A hypothesis is a conjecture, a speculation, a hunch, framed in such a way that it can be tested. The result from this test is accepted, until new evidence is available and regarded (Gee, 1999). But, testing a hypothesis requires that it makes a prediction that can be checked by observation, and we make observations in order to learn about things that we do not observe (Sober, 1999). Our observations, then goes far from the mere action of sensing an object, and that's why our observations are always full of theory. Although, when we test competing hypotheses we use the same tools for doing observations, and the theoretical machinery could be regarded as auxiliary hypotheses that do not alter the test result.
Cladistics, is a historical science, then it emphasis in analyzing and shapering traces at the light of a hypothesis (Cleland, 2002). In cladistic analysis a hypotheses is a presume relationship between taxa (a monophyletic group), and the evidence that 'corroborates' a group, are the synapomorphies (Patterson, 1988). But, at the beginning of the analysis there are not synapomorphies, just characters, observations that need a theoretical context. However, that a theoretical construct is necessary for evidence interpretation that is not an excuse for not seeing it with out bias about what it tells (i.e., adaptation, change sequence).
Because our hypotheses and methods always could be refined, and we would never have all the information any presumption is susceptible of being accepted or refuted in the light of new evidence.