Since this is multiple choice, I'll choose (b) and (c), with misgivings.
It is not hard to locate paradigms of what is meant by "rationality." At the etymological and stripped-down, Hobbesian level, rationality is a coherent, reliable system of input-output calculations. The model for Hobbes and many others was Euclid's axiomized geometry. It is "in the eye of the beholder" in the sense that the correlating axioms, assumptions, applications, and social contexts of such systems will vary.
The problems in arriving at one, "unified account of rationality" are numerous, famous, and fraught, constituting a virtual history of philosophy.
First, there is the problem of the axioms, as the case of Euclid's fifth postulate famously illustrates. Axioms must ultimately be grounded in consensus and intuitions, which are unreliable leaps of faith, historically variable, and always susceptible to skeptical demolition, as Berkeley, Hume, and others demonstrated. And one can always reason from "biblical axioms" or "racialist axioms" and be perfectly rational, in the eyes of some beholders.
Second, the reliable symbolic systems of mathematics and logic must be translated into value-laden, oral languages, where all sorts of "bewitchments" await. This exorcising task inspired the logical atomism of Frege, Russell, Wittgenstein, and others, all of whom conceded defeat, some of whom also conceded that the "irrational," evolving plasticity of language was perhaps just as well.
Third, any rational system that wants to advance beyond tautology must attach itself to the slippery contingencies of experience. Carnap and others attempted to stick manipulable symbols to sense data. Kant demonstrated how the analytical concepts of reason could produce synthetic conclusions about experience, but only at the sad cost of truncating reason and possible knowledge.
Fourth, all rational systems that extend their domains must live in doubt of their own internal coherence. Such doubts were put to rest by Godel's mapping of Russell and Whitehead's great attempt at a "universal account of rationality." No rational system, even cleansed to consistent symbols, could ever "prove itself" or ensure it own completion and consistency. Game over.
So, even apart from vague, colloquial uses of the term "rationality," we are a left with a historical scrap heap of "expert-built" rational systems, which remain perfectly useful, but are unlikely to be unified into a single account to everyone's satisfaction.
But people never give up. Set theory is said by some to be the most fundamental, unified account of various rational systems, though I wouldn't know. Efforts are underway to employ a Quinean naturalized epistemology linking set theory with perception-based "cell assemblies" theoretically accessible to neural-cognitive tracking. Game reboot.
