Hopper’s practical approach is of course the appropriate one … if this is a repair job.
If the objective is to produce a functional bicycle, then the practical man uses whatever tools and technique are available … to produce something that works [possibly better than the original].
If, however, this is a restoration job on a ‘significant’ bicycle, then it may be appropriate to identify exactly what the original maker used in the first place.
[ the male threads on the pedals could presumably be measured for starters ]
Perhaps the most academically interesting possibility [as yet unmentioned] is that someone has reason to want a Failure-Mode Analysis … in which case the whole can-of-worms is open!
MichaelG.