Another Small Point in Republic I

Here's another passage which caught my eye this time in Republic I. This one is right after Thrasymachus enters the discussion for the first time. You will recall, he tells Socrates not to define justice as what is advantageous or profitable, and then, with derision, he predicts that Socrates will not, after all, attempt to give a definition but will merely ask questions and refute answers that other people give. Once again, the Griffith translation:

"Anything to allow Socrates to play his usual trick--not answer the question himself, but wait for someone else to answer it, and then take what he says and try to prove it wrong."
"Really, my dear fellow!" I said. "How could anyone answer the question if for a start (prw=ton me\n) he didn't know the answer--didn't so much as claim to know it (mh\ ei)dw\j mhde\ fa/skwn ei)de/nai)--and on top of that, even supposing he did have some idea on the subject (ei)/ ti kai\ oi)/etai), if he'd been told by a man of some authority not to say any of the things he thought (mhde\n e)rei= w(=n h(gei=tai )? No, it makes much more sense for you to speak. You're the one who claims to know the answer (fh\|j ei)de/nai) and have something to say." (337e4-338a1)

This puzzles me because it seems as if Socrates is saying that it is not possible to try to answer a question, without being convinced that one knows the answer. And yet this seems wrong: Couldn't one answer, rather, with what seems likely? If, to save the claim, one wants to insist that someone who says what seems likely to him knows that it seems likely, then the claim becomes empty, because presumably everyone always knows what seems true to him.

Also, is the 'method of hypothesis' of other dialogues perhaps developed with a view to this? That is, someone who thought that one could genuinely answer a question only if one knew the answer might think that a hypothetical answer would be a way of loosening, in a sense, the restriction. (One knows that the hypothetical claim is true but cannot assert the antecedent or consequent.)

Also, Socrates claims elsewhere that he knows that he does not know. Does that imply, then, that there are only a few questions that he could answer, but these are only about his own knowledge? (Presumably he could say, "No", if asked, as here, whether he can give a definition of justice.)

Or am I misreading the passage because "for a start ... to know" and "on top of that ... have some idea" is an example of what lawyers call "pleading on a hypothesis" and what Socrates is saying is that he could answer, if he had some idea (which falls short of knowledge), but even then he couldn't give his answer to Thrasymachus, who has ruled out the alternatives he would be likely to offer?