remains in principle at first undetermined. One can thus with some justice see in the observation statements the ultimate origin of all knowledge. But should they be described as the basis, as the ultimate certain ground? This can hardly be maintained, for this "origin" stands in a too questionable relation to the edifice of knowledge. But in addition we have conceived of the true process as schematically simplified. In reality what is actually expressed in protocols stands in a less close connection with the observed, and in general one ought not to assume that any pure observation statements ever slip in between the observation and the "protocol."
 
But now a second function appears to belong to these statements about the immediately perceived, these "confirmations"á as we may also call them, namely, the corroboration of hypotheses, their verification. 

Science makes prophecies that are tested by "experience." Its essential function consists in making predictions. It says, for example: "If at such and such a time you look through a telescope adjusted in such and such a manner you will see a point of light (a star) in coincidence with a black mark (cross wires)." Let us assume that in following out these instructions the predicted experience actually occurs. This means that we make an anticipated confirmation, we pronounce an expected judgment of observation, we obtain thereby a feeling of fulfillment, a quite characteristic sat?isfaction: we are satisfied. One is fully justified in saying that the confirmation or observation statements have fulfilled their true mission as soon as we obtain this peculiar satisfaction. 

And it is obtained in the very moment in which the confirmation takes place, in which the observation statement is made. This is of the utmost importance. For thus the function of the statements about the immediately experienced itself lies in the immediate present. Indeed we saw that they have so to speak no duration, that the moment they are gone one has at one's disposal in their place inscriptions, or memory traces, that can play only the role of hypotheses and thereby lack ultimate certainty. One cannot build any logically tenable structure upon the confirmations, for they are gone the moment one begins to construct. If they stand at the beginning of the process of cognition they are logically of no use. Quite otherwise however if they stand at the end; they bring verification (or also falsification) to completion, and in the moment of their occurrence they have already fulfilled their duty. Logically nothing more depends on them, no conclusions are drawn from them. They con?stitute an absolute end. 

Of course, psychologically