Truth without interpretation. It seems you are questioning the reliability of using 'testimony' or 'observation' as an evidence for an argument/event. I I will use my own 'observation' to offer a possible solution to this question.
According to my experience, there are situations where you could only know the full truth, if you at least believed/assumed the partial truth. For example, in a mathematical inequality equation, you would substitute a number and assume it to be true; and if x is true, then x must be true for all situations (that satisfy the conditions) else x is false.
Now I'm going to answer, using deductive reasoning derived from sensory observation of partial/uncertain truths:
Empirical data based from observation is reliable and strong evidence.
There is limited things/events which a person can observe within his field of sensory-experience.
A person can't observe everything, sometimes he must rely on others observation to get information.
This is secondary information, which 'truth' can get stained by other person's personal interpretation.
Assume: We need all information we can get for humanity to progress and for civilization to evolve.
Conclusion: We need to trust other person's observation, test its truthfulness, build new ideas around it. Even if it means accepting 'truth' that is uncertain, even if 'truth' is primitive or imperfect, even if it means getting the 'truth' contaminated by bias and subjective interpretation. As long as this 'truth' can be useful, we can rely on other person's 'truth' and learn new things.