In his lectures on the history of philosophy, Hegel said that beyond all others, Plato and Aristotle deserve the title of educators of the human race. A big part of what makes this true is what I will call their epistemic modesty. In contrast to the sweeping and very strong claims of many later philosophers, Plato and Aristotle were both masters of careful understatement.
Plato developed a very sharp contrast between knowledge and opinion. No opinion counts as knowledge, period. Arguments from authority or tradition may yield true conclusions, but if so it is not the authority or tradition that makes them true. To assert anything at all is implicitly to take responsibility for the assertion, and therefore to invite dialogue. For Plato, there is no knowledge in the strict sense (episteme) of anything that becomes. The most important component of wisdom is knowing what we do not know. Plato’s great literary homage to Socrates makes the latter a moral hero whose honest pursuit of truth got him executed, when his only real offense was embarrassing powerful people with questions they could not answer without admitting that they could not justify their positions. Some of his later students in the New Academy even pushed Plato’s epistemic modesty to the point of generalized skepticism.
Aristotle said that to know something (episteme) is to be able to give adequate reasons why it is true. Many things initially seem clear to us from intuition or opinions we have learned to accept, but this is only an apparent clarity. No immediate awareness counts as knowledge. We should treat the opinions of the wise with respect but, as with Plato, opinion can never be knowledge.
This last is a bit controversial, due to traditional interpretations of language in the Posterior Analytics about presuppositions and first premises. Aristotle recognized that you cannot rigorously prove anything without some starting point. But he explicitly uses a different word for knowledge (gnosis) to say we start from premises that we are better acquainted with, and work toward conclusions we are less well acquainted with. This is appropriate, because in modern terms his description of episteme in part makes it a function from premises to conclusions, but here he is talking about a beginning. One episteme (“science”) may prove premises of another, yielding episteme on a larger scale via a sort of function composition, but we still have to begin somewhere.
Aristotle is very keen to make distinctions, and to point out when the same word is “said in many ways”. Here he just uses a different, more informal word (gnosis) that typically has a connotation of personal acquaintance, as opposed to the technical concept of episteme. Unfortunately, Posterior Analytics often seems to have been read as meaning or implying that immediately accepted premises can be more certain than the reasoned conclusions. But there is no textual evidence that Aristotle considers gnosis to be in any way more certain than episteme. The imputation of an argument about certainty to Aristotle at this location rests on a circular assumption that certainty is required here. That sort of thinking belongs to a foundationalism that is utterly foreign to Aristotle. The only kind of necessity Aristotle recognizes is what Leibniz called hypothetical necessity, which is the if-then variety. At the beginning of the Topics, Aristotle explicitly says it is what he calls dialectic that evaluates first premises. Even though it employs the same structures of necessary reasoning as episteme or “science”, dialectic is expressly said to yield only “probable” conclusions (precisely because first premises are inherently uncertain).
The misinterpretation of Aristotle on first premises is partly due to the influence of Stoicism on nearly all Western philosophy after Aristotle. Stoicism is fascinating and original in its own right, but where Plato and Aristotle cultivated epistemic modesty and left many questions open, the Stoics claimed to have all the answers, to have unproblematic direct access to reality, and to have formulated it all in a complete, final system. Stoicism was the first philosophy to have significant diffusion in society at large. This was possible in part because overly strong, reassuring claims made for easier marketing. The dogmatism denounced by Kant did not actually infect all previous philosophy as Kant implied, but it was extremely influential, and Stoicism was its most important historical source. (See also Stoicism and Skepticism.)