social disciplines. It may be that the intrinsic complexity of social phenomena impedes rapid scientific progress, but modern physics is very complex, and its findings were unearthed by procedures relatively unimpeded by self-deception. The study of history seems to be a conflict between a few honest historians trying to gain a true picture of the past and the greater number, who are primarily interested in promoting an uplifting view of the group past—in short, a false historical narrative.
Another possibility regarding the development of social disciplines is that a prior moral stance regarding a subject may influence the development of theory and knowledge in that subject—so that, in a sense, justice may precede truth (and false justice, untruth). Let us begin with this topic.
PRECEDENCE OF JUSTICE OVER TRUTH?
The usual assumption within academia is that we will derive a theory of justice from our larger theory of the truth. But what if our prior stance regarding justice impedes our search for the truth? For example, an unconscious bias toward an unjust stance will invite cognitive biases in favor of this stance. The “truth” that one produces on the justice of a situation will have been distorted by the prior commitment to an unjust position. In short, injustice invites self-deception, unconsciousness, and inability to perceive reality, while justice has the opposite effect. This can be a very pervasive effect in life. That is, we can construct social theory—at the microlevel, marriage, family, job; at the macrolevel, society, war, etc.—and think we are pursuing the truth objectively, but we may only be fleshing out our biases. This suggests that an early attachment to fairness or justice may be a lifelong aid in discerning the truth regarding social reality. Of course, if your attachment is to pseudo-justice, one may have exactly the opposite effects. It is possible to use an alleged attachment to justice defensively—for example, to prohibit outside knowledge from entering your discipline—which may lead you far from truth, as we shall see for cultural anthropology. Behavior may cause belief, as I have been arguing, but that still leaves open the question of what causes the behavior in the first place, that is, the just or unjust stance.
SUCCESS OF SCIENCE IS BASED ON ANTI-SELF-DECEPTION DEVICES
The success of science appears in great part to be due to a series of built-in devices that guard against deceit and self-deception at every turn. First, everything is supposed to be explicit. Famous mathematical proofs (Godel’s theorem) begin with a set of all the symbols used and what they mean. By contrast, in the social sciences, entire subdisciplines may flourish in the interstices of poorly defined words. Scientific work is supposed to be described explicitly in detail, with terms and methods defined to permit the work to be repeated exactly in its entirety by anyone else. This is the key guard against untruth: repeating work to see whether the same results emerge. Think of the number of tantalizing hoaxes that are dismissed because they can’t pass this first hurdle—for example, achieving atomic energy via cold fusion. Of course, full-time hoaxes, such as psychoanalysis, preclude experimental tests at the outset (in favor of such bedrock data as clinical lore). The requirement for exact description permitting exact repetition applies not just to experimental work but also to any way of gathering data that reveals patterns of interest.
Experiments are conducted under controlled conditions—that is, with certain key variables held constant and/or varied in a logical and systematic manner. The results are then subjected to a statistical apparatus that has grown very sophisticated in the past one hundred years. Very complex sets of data can now be rigorously searched for information regarded as statistically significant. By convention, data that can be generated by chance more than 5 percent of the time are rejected as unreliable. For important results, such as medical findings, we prefer an error rate of 1 percent or less. Finally, meta-analyses can be performed on large numbers of related studies to see what statistically valid generalizations can be made across the full range of evidence. Every single one of these advances tends to minimize the opportunities for deceit and self-deception. They also permit us to rank information by degree of reliability (statistical significance) and effect size (weak or strong).
The acid test of science is its ability to predict the future, in particular, hitherto unknown facts. Yes, light really is bent by gravity (per Einstein); in an eclipse of the sun, the apparent position of stars in the nearby background was