May, might, can, could, seemingly, apparently, perhaps, possibly, proposed, expectedly, unexpectedly, surprisingly, unsurprisingly, presumably, putative, ...litter scientific literature. If we remove these ambiguous terms, the length of most article would be halved, and over 95% would never be published.
These terms are never used for expressing statistics, because statistics have anyway overpowered biology in medicine. Because numbers never lie, like Swiss Bank account passwords, a halo surrounds numbers and their unholy manipulations. Four P values ensure publication of an article, whether or not most readers and editorial teams (to the tune of 90%--barring speakers in Medical Conferences where eminence scores over pure nonsense, eg., the crush to somehow embalm Cortical Spreading Depression or genetics or neuropeptides into migraine pathophysiology and neurotherapeutics without considering lateralization of headache and the essence of the blood-brain barrier) (see attached files).
I reproduce one of the finest quotes I have come across in 50 years of scientific undertaking: "Science is an ongoing race between our inventing ways to fool ourselves, and our inventing ways to avoid fooling ourselves"--Regina Nuzzo, 2018. The scientist can win both ways, and, the discipline will also pay the price either way. Many times the idea is so big, that even if proven wrong by the sword of logic, it is not possible to backtrack, the classical examples being Cortical Spreading Depression or magnesium supplementation in neurovascular catastrophes or non-invasive vagal nerve electrical stimulation for treating migraine. The price then paid by society-at-large is catastrophic-beyond-imagination. "In God We Trust" turns to "In Digital We Trust" (Jon Baldwin, 2018).
Digital concealment and its compounding damage -- to paraphrase Einstein -- is a truly trans-galactic swallowing by a series of immense limitless black holes.
That is why in every age there is need for an Einstein as well as an Aldous Huxley.