of error? But different sensations do not refer to the same object, and similar sensations have equal value. Nothing remains, therefore, but to attach implicit belief to every impression of the senses. Every such impression is directly certain, and is accordingly termed by Epicurus clear evidence (ἐνάργεια). [1130] Nay, more, its truth is so paramount that the impressions of madmen, and appearances in dreams, are true because they are caused by something real, [1131] and error only becomes possible when we go beyond sensation. [(2) Notions.] This going beyond sensation becomes, however, a necessity. By a repetition of the same perception a notion (πρόληψις) arises. A notion, therefore, is nothing else than the general picture retained in the mind of what has been perceived. [1132] On these notions retained by memory depend all speaking and thinking. They are what commonly go under the name of things; and speech is only a means of recalling definite perceptions [1133] to the memory. Notions are presupposed in all scientific knowledge. [1134] Together with sensations they form the measure of the truth of our convictions; [1135] and it holds true of them as it did of sensations—that they are true in themselves and need no proof. [1136] Taken by themselves, notions, like perceptions, are reflections in the soul of things on which the transforming action of the mind, changing external impressions into conceptions, has not as yet been brought to bear. [(3) Opinion.] For this very reason notion are not sufficient. From appearances we must advance to their secret causes; from the known to the unknown. [1137] Far too little value was attached by Epicurus to the logical forms of thought, or he would have investigated more accurately the nature of this process of advancing. [1138] Thoughts, in his view, result from sensations spontaneously, and although a certain amount of reflection is necessary for the process, yet it requires no scientific guidance. [1139] The thoughts arrived at in this way do not stand as a higher genus above perceptions, but they are only opinions (ὑπόληψις, δόξα) without a note of truth in themselves, and depending for their truth upon sensation. That opinion may be considered a true one which is based on the testimony of the senses, or is at least not contrary to the senses, and that a false opinion in which the opposite is the case. [1140] Sometimes we suppose that upon certain present impressions other impressions will follow: for instance, that a tower which appears round at a distance will appear round close at hand. In that case, if the real perception corresponds with the assumption, the opinion is true, otherwise it is false. [1141] At other times we suppose that certain appearances are due to secret causes: for instance, that empty space is the cause of motion. If all appearances tally with their explanations, we may consider our assumptions correct; if not, our assumptions are incorrect.