Genuine faith in the Christian sense is that act of mind and will by which we discover all that we ever can come to know. Faith means trusting, and holding firmly to, what we have good reason to believe is true, in the face of trial. In that sense, no science, no history, not even the most platitudinous reasoning, would be possible without faith.I'm not sure what compels Marshall to assume this definition is ubiquitously held by Christians, but whatever. The problem with this definition is that it erroneously and needlessly conflates the concept of a provisional assumption with religious faith.
Epistemological humility – the idea that we could be wrong – is a necessary component of any reasonable epistemological framework. The alternative is rigid dogmatism, in which claims of absolute truth are asserted with absolute certainty and it's decided that no amount of yet-unknown evidence can even in principle undermine the validity of a belief. Provisional assumptions allow us to hold varying degrees of confidence in our knowledge depending on the evidence available while also allowing us to amend or even discard our current knowledge should sufficient evidence arise.
Let's take a simple proposition – I exist. My sensory perception gives me reasonable grounds to make this assumption, as does my consistent interaction with the physical world and other rational agents. Now, it's possible I do not exist. Perhaps I am the alter-ego of a dreaming alien, or a mere computer simulation in some vast Matrix-like computer network. But given that I have never seen, and likely in principle could not see, any evidence of such things, I'm justified in holding a very high degree of confidence in the assumption that I exist.
Another proposition might be the sun will rise tomorrow. This provisional assumption can also be held with a very high degree of confidence due to thousands of years of human experience as well as mountains of astronomical data collected on the Sun and Earth. Perhaps there is some yet unknown physical law that will cause the Sun to implode in the next 20 minutes. Given our thorough knowledge of these subjects – right down to the forces in quantum mechanics – such an event is so improbable that colloquially we deem it "impossible". We don't need "faith" that the sun will rise; we have overwhelming evidence that it will.
This is nothing like belief in the supernatural, for one extraordinarily important reason: the provisional assumptions of empirical experience are buttressed by a methodology that allows us to discern the validity of one or more alternative possibilities. If new evidence arises, it may force us to reconstruct or even discard our current models of reality. That's the caveat of an evidence-based world view – it's contingent on evidence!
But with religious claims to knowledge – Jesus is God, Muhammad is God's prophet, Buddha experienced enlightenment, whatever – there is no methodology by which to ascertain the validity of alternatives. This is precisely why religion's growth is accompanied by more and more schisms, while the growth of science is accompanied by a growing consensus. Science, being evidence-based, has a methodology for identifying and discarding erroneous claims. And while theologians talk of "evidence for God", absent any methodology to weed out invalid claims the faithful are left with nothing more than tautologies that can hardly be called "evidence" at all. The only way to find out whether you're wrong is to die.