2.1. A claim about commercial background

What background conditions does Nozick’s possibility proof presuppose? On the Lockean interpretation that Nozick explicitly inherits, agents already recognise norms of property, promise-keeping and mutual restraint, and they are capable of converging on conventions that regulate interaction in the absence of a political sovereign. In this respect, Nozick’s setting is closer to a world of evolving conventions than to a primitive war of all against all. Geography matters too: Nozick discusses local dominance and the declining effectiveness of enforcement over distance. I leave those spatial elements aside, not because they are unimportant, but because the question here is narrower: whether the commercial background suggested by division of labour and market pressures changes the logic of affiliation. The burden of the argument is accordingly limited: the paper need not show how every part of Nozick’s story works, only what happens once protection is placed in a setting where specialised exchange already matters.

The claim I need is modest. Once protective services are supplied by specialists and entrepreneurs, individuals must face sufficiently high opportunity costs of self-help for it to be rational to purchase protection. This requires some degree of division of labour: individuals earn more by applying their time and capital to their own trades than by personally adjudicating disputes. In addition, where specialised producers transact frequently and with many partners, they derive incremental benefits from sharing common contract forms, evidentiary standards and remedial procedures. Protection thereby operates not only as a shield against predation but also as a way of making exchange more regular and reliable. In a sufficiently rich state of nature, repeated interaction and mutual advantage may sustain conventions that function as proto-legal institutions, even before a state exists. This is the sense in which Nozick’s references to division of labour and market pressures do more than set the scene. They suggest a world in which the organisation of protection can affect the productivity of exchange itself.

This is best read as a comparison with thinner baselines like Bruner’s, not as an attempt to say everything relevant about Nozick’s state of nature. Bruner asks what follows if protection serves only to reduce conflict losses. I ask what changes if one also takes seriously Nozick’s emphasis on specialisation and market pressures and allows productive gains from shared procedures. Section 3 spells out that difference without claiming to reproduce the whole of Nozick’s emergence story. The point is not that Bruner’s approach is incomplete, but that a thinner baseline and a richer commercial baseline generate different dynamics, and that difference matters for how we read Nozick’s invisible-hand argument.Footnote ¹

2.2. Side constraints and the economic model

How do Nozick’s side constraints interact with the economic forces he invokes? The model developed in Section 3 treats agents as expected-payoff maximisers over a limited set of institutional affiliations (joining a particular agency or remaining independent). This is an idealisation designed to track the economic channel in a background where normative constraints have already filtered out impermissible strategies. The relationship between the philosophical and formal levels can be stated succinctly. At the philosophical level, side constraints and compensation requirements determine the admissible institutional paths: for example, a dominant association may neither coerce membership nor deliberately worsen the position of independents. At the formal level, these constraints appear as parameter restrictions and modelling choices: the dispute payoff of independents is weakly increasing in the share of agency members, ensuring that consolidation does not harm nonmembers; transitions are always voluntary, with no forced membership; and the agency is assumed not to employ exclusionary tactics that would increase its profits at the expense of independents.

The model does not attempt to reproduce Nozick’s ethical reasoning directly. Instead, it asks a conditional question: given a set of institutional profiles that pass Nozick’s normative filter, how do division-of-labour considerations and institutional compatibility shape the relative attractiveness of those profiles? This ‘economic within the normative’ orientation motivates the compatibility-augmented analysis. It allows us to separate two issues that are often run together in discussions of Part I of ASU: whether a minimal state could in principle be morally permissible and whether the particular invisible-hand mechanism Nozick describes is a robust route from morally constrained anarchy to such a state once its commercial background is made explicit.

3.1. Motivation: beyond violent conflict

In Bruner’s environment, individuals are self-biased in disputes, violent conflict is costly and third-party adjudication is valuable insofar as it suppresses retaliatory feuds. When protection is modelled purely as loss avoidance, agency membership takes the form of a public good: nonmembers can free-ride on reduced violence without paying the agency fee. When correlated conventions coordinate agents on peaceful behaviour, demand for affiliation may vanish. Bruner’s analysis thus shows that, in a thin Lockean setting where protective associations only reduce conflict losses, there is a structural pressure toward free-riding that undermines the emergence of a dominant protector.

But in a commercial pre-society of the sort suggested by Nozick’s own references to division of labour and market pressures, the primary frictions often arise not from violent confrontations but from trading disputes – failures of delivery, ambiguities in contracts, contested titles, disagreements over evidence and remedies and so forth. Protective and judicial institutions play two conceptually distinct roles in this environment. They suppress violent escalation in disputes, and they standardise commerce by providing common adjudicators, shared evidentiary standards, predictable remedies and recognisable contract forms. The second function has a direct productive effect: when two specialists operate under the same protective institution, their legal expectations are aligned ex ante, and disputes are resolved in a timely and legible manner ex post, raising the likelihood that joint projects succeed and that the gains from specialisation are realised. The formal move in this section is to give this second, productive role a simple probabilistic representation.

3.2. The compatibility premium

To capture this commercial standardisation, I introduce a compatibility premium in the production stage: an incremental return that arises when parties transact under a shared institutional framework of common procedures, evidentiary standards and enforcement rules. Let $\alpha$ denote the probability that a joint project between two agents succeeds when they share the same protective institution, and let $\beta \lt \alpha$ denote the probability when they do not (either because one or both are independent or because they belong to different institutions). The gap $\alpha - \beta$ is the compatibility premium.

The assumption is modest. Nozick’s remarks about division of labour and market pressures do not by themselves prove that shared legal rules function like network goods, but they make it natural to ask what follows if common procedures have productive value. The premium is therefore an additional assumption, prompted by Nozick’s commercial remarks rather than extracted deductively from ASU. Nor does $\alpha \gt \beta$ smuggle in a utilitarian calculus: it simply encodes that among institutional profiles that satisfy Nozick’s side-constraint filter, those with shared procedures can yield higher expected productive success. It is also distinct from the coordinating role of conventions. Correlated conventions may eliminate violence-related losses, but they do not by themselves generate a full legal infrastructure of evidence, remedies and contract forms that systematically reduce bargaining breakdown and enforcement delay. The compatibility premium captures a channel that conventions alone cannot supply.

The size of $\alpha - \beta$ should vary with the setting. Where agencies recognise one another’s judgements, share arbitral forums, or operate under appellate arrangements that harmonise procedures, the gap narrows – and in the limit of full interoperability it vanishes, pushing the model back toward Bruner’s environment. Where each agency instead uses idiosyncratic evidentiary standards with no mutual recognition and where exchange is frequent and time-sensitive, the premium will be correspondingly large. The magnitude of the premium is not empirically calibrated; what matters for the mechanism is its sign and its interaction with the fee $e$ and the dispute-avoidance benefit $B$ in shaping incentives to affiliate.

3.3. Environment and timing

Time is continuous. In each short period, a large anonymous population of agents is randomly matched in pairs. Each match gives rise to an economically salient interaction with probability $q \in \left( {0,1} \right]$ . When such an interaction occurs, the pair faces a productive opportunity that yields surplus $Y\gt 0$ if cooperation succeeds and $0$ otherwise; when successful, this surplus is split equally between the partners. If the partners share the same protective institution, they achieve success with probability $\alpha$ ; otherwise with probability $\beta \lt \alpha$ . If both partners are independent, the same interaction may also generate a standard Lockean conflict game, where each is self-biased and conflict is costly, yielding the familiar polymorphic average payoff $ - {b^2}/\left( {4c} \right)$ for each independent.Footnote ² If at least one partner belongs to the agency, the agency’s procedures prevent violent feuds and the expected payoff from this channel is normalized to zero. Membership in the agency costs a flat per-period fee $e\gt 0$ .

Agents are ex ante identical and choose a long-run type: either a $J$ -type who joins the agency, or an $I$ -type who remains independent. Let $\pi$ be the share of $J$ -types in the population. Matching is random: a $J$ meets a $J$ with probability $\pi$ and an $I$ with probability $1 - \pi$ ; an $I$ meets a $J$ with probability $\pi$ and another $I$ with probability $1 - \pi$ . This evolutionary or population-level representation is not intended as a demographic claim about Locke’s or Nozick’s state of nature, but as a way of tracking how the relative attractiveness of institutional types changes as the composition of the population shifts.

3.4. Expected payoffs and join incentives

Given the population share $\pi$ of agency members, expected payoffs are

${U_J}\left( \pi \right) = - e + q\cdot{Y \over 2}\left[ {\pi \alpha + \left( {1 - \pi } \right)\beta } \right]$

for a member, and

${U_I}\left( \pi \right) = q\left[ {{Y \over 2}\beta - \left( {1 - \pi } \right){{{b^2}} \over {4c}}} \right]$

for an independent. Define

$\begin{align}K: = {{q\left( {\alpha - \beta } \right)Y} \over 2} & \gt 0,\;\;\;\;\;\;\;\;B: = q{{{b^2}} \over {4c}} \ge 0.\end{align}$

Then the incremental incentive to join is

${\rm{\Delta }}U\left( \pi \right) = {U_J}\left( \pi \right) - {U_I}\left( \pi \right) = - e + \pi K + \left( {1 - \pi } \right)B.$

Here, $B$ is the gain from avoiding independent-independent disputes and $K$ is the compatibility premium: the extra expected return from transacting within a shared framework of adjudication and enforcement. Because ${U_I}\left( \pi \right)$ rises weakly with $\pi$ , nonmembers are not made worse off as others affiliate: they face fewer $I$ – $I$ encounters and hence fewer violent disputes. That monotonicity keeps the model within Nozick’s non-worsening constraint.

Proof. Differentiating yields $\partial \Delta U/\partial \pi = K - B$ .

Lemma 1 captures the key strategic distinction between a pure loss-avoidance environment and a compatibility-augmented one. The compatibility premium does not simply wipe out free-riding. If $K\lt B$ , the familiar free-rider logic remains: as more agents join, the marginal attraction of membership may actually fall. If $K = B$ , the join incentive is flat in $\pi$ . Only when $K\gt B$ does the incremental benefit from shared institutions outweigh the free-rider attraction of staying outside and the private incentive to join increases in the extent of membership. This is the structural condition under which the decision to join takes on an assurance-game character.

3.5. A tipping threshold and assurance dynamics

The most instructive case occurs when protection cannot invade a purely independent population on dispute-avoidance grounds alone (so $e\gt B$ ) but compatibility benefits are nonetheless large enough to make membership attractive in a sufficiently ‘thick’ regime – so $K\gt e$ . Under these conditions, there exists a unique interior threshold ${\pi ^{\rm{*}}}$ at which the sign of ${\rm{\Delta }}U\left( \pi \right)$ changes.

$\pi^{*} = {{e - B}\over{K - B}} \in \left( {0,1} \right)$

such that $\Delta U\left( \pi \right)\lt 0$ for $\pi \lt \pi^{*}$ and $\Delta U\left( \pi \right)\gt 0$ for $\pi \gt \pi^{*}$ .

Proof. Because $\Delta U\left( 0 \right) = B - e\lt 0$ and $\Delta U\left( 1 \right) = K - e\gt 0$ and ${\rm{\Delta }}U$ is affine, the expression follows.

Proposition 1 shows that under these parameter conditions the decision to join acquires a tipping-point structure: below ${\pi ^{\rm{*}}}$ , membership is privately unattractive; above ${\pi ^{\rm{*}}}$ , it becomes privately worthwhile. This is the formal counterpart of the intuitive idea that in a world with division of labour and compatibility gains, protective membership may only be appealing once a sufficiently thick network of participants already exists.

To capture how $\pi$ evolves, suppose it follows the two-type replicator dynamic

${\dot \pi} = \pi \left( {1 - \pi } \right){\rm{\Delta }}U\left( \pi \right).$

Under the conditions of Proposition 1, ${\rm{\Delta }}U\left( \pi \right)$ crosses zero exactly once in $\left( {0,1} \right)$ , changing sign from negative to positive. A standard phase-line analysis yields the classic assurance pattern.

Proof. The sign of $\dot \pi$ matches the sign of ${\rm{\Delta }}U\left( \pi \right)$ . By Proposition 1, ${\rm{\Delta }}U\left( \pi \right)$ is negative below ${\pi ^{\rm{*}}}$ and positive above it, establishing the stability structure.

Figure 1 illustrates the one-dimensional phase line: on $\left( {0,{\pi ^{\rm{*}}}} \right)$ the arrows point toward $0$ and on $\left( {{\pi ^{\rm{*}}},1} \right)$ toward $1$ , making the assurance structure immediately visible. From a philosophical standpoint, Propositions 1 and 2 together say that once compatibility gains are sufficiently large relative to both the fee and the dispute-avoidance benefit, there is a basin of attraction around a high-membership regime: if enough agents join, the invisible-hand mechanism tends to complete the transition to a dominant protective association.

Figure 1.

Phase line for the replicator dynamic with tipping point $\pi {\rm^{*}}$ .

At this point, however, the obstacle to consolidation has not disappeared; it has changed form. Once $K\gt B$ , the problem is no longer whether membership has the public-good structure identified by Bruner at every population share, but whether the population can move into the basin above ${\pi ^{\rm{*}}}$ . This is an equilibrium-selection issue of the kind emphasised by Skyrms (Reference Skyrms2004) in his discussion of stag-hunt dynamics. In a commercial setting of the kind described here, several devices could help a population cross that threshold: clustering around commercial hubs where the same traders interact repeatedly, the gradual buildup of trust through repeated supplier relationships that make the advantages of common procedures tangible and the convergent expectations that arise when successful coordination is observed and imitated. None of these are modelled, but they suggest that a division-of-labour environment is more hospitable to threshold-crossing than a thin Lockean baseline would be. The model identifies the assurance structure; it does not claim to solve the selection problem in full.

3.6. The role of conventions

Bruner (Reference Bruner2020) emphasises that correlated conventions may coordinate agents on peaceful behaviour, effectively eliminating violent feuds. In my notation this corresponds to setting $B = 0$ . In that case,

${\rm{\Delta }}U\left( \pi \right) = - e + \pi K.$

If $K\gt 0$ , joining is privately optimal exactly when $\pi \gt e/K$ . Thus, even when conventions replicate the peace-preserving role of protective procedures, they do not fully substitute for the production-enhancing effects of shared institutions, and demand for affiliation survives whenever the compatibility premium is large relative to the fee. Philosophically, this shows that conventions can repair one part of the Lockean/Nozickian story (the suppression of retaliatory feuds) while leaving another part (the productive advantages of shared institutions) untouched. The viability of the invisible-hand path thus depends on more than what conventions alone can deliver.

3.7. Comparative statics and robustness

The comparative statics of the threshold ${\pi ^{\rm{*}}} = \left( {e - B} \right)/\left( {K - B} \right)$ under $e\gt B$ and $K\gt e$ are intuitive: ${\pi ^{\rm{*}}}$ rises with $e$ , since higher fees make affiliation harder to sustain, and it falls with $K$ and therefore with $q$ , $\alpha$ and $Y$ , since more frequent production opportunities, larger compatibility gains, or higher cooperative surplus make it easier to cross the threshold. These comparative statics connect the formal analysis back to the underlying Nozickian thought that a deeper division of labour can change the economics of protection.

This qualitative picture proves robust across several natural variations. Alternative adjustment dynamics that are payoff-monotonic preserve the same stability structure. Allowing multiple protective associations, with same-agency success ${\alpha _s}$ , cross-agency success $\gamma$ and independent success $\beta$ satisfying ${\alpha _s}\gt \gamma \gt \beta$ , yields fragmented equilibria when ${\alpha _s} - \gamma$ is small and consolidation when it is large; appellate mechanisms and procedural harmonisation can be interpreted as raising $\gamma$ . Allowing endogenous parameters – for example, letting production frequency $q$ or compatibility $\alpha$ rise with $\pi$ – typically amplifies the assurance structure. Even more complex environments in which $B$ falls as institutions spread while $K$ rises do not overturn the basic tipping logic, provided that $K$ remains sufficiently large relative to $e$ .

3.8. What the model shows

Relative to Bruner’s thin loss-avoidance model, the result is simple. If protection only reduces conflict, incentives drift toward free-riding. If shared procedures also make cooperative projects more likely to succeed, affiliation can instead take on an assurance structure with a tipping point and a basin of attraction around dominant-agency membership. That is the main formal payoff of adding the compatibility premium.

The comparison also helps clarify what has and has not changed. Bruner’s difficulty does not disappear; it survives whenever the compatibility gain is too small relative to fees and dispute-avoidance benefits. And even when the premium is large enough to create assurance dynamics, the model still leaves open how a population gets into the high-membership basin in the first place. What changes is the form of the problem. The question ceases to be whether protection is always undermined by free-riding and becomes, instead, which commercial environments make shared institutional affiliation self-reinforcing.

From a philosophical point of view, this result sharpens the familiar worry that Nozick’s story trades on an overly frictionless picture of economic interaction. The viability of the minimal state now depends not only on rights-respecting individual choices but also on the structure of institutional complementarities encoded in the compatibility premium. In this sense, the model can be read as a bridge between purely normative critiques of Nozick and the literature on the emergence of conventions and social institutions in evolutionary game theory, including Lewis (Reference Lewis1969), Skyrms (Reference Skyrms1996) and Vanderschraaf (Reference Vanderschraaf2006): it shows that invisible-hand political philosophy must attend not only to side constraints but also to the economic and institutional background that makes emergent authority dynamically plausible.