British physicist David Tyler points to a recent paper by L. Vogt in the journal Cladistics which explains why attempts to construct a tree of life are generally unfalsifiable:

Putting this in more popular language, cladists have adopted a variety of rationales to justify giving weight and credence to their evolutionary trees, but these rationales do not survive critical scrutiny if the test is Popper's demarcation criterion for science.

[ ... ]

The debates within evolutionary circles are always about specifics: the broader issues are not debated because they have an axiomatic status. So, evolutionary theorists do not have the mental tools that would allow them to disprove common ancestry, or whether design inferences are warranted. Consequently, it is not unreasonable to conclude, from the perspective of empirical science, that proposed evolutionary scenarios represent not "scientific but metaphysical hypotheses".

The Vogt paper suggests that the common science criterion that a theory should be falsifiable (able to be shown to be incorrect) be abandoned where evolution is concerned.