The type of universe whose investigation requires the methods of modern science must, I would suggest, have two characteristics: contingency and rationality. If rationality were absent, there would be no laws for science to discover; if contingency were absent, there would be no need for empirical observation and experiment, for every truth about the world could be deduced from first principles. The combination of the two characteristics is precisely correlative to a technique which believes that there are uniformities in nature and yet that these uniformities need to be discovered.[E. L. Mascall, Existence and Analogy, Longmans, Green and Co(New York: 1949) p. 9.]
Of course, since either one, rationality or contingency, in the robust senses Mascall has in mind, tends to be theism-friendly and uncongenial for the more obvious kinds of naturalism, it is unsurprising that a great deal of ingenuity in the past sixty years has gone into trying to find weaker proxies (purely epistemic analogues, pragmatic analogues, and the like) and, more recently, to find less limited naturalisms. I imagine that it is this that Mascall would write a book on today.
Needless to say, Mascall isn't the only person to note the point; Max Planck made it before him, noting that it was positivism's weakness, and Jaki argued it after him in his Gifford Lectures, and many others have made the point. But it does seem to require more systematic study.