Free and Frank Advice: Contested concepts, the remainder and agonism

06/06/2026

I explain why free and frank advice remains undefined, drawing on the essentially contested concept, the remainder and agonism to argue that a definition would take a side in the contest it claimed to referee.

This is the seventeenth of twenty posts drawn from my doctoral thesis on free and frank advice in Aotearoa. In it I take up a question the four fault lines leave behind: why the literature that names those tensions cannot tell us which side of each is correct, and why leaving free and frank advice undefined is a requirement of democratic administration and not an omission waiting to be repaired.

The four fault lines I described earlier in this series are structured, recurrent and attached to normative ideas that matter to governing, and that raises a fair question about why scholarship cannot simply say which pole of each fault line is right and which is mistaken. My answer in the thesis is that scholarship cannot declare a winner without dismantling the practice it sets out to explain, and I draw that answer from three theories that form a whakapapa, an intellectual lineage of teachers and students, questions and responses passed from one generation of scholars to the next. In this post I follow the order of that lineage: first, the claim that some political concepts remain essentially contested; second, the claim that they remain contested because a definitive closure carries costs a democracy cannot bear; and third, the claim that the continuing contest can nevertheless be given a durable and workable form.

Free and frank advice as an essentially contested concept

The first claim starts from the idea of the essentially contested concept, a class of concepts that work as terms of accreditation, awarding a valued status to a practice more than describing a neutral fact (Gallie, 1956). Because the concept awards a status, a dispute follows over what earns it, and the dispute is a knowing one, in which each party agrees the term names something worth having and recognises that its own use is contested by the others, so that to use such a concept means “to use it both aggressively and defensively” (Gallie, 1956, p. 172). In my reading free and frank advice behaves in exactly this way, since everyone who gives or receives it agrees that it matters while disagreeing, often quietly, about which advice deserves the name.

The fit between free and frank advice and the conditions for membership of that class is close, and in the thesis I test it condition by condition, taking the bounded plurality whole. Free and frank advice accredits, because all six postures use the term for advising done properly and the statute names it as a principle of the public service; it is an internally complex achievement whose worth practitioners attribute to the whole; and that worth is describable in rival orderings, which is what the six postures are. Each factor array is an ordering of the same 34 statements, each gives different priority to the parts of the obligation, and here the methodology and the philosophical condition meet, because a Q-sort is an ordering of parts and the factor solution is the empirical record of rival orderings of one valued practice. The concept is also open, as both its four decades of statutory evolution and practitioners’ daily passage between postures suggest, and the validation stage showed practitioners recognising the rival orientations at once and defending their own while acknowledging the pull of the others. I find that correspondence one of the more satisfying results of the study, because it came out of the data and was never designed into it.

From that correspondence follows the reason scholarship cannot rule between the poles, which the carriage of the idea into political theory makes explicit. An authoritative definition of an essentially contested concept states the tests a practice must pass to earn the award, and because rival sets of tests exist, each expressing a judgement about which parts of the whole matter most, choosing the tests is a choice among the values the concept carries (Connolly, 1993). To rule for the mandate pole of the first fault line would define free and frank advice as responsiveness to the minister’s programme, and to rule for public reason at the second would hollow out the protected space in which the hardest reasoning between ministers and officials takes place, and in each case the ruling would sever a part from a whole whose worth practitioners attribute to the whole. That is why, following the tradition of the sensitising concept (Blumer, 1954), I offer the bounded plurality as a direction in which to look and a shape to expect there, and never as a definition stating what must be found in every case.

The remainder and the cost of closure

The second claim explains what the first leaves open, namely the mechanism that makes leaving free and frank advice undefined an advantage, and the harm that would follow if a definition were imposed. The remainder first appears in moral philosophy, where a choice between two conflicting obligations leaves the claim of the unchosen course present, producing regret and a demand for reparation (Williams, 1973). Carried into political theory, the remainder becomes something institutions produce, because every settlement engenders resistances, even a settlement that is relatively empowering (Honig, 1993). In my reading the shift matters for free and frank advice because it locates the cost of closure in the institutional act of ordering, which is exactly where a statutory definition would sit.

Applied to the bounded plurality, a definition imposed by Parliament or by executive guidance would close the concept across all three dimensions of its structure at once. Fixed around one posture, it would brand the other five illegitimate, and fixed around one pole of each fault line, it would turn the claim carried by the opposing pole into a remainder pressing back against the order from outside it. A definition centred on responsiveness would leave principled independence without formal standing, one centred on public challenge would dismantle the confidentiality frank counsel requires, one centred on routine candour would disable the protective failsafe, and one centred on an unyielding duty to speak would outlaw the craft of timing. My view is that a closure of this kind would make enemies of politicians and officials, because it would turn a legitimate advisory tension into a contest for institutional supremacy in which one side’s claim is cast as obstruction and the other’s as interference.

The remainder also answers the constitutional puzzle that opened the thesis, which is how a convention can be recognised as obligatory while its content remains uncertain (Marshall, 1984). Later scholarship carried that puzzle forward as two gaps, one concerned with identifying what conventional norms are and one with explaining how they bind without codification (McLean, 2024), and the thesis meets the first empirically, through the bounded plurality, and the second through the avoidance of closure. The Public Service Act 2020 shows that avoidance at its most legal moment, because the Act named free and frank advice as a principle and stopped short of defining it, and there “is no public sanction if the advice given is not free and frank” (McLean, 2024, p. 284). I read that naming without closure as the system keeping every legitimate claim the convention carries inside the plurality, where the claims meet each other as recognised rivals instead of pressing on the order from outside it.

The contest is real enough to show in closures below the level of statute, and the Stage Three discussions gave numerous examples of attempts to make one pole of a fault line the working rule of an agency, a government or a period. A working rule of that kind is a closure in miniature, and the advisory literature records how the claim set aside tends to return: where responsiveness dominates, independent judgement returns through functional politicisation, as officials anticipate preference and adjust advice before any direction is given (Eichbaum & Shaw, 2019), and where openness demands a written record at every stage, protected counsel can return through oral channels and thinner papers (White, 2007a). I take those returns as evidence that the remainder is a working feature of advisory systems in Aotearoa and more than a theoretical possibility.

Agonism and the shared frame

The third claim explains how a permanent tension can operate in ordinary governing without collapsing into paralysis or open conflict, and agonism supplies the explanation. Democratic institutions carry disagreements over power and authority that no final rational consensus can reach, and the democratic task is to shape the form conflict takes, which turns on the distinction between antagonism, in which opponents treat each other as enemies whose claims lack legitimacy, and agonism, in which they treat each other as adversaries whose right to contest within a shared enterprise is recognised (Mouffe, 1993, 2000). Agonism requires both genuine disagreement and a shared frame, the condition known in that tradition as a conflictual consensus. In my view the bounded plurality works as a conflictual consensus of exactly this kind, with the six postures as the genuine disagreement and practical statecraft as the frame they share.

Practical statecraft gives the contest its common test and its role clarity, since ministers carry democratic authority, electoral legitimacy and the right to decide, while officials carry institutional memory, analytical expertise and knowledge of implementation consequence, and neither side can absorb the other or govern alone. The test applies in a fixed order: its content is the integration of evidence, expertise and practical knowledge of consequence, especially knowledge of how policy will work once it meets the world, and that content must be brought together between ministers and officials out of their different knowledge, authority and responsibility. Free and frank advice is the means by which those different authorities meet. I read that two-part test as the reason the contest among the postures stays agonistic, because it keeps the exchange tied to the consequences of governing and away from both ideological compliance and technical abstraction.

Agonism stops short of relativism, and the six postures remain legitimate orientations only while two conditions stand. Each posture must continue to satisfy the shared test of practical statecraft, which protects the quality and governing use of the advice, and each must continue to recognise the legitimate standing of the others, which protects the plurality itself. Closure begins where either condition fails, as when a ministerial office insists on responsiveness to the exclusion of everything else, or an agency treats its own policy preference as superior to the elected mandate, and in either case adversaries become enemies and the remainder names the democratic cost that follows. In my view this is why the contest cannot be ruled on, must not be closed, and does not need to be, because practical statecraft already gives it form, and that balance lasts only while the plurality remains open.

Next, I take up what happens when one posture attempts to occupy the whole space, and argue that the work of free and frank advice is more political than the ordinary language of the public service admits.

References

Blumer, H. (1954). What is wrong with social theory? American Sociological Review, 19(1), 3–10.

Connolly, W. E. (1993). The terms of political discourse (3rd ed.). Blackwell.

Eichbaum, C., & Shaw, R. (2019, June 26–28). Two faces of politicisation: Administrative and functional politicisation within the New Zealand core executive [Conference paper]. International Conference on Public Policy, Montréal, Canada.

Gallie, W. B. (1956). Essentially contested concepts. Proceedings of the Aristotelian Society, 56, 167–198.

Honig, B. (1993). Political theory and the displacement of politics. Cornell University Press.

Marshall, G. (1984). Constitutional conventions: The rules and forms of political accountability. Oxford University Press.

McLean, J. (2024). Conventional norms and the comparative constitutional law project. Comparative Constitutional Studies, 2(2), 272–291.

Mouffe, C. (1993). The return of the political. Verso.

Mouffe, C. (2000). The democratic paradox. Verso.

White, N. (2007a). Free and frank: Making the New Zealand Official Information Act 1982 work better. Institute of Policy Studies. Williams, B. (1973). Problems of the self: Philosophical papers 1956–1972. Cambridge University Press.