I’ve returned to some of Richard Primus’s work, and want to note the succinct summary of two different models of constitutional decisionmaking given in the generally excellent “When Should Original Meanings Matter?”
The first model seeks to connect constitutional decisionmaking to the values that constitutional law seeks to promote. On this account, “[t]he validity of [constitutional methods of reasoning] as aids to constitutional decisionmaking is a function of their relationship to the [constitutional] values . . . More completely, it is a function of their relationship to [constitutional] values . . . as refracted through the institutional roles of the decisionmakers.”
This focus on constitutional values highlights the contingency of different modes of constitutional reasoning. Different constitutional values – assuming there’s more than one – can be served by different kinds of adjudication (e.g. stability might be served by irrepealably judicial codification of a well-developed body of law, whereas democracy is obviously hindered by such a maneuver). Additionally, even the same constitutional value can be served by different modes of reasoning at different times. To take a contrived example inspired by Primus: if all citizens unanimously pass a law depriving a subset of citizens of the right to vote, democracy might best be served by respecting that legislative choice . . . for a time. A hundred years later, however, it is less clear that a commitment to democracy requires respecting the antidemocratic wishes of a long-dead generation.
The second model, which Primus credits to Phillip Bobbitt, is the obverse of the first. Whereas the first says that the goal of constitutional decisionmaking is to respect constitutional values, the second says that constitutional values are determined by constitutional decisionmaking. Constitutional law just is a set of rhetorical practices, or modes of reasoning, many of which include appeals to certain constitutional values. But the validity of methods of constitutional reasoning does not depend on how well they serve constitutional values. Rather, “it is the forms of argument that supply legitimacy to the values and not the other way around . . .”.
It might be thought that these two models of constitutional decisionmaking are not actually at odds with each other. Rather, the first model might be thought to provide a prescriptive account of how constitutional decisionmaking ought to proceed, whereas the latter is a descriptive account of what actually happens.
There are two complications. First, prescriptive constitutional theories must always be reformist, otherwise they amount to recommending the abandonment of constitutional law. If Primus’s prescriptive model cannot vindicate a sizeable swathe of constitutional practice then it is no longer a model of constitutional law. Provided Bobbitt’s model is an accurate description, then, it sets limits on the prescriptions available to a constitutional theorist. (To be concrete, if, for example, decisions by a Primus-style decisionmaker would only infrequently advert to precedent then Primus’s model would cease to be a prescription within constitutional theory, and become a call to do away with our Constitution altogether. This is an unlikely result, however, as Primus’s model has built into it a respect for the status quo.)
Second – and more importantly – Bobbitt’s model can itself be reformulated prescriptively (though Bobbitt does not do so himself). It is arguable that decisionmaking that only contingently connects to constitutional values is preferable to decisionmaking that flows from them, because it makes it easier for judges to arrive at results that accord with particularistic evaluations of fact patterns (as opposed to evaluations rooted in necessarily general, system-wide, constitutional values). Unsurprisingly, people who are antecedently committed to constitutional values will find this approach distasteful, but those who like the system precisely because of its normatively flexibility won’t be disturbed. But this is just to say that there can be partisans of Bobbitt’s model, as of Primus’s model, so the two can potentially conflict.
 Primus explicitly declines to articulate either a comprehensive list of constitutional values, or a methodology for determining which values go on the list. He more or less takes as a pretheoretical given that there are multiple constitutional values, though he also argues against Keith Whittington’s suggestion that fidelity to the constitutional text is, at least, lexically prior to all other constitutional values. (Whittington does not think it is the only constitutional value – witness his interpretation/constuction distinction – but the argument in Chapter 3 of his book Constitutional Meaning assumes that, properly understood, a written constitution directs constitutional decisionmakers to “interpret” before anything else. Accept the written constitution and you are committed, first and foremost, to fidelity.)