there are so many variables that may or may not improve readability that it’s probably next to impossible to isolate them to measure them (...). all this talk about readability might possibly be a red herring which detracts from the amazing ability of the reader to adapt to a wide range of reading environments.

Absolutely. We had a similar discussion on the ATypI list recently, and if you don't mind the cross-posting, I think part of the discussion fits here perfectly. I remarked that "I believe graphic and type designers can learn quite a bit from complex systems theory, aka chaos theory."

David Berlow's reply (hope he doesn't mind my quoting him here) was: "What is/are the initial condition(s)?"

And this was my subsequent post:

And can you control all initial conditions so that they're all exactly the same (except for the one you're trying to test) every single time? In our specific case here, "hardly" is probably the answer, especially if we're trying to arrive at absolute laws of legibility.

But if we reduce the number of variables by dealing with one specific target audience (one culture, a limited age range etc) and one medium, then I believe that a well designed experiment can make a whole lot of difference to a design process. You're not looking for a single universal "objective" truth -- an immutable law -- but merely trying to determine how this or that design will work for a given target audience, on average.

As Paul Hunt pointed out, a very large number of variables would be involved in a purely objective analysis of whether or not ligatures contribute to legibility/readability, especially if the goal is to arrive at absolute laws that apply to the whole human race. And right now not enough research has been conducted to warrant definite conclusions as to whether ligatures help the human brain process printed text more efficiently/effectively.

So what? Does this mean typographers should stop using ligatures, or not trust their instincts? I don't think so. Perhaps what we can do is have a healthy scepticism (not paranoia, though) of our intuition and every now and then check if a specific design works for its target audience.