Answer

According to my explanation of establishing interpretive cases, here too we need to look for an explanation according to which the seizure is what is under discussion, and the silence is only a condition without which the question would not arise. The Talmud rejects the possibility that he remained silent, because in such a case there is an admission here and the seizure has no significance at all. It therefore reaches the conclusion that he was silent at first and afterward protested / cried out. In that situation, the discussion is about whether his silence is significant or not. This can be understood in my approach in two ways, and at the end I will add a note about a third direction:

1. Even if there was silence here, the Talmud does not mean to say that the silence is an admission regarding the subject of the discussion itself (the cloak), but rather acceptance of the seizure itself as legitimate. For example, the one who was attacked might claim that the cloak really is his, but in fact the one who seized it is the current possessor (and perhaps was already the possessor beforehand, and the attacked party himself had seized it from him). The assumption is that in a situation where the seizure is not accompanied by immediate protest, it succeeds in creating possessory status for the one who seized it. Now the attacked party cries out, but he is already in the position of one trying to extract property from a possessor. And if this is not considered silence, then the seizure is ineffective and the usual rule of “two are holding” applies. According to this, the discussion really is about whether such a seizure creates possessory status or not, in accordance with the plain meaning of Rabbi Zeira’s wording.
2. Once there are two possibilities regarding the silence, we are in doubt as to whether there was meaningful silence here or not. I am now assuming that the silence does relate to the cloak itself, and not to accepting the seizure as in option 1. We are now in a state of doubt, and Rabbi Zeira’s question was what we do in such a case. That depends on whether the seizure created possessory status or not. According to this suggestion, Rabbi Zeira’s doubt is not about which of the two sides is correct; rather, he assumes that this is a case of doubt. The discussion is whether, in such a doubtful situation, the seizure creates possessory status, and therefore despite the doubt the burden of proof rests on the attacked party, or not. And again, this is a question about the seizure. I should note that this does not seem to fit the plain wording of the Talmud, since the Talmud presents the two sides of the doubt as an explanation of Rabbi Zeira’s doubt, and not as the basis upon which his doubt is formulated, as I suggested. But the text can be stretched that way.
3. In the case of a question raised by an Amora, this may be different than in the Mishnah. Rabbi Zeira asked a question. We do not have its exact formulation before us, since this is not a canonical formulation like in the Mishnah, and therefore here we may be uncertain what exactly his question was. In such a situation, it is possible that Rabbi Zeira’s original doubt was about the nature of the silence and not about the seizure, and the initial formulation of his question was simply not precise.