Francis Galton originally used the term "regression to mediocrity" to refer to the phenomenon that children of very tall parents were on average less tall. More generally, the heights of children over multiple generations would regress toward the mean, or average, height of the population.
Today, "regression" is used to describe a particular statistical model in which data is distributed around a mean.
Galton writes,
The mean regression ... is easily ascertained
intending to mean something like
It is easy to ascertain the mean amount of regression
but a careless reader could instead understand
It is easy to ascertain this thing called a "mean regression"
of which the contemporary meaning of regression is a natural extension. A phenomenon ("regression") was described with a statistical model, but the name of the phenomenon came to be used to refer to the model itself.
I'm struggling to think of a less esoteric example, although I feel like there should be more of these. Is there a name for this kind of error? Are there more familiar examples of it, ideally ones that don't rely on somewhat archaic language?