On est toujours libre de ne rien comprendre ` rien—Gabriel Marcel

Experimental natural science is grounded in careful observation. Each investigation must proceed from observed facts. In physical and biological science these observed facts are usually inert facts, that is to say they are grasped from the exterior by an observer who is not disturbed by them and does not disturb them by his process of observation. Even in micro-physics where the uncertainty principle tells us that the observational procedures disturb the field of the observed there are mathematical techniques which maintain the observer in some sort of relation of exteriority to the observed and indeed to his observing techniques themselves. In a science of personal interaction, on the other hand, mutual disturbance of the observer and the observed are not only inevitable in every case but it is this mutual disturbance which gives rise to the primary facts on which theory is based and not the disturbed or disturbing personal entities. The facts which are the observational data of anthropological sciences are not different from the facts from which science proceeds in the same sense that facts for biology are different from facts for physics: they differ in ontological status from natural scientific facts. Put another way, the observer-observed relation in a science of persons is ontologically continuous (subject-object vis-`-vis subject-object), whereas in natural sciences it is discontinuous (subject vis-`-vis object) permitting a purely exterior description of the field of the observed.

From statements of observed fact the natural scientist proceeds to conjectural statements which assume the conditional form; ‘if such and such conditions obtain we may expect so and so in the observational field’. If these predictions expressed in the hypothesis are experimentally verified we are in a position to form a theory. In the sphere of personal action, however, conditional statements are complicated in this way: given these specifiable conditions we may expect this person, on the basis of everything we know about him and his past, to behave in this particular way; however, personal action in its essence is the possibility of surpassing all determinations of what it is to be and proceeding perhaps in the direction opposite to that expected—unless there is a locatable choice to conform with such expectations, a choice not to choose. Certainly the field of human actions is readily seen in probabilistic terms. What cannot be left out of the question, however, is the possibility of the subject realizing this probabilistic structuring of the field in which he is situated and through this realization destructuring the field and acting ‘improbably’. This possibility, which the subject always has, of altering his conduct from the expected, through reflective awareness of the factors that are conditioning him at a certain time, really constitutes a crucial difference.

In short, while we are entitled to, and in any practical context must, have expectations (which we must expect to be disappointed) about a person’s behaviour, natural scientific prediction must be seen to be neither possible nor impossible in the sciences of persons but as inappropriate to the field of discourse.

In the natural sciences verifiability and falsifiability of hypotheses depends on the repeatability of situations. In the sciences of persons, however, we note that repetition of an individual or group life-historical situation is, in principle, impossible. There are certainly all the appearances of repetition but in each case we discover that this ‘repetition’ is the product of an illusory project of self-dehistoricization. A person dehistoricizes himself when he chooses (however unknowingly) to deny that by a prior series of choices he has moved his life on from an earlier situation to a new situation: this denial (an act which, by a further act, he denies in turn and so ‘does not know’) allows an illusion of a historical fixity and substantiality. This is the principal mode in which a person rids himself of the anxiety that issues from a recognition of his responsibility for himself. It is remarkable to find scientific theory sometimes effecting the same evasion.

If repetition of life-historical situations is impossible, then natural scientific criteria of verifiability and falsifiability of hypotheses become irrelevant and we must find other criteria by which we may know that we are speaking ‘the truth’. To do this we have to distinguish between two types of rationality which are each appropriate to a field of discourse different from but inter-related with that of the other. These types we call analytical and dialectical rationality.

By analytical rationality we mean a logic of exteriority according to which truth lies, according to certain criteria, in propositions formed outside the reality with which they are concerned. The epistemological model here is characterized by a dual passivity: the observed system is passive with respect to the observer (whatever actions and reactions go on within the system); the observer is passive in relation to the system he observes, such activity as he appears to manifest being limited to conceptual re-arrangements of the facts, which are registered on him from the outside, and to inferences he makes from these facts.