Maybe the problem is logic? Logic is quantized, usually binary yes/no.
A fetus is either a person or it is not.
Maybe if we instead of starting from absolutist principles we started from Bayesian priors?
For some people the start of personhood would be narrowly centered on conception and for others on birth, but many would be not so sure where it is.
We can combine all these distributions and find a peak consensus of when personhood starts.
In much of Europe the term limit for abortion is 12 weeks. There is no rationale for a switch from fetus to person at this time point but it is a solution of sorts.
Not sure about your science example. The new paradigm gives measurably improved prediction of the results of physical experiments. So these are not equal paradigms.
Also what has science to do with morality?
We take it is a moral principle now that people should be treated equally, but before WWII science was enamoured with eugenics.
I have worried over this issue for a while.
One approach is to consider that equality is a firm moral principle and therefore to study the question at all (as in Charles Murray's notorious Bell Curve) is the act of an immoral person. The problem with this stance is that it is anti-science (some areas of study are banned)
The other approach I am considering is that of a strong Bayesian prior. We can consider that genetic similarity between races is >99.9% and therefore it is unlikely that there is any difference in capabilities. Murray's evidence must be due to confounding factors such as poverty etc and our opinion is not changed.
An oddity with this approach is that where one positions one's prior seems to have a moral dimension, those that position their prior most narrowly on the equality point being the most moral and those that position their prior more broadly or away from this point being less moral. A kind of continuous morality spectrum which seems strange. We think of priors as being a subjective synthesis of a persons knowledge about a subject, but maybe they can have a moral dimension too?
