- observations lead to conclusions
- conclusions, based on observations, lead to new conclusions
- conclusions do not lead to themselves
- conclusions do not lead to observations
You might suspect something to be true (a conclusion). You question why you believe it to be true -- hypothesis formation. When you fail to find reasons to believe it to be true, you abandon the hypothesis. It could be that:
- The hypothesis is indeed true, but you have failed to find reasons, for now.
- The hypothesis is not true.
If you can support your hypothesis with reasoning that leads to its suspicion, you then perform carefully designed experiments to support it further. If these experiment falsify the hypothesis, you abandon it.
However, abandonment of a hypothesis does not mean to throw away all your work. For example, you might modify your hypothesis -- this is the same as throwing your old hypothesis away and developing a new hypothesis. You then repeat more carefully designed experiments until you have failed to demonstrate the hypothesis to be false.
You might have a hypothesis for which it is not possible (even given unlimited resources) to develop an experiment for. An unfalsifiable hypothesis should be discarded immediately as it has no practical purpose. A hypothesis may have the ability to develop the experiment for, but it is not tractable e.g. it would cost six zillion dollars to execute the experiment -- this is a falsifiable hypothesis, but it is unable to develop any further while the tractability barrier persists.