity of the two reports in the examples above. For example, reports by independent observers with no propensity for belief in the paranormal may surface, along with physical evidence on the ground of extra-terrestrials or the Yeti. The point of Criterion 7 is not to rule out all observational reports. Its purpose is to exclude only those for which there is a possibility that the report offered is tainted with one or more of the subjective biases discussed above.

Psychological considerations such as those that underpin Criterion 7 above have led researchers to develop and refine sophisticated test protocols. The most rigorous of these, now commonplace in many research labs, is controlled double-blind testing. Through devising and using such procedures during the design and test phases, extraneous external causal influences and the subjective biases of the experimental subjects and researchers can be identified and eliminated.

In tests employing double-blind procedures, the value of one variable is under experimental control, with the experimental subjects being kept uninformed (blind) as to the value of that variable in order to eliminate or later isolate distorting psychological influences on the part of the subject. The judges of the value of a second variable are also kept uninformed (blind) as to the value of the first variable, once again, in order to eliminate their personal biases. How the control of the first variable is established will vary with the type of theory under test. Such details are not important here. However, what is important is that this type of procedure is a natural development not only from the psychological considerations discussed above, but is also necessitated by my general Criterion 1 to Criterion 5 if knowledge is to be systematically developed.

Controlled double-blind testing works to test theories using objective criteria not by testing a single theory in isolation, but by testing it against a second mutually inconsistent theory. The first theory is of a type that postulates that a variable x has a statistically significant causal effect on another variable y. An example of this type of theory is one that states that there is a causal connection between the chemical properties of a particular drug and the disappearance of symptoms in a group of diseased patients. The second theory under test denies that there is such a causal relationship between the two variables.

During the test design phase, the logical relationships between the two theories and the probability distributions of test results are calculated (Criterion 1 to Criterion 4). By holding all other known causal factors constant, or randomly distributing them throughout the samples, novel predictions (Criterion 5) concerning the statistical significance of the correlations between the two variables are made for each theory. The test is then performed. If a statistically significant result eventuates, this is counted as independent objective evidence for the first (causal) theory and counter-evidence for the second (non-causal) theory. If the ensuing result is not statistically significant, this is regarded as evidence against the first theory and evidence for the second.

Book cover: Against Method by Paul K. Feyerabend

So far, I have discussed the use of double-blind protocols only as they relate to observation statements made about the external world; that is, about the world of mind-independent physical objects and their properties. In this context, double-blind procedures are used extensively. A couple of further examples are their use in the analysis of data in particle physics experiments and in police photo lineups of crime suspects.

I want now to extend the application of Criterion 7 to the domain of private, introspective reports. Just as observers report events in their external world, they also report their internal mental states, such as feelings of relaxation and sensations of pain. Unlike observations about external objects, investigators are not privy to checking directly the observations of an individual's private mental states. The challenge arises in that such private, introspective reports are used to test theories in a variety of domains. The testing of new medicines and studies on human perception, for example, regularly rely on such introspective reports about subjects' phenomenological experiences.

Here again, controlled double-blind test protocols serve to eliminate distorting influences from the test setup itself. Biases can be introduced by the subject's or the tester's belief or hope that the theory under test is true (placebo effect) or by the extra attention that the researchers pay the subject (Hawthorne effect), or both. As with observation statements about external physical entities, introspective reports distorted by environmental or psychological factors cannot be accepted as support for a theory undergoing critical appraisal. The upshot here is that our use of Criterion 7 in barring observation statements that are the result of such distorting subjective influences applies to both types of observation statements. This criterion applies to observer reports about external physical entities as well as reports about the observer's internal mental states.

I have now discussed the criteria for the acceptance of observation statements tendered in support of explanatory theories. But what of evidence-statements that are not observation reports as such? These kinds of evidential support include statements about the past and future, such as the statement that an asteroid impacted the earth 65 million years ago and the statement that our galaxy will collide with the Andromeda Galaxy in another four billion years. It also includes statements about objects and properties that cannot be observed, such as the statement that the temperature of the Earth's inner core is 5400 °C. How are evidence-statements such as these rationally acceptable when used in support of other theories? In a nutshell, they earn that honour because they are deducible from theories satisfying acceptance Criterion 1 to 5 in conjunction with statements of initial conditions.

In fact, the complex of accepted highest-level theories (which some philosophers refer to as a 'paradigm' or 'world view') serves as the touchstone for determining what constitutes a 'rational reason' for believing a lower-level theory or a statement about the future or past or an unobserved event. It provides the 'rules of evidence', stipulating what type of statement can or cannot be regarded as evidence for such. Within the twenty-first century scientific paradigm, for example, the particular details of a newly-born child's horoscope is not allowable as evidence for statements about his future prospects. Similarly, that a person is left-handed is not admitted as the type of evidence required to demonstrate that he is possessed by a demon. Alternatively, within a medieval paradigm that incorporates an astrological cosmology and a demon possession theory of abnormal behaviour, such statements are accepted as rationally relevant to the appraisal of the statements in question.

However, we must not confuse these intra-paradigm standards of rationality with the extra-paradigm standards defended in this essay. We may, for example, criticise the New Age standard of rationality that stipulates that horoscope readings are rationally relevant to the determining of people's personalities by criticizing the astrological theories in which this standard is embedded. It is this confusion between intra-paradigm and extra-paradigm standards of rationality that has led Barnes and Bloor [1982] and Feyerabend [1978: 22f, 28–30, 32–4, 81–3] to lapse sometimes into a radical version of relativism that maintains that all standards of rationality are culture dependent.

Copyright © 2016

You will be interested in
Book cover: The Man Who Wasn't There: Investigations into the Strange New Science of the Self by Anil Ananthaswamy
Book cover: An Idealist View of Life by  S. Radhakrishnan
Book cover: The Better Angels of Our Nature: Why Violence Has Declined by Steven Pinker
Book cover: The Logic of Scientific Discovery by Karl Popper
Book cover: The Most Good You Can Do by Peter Singer
Book cover: Dialogues Concerning Natural Religion by David Hume

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: