[Home] [Help] [Databases] [WorldLII] [Feedback]
	Macquarie Law Journal

ON THE POSSIBILITY OF IMPARTIALITY IN DECISION-MAKING

JOHN C W TOUCHIE[∗]

INTRODUCTION

Can we make impartial decisions anymore? Could we ever? It seems that wherever we turn, we are told that impartiality just isn’t possible. We are told that things are much more complicated than they once seemed, that the idea that we could ever be impartial was a mischievous intellectual error and that impartial judgments were and are just a front, a façade, for the harsh reality of some imposing power over others.

But we wonder, sometimes, whether this can be the case. Are things so complicated that impartiality is always impossible? If not, when might impartiality be possible, at least to some degree? If impartiality is an intellectual error, revealed by laborious scholarly insight, are we to believe that this insight is somehow impartial? Can’t some insights escape partiality? Similarly, are all decisions by some over others simply the imposition of power? If not, why not? And for what types of decisions?

What is needed, I think, is a comprehensive re-examination of the meaning of, arguments for, and possibility of impartiality. If it is a fact that impartiality is no longer an attractive ideal – and there have been arguments to that effect circulating for some time now – we should want to know why this is the case. Moreover, we should want to know if it were ever possible to be impartial and, if so, in what circumstances. Has impartiality always been a mirage? Perhaps it has, but before coming to this conclusion one might be curious to know why one was fooled into believing that it was at least possible in principle to be impartial. In short, we must ask: what are the pre-conditions for impartiality? What features of the world make it more difficult to be impartial? Which make it easier?

This article aims at a critical analysis and re-evaluation of the notion of impartiality. My goal here is to clarify this ideal through analysis of the arguments concerning impartiality, and in doing so to put forward a circumscribed but defensible notion of impartiality. The first step towards this goal is an examination of the meaning of impartiality. Here, the article puts forward three criteria under which the impartiality of decision-making can be assessed: replicability of decision, commonality of criteria of validity and similarity of rank-order over such criteria. These criteria then act as touchstones throughout the article as I reframe and re-evaluate discussions of impartiality.

Having clarified the definition of impartiality, the next step is to map out the territory over which the article will range. The discussion moves in two general directions. First comes an examination of a static notion of impartiality where discussion divides up into minimalist and aspirational aspects of impartiality. Here the article examines minimalist arguments that we must, to some degree at least, be impartial in our judgments. I outline arguments from cognitive science and the philosophy of mind and psychology that support such a position, and analyze in some detail the qualifications to, and limitations of, such arguments if they are to make sense.

Next, the article turns its focus to impartiality as an aspirational ideal. It argues that the aspirational ideal of impartiality in fact has two different faces, predicated upon different views concerning the source and hence the meaning of ‘objective’ normativity, and that confusion between these two views can spill over into confusion as to the possibility and meaning of impartiality. Discussion focuses upon two idealised extremes. The first stresses a commonly shared ‘grass-roots’ normativity, emerging ‘spontaneously’ from social interaction, whose objectivity lies in the independence of its existence on our choices and decisions. The second, by contrast, emphasises a deliberately created normativity, whose objectivity stems from the deliberate attempt to keep separate the act of creation and the act of application of the norms. The implications for discussions of impartiality are discussed and the weaknesses of each view are highlighted.

Following on from this discussion, the article then turns to an explicit examination of what might be termed the micro-analytics of impartiality. This discussion runs in three directions. First comes an examination of the micro-dynamics of impartiality using the surprisingly contemporary discussion in Adam Smith’s Theory of Moral Sentiments as a focal point. Smith presents a theory of the bottom-up growth of normativity that has several interesting properties, as we shall see. The article devotes some time to outlining the operation and implications of Smith’s model, focusing attention on his explanation of the dynamics of impartiality and on his conjectures as to how impartiality emerges and is sustained over time.

Second comes a deeper examination of the micro-foundations of impartiality. Here, I argue that grounding impartiality on the bottom-up normativity of Smith’s model, while appealing, in fact begs the question in many places, but can be rescued from its prima facie circularity through a resort to recent insights in neurophilosophy, some of which are summarized in the article.

Third, and finally, comes an examination of the micro-context of a decision and its effect on the achievement of impartiality. Here, the article briefly examines those factors and circumstances that impact on impartiality and either enhance or diminish its chances of being realised.

The article concludes by addressing the question of why we should be concerned to focus on bottom-up, micro-analytic models of impartiality and normativity. Here, I critically evaluate the widespread assumption that there does not exist a value consensus upon which we could ground impartiality other than one of our own deliberate creation. I argue that this obvious truth rests on a fundamental mistake, this being the failure to differentiate between minimalist and non-minimalist forms of normativity. I further argue that the implicit denial of ‘normative foundationalism’ rests upon surprisingly tenuous intellectual foundations, and that the grounds for this belief should be subject to a close and critical scrutiny. The article concludes with an examination of the problems created by differences in material circumstances and power and the implications this has for arguments concerning minimalist normativity.

THE MEANING OF ‘IMPARTIAL’

One thing is certain – if we are going to talk about ‘impartial decisions’, we had better be clear about what it is that we mean. So what is meant by the term ‘impartial’? There seem to be at least three inter-related notions at play here. First comes the idea that for a judgment to be impartial, it has to be in some sense replicable by other decision-makers. That is, if the decision were given to other decision-makers, it has to be likely that these decision-makers would come to the same judgment as the original decision-maker.[1] This can be characterised as a result – oriented notion of impartiality. An example might be a teacher marking a multiple choice examination – in most situations, the decision as to the mark to award to each question is replicable by other markers of those same questions. The issue here is whether decision-makers are able to reproduce the same result, in the sense that they can all consistently come to the same decision.

Now, while this seems to capture an element of what it is to make impartial decisions, this can’t be all there is to an impartial judgment, for one can make judgments that can and would be replicated by others facing the same circumstances and yet be subject to the criticism that the judgment was not impartial. As an example, consider an individual asking for a decision from a particular corporation regarding, say, their refund policy. Although the individual might receive the same decision from each and every individual within the corporation, we still might be able to coherently claim that the judgments of all of these individuals were not impartial, in the sense that the judgments were biased towards the interests of the corporation. Thus, consistency and repeatability in decision-making, while an important element of impartiality, are not sufficient for impartiality.

From this example, we can see that the second aspect of an impartial decision seems to be that such decisions are not ‘based upon’ the decision-maker’s particular desires and wishes, but rather upon considerations taken more generally to be valid. In this light, a decision is impartial if it is ‘inter-personally valid’, that is, if the grounds (or criteria) of the decision are accepted[2] as valid by those making the impartiality attribution. As an aspect of this, it might be pointed out that these inter-personally valid criteria might extend to include the methods by which decisions are made, for example, the rules of inference used. Now, while strictly speaking this commonality of method is not a necessary requirement for impartiality (it is possible to imagine different decision-makers making their decisions in different ways and nevertheless arriving at similar, or the same, results over time) it does seem unlikely that, at least in its general outlines, the methods used by decision-makers who made a large number of similar decisions would have no similarities. It is reasonable to argue that similar, if not identical, rules of inference will emerge among decision-makers producing similar, replicable, decisions based on inter-personally valid criterion, even if it is not necessary to go all the way to the stronger condition that all decision-makers actually resort to the same method for making decisions for their decisions to be considered to be valid.[3]

These two considerations – replicability and inter-personal validity – lead into what might be considered a third element of impartial decision-making: this being that the rank-ordering of that which is considered to be inter-personally valid, is similar (or the same, depending on the degree of strictness we are aiming for) across those making the impartiality attribution. It should be obvious that this requirement is closely connected to the requirement of replicability of decision-making, for the requirement of consistency applies to at least some of the implications of one’s decisions. This implies that if there are differences in the rank-ordering of the decision-makers’ values these will probably emerge as differences in decisions – as inconsistencies or incompatibilities – and hence to divergences in the replicability of decisions made over time.

THE PRE-CONDITIONS FOR IMPARTIALITY

From the above discussion there emerges an interconnected conceptual framework for considering the impartiality of decisions. Firstly, impartiality presupposes a minimal degree of regularity in decision-making, such that it is possible to distinguish between partial and impartial decisions at all. This is the question of replicability. Secondly, impartiality demands a resort to inter-personally valid decision-making criteria, at both the general level of method, or at the more specific level of content and tests of relevancy for a decision’s content. This is the question of inter-personal validity. Thirdly, and finally, impartiality requires that decision-makers, to varying degrees of similarity, can, in case of incompatibility, order in importance the criteria they use (or the content these criteria pick out). This is the question of rank-order.

Having said all of this, it must be asked whether there is any reason to expect that these conditions would ever be satisfied. If so, in what circumstances is this to be considered likely? And what circumstances affect the degree to which this is likely? In other words, if impartial decision-making is a possibility, when is this so, and why? The first, and perhaps the most obvious, way to respond to these questions is to argue that we must of necessity be at least to some degree impartial in making certain types of judgments. This is an argument by certain commentators in the philosophy of mind and psychology when discussing the rational interpretation and evaluation of the conduct of others, to be discussed in the sections below. Another way to address these questions would be to argue that, regardless of whether or not a minimalist impartiality is a necessity, we should strive for impartiality in making certain types of judgments. Later on in the article I explore the idea that there are in fact two quite different notions of impartiality at play in aspirational discussion, and that while there is some overlap between them, it is of importance to clearly distinguish their different implications for arguments of impartiality.

For the moment, however, let us turn to what might be called the argument from necessity.

ARGUMENTS FOR IMPARTIALITY: THE ARGUMENT FROM NECESSITY

Arguments for impartiality come in a variety of types. One of the most persuasive families of these is one that argues that we must, of necessity,[4] adopt some degree of impartiality in our understanding or evaluation of the conduct of others. One particularly persuasive member of the family of arguments for the necessity of impartiality emerges from a large and interesting literature in the philosophy of mind and psychology that discusses how we go about understanding, explaining and predicting the behaviour of others. The question here is often framed as one of interpretive practice, about how we go about interpreting the behaviour of others.

The discussion tends to head down two rather distinctive paths. On one path, occupied by Daniel Dennett,[5] Donald Davidson[6] and others,[7] the argument is that in interpreting the behaviour[8] of another, we must of necessity adopt what is termed a ‘principle of charity’,[9] this being that we assume that the individual is a rational agent, and ask what a rational agent would do in that particular situation.[10] This path tends to emphasize impartiality in interpretation, where the actual interests of others are separated from the ideal of rationality – although it is important to note that the principle of charity is not necessarily impartial, as there might well be differences between different ideals of rationality.

On the other path, elaborated upon by Steven Stich[11] and others,[12] it is argued that we need make no such assumption, but rather that we should instead adopt a ‘principle of projection’, and ask what we would do if placed in a particular situation. This position tends to emphasize the partiality of an interpretive position, although again it is important to note that it is not at all clear that the projection principle is necessarily partial. It could be the case, for example, that some of the goals and methods that we project in making interpretations are also projected by other individuals, implying the possible existence of shared, inter-personally valid, decision-making criteria.

The question in both of these theories of interpretation is how to draw the line between ourselves as we are and a completely idealised agent that embodies ‘perfect’ rationality. Those adopting the principle of charity argue that we must attribute rationality to any agent the actions of which we try to interpret (though precisely what this means, and how much rationality we must attribute, are open questions), while those advocating the projection principle claim that such an attribution of rationality is not necessary and that we must instead ‘project’ ourselves into the position of the agent we are observing.

Both positions have their strengths and weaknesses. For instance, those that advocate the principle of projection often do not acknowledge that instead of imagining that we are projecting ourselves into the position of another, we might instead imagine that we are projecting someone better than ourselves (in certain respects), a point made abundantly clear by Daniel Dennett.[13] On the other hand, those that adopt the principle of charity are often not clear about the fact that there is an important difference between minimal degrees of rationality and going all the way to an assumption of perfect rationality.[14]

What is needed here is a balanced approach that accepts a minimal amount of rationality, but which points out that there is no reason for us to assume hyper-rationality, that is, reasoning as it would be in some Panglossian, best of all possible worlds, sense. However, I do agree with the view that at the least we must assume a minimalist degree of rationality. We must assume, as a provisional starting point for any critical analysis, that what we do think is in conformity to some degree with what we ought to think. If this were not so, we would be unable to get any rational analysis off the ground.[15]

Similarly, a balanced perspective would point out that in the interpretation of the conduct of others we must assume that we project at least some aspects of our own perspective into the circumstances. At the same time, there is no need for us to assume a completely factual starting point, that is, we might imagine an agent who might not have our particular foibles. For instance, they might have knowledge that we, at the moment, do not possess.[16] It is possible to balance ourselves on a tightrope of the middle ground between the stringently normative (‘assume perfect rationality’) and the strictly actual (‘assume the person we are now is projected into the position’), just as it is possible to assume some form of impartiality is possible without going all the way to an unrealistic form of excessive impartiality or partiality. In terms of the three requirements of impartiality discussed previously, the idea is that it might be possible to strike a balance between the degree of replicability in decision-making produced by an ideally rational agent and the more idiosyncratic decisions emerging under a projection principle. Similarly it might be possible to come up with a balance between the degree of inter-personal validity and similarity of rank-order under the principle of charity and the projection principle.[17]

Thus far the discussion has focused on interpreting conduct for the purposes of explanation and prediction. Suppose that we were interested in evaluating the conduct of others (as when, for example, we apply our moral beliefs to the conduct of others). Is there any substantial difference between interpretation for the purpose of explanation and prediction and the evaluation of conduct?

Some authors have argued that interpretation for the purposes of explanation and prediction operates on fundamentally different principles from normative (moral) evaluation.[18] Hence, they conclude that much of the above discussion is of little interest to those focusing on normative, and more particularly, moral judgment.

For my part, I believe this conclusion to be a mistake. The difference that is claimed to exist between the two types of purposes is that, whereas for explanation and prediction we want to make the observed individual’s conduct as coherent and rational as possible, whereas for normative (evaluative) purposes we focus our attention on evaluating conduct, however incoherent or irrational it may be. The idea seems to be that in the former case, we must adopt some minimal degree of rationality to get an investigation off the ground, whereas with the latter this seems not to be the case.

It strikes me that in many cases of interest, this is not an accurate description of what goes on in the process of evaluating conduct. The primary difficulty here is that it is not an obviously simple task to separate explanation and prediction from evaluation in many situations of interest. In many cases it seems that at least a degree of understanding is required before adequate evaluative judgment can be made. But for understanding to be a possibility, it is arguable that at least a minimal degree of rationality must be assumed for the party whose conduct you wish to evaluate. I suspect that in undertaking normative judgment we are often concerned, explicitly or implicitly, with whether there has been conformity with the norms of rationality, at least in their minimalist manifestation and that therefore it does not make sense to dismiss the similarities between explanatory/predictive and evaluative judgments.

This idea is familiar in many cases of day-to-day normative judgment, where we often simply presuppose that conduct accords with at least the minimalist elements of rationality, and then proceed from there to evaluate the conduct at hand under our particular evaluative normativity. Similarly, when we turn to spheres of more formalised normative reasoning, we can see the same connection being made between an assumption of minimalist rationality and normative judgment. For example, some moral systems hold that the norms of rationality must in principle be satisfiable by the observed agent as an antecedent condition for judgments of moral merit. As another example, it might be noted that many legal systems impose the same pre-condition in various ways, although it must be added in qualification that this depends upon the historical period and functional complexity of the legal system to which one is referring.

It is important to be clear that in making normative judgments such a pre-condition of rationality is not always satisfied. Nor is it the case that we must necessarily presuppose complete and ‘perfect’ rationality to be able to apply evaluative normative standards, as the disagreement between those advocating the principle of charity and those pushing for the projection principle makes clear. But the important point is that in many cases of interest we simply presuppose a minimal degree of rationality, or make such rationality a pre-condition for our normative judgments. Hence, one should be careful not to exaggerate the differences between evaluative and explanatory/predictive judgments.

How does this debate in the philosophy of mind and psychology help us to answer the question of whether we must be impartial in interpreting and evaluating the conduct of others? To my mind, the most important lesson to be drawn from this debate is that if we must be impartial in interpretation, this remains true only for minimalist aspects of interpretation. As we stray from minimalist aspects, we move much more into matters of degree, debate and disagreement. Put differently, the lesson that the discussions in the philosophy of mind and psychology have made particularly clear is that we must acknowledge that impartiality has its limits. As far as normative systems go, it is arguable that decisions based on the logical and rational systems of normativity are the most impartial, that is, they command the most widespread replicability, inter-personal validity, and rank-order similarity.[19] Indeed, as we have seen, some of the advocates of the principle of charity argue that we must assume an idealised model of rationality if we want to interpret the conduct of others.[20] Yet even when considering the normative standards for logical and rational conduct, there is disagreement as to the detailed content of the norms of rationality. When moving to other areas of normative evaluation where there is a wider diversity of normative systems present than is typical in the normativity of rationality, we should expect, ceteris paribus, a concomitant reduction in the possibility of impartiality.

ARGUMENTS FOR IMPARTIALITY: IMPARTIALITY AS A NORMATIVE ASPIRATION

The discussion in the previous section focused on an argument that emphasized the necessity of impartiality in interpreting the conduct of others. There, an argument was made that we needed to assume rationality when interpreting the conduct of others, with the implication being that a minimalist degree of impartiality flows from the adoption of such an assumption. Suitably qualified, this was seen to be a rather persuasive argument, assuming that such impartiality was a possibility at all – a point to be picked up later in this article.

The argument for necessity is, however, not the only argument that one could imagine being invoked in support of impartiality. Perhaps the most important point to stress concerning impartiality is that it is not something that just happens. Rather, impartiality, if it is achieved at all, is often realised through a striving towards a normative ideal. If the issue of impartiality were simply a matter of assessing the degree to which the application of normative systems conformed to the criteria discussed earlier in the article, one might be inclined to argue that the issue of impartiality is primarily an empirical matter. But to do so would be to ignore how we arrive at impartiality, and would overlook the fact that impartiality operates interactively and dynamically, in that impartiality is an ideal towards which we strive, and not merely a result that happens by chance or through some random process. It is not enough to talk about the properties of impartial decision-making at some end-point, independent of how we arrived at this point. For it is precisely how we arrive at this end point that contributes to characterizing impartiality as a process.

These considerations are central to a second family of arguments for impartiality. Unlike the argument from necessity, the focus of attention is typically not on the claim that we must act to some degree impartially. Rather, the idea put forward is that impartiality is an ideal towards which we should strive. This family of arguments, then, makes explicit the aspirational element of impartiality. This much they have in common. Underlying this commonality is, however, a fundamental difference in approach that has some rather important implications for approaches to characterizing the processes preserving and sustaining impartiality, as we shall see.

DISAGREEMENTS OVER THE APPLICABILITY OF IMPARTIALITY

It is arguable that most of us encounter no small difficulty in striving to be impartial in our personal dealings. Indeed, for some it seems that the very ideal of impartiality in one’s personal life is unintelligible, for we are always and necessarily inseparable from our own personal interests, goals and values. The idea might be stated as follows: impartiality is perhaps possible in public dealings, where we often have to adopt a ‘role-ethic’ that is different from our personal one; but when we are talking about our own moral positions, where we have not agreed to adopt or enforce the morality of others, how can we ‘separate’ ourselves from our own interests? The very idea seems to be nonsensical.

Further, it is sometimes asked why one would want to be impartial with respect to one’s own interests in one’s personal dealings. Is there any point to being impartial regarding one’s own interests?[21] Some see the answer to this as a clear and obvious ‘no’. But the answer, as one might expect, is a bit more subtle than this.

Is it meaningful, then, to talk about impartiality with respect to our own interests when we talk of our personal interests? Is it possible to be impartial in one’s own personal dealings? And, if so, how does one come to be impartial? As we shall see, the answer to these questions seems to hinge upon one’s view regarding the meaning and possibility of objective normativity.

That there is an important difference between interpreting the conduct of others as compared with judging our own conduct seems to stem from the very notion of impartiality itself. When judging the conduct of others, we take our own normative system as a ‘given’, and then apply it to the conduct of the other. In this situation, there is a relatively clear separation between the referent conduct and the activity of applying the normative system in question. The question of impartiality in such cases consists primarily in the degree to which the operation of a clearly separated normative system approaches the ideal outlined earlier in the article.

When judging our own conduct, on the other hand, we find ourselves in a somewhat different position. Here, we can be in the precarious position where the boundary between the conduct in question and the activity of applying a normative system becomes much more difficult to sustain. The idea here is that it is impossible to be impartial with respect to your own conduct because you are using your values to judge your own conduct which, by assumption, is predicated upon the fulfilment of your own goals and values. In other words, and put starkly, does it make any sense to make value judgments over conduct that is motivated by these same values?[22] Surely, it is argued, this is a biased position from which to judge one’s own conduct. Surely we will be partial to our own interests. How could we ever hope to overcome this partiality?[23]

What is needed here, it seems, is a ‘separation’ of the actor from the judge of the actions of the actor. The question is how one would go about putting such a separation into operation and what, if anything, such a separation would entail.

TWO NOTIONS OF IMPARTIALITY

There seem to be two ‘solutions’ to this problem,[24] depending upon one’s view regarding the possibility of the existence of a decentralized, objective form of (moral) normativity. The first position assumes that we can’t be impartial with respect to our own positions, interests, values, and the like.[25] Given this assumption, this position tends to focus on issues of public impartiality, and upon methods of ensuring that this is a possibility. Here, we tend to find an emphasis upon the doctrine of the separation of powers, upon the importance of the democratisation of the processes whereby content is given to public norms and the importance of judges impartially and strictly following set down rules.[26] Each of these themes is related to the initial assumption that there is no widely-shared objective normativity that exists independent of our deliberate creative efforts and upon which we could ground impartiality. The separation of powers doctrine and the emphasis on objective standards in public judgments highlight the importance of separating the creation of the standards from their application, while the democratisation argument focuses on the need to ensure that everyone have a say in the content of these norms. The former doctrines, then, attempt to ensure an impartial application of public norms, while the latter doctrine attempts to ensure that the content of these norms has some degree of inter-personal validity embodied within it.

There are a couple of points that bear mentioning. Firstly, it should be obvious that the democratisation argument presupposes the need to create inter-personally valid content. In other words, it is simply assumed that there does not exist any such inter-personally valid content that could be embodied in such norms independent of their creation by somebody and the assent, however imperfect, of the governed individuals. Secondly, it should be equally obvious that these mechanisms have their weaknesses. That it can be notoriously difficult to separate the acts of creation and application of norms has been made clear by systems of case law that seem to combine the two.[27] Thirdly, and finally, the fact that standards are publicly created says little about whether they allow for replicable decision-making. Nor does it imply that such standards are inter-personally valid to or ranked similarly by those to whom the standard is meant to apply, as numerous criticisms, such as those made of the reasonable man standard,[28] as well as the historical experience of this century,[29] have made clear. Moreover, I think it is arguable that it is not at all obvious that the move to democratisation has in substance made this criticism any less applicable, for while partiality has been stripped away in some areas, it has been introduced in others. Democratisation, important as it is, is not synonymous with a striving for impartiality.

The second position, by contrast, assumes that it is possible to be impartial to our own positions, interests and values, at least to a degree.[30] Such standards are then imposed over our own conduct, the conduct of others, and enshrined in, and applied to, various public role-ethics, including the standards embodied in the ‘reasonable person’ standards in law, and the content embodied in various social institutions. On this view, an objective (moral) normative content can and does exist independent of our deliberate choice and creation, and often times it is argued that it is more a matter of discovering the content of this normativity and embodying it in the various institutional structures of society than it is a matter of creating such normativity through such institutional arrangements.[31]

The prima facie strengths and weaknesses of this position are relatively well known, but for the moment I will only touch selectively on some of the more important ones.[32] The most significant problem for this view is that while it presupposes the existence of an objective content, it does not say how such content is to be discovered, nor how it is that such content comes into existence or is sustained. Furthermore, it is not clear that an objective content, assuming it existed, would extend in a straightforward fashion to matters in the institutionalized public realm, or to matters that existed in circumstances characterised by substantial complexity. Moreover, all of this presupposes, for the sake of argument, that such content exists – which is precisely what is denied in most of the substantive critiques of this position.[33]

There is much to discuss here, and I will return to this important issue later on in the article. But first, let me turn to a brief comment on the similarities and differences between the two positions.

OVERLAP AND DIVERGENCE

The mere fact that these positions have been presented as alternatives does not mean that there are no similarities between them. Indeed, there does seem to be an overlap of sorts of the two positions with respect to judging public normativity and its conformity to the norms of impartiality. But any convergence is, it would seem, only partial.

Consider why this might be the case. For the first position (which, it might be noted, seems to bear a striking similarity to the position adopted by neo-classical economics, in that society is assumed to be populated by a number of rational agents, each with their own idiosyncratic/subjective preferences), ‘impartiality’ consists of the uniform and standardised imposition of norms created by some and applied by others. This is, in essence, procedural impartiality, as opposed to a more substantive form of impartiality that is embodied in the content of the standards themselves – a more substantive form that is ruled out by assumption, that is, by the assumption that individuals have idiosyncratic/subjective preferences and hence there could be no content that could be objective for each and every one of them. On such a view, ‘impartiality’ refers to (a) impartiality in norm application, and (b) the consensual aspect of the democratic filling-in of the norms’ content.

By way of contrast, the second position, while not ruling out the importance of the doctrine of the separation of powers or the adherence by public officials to objective standards, focuses more on the embodiment of an already existing objective moral content and less upon processes through which such objectivity can be created. The idea here is more of a bottom-up development and application of morality than of a more institutionalised and proceduralised form of objectivity and impartiality.

The contrast seems to be between a position that holds to the view that there exists no society-wide, sufficiently objective basis for normative (moral) judgment upon which we could ground impartiality, save for what is constructed in the public domain through institutional arrangements and design, and a view which holds that there does exist a shared-in-common moral order which can act as the basis for impartiality, the content of which can and should be discovered and then embodied in the content of the institutions governing our society.

This way of putting it makes it clearer that the conclusions reached by each position regarding the meaning of impartiality flow on from their original assumptions regarding the possibility of objective (moral) normativity. The first of these assumes that there is no way to avoid partiality of our particular positions, and focuses its attention on creating impartiality in institutionalized spheres. The idea here is of grounding impartiality - regularity in decision-making, shared inter-personal validity, and similar rank-orderings - on externalized, publicly institutionalized forms of criteria that have been deliberately created. The second assumes it is possible to avoid partiality towards our particular interests (to some degree), and then use this impartiality as the cornerstone for other sites of impartiality. The former view becomes a procedural form of impartiality, because on its assumptions all content is partial towards someone’s interests. The latter becomes a more substantive form of impartiality, because it assumes that, at least to a degree, it is possible to be impartial towards one’s own interests, and to the interests of others, by basing one’s judgments on an objectively existing, commonly-shared, normativity.

THE DYNAMICS OF IMPARTIALITY

One fundamental weakness common to both of these positions is that, in their arguments for impartiality, they assume that there is a commonality of values and normativity at some level of social interaction without explaining how this commonality comes into existence or how it is sustained. Thus, the first position discussed above, which explicitly denies the existence of objective normativity, is flawed in that it cannot explain how it is that social order exists if there is no shared normativity at some level of generality. It seems that the first position simply takes social regularity for granted, as if it were a given, when in fact it is arguable that such macro-regularity is only possible if individuals are guided by certain micro-level normative principles of interaction that are shared in common. The second view seems to be in a better position, in that it presupposes that an objective normativity exists, but in fact it too suffers from the weakness that it does not explain why there is commonality between views, nor how this is sustained. Rather, this view simply assumes that such commonality exists. This claim, while perhaps correct at some level of description, is of little value if we are given no insight into how it is that such commonality can come to exist.

If we return, briefly, to the arguments made concerning the necessity of impartiality, we can see that the same can be said for some of the arguments there. A resort to the projection principle assumes a commonality of sorts in interpretation – otherwise, communication would inevitably fail. Similarly, interpretation made through resort to the principle of charity presupposes a commonality of interpretation in that we must attribute a common rationality to the others when interpreting their conduct. What is the source of this commonality? Davidson seems to argue that it is ‘the nature of things’ that forces us to adopt the principle of charity. Why this might be the case is not addressed, however – it is simply ‘the way things are’. Dennett and Stich see things a bit differently. Dennett makes explicit evolutionary arguments as to why one might expect to see the emergence of commonality in rationality. Stich makes counter-arguments that one might quite reasonably expect diversity in the particular results of evolutionary processes. Neither author, though, makes specific a proposed model tying together the generation of normativity with the creation and sustenance of impartiality. In sum, none of the discussion thus far puts forward much micro-level detail of how impartiality is generated and sustained over time, nor its connection to the normativity upon which it is based.

The sections that follow attempt to fill this gap in the discussion. They discuss a model of moral judgment put forward some two centuries ago by Adam Smith in his Theory of Moral Sentiments (1759). Surprisingly contemporary, Smith’s discussion forms an interesting starting point for a micro-analysis of impartiality and normativity. There are a couple of features of this model that merit a mention before we turn to its details. First, while Smith’s model of impartiality has many interesting points, perhaps the most important of these is the centrality of minimalist considerations in sustaining his argument. Much hinges upon this, particularly when one turns to an examination of the jurisprudential and political aspects of his thought, as has been noted elsewhere.[34] I will focus attention on this point throughout the remainder of the article, and bring it into explicit focus in the final sections of the article, in the discussion of its uneasy co-existence with some of the core doctrines of legal positivism. Second, and as we will see, Smith’s discussion is not without its problems, perhaps the most important of these being its circularity of argument. I will turn to a discussion of this difficulty, and how it might be circumvented, in the sections that follow the outline of Smith’s theory.

ADAM SMITH AND THE ORIGINS OF MORAL JUDGMENT

Consider the explanation put forward by Adam Smith, in his Theory of Moral Sentiments, first published in 1759, and running through six editions, for the origins of normativity and impartial judgment. Smith argued that the basis of normativity rested on what he termed the process of ‘sympathy’ (a technical term to Smith which is in no way limited to an evaluative judgment,[35] as the modern usage might suggest). In contemporary terms, Smith’s ‘sympathy’ would accurately be termed the process of simulation[36] of what was running through the mind of another. Smith argued that normativity at its earliest stages stems from our ability to simulate the mind processes of others and to evaluate these processes using our own, pre-existing, valuations attached to these processes.[37] We simulate, he argued, based on our knowledge of the circumstances[38] that led to the processes being caused in the mind of the other. Thus, on Smith’s view, we are able to distinguish between how another individual actually reacts to a circumstance, and how they should react, given the results of our simulation based on the circumstances we have observed. Hence, ‘to be able to judge is to be able to know the situation’.[39] In this way, Smith manages to create a gap between factual normativity (how individuals in fact evaluate) and ideal normativity (how they ought to evaluate, given the circumstances).[40] What is of decisive importance here is that situations will be knowable to different degrees, depending on the type of knowledge and information that is available. Thus, in some situations one will be able to judge in greater detail than in others.[41] The important point is that normative judgment is dependent upon knowledge and information, and upon contextual conditions that impact upon these. It is also dependent upon the existence of the ability to replicate, to greater or lesser degrees, the reaction of others to certain situations.[42]

This is the first step in his explanation. The second is the postulation of what Smith termed an ‘impartial spectator’, which is directly connected to Smith’s explanation of how ideal morality becomes detached from actual social morality. How does such a separation come into existence? Smith argued that when we want to consider our own actions, we first consider the judgments of others. But we can also ask the question of whether these judgments are in fact correct by imagining the judgment of a spectator, and asking what their judgment would have been. In this way, Smith is able to distinguish between the judgment of actual spectators and the idealised judgment of an impartial spectator, and hence is able to distinguish between actual and ideal moral judgment.

It is important to note what is not being claimed. Smith is not claiming that the impartial spectator, generated by the process of sympathy, would come to the same judgments as an actual spectator. Sympathy is simulation, an act of imagination, and thus depends upon the particular interests, passions, knowledge and information of the judging individual. It is possible, then, that different individuals might, if faced by the same environment, generate many different impartial spectators.

It is important to be clear on this point. One of the interesting features of Smith’s model is that it explicitly takes into account, and highlights the differences in, and the connections between, what we know and our location in space-time. In a sense, Smith’s model emphasises the limitations of existence in space-time, that is, it highlights the fact that by not being where you are (and, by implicit extension, by not having been where you’ve been), I am limited in how much I can know about what you know. It also allows us to see that there might be some forms of knowledge that are more difficult for me to have access to (such knowledge being, in the term of the analysis of this article, relatively subjective) or, conversely, that there might be forms of knowledge which are easier for us to know in almost the same way as the individual that we observe (such knowledge being relatively objective).

Two final features of Smith’s theory bear mentioning. First, Smith argues that general rules of judgment are the product of the interactive feedback system and selection process of Smith’s notion of sympathy. As Smith puts it:

We do not originally approve or condemn particular actions, because, upon examination, they appear to be agreeable or inconsistent with a certain general rule. The general rule, on the contrary, is formed by finding from experience that all actions of a certain kind, or circumstanced in a certain manner, are approved of.[43]

Put differently, Smith is claiming that abstract rules of morality grow from numerous particular instances of evaluation.[44] Abstract rules do not pre-exist these evaluations but instead grow up out of them, in a social environment in which one generates moral rules based on the evaluations of others. Second, and given this assumption about the bottom-up growth of moral rules, Smith also claims that the rules of judgment which evolve under the process of sympathy will be generally well-adapted to their particular environment.[45] What this means is that ‘behaviour which is not so fitted will tend to be weeded out by means of antipathy conveyed through the mutual sympathy mechanism, whereas behaviour which is fitting will tend to be reinforced by approval conveyed in the same way’.[46]

SMITH ON THE GENERATION OF IMPARTIAL JUDGMENTS AND OBJECTIVE NORMATIVITY

These last two points are of particular importance when it is asked how Smith’s model can be viewed as an attempt to explain how the three elements of impartiality are generated and sustained. In his model, the three elements of the definition of impartiality discussed previously – replicability in decision-making, a commonly-shared inter-personal validity and a similar rank-ordering – are intimately intertwined. Consider first the requirement for regularity and replicability in judgment. While regularity of an individual’s conduct is simply assumed, and must be explained outside Smith’s theory (this explanation is, in fact, the subject of the section that follows), Smith’s theory claims that inter-personal regularity in decision-making comes about through the weeding out of ‘inappropriate’ judgments through a process of antipathy, and through the reinforcing of approved judgments through approval. These same processes support the social standardization of the criteria of validity and the similarity of rank-ordering of these criteria. This process of standardization was, to Smith, always an imperfect one, always in danger of being overcome by our partiality to our own interests, and by a variety of other flaws and foibles.[47] Nonetheless, Smith did view the process of inter-personal sympathy as one means whereby the judgments of the various impartial spectators could come to converge, to a degree at least.

What this means is that objective normativity in Smith’s scenario – ‘objective’ in the sense that its existence is independent of the desires, choice or particulars of the individual making the judgment – is an unintended by-product of the sympathy process.[48] In Smith’s theory, moral objectivity arises through the stripping away of arbitrary, person-specific detail that occurs at each step of the sympathy process.[49] By trying to view yourself as others view you, you must eliminate any knowledge that is known only to yourself. This implies you must consider the degree to which others could enter into your situation as spectators (that is, without your specific knowledge). Hence, the judgment which is made, based on the idea of spectators who can only to a limited degree enter into your situation, is objective to the degree that the knowledge and information upon which a decision would be based would be that which could be shared by the participant and a potential observer. Moral judgment is thus different from the actual judgment of existing spectators in that the impartial spectator bases judgment upon what a potential spectator could know, and not upon what actual spectators do know. By this process, then, information which is idiosyncratic, individual-specific, and known only by particular individuals is eliminated as grounds for objective moral judgment.[50] To be impartial, on this model, is to base conduct on a normativity that has grown from the bottom-up, a normativity that has a degree of objectivity, albeit imperfectly and with difficulty, even though this was not the intended result of any actor or group.

Of course, none of this implies that actual moral judgments will be (or are) objective or impartial. Rather, it merely provides one mechanism whereby a convergence to an objective normativity might come about,[51] and an explanation of what impartiality would mean under such a mechanism. In Smith’s model, it is the differing information/knowledge requirements and the assumptions of individuals about the knowledge of the impartial spectator that spurs on the ground-up growth of morality. Moreover, it is the relative degree to which individuals hold different views of the knowledge of the impartial spectator that determines the relative objectivity of the judgments of that spectator. It is this difference between each individual’s notion of the impartial spectator that makes it possible for an individual to evaluate the judgments of others and to say what they should have been (relative to each individual’s more-or-less idiosyncratic notion of the impartial spectator, of course).

Smith’s sympathy mechanism thus, in some senses, imposes a criterion that attempts to separate objective from subjective knowledge. The striving to achieve an impartial spectator perspective acts as a filter over person-specific knowledge, leaving objective knowledge as the basis of a judgment or, to be more accurate, knowledge that individuals imagine to be objective. Whether such knowledge would be actually objective would, of course, depend upon the similarities between, and replicability of, the judgments of different individuals, and this would, one would imagine, be a rather difficult thing to quantify with any degree of precision.

PROBLEMS WITH SMITH’S EXPLANATION

This explanation of the bottom-up emergence of a relatively objective normativity is an interesting one, but it would be a mistake to believe that it did not have its own difficulties. First among them is the fact that Smith’s theory, in common with other theories of this type,[52] seems to presuppose the very thing he sets out to discuss, this being the existence of normative judgments.

Consider the following. Smith’s theory seems to assume:

The problem here is that Smith argues that rules of judgment grow from particular instances; yet it would seem that, on his theory, there must already be a pre-existing systemic regularity to be able to imagine, judge, and select rules of conduct. Moreover, it would seem that Smith must presuppose an already existing commonality if he wants to get his model of the impartial spectator off the ground, for he presupposes that an individual is able to imagine the degree to which an imaginary spectator could ‘put themselves in my position’. This presupposes both the existence of a degree of commonality between agent and spectator and a knowledge, on the part of the agent, of the content of this commonality.

How can Smith’s model escape from this circularity? The answer proposed here is that an escape is possible through the adoption of a model of normativity that allows for a distinction between the type of judgments that Smith has in mind (conscious, intentional, on purpose judgments) and the foundations that such judgments presuppose. To do this, the article must briefly venture into the deep waters of the philosophy of mind.

THE NEUROPHILOSOPHICAL FOUNDATIONS OF NORMATIVITY

Smith’s ventures into moral philosophy took place in the 1750s. Much has changed in the intervening years in moral philosophy, particularly with respect to work on the foundations and what has been termed the ‘naturalisation’ of morality. As will be seen, some of this work can be of great value to Smith’s theory of normativity, particularly theorizing on an alternative level of explanation of normative systems, a level I term ‘neurophilosophical’. In particular, such work allows us to see a way out of the vicious circularity that is seemingly present in Smith’s theory of normativity.

As discussed above, the general difficulty with Smith’s theory is that it seems to require the existence of some sort of normativity in order to argue for a mechanism governing the emergence of normativity. If this were correct, Smith’s theory would surely be of little use. Fortunately, there does seem to be a way out of this dilemma, this being to argue that while Smith’s theory is concerned with the emergence of normativity at a reflective, conscious, level, the ‘normativity’ required by such a reflective level would not be itself made operative by conscious reflection, but rather would operate at more subterranean levels. Put simply, the ‘normativity’ that Smith’s theory requires to get its conscious and reflective normative judgments off the ground is neither necessarily conscious nor reflective, but rather is built into the foundations that support consciousness and the ability to reflect.

What, precisely, are the required foundations? To summarize the list presented above, it would seem that Smith’s discussion of normativity requires that there be an ordering mechanism that selects (classifies) and weights (for importance) various aspects of the organism’s environment, allowing a role for ‘imagination’, namely, to the consideration of potential conduct before such acts are manifested in actual conduct, and for growth, learning, and the active participation of the organism in making decision. Finally the mechanism must be such that it contains a common, shared, content with other such mechanisms.

These are precisely the properties of mind that are present in a particular tradition within the neurophilosophical approach, one that can be termed the ‘neural networks’, or connectionist, approach. The modern variants of this approach date back to the early 1940s, and is a familiar one in psychological, cognitive science, and philosophical circles, summarized in standard textbooks,[53] and popularized in a variety of recent books.[54]

The general outlines of this approach are as follows. The approach emphasises the systemic nature of mind, and places great weight on the interactional aspects of mental life. This systemic emphasis is of importance because of its intimate association with issues of meaning and interpretation,[55] as well as with arguments concerning the degree of evolutionary adaptiveness of the activities of our minds, and more particularly the degree to which these activities are rational (or have rationales). According to this approach, the systematicity of mind stems in large part from the mutually adapted nature of a mind’s activities.

Being mutually adapted has some important implications. First, it is this adapted connectedness, flowing from a selectionist history, that is of decisive importance in determining the meaning of activities and their effects.[56] On this view meaning must be, to some degree at least, holistic and systemically interconnected.[57] At the same time, the way of interpreting the systemic quality of mind implies that meaning will be dependent on functionality in a variety of ways, such as in the distinction between the functions of various activities and other, accidental, effects.[58]

The systemic quality of mind is thus intimately matched to its environment through the selection of functionally adapted neural networks. Such networks are not necessarily already in existence, however. Rather, such networks can be formed through an interaction between the organism and its environment. The idea here is that the activities of mind replicate aspects of the external world through systems of neural networks and the intensities associated with these networks that have gone through a process of evolutionary selection. Thus, the connections between events in the world external to mind can, to some degree (and always imperfectly),[59] be reproduced within mind by classifications which are associated with weights that attach to these particular aspects of the world.

How, then, does this replication take place? The general idea is that simultaneous events in the environment external to mind become connected to (are capable of initiating the activities of) certain neural activities, and that certain neural activities which occur simultaneously become connected. This beginning, when coupled with selectionist survival pressures,[60] tend to ensure that such simultaneity endures through time.[61] In other words, the simultaneity of events external to mind can become to some degree replicated in the neural connections of mind when simultaneous external events initiate simultaneous weighted neural activities which, owing to evolutionary selection pressure, remain connected within mind.[62]

Note that this is merely the first stage in a cascade of connected simultaneous events, some of which involve events within mind replicating other events within mind. The idea here is that some simultaneous neural activity can be viewed as forming the environment for still other neural activity; thus, simultaneous events are in this case replicated by neural connections, but these simultaneous events are themselves neural events. This process can occur over and over again, resulting in what might be thought of as an enormously complex ‘classification’ of both the world external to an organism and aspects of its own neural mechanisms. It should be clear, then, that on this view mind is not merely a (rather imperfect) ‘mirror’ [63] of the adaptively significant causal networks of the external world. Rather, it is also, to some degree, a mirror of itself (in the sense that some neural networks replicate events which occur within mind, in the sense of being a re-classification, and a re-re-classification, and so on to varying degrees of complexity).

Thus, simultaneity in external events can lead to the formation of weighted neural simultaneity (which, when one adopts a system’s perspective, form neural networks).[64] Moreover, these causal connections which are replicated to a degree within mind differ from causal connections external to mind in that there are intensities (weights) embodied in neural events and the connections between them, with the weights depending on how the external events map onto the ongoing processes of mind, which in turn depends upon both the quality and quantity of the processes which enter into each aspect of the internal connections.

IMPLICATIONS OF A CONNECTIONIST APPROACH

The reason that this insight assumes an importance is its connection with evolutionary arguments, both biological and cultural, concerning the origins and development of mind and normative systems. On this way of thinking about normativity, mind has been viewed as an expectation-generating and filtering mechanism. On this view, mind is a system in which various ‘expectations’ [65] as manifested in the interaction between neural network activities, their associated intensities, and their environments, are weeded out or reinforced by complex evolutionary selection processes of feedback loops.[66] This selection process over ‘expectations’ (‘anticipations’) is an ongoing activity within mind, and represents a complex interaction between the environment of an organism and the mechanisms of mind which generate, reinforce and filter out expectations. Another way of putting this would be to say that mind is a mechanism of classification, a sphere within which ‘expectations’ are generated and run through a process of selection for their compatibility with their environment, with other ongoing processes of mind, and with other ‘anticipations’ within mind. In other words, mind is an anticipation mechanism.[67] Such anticipations are intimately related to conduct in that potential paths of action are ‘tested out’ in the ‘expectational’ sphere before being selected as appropriate for manifestation as actualized performance.[68]

This has important implications for a theory of normativity, given that normativity is one mechanism whereby potentiality (in the form of ‘anticipations’, ‘expectations’ or some other term) is ‘narrowed down’ and ‘mapped onto’ actuality through a filter of ‘oughts’ and ‘ought nots’, ‘shoulds’ and ‘should nots’. In a sense, normativity fulfils the same function as a neural network based mind, weeding out certain types of potentiality (‘anticipations’, ‘expectations’) in the connection between the potential and the actual. In other words, the filtering activities of a neural network mind perform the same function as more conscious forms of normativity at more articulated, reflective, levels of description. Of course, both of these levels are intimately intertwined in an extremely complex manner, the details of which shall not be discussed here. But for the moment, the point to be stressed is that there are important parallels between reflective, conscious normativity and its less reflective, or wholly unconscious, forms.

That all of this is important to Smith’s model of normativity should be obvious. Smith’s model of reflective normativity seemed to predicate itself upon the existence of some other form of normativity. Here, I have outlined the workings of just that form of normativity. Thus, what we have here is a mechanism that is capable of supporting Smith’s model of reflective normativity. Smith’s sympathy mechanism and the resultant impartial spectator are predicated on the assumption that there exists a mechanism that is (a) capable of classifying aspects of the world, and attaching varying weights to the aspects so classified, (b) capable of imagining – simulating, if you will – a variety of scenarios and being able to select the most ‘attractive’ of them, (c) capable of resolving incompatibilities between alternative ways of classifying the world. All of these features are present in the model outlined above. In addition, the conscious, reflective and intentional focus of Smith’s model is complemented by the more foundational, and non-conscious substratum provided by a neural networks vision of mind.

The connectionist model of mind can provide the foundations for the presuppositions of Smith’s theory of normativity, including his assumption of the regularity present in moral judgment and reasoning. On a connectionist model, individuals act on the basis of weighted ‘maps’ which replicate to some degree regularities which occur in their environments (and within their own minds). These maps are constructed by connections being formed between activities of mind, and are subjected to evolutionary selection for their conformity with both the ongoing activities within a particular mind and with events external to mind which initiate their own activities within mind.

The implications for Smith’s notion of impartiality are equally important. The three properties of impartiality discussed earlier in the article – regularity/replicability in decision-making, common inter-personal validity, and a similar rank-ordering, are each present in, and highlighted by, the connectionist model.

Consider the issue of regularity and replicability in decision-making. The model of mind outlined above is capable of replicating elements of the simultaneous causality present in its environment. In other words, such a model provides a framework under which individuals are capable of absorbing (not necessarily in a passive manner) environmental regularity, and of generating and sustaining regularities of conduct. Perhaps the most important element which can be replicated is the performative environment of a developing individual, for this can provide a foundation for the transmission and persistence of culture within individual minds. It can be argued that many of the elements of culture, and many of the abstract rules of conduct which govern individuals’ performance in their day to day lives - such as those of language,[69] morality,[70] and rationality[71] - were at least in part transmitted across and embedded within individual minds in precisely this way.[72] Such a way of viewing mind and normativity stresses the inter-dependence between the growth of such systems and the environment in which they were, and are, embedded.

Now consider the issue of the commonality in inter-personal validity and the similarity in rank-ordering in decision-making. This is intimately related to the connectionist insight that mind embeds, to some extent, a weighted causal map of its environment.[73] There are many important implications that flow from the presence of intensities (weightings) associated with the activities of mind.[74] One of the most important of these come from their association with what might be termed the closure properties of mind. Closure within mind arises from conflicts between different ongoing activities of mind. One activity closes down another in a conflict when that activity occurs and the other does not, this taking place in two general ways: as an all-or-nothing closure, or as a matter-of-degree.[75] Within mind closures are the instantiation of the implicit importance-order associated with mind’s activities. This rank-order is embedded in the evolved capacities of mind, and is generated in its particularity by the interaction between mind and its environment over the life span of the individual. One could say, then, that there are two aspects of rank-ordering which play a role in mind: at the species level and at the individual level. The individual aspects are developed during the lifetime of particular individuals, while the species level rank-ordering manifests itself in the existence of evolved activities of mind which are transmitted across generations. Species level activities manifest an evolutionary importance-ordering in that some of these activities have aided the organism in adapting to its environment in the past. On this way of viewing mind, one necessary (but not sufficient) condition for an attribution of evolutionary adaptedness would be the persistence of capacities over time which have contributed to the persistence of the organism in which they are presently manifested. This insight is crucial to Smith’s theory of normativity and, more broadly, to his jurisprudence and political theory, in that he makes both explicit and implicit assumptions concerning the structure of normativity and its relationship to issues of law and politics. In particular, Smith assumes that there exists a minimalist, commonly shared, all-or-nothing importance-ordering, the content of which refers to harm rather than pleasure, a topic to which I shall return later on in the section that follows.

Another feature of connectionist models important to our discussion of impartiality is their emphasis on both the importance and the irrelevance of perspective. For some issues, an individual’s perspective can be of decisive importance, depending upon what it is that is known. For example, an individual’s emotions are typically known in a different way by that individual than they are by others. In this sense, the particular space-time relationship between the knowing mechanism and that which is known - one’s perspective, as it were - is crucial in determining both what is known and its qualities. Furthermore, it is difficult for the other to know what you know in the same way (because you exist in space-time, and they cannot be where you are and hence know your emotions in the same way that you do). Such knowledge would be subjective, in that it would depend upon the particulars of the individual not shared by other individuals.[76]

Now, this dependence on perspective does not extend with equal force to all forms of knowing and knowledge, a feature examined in some detail by Smith in his discussion of the impartial spectator.[77] For example, the capacity to act in a certain way is one example of a perspective dependent form of knowledge that can vary by degrees, depending on the degree to which such performative knowledge can be ‘mechanized’ and replicated across different individuals. For some particularly talented individuals - some musicians, for example - their perspective always remains to some degree theirs and theirs alone. But to a degree at least, their performative knowledge can, through training, be shared by other individuals. At the other extreme are matters where issues of perspective are much less important, and where it is the overlap between perspectives that typically come to the fore. Seeing colours and hearing sounds are instances of types of activity that seem, in large part, to be relatively standardised across individuals.[78] Again, all of this has important implications for the idea that there exists a certain core of common normativity, and the related idea that difficulties in perspective might emerge if and when we tried to extend impartiality in decision-making beyond this core commonality.

THE LIMITS TO IMPARTIALITY

From Smith’s model of impartiality, down through the neural network foundation, I have argued that the focus of attention has remained on the decision-maker. Indeed, the very definition of impartiality introduced at the beginning of the article focuses its attention on the properties of the decision-maker. One might gain the impression from all of this that an investigation of impartiality could centre almost exclusively around decision-makers.

This would be a serious mistake. What is missing from all of this is a consideration of the relationship between impartiality in decision-making and the environment in which decisions are made. As the discussion of the connectionist model should have made clear, normativity and its environment are tightly tied together. Impartiality is similarly tied, in that the possibility of realizing the pre-conditions required for impartial decision-making are not independent of the state of the decision-making environment. What, then, are the general circumstances that impact upon impartiality? In other words, what are the conditions and circumstances under which impartiality becomes easier, or more difficult, to attain?

Recall for the moment the constitutive elements of impartiality. Impartial decision-making requires three conditions to be satisfied as a matter of degree[79] (although the second and the third can tend to merge into each other). These conditions concern the replicability of decisions, the degree of overlap of the inter-personal criteria (the grounds, procedures, and values) underlying the decision, and the degree of similarity between the rank-orders of criteria. Difficulties in any of these areas spells difficulties for the possibility of achieving impartial judgment.

Now, there are a variety of circumstances that impact in varying degrees upon the requirements for impartiality. Here, I will only focus on some of the properties of the more generally applicable circumstances, these being:

Properties of the interactional environment

One of the most obvious circumstances impacting upon impartiality is the degree of complexity of the environment to which normativity is applied. Increases in complexity make the achievement of impartiality more difficult by enhancing the epistemic difficulties inherent in applying normativity. Consider the three properties of an impartial decision, and ask yourself the following questions: can I know that I am acting with regularity? That the same result has been obtained across decision makers? That there exist inter-personally valid decision criteria? That they are rank-ordered the same? The difficulties here should be obvious.

This ties into a second general circumstance relating to the environment in which normativity is applied. Generally speaking, if we are faced with circumstances that require rapid and numerous adaptations between individuals, it becomes more difficult to achieve a pattern of regularity of decision-making, and in maintaining a similarly rank-ordered system of inter-personally valid criteria, or similarly, it becomes difficult to know when these conditions hold and when they are violated. A principled adjustment from one perspective can be seen as an arbitrary choice from another if changes become frequent and numerous enough.

The third general circumstance impacting on impartiality is the form of social relation present in the environment in which normativity is applied. This is an extremely complex issue, and it is difficult to avoid over-simplifying here, but a significant issue relates to the effects of the relative complexity of Gemeinschaft-type social relations as compared to Gesellschaft-type relations and the way this impacts on decision-making. On the one hand, the greater complexity and degree of mutual adaptation, both to others and to the environment, that is present in the former but less so in the latter would lead one to conclude that it would be more difficult to produce a replicable decision in intimate-to-intimate environments than in stranger-to-stranger environments. This tight degree of mutual adaptation would mean that there is close connection between a relationship and its context, implying that ‘detaching’ the basis for judgment from its context, implying that such judgment will be less portable in circumstances of ‘heavy and complex interdependence’,[80] and that drawing out a general principle to carry over to other cases will be possible only with great difficulty. Moreover, there might a greater partiality towards one of the positions than to another, particularly in intimate relations where the degree of attachment to another is greater than it might be with strangers. All of this implies that impartiality in decision-making will be more difficult in such circumstances than it would be in Gesellschaft-type environments. Going against this, however, one might expect to find a greater degree of similarity in the criteria of inter-personal validity and the rank-ordering of these criteria in intimate-to-intimate relations relative to stranger-to-stranger relations, leading to a potential enhancement of the conditions of impartial decision-making. How these considerations end up impacting in practice would be dependent on the significance of these opposing tendencies.

Properties of the content of the normativity

Another factor that affects impartiality is the degree of formality of the normativity upon which decision-making is based. It is sometimes assumed that the more formalised normativity becomes, the better the chances for achieving impartiality. This obviously has its limits, for it is not obvious that formalizing necessarily enhances each of the conditions for impartiality – particularly if the primary result of the formalization is an ill-conceived linguistic formulation.[81] Sometimes, it might be that ‘a specious clarity can be more damaging than an honest open-ended vagueness’ to the cause of impartiality.[82] Moreover, an unrealistic degree of formalization and an overabundance of formal rules can lead to a divergence from governance by formalised means, and can actually enhance partiality.[83] That being said, however, it is arguable, I think, that formalization does have its benefits, particularly in the enhancing of memory and standardization. Whether this has an unambiguous effect on impartiality depends, in large part, on the interaction of these factors with the other circumstances discussed here.

Another, often misunderstood, factor that impacts on impartiality is contained in the idea that objectivity and impartiality in normative judgment are enhanced when focusing on issues of harm, detriment, or loss rather than on situations of pleasure, benefit or gain.[84] This is an idea with a long history, much of it controversial.[85] Yet there is, I believe, a core of useful insight to be gained here.

Return for the moment to Adam Smith’s theorizing. As we have seen, the idea that harm was more objective and more amenable to impartial decision-making was a powerful influence in Adam Smith’s theory of moral sentiments, as well as on his legal and political theorizing.[86] This belief can be dissected into two general themes. First comes the claim that, in general, pain and displeasure are stronger and longer lasting feelings than pleasurable ones.[87] Second is the claim that in many cases pain and displeasure are more universally shared than are feelings of pleasure.[88]

The justification Smith puts forward for these two points, put into modern terms, is evolutionary: his argument is that in general it is pain and not pleasure which has the highest survival value for us.[89] The more extreme forms of pain deter us from that which might end our lives and hence is very important from the standpoint of survival,[90] whilst pleasure is in many cases an addition to survival and hence not as valuable from an evolutionary perspective.[91] These types of argument are intimately connected to the idea that minimalist obligations, obligations that refer to the minimal (functionally or contextually necessary) conditions for social interaction, are more precise than non-minimal obligations, and to the related idea that such minimalist obligations are more amenable to formulation as articulated rules than are non-minimal obligations.[92] The idea here is that rules formulated to govern minimalist circumstances will tend to be more precise than rules governing less minimal circumstances because the criteria of validity and the rank-order of such criteria will be more commonly-shared and more strongly-held than would the criteria attaching to non-minimal circumstances.[93] The reason for this appears to be that feelings of harm are stronger, of longer duration, and more universal because they arise from the violation of regularities which constitute minimal, functionally necessary, conditions for ordered social life. All of this, if correct, has significant implications, particularly for jurisprudential theorizing. A core tenet of legal positivism, for example, is that there is no relationship between legal validity and moral merit. Yet if there is relationship between moral merit and precision in articulation, in that ‘the bad’ (harm) can be formulated with greater precision than can ‘the good’, and if the ability to be precise is itself a pre-condition for legal validity – as was generally acknowledged by legal positivists in their debate with Lon Fuller[94] – then there is in fact a connection between moral merit and legal validity. Put differently, if the satisfaction of Fuller’s ‘principles of legality’ are prerequisites for legal validity, (a point even the most ardent of defenders of a separation between legal and moral validity, the legal positivists, have acknowledged) and our ability to satisfy these principles does vary according to whether we are talking about minimalist situations of harm or aspirational situations of benefit, then there clearly is a relationship between the moral merit of the content of law and legal validity. If true, this relationship would in turn imply that what is perhaps the most fundamental tenet of legal positivist theory and the one point upon which almost all positivists, are in agreement, is factually false.[95] Thus, if there is a relationship between harm and verbal/referential precision, then there is in fact a relationship between the moral merit and legal validity, in that the bad (harm) is more precisely definable and ascertainable than the good, and hence more amenable to inclusion in legal rules which are part of a process, the validity of which is dependent upon a striving for precision.[96]

All of this highlights the importance of a more detailed investigation into the properties of minimalist obligations,[97] a theme to which I will return in the concluding sections of this article. For the moment, however, let me highlight one more consideration that impacts upon impartial decision-making. The discussion concerning the relative precision of minimalist and non-minimalist obligations is closely connected to another blindspot of contemporary legal theory, this being the distinction between prohibitions and performance-demanding obligations. How this impacts on impartiality in decision-making is a complex issue, which merits a more careful exposition than I can put forward here,[98] but the general idea is that when focusing on minimalist considerations, it can be easier to be impartial when dealing with prohibitive obligations than with performance-demanding ones.

Why might this be the case? The reasons are numerous, but there are three in particular that merit a mention. First comes the insight that certain types of prohibitions are universalizable and consistent to a greater degree than are performance-demanding obligations, in that such obligations are more amenable to forms of reasoning that do not require explicit trade-offs to be made between obligational demands that are incompatible. Second comes the point that in certain circumstances, the evidentiary requirements of the differing types of obligations (depending on the degree of precision with which the obligations are defined) are easier to satisfy for prohibitions than are the requirements for performance-demanding obligations. Third, and finally, is the idea that certain types of prohibitions can be more easily interpreted in a separable and individualistic fashion than can certain types of performance-demanding obligations which demand a more holistic approach that can add to operational complexity. Each of these factors can contribute to impartiality through an enhancement of the three conditions for impartiality – replicability of decision-making, commonality of inter-personal validity, and similarity in rank-ordering – discussed earlier in the article.

CONCLUSION

The article began with a brief introduction to various critiques of impartiality as an ideal and as a possibility. I have come a long way from that discussion in this article. Let me return to it now. One of the chief factors motivating the analysis of this article is my belief that impartiality as an ideal needs to recover, or perhaps to uncover, its sense of direction. Having looked on from the sidelines of many an academic debate, I now have a strong suspicion that many members of the academy have, to some degree at least, lost their way as to the meaning, the value, and perhaps most importantly, the scope for the applicability of impartiality. Here, I would like to return to the general issues that motivated my analysis in the first instance and discuss how this loss of direction came about, in the process turning attention to some of the more widely held, and often times unquestioned, assumptions that underlie contemporary academic discussion.

Let me begin with a focus on the limitations of the ideal of impartiality in decision-making as discussed in academic circles. These seem to come in at two main varieties. The first of these argues that human beings are partial beings, partial to their own interests, in particular, partial to significant others, whose interests they take to be more important than the great masses of humanity for whom they have no particular attachment. The second form of critique extends beyond this, to an institutional or systemic level, and argues that at this level certain interests are privileged over others in decision-making, and that impartiality, even if it is possible at the individual level, is nonetheless undermined by systemic bias.

Now, as a statement of social fact, these are in my view valid criticisms, for there certainly seems to be ample evidence of partiality in decision-making. From the point of view of this article, however, the issue is somewhat different, for what I am interested in is whether impartiality is possible at all. For the purposes of this article, then, these critiques become interesting when joined with the argument that there is no way to overcome the partiality of individual, of institutions or of systems. This argument is that partiality is simply a fact of life, the way things are, and the sooner we come to accept that, the sooner we will stop deceiving ourselves, or being deceived, with the false ideal of impartiality. On this view, impartiality is a mirage, or worse, a façade behind which particular groups have tried to dress up their own interests in a more universalistic garb so as to disguise their own particularistic interests.

These are important critiques, and I have few doubts that they embody an important insight. But it is important, I think, to carefully qualify these criticisms in certain directions. First, when discussing impartiality in decision-making, it can be useful - indeed, crucial - to separate the impartiality of the decision-maker from that of the content of the normative system that the decision-maker brings into the decision-making process. It is useful to note that one can quite sensibly talk of the partiality or impartiality of a decision-maker as separate from the partiality or impartiality of the system of norms that they apply in their decision-making. Thus, we can sensibly say, for example, that a judge was impartial in their judgment and yet at the same time critique the decision as being partial to particular interests.

But can an individual ever be impartial in any decision? I believe the answer to this is yes, for the second point to note is the rather obvious one that impartiality does not require that a decision-maker have no interests whatsoever. Rather, impartiality in decision-making demands that a decision-maker exhibit regularity and replicability in their judgments based upon shared criteria of validity and a similar rank-ordering to those making the impartiality attribution. Impartiality refers to shared criteria, ranked similarly in importance, and not criteria that are valid in some ‘absolute’ sense.

But are there any such shared criteria? This seems to underlie a similar concern, this being that interests at a systems or institutional level diverge for different groups of individuals. Don’t social and institutional systems systematically privilege the interests of some and downgrade the interests of others?

There are two main lines of response to this concern that I would like to highlight. First comes a response that strives to overcome such partiality by construction, primarily through a combination of procedural mechanisms. The ideal behind the separation of powers doctrine and the arguments for enhancing the democratic input into the content of normativity put into effect institutionally, whatever their difficulties in practice and however imperfectly such impartiality would be overcome, seem to motivated by this impulse. The idea here, as discussed earlier in the article, seems to be that we should strive to create an externalized and public form of normativity that, while perhaps not ideally impartial, allows for some degree of impartial decision-making, dependent of course on the degree of input into the creation and application of the systemic normativity. Note carefully that this position seems to assume that there is a need to create a shared normativity, presupposing in effect that there is not already in existence a shared normativity to which we can resort in decision-making.

Second comes a response that assumes that impartiality based on commonly-accepted, society-wide normativity is not possible. This assumption seems to me to be one of the central presuppositions of (post)modern thought. At its core is the denial of society-wide common values which could support, or be embodied in, commonly accepted inter-personal criteria of validity. There are, it has been said, no ‘meta-narratives’,[99] no commonly-shared, society-wide values that can be said to underlie reasoning and decision-making. Closely related to this are discussions in political theory that assume that there does not exist a value ‘consensus’ in modern Western society, and the further idea, shared in common with moral theory, that there has been a marked fragmentation of values in modern society.[100] Branching out from this comes the now common denial of foundationalism that stands at the centre of modern philosophy, particularly the epistemological and moral branches.[101] And, of course, all of this is related to the common assumption of neo-classical economic theory that there exists a fundamental incommensurability of values; an assumption that underlies claims that it is therefore necessary to resort to market exchange to resolve differences that cannot be dealt with in an impartial and objective manner.[102]

For my part, I cannot help thinking that this consensus is based on an error, on a failure to distinguish between the minimal foundations of life and social interaction, and that which presupposes this foundation as a given, and is built up on it. Given the increases in the division of labour and knowledge in modern societies, it is not all that surprising that this might lead to a division of normativity in the same way, with a fragmenting of normative views as social contexts become increasingly complex. In this respect, the existence of value fragmentation is unproblematical, even though it does pose new challenges for interaction in such a complex society and for the ingenuity of those who will strive to overcome it. What is surprising, however, is that it might be thought that such fragmentation would extend to the minimalist normativity associated with the common minimum that faces all of humanity, no matter how complex the social environment. No matter how far we have moved out onto the diverse branches of social interaction, every member of humanity has to protect the same basic roots of the tree of life, and hence there might be good reason to think that a very similar minimalist normativity would arise from this common need.

This failure to distinguish between minimalist consensus and aspirational diversity undermines contemporary claims regarding the collapse of a value consensus. The quality of the arguments and evidence offered in support of the proposition create further problems. It is often simply assumed that we live outside a value consensus, with little factual evidence offered to support this claim, other than the ubiquitous claim that a modern division of labour, knowledge and normativity make such a fragmentation inevitable.[103] Yet I would have thought it obvious that the questions of the precision and the distribution of normativity – both key factors in determining its objectivity, and hence the existence, or lack thereof, of a value consensus – would be at least to some degree empirical questions. I would have also thought that the widespread abhorrence of murder and torture as breaches of minimalist normativity might have called for some explanation as to its compatibility with a no-consensus position. Similarly, some of the arguments offered up in support of the view of a consensus disintegration are of dubious merit. Consider, for instance, the argument based on the proposition that existence of disagreement over the content of such normativity leads to the conclusion that therefore no consensus exists.[104] This is unconvincing for a number of reasons, for it fails to distinguish between minimalist and non-minimalist normativity, confuses disagreement over the content of a normative consensus with the non-existence of a normative consensus itself, and is grounded upon a suspect epistemological perspective. What is needed here is a carefully investigated position resting on sound intellectual and empirical foundations, rather than arguments for the lack of a value consensus that seem to be based on the barest form of a priorism.

This article argues that it is important to be clear on the factors that affect the degree of objectivity of normativity. As discussed in previous sections, it is arguable that there can be substantial differences in the degree of commonality for different types of values, or when faced with different types of circumstances. It is important, I think, to ground claims of ‘value fragmentation’ on an understanding of those factors that can exacerbate or ameliorate such differences. At least part of the reason for this comes from the tight connection between the degree of value consensus and the degree of domination in conduct governance. The extent and scope of such domination, if one cares about such matters, is in turn connected to a variety of factors, some of which are discussed in this article. Thus, if we wish to increase the degree of legitimation of governance, or conversely, to decrease the problem of domination, then the article would argue that we should take note that there are some general directions in which we should want to proceed. The important point here is that changes to enhance legitimacy might be more likely to be achievable in certain general directions than in others.

The goal of enhancing legitimacy is intimately connected to the implicit argument of this article concerning the possibility and importance of a suitably qualified and hedged-in notion of impartiality. Both goals are also closely connected to, though not identical with, the aspiration of striving for the achievement of an impartial and objective justice. Some have thought this to be a fruitless task. It has been argued that from the perspective of ‘rational cognition there are only interests of human beings and hence conflicts of interests’,[105] these conflicts not being resolvable by resort to objective criteria that might exist independent of the desires and wishes of human beings.[106] Justice, on such a model, is an ‘irrational ideal’, the idea being that ‘norms prescribing human behaviour can emanate only from human will, not from human reason’.[107] Still others seem to view law as necessarily partial to certain interests, implying that domination by some over others is in some sense inevitable, although it might be arguable that this point is meant to apply solely to societies the governance of which is based upon formal legal mechanisms, and not necessarily to other forms of social organization.[108]

Against this, I can only repeat my suspicion that these claims are themselves based on a confusion, on the failure to distinguish between minimalist and non-minimalist considerations, combined with an untenable, excessively particularistic epistemology that ignores the possibility of consensus at abstract levels of generality. To a degree, it is true that the particular details of institutionalized systems are at least in part attributable to the arbitrary acts of will of particular individuals, representing the particular interests of select groups. But this does not imply, however, that all institutionalized content is necessarily partial to some group’s particular interests. It is possible, I believe, to have normative content that is partial to all of humanity, not merely to particular groups. If this is the case, then it ought to be possible to extend the applicability of such content keeping in mind that such an expansion depend at least in part on the properties of the obligation that is in question here, as discussed in the text above. Thus, obligations relating to harm, and in specific obligations relating to prohibitions over conduct, should be extendible to a greater degree than is the case for more complex and contextually dependent[109] forms of obligations, such as performance-demanding, ‘transfer’[110] obligations.

Now, whether such an extension is desired by all is another matter. The grain of truth in the objections to an objective and impartial justice is the insight that not everyone necessarily views the enhancement in the legitimacy of governance and the striving for objective and impartial justice as uncontested goods. There are those who might view things differently, who even though they too must satisfy the minimalist requirements facing humanity, and even thought they too are subject to minimalist normativity, might be unconcerned with the minimal requirements of others. In fact, the primary beneficiaries of a view that dismisses a commonly-shared, society-wide notion of minimalist normativity, a view that argues that there only exists interests and conflicts of interests that cannot be resolved by rational argument and objective criteria, are the dominant elites of the status quo, who have perhaps grown comfortable with the idea that it is their interests that are dominating the interests of others, and that their minimalist requirements are met as a matter of course. Why would they want to promote the minimalist normativity of others?

There are, I believe, a variety of grounds, although only the barest of sketches can be laid out here. Consider some instrumental arguments that focus on some of the consequences of such breaches of minimal normativity. Many of the effects would seem to be long term, centering around the general idea that a breach of the minimal conditions of others has serious consequences for the elites themselves, particularly where the breaches are not geographically remote from the elites themselves. That the breaching of minimalist conditions of others, particularly through the violation of prohibitions on particular forms of conduct, such as murder and torture, has consequences for the elites themselves, through the barbarization of those in the dominating countries, deleterious effects of such conduct on those within the sphere of elites, or upon the ‘moral capital’ of society, particularly among the young, is not a novel view. How long it will take for substantial feedback effects to be felt in the home countries, assuming they are in fact felt is, of course, another matter.

There are other arguments that could be made. For instance, it could be argued that intellectual consistency demands a symmetry in our normative views, particularly when demanding conformity to minimalist normativity by others. Thus, if elites argue that one ought not to murder and torture, symmetry, in the form of intellectual consistency and coherence, demands that the elites conform to this themselves. Of course, if the argument is simply that others ought not murder, then this argument is not necessarily applicable, although the grounds of such a privileging would be open to critical examination. One could also make an argument based on altruism, which assumes that there is normative force to arguments that we should not only conform to minimalist prohibitions, but also should help out others in need or in danger of suffering harm from a breach of their minimalist requirements.

One could make argument after argument, giving reason after reason for why the minimalist normativity of others ought to be respected. But what is needed here are qualities that are difficult to argue for, because they seem to lie at the level of providing reasons for arguing at all. What is needed here, I believe, is a generous answer to the question of ‘who is my neighbour?’, in keeping with the spirit of the good Samaritan. What is needed, in the final analysis, is partiality towards humanity, an aspiration for justice, a desire for intellectual honesty and degree of humane decency. What is needed, in short, is the recognition of our duties to others, generously interpreted, and the aspiration and the will to fulfil them.

[∗] Lecturer at Griffith University Law School, Brisbane, Australia. Thanks go to Linda Bull, Neil MacCormick, Justin Malbon and Scott Veitch for useful discussions, and a particular thanks to Valerie Kerruish for ongoing insights and critique. The usual proviso applies.

[1] Ignoring for the moment the coherence of the notion of the ‘same’ judgment, and simply assuming that judgments come in a variety of degrees of similarity, some more similar and others less. Note that this issue will tie into later discussions, in that the move from simple decisions, characterised by sharp boundaries, to more complex decisions, which typically have vaguer, or fuzzier, boundaries, is intimately related to the ability to replicate the judgment of others.

[2] Conceptualising the ‘acceptance’ requirement has produced a voluminous literature on validity, legitimacy, etc. in legal and political theory. For my purposes here, though, I simply want to point out that this familiar requirement does not necessarily entail a conscious acceptance of these criteria as being valid – although authors differ upon what exactly are the requirements for someone to have ‘accepted’ a norm as valid. For more on some of the difficulties here, see the discussion of H L A Hart’s views of acceptance in N MacCormick, H L A Hart (1981).

[3] It might be noted that this way of putting the matter means we can distinguish between a judge’s impartial conduct and the partiality of the content of the criteria that are applied in particular cases. It is be clear on the point that the degree of impartiality of a decision will hinge upon the extent to which the grounds/criteria for the decision are shared, and might differ between observers, or between the decision-makers and those subject to their decisions.

[4] Is this a logical or a practical necessity? This differs for different authors. Some, such as Donald Davidson, suggest that it is a requirement of logical normativity; see D Davidson, ‘Reply to Essays’ in B Vermazen and M Hintikka (eds), Essays on Davidson (1985) 245. Others, such as Daniel Dennett, seem to argue on both logical and practical grounds. See generally D Dennett, The Intentional Stance (1987) and D Dennett, Darwin’s Dangerous Idea (1995).

[5] Primarily elaborated on in Dennett (1987), above n 4 but also developed in his Brainstorms (1978).

[6] As discussed in his collected essays Essays on Actions and Events (1980) and Inquiries into Truth and Intepretation (1984).

[7] See, for example, W V Quine, Word and Object (1960).

[8] Note that most discussions limit this behaviour to speech-acts. The interpretation I give to this principle is a broader one, encompassing the wider realms of human action and behaviour. For a discussion of the issues in delimiting the boundary between behaviour and events, see F Dretske, Explaining Behavior: Reasons in a World of Causes (1988); R Millikan, Language, Thought and Other Biological Categories (1984) and White Queen Psychology and Other Essays for Alice (1993).

[9] Dennett (1987) above n 4. So named because it grants a ‘charitable’, namely, rational, interpretation to the utterances (here, the actions) of others.

[10] Although there has been some dispute concerning the meaning of the principle of charity, and there does seem to be evidence that Davidson has changed his views on its content as time passed, here I will be focusing solely upon the more limited version of Davidson’s enunciation that focuses solely on the requirement that we must assume rationality when interpreting.

[11] The constructive argument is primarily in his From Folk Psychology to Cognitive Science (1983) , while the criticisms of the arguments of those advocating the principle of charity are located in his The Fragmentation of Reason (1990).

[12] See the references in Stich (1983) and (1990), above n 11.

[13] Dennett (1987), above n 4, 99-100.

[14] As argued, quite effectively in S Stich, ‘Dennett on Intentional Systems’ (1981) 12 Philosophical Topics 38. It should be pointed out that Dennett does point out that drawing the line between the two is no simple task given the definition of rationality is so difficult to draw. See Dennett (1987), above n 4, 94-98.

[15] Dennett attributes this argument to Roger Wertheimer, who dubs it the ‘Factunorm Principle’; see the discussion in Dennett (1987), above n 4, 98-99 and Dennett (1995), above n 4, 504.

[16] A point made quite effectively by Dennett (1987), above n 4, 83-116.

[17] How this might be possible, however, is another matter, to be reserved for discussion later on in the article.

[18] See, for example, the views of various of the commentators in L May et al (eds) Mind and Morals: Essays on Cognitive Science and Ethics (1996).

[19] Though one could, perhaps, make a convincing case that, in some countries, these were rivalled, if not surpassed, by religious normative systems.

[20] Donald Davidson puts it bluntly: ‘[the principle of] charity is forced on us’, ‘it is not an option’, in Stich (1990), above n 11, 44. Such an idealised notion of rationality would force impartiality on us, in that it demands and presupposes a standardised template for decision-making, thus satisfying the replicability, inter-personal validity and rank-order similarity requirements required for impartiality.

[21] A point often made by Nietzsche and those of similar bent of mind; for instance, see Nietzsche’s interesting discussion of this point with reference to categorical imperatives in The Gay Science (1974) 263-266; or his discussion of ‘objectivity’ in ‘The Genealogy of Morals’ in Basic Writings of Nietzsche (1968) 554-555.

[22] Impartial judgment is, from our previous discussion, judgment that is replicable, that resorts to ‘valid’ criteria for judgment, and to similar rank-orderings. Here, the notion of validity is sometimes called into question. Similar comments apply to the rank-ordering over these valid criteria.

[23] Similar quandaries arise when considering the principles of interpretation discussed above. When considering interpretation by adopting the projection principle, one might wonder how we know what we should do, given that this view seems to imply that we would answer this question by asking what we in fact would do. Does this mean that in adopting the projection principle we equate what we should do with what we do in fact do? Doesn’t this place too much weight on what is the case, and not enough on ideals towards which we might be striving? Doesn’t this, in fact, eliminate altogether the role of normativity as a guide to future conduct? Or consider the position of those who adopt the ‘principle of charity’. On this view, when judging our own conduct we would assume that we are a rational agent and then ask what such an agent would do in our circumstances. The question this leaves, of course, is how the assumption of rationality manages to ‘fill in’ any actual gaps in our rationality (and if it does not, how then are we supposed to act?). In other words, if we aren’t in fact rational (and don’t in fact live up to our ideal of rationality), does it make any sense to simply assume that we are? Such a move seems to place too much weight on the ideal, and not enough on the actual circumstances limiting our judgment – circumstances that we cannot overcome simply by assumption.

[24] These ‘solutions’ are, of course, ideal types, and as such are put forward as expository devices rather than as empirical summaries. Nevertheless, I do believe that these types embody positions that are present in discussions of impartiality, albeit usually not with the clarity and mutual exclusivity found in the ideal types.

[25] As noted above n 21, probably the leading figure(head) in contemporary thought advocating this line of thought is Nietzsche, although it would be a (serious) mistake to claim that he then advocated a move towards impersonal public duties. Nothing could be farther from his actual position, this being that ‘[n]othing ruins us more profoundly, more intimately, than every ‘impersonal’ duty, every sacrifice to the Moloch of abstraction’; Nietzsche, The Portable Nietzsche trans W Kaufmann (1954) 577. What should perhaps be emphasized here once again is that I am writing in terms of ideal types and hence abstracting from a (much) more complex reality which does not necessarily share all the features I am discussion.

[26] That these features are not incompatible with a belief in objective normativity should be obvious. For example, Robert Bork espouses these principles (as do a good many of what are termed ‘conservatives’ in the United States); for more, see R Bork, The Tempting of America: The Political Seduction of the Law (1990).

[27] This comment applies, as well, to statute based regimes, for statutes too have their content ‘filled in’ by judges.

[28] For more on this from a feminist perspective, N Naffine, Law and the Sexes (1990) provides a useful point of reference. From a critical race perspective, P Williams, The Alchemy of Race and Rights (1991) is a notable, if flawed, point of reference.

[29] One would think that the legislation of progressive income taxation provides a rather banal example of this; while many of the more insidiously discriminatory laws, such as the Third Reich’s infamous Nuremburg laws of 1935, or the laws establishing and maintaining apartheid in South Africa, provide a more disturbing case in point. In the former case, it might be noted here that it was not merely the content of the formal rules that was discriminatory; rather, numerous judges seem to have acted with rather extreme partiality against the Jews as well. For more on this, see I Müller, Hitler’s Justice: The Courts of the Third Reich (1991).

[30] The discussion of this paragraph is put in all-or-nothing terms when in fact the situation is (much) more complicated than this. Nonetheless, for expository purposes it does contain a germ of what thinkers such as Hayek, in his Constitution of Liberty (1960) and Law, Legislation and Liberty (1973, 1976, 1979) shares in common with the views of what have been termed ‘natural law thinkers’, the most notable of these in recent times being John Finnis and Lon Fuller. For the former, see his Natural Law and Natural Rights (1980); for the latter, see either his Morality of Law rev edn (1969), The Principles of Social Order (1981) or his often overlooked but consistently interesting Anatomy of Law (1968).

[31] Ibid.

[32] For a representative sample, see the interesting discussion in J Ely, Democracy and Distrust: A Theory of Judicial Review (1980) or A MacIntyre, After Virtue (1981).

[33] Ibid.

[34] A point made throughout Knud Haakonssen’s The Science of a Legislator: The Natural Jurisprudence of David Hume and Adam Smith (1981); the discussion of the precision of justice at 83-87 is of particular interest in this regard.

[35] An important point emphasized by Haakonssen, ibid 45-49.

[36] A point made by other commentators; see, for example, Robert M Gordon, ‘Simulation and Interpersonal Utility’ in L May et al (eds) above n 18.

[37] The source of which is not clear in Smith’s theory, a weakness we will discuss, and attempt to rectify, in later sections.

[38] And not merely upon observing another individual’s response to a situation. That there is a distinction between the individual to whom our simulation is directed, and the causal effects we take into account in our simulation, is of crucial importance to Smith’s theory.

[39] Haakonssen, above n 34, 47.

[40] Ibid 54-61.

[41] Smith runs through a variety of circumstances that impact on our ability to simulate; see, for example, A Smith, The Theory of Moral Sentiments., D Raphael and A Macfie (eds) (1984) 27-43. Not all of Smith’s discussion, it might be added, is equally plausible.

[42] Ibid.

[43] Smith, above n 41, 159.

[44] Of course, this presupposes that, at a minimum, general values and beliefs pre-exist such particular acts of evaluation and provide the foundations upon which such evaluations are made (though it is important to keep in mind that the terms ‘values’ and ‘beliefs’ are not necessarily being used in their conscious sense. How these values and beliefs might come to exist and are sustained are discussed later in the article.

[45] See Haakonssen, above n 34, 56-61.

[46] Ibid.

[47] See the discussion, above n 41.

[48] A point stressed by Haakonssen, above n 34, 54-61.

[49] While Smith laid stress upon the reflective and imaginative aspects of this stripping away process, it should be clear that such a process is not necessarily deliberate, reflective, or even within the scope for choice by an individual. As the discussion in the sections that follow should make clear, to insist on deliberate choice at every level of the process of abstraction is to strive for the impossible, as the mechanism underlying deliberate choice itself embodies valuations and abstractions which are not themselves deliberately chosen: Smith, above n 41.

[50] Ibid.

[51] Subject, of course, to the requirement that an individual could imagine that others see things differently. Given this ability probably varies across individuals, we might then expect that individuals would differ both in their abilities to put oneself in the position of another and in the consequent impartial spectator position that would emerge as a follow-on from this act of imagination.

[52] Hume’s theory, as elaborated in his Treatise of Human Nature (ed) L A Selby-Bigge (2nd ed 1978), would stand as one example.

[53] Such as M Rosenzweig, A Leiman and M Breedlove, Biological Psychology (1996) and B Kolb and I Whishaw, Fundamentals of Human Neuropsychology (1996).

[54] Such as D Hebb, Essay on Mind (1980); J Searle, Minds, Brains and Science (1984); P M Churchland, Matter and Conscious ness (1988); D Dennett, Consciousness Explained (1991); O Flanagan, Consciousness Reconsidered (1992); G Edelman, Bright Air, Brilliant Fire (1992); and A Damasio, Descartes’ Error (1994).

[55] For more, see Dennett, above n 4 and Content and Consciousness (1969); F Dretske, Naturalizing the Mind (1995), and Millikan, above n 8.

[56] As introduced by D Hebb, The Organization of Behavior: A Neuropsychological Theory (1949); F A Hayek, The Sensory Order: An Inquiry into the Foundations of Theoretical Psychology (1952) 52-53; Dennett, above n 4 and n 55; Millikan, above n 8.

[57] A notable feature, shared in common by Quine, Dennett and Davidson’s theorizing on interpretation and meaning.

[58] As discussed in the references in n 54.

[59] See Hayek (1952) above n 30, 108-109 for a brief overview of some of the limitations which might exist on this process.

[60] A feature emphasized in G Edelman (1992), above n 54, which summarized his earlier work Neural Darwinism (1987).

[61] This is one of the most influential insights to be found in Donald Hebb’s influential work, The Organization of Behavior (1949). This notion of a relationship between causal and neural connectivity underlies much of the recent literature on the neurophysiological mechanisms providing the foundations for memory and learning. For a textbook introduction to this literature, see the corresponding topics in Rosenzweig, Leiman and Breedlove (1996), above n 53 and P S Churchland and T Sejnowski, The Computational Brain (1992).

[62] Now, while the details of both how these activities come to be joined together and why they endure over time is a very interesting one, it is a matter of more physiological detail than I will examine within this article. This is because the precise details of such a replication, while of great interest, are for the purposes of this study of secondary importance. What is important, however, is the two pronged insight that (a) it might in some cases be possible for simultaneity in events external to mind to be replicated by a weighted simultaneity within mind in the form of neural networks, and that (b) such simultaneity can have a functionalist aspect, owing to the (assumed) presence of evolutionary selection pressures.

[63] To borrow terminology from Richard Rorty’s Philosophy and the Mirror of Nature (1980).

[64] This is also a dominant theme in the neurophysiological approach to ethology, as in J Camhi, Neuroethology: Nerve Cells and the Natural Behavior of Animals (1984).

[65] The quotation marks are used to emphasize that this usage of ‘expectations’ (or ‘anticipations’) are not the conscious and conceptual form of expectations that we are acquainted, but are rather non-conscious activities of mind through which ‘future is simulated’ within mind.

[66] This selectionist interpretation of mind has been developed by a number of authors, including (recently) Dennett (1991), above n 4; and G Edelman, above n 54. For further references, see Dennett, above n 54, at 184 and above n 4, at 397.

[67] See Dennett’s similar claim that brains are ‘anticipation machines’: above n 54, 177.

[68] Or non-performance, as it were. It should be stressed that although this discussion seems to focus solely upon events which occur, this does not imply that events which do not occur are of no importance. In fact, from the point of view of this work as a whole, the lack of occurrence of simultaneous events are of equal, if not greater, importance. The reason for this is that simultaneous events in the environment external to mind which are reproduced within mind form the environment within mind for events which are unfamiliar; hence, the creation and existence of connections also at the same time forms the environment in which a mind registers what is not familiar. Thus, while the replicated connections set the pattern for what is ‘normal’, it also sets the pattern for what is not normal, and hence for what is discordant with what is typically taken to be the ‘norm’. This fact becomes of great importance both when turning to a consideration of the structure of obligations and when examining the baseline that is assumed by such obligations..

[69] For more on this, and the interaction between cultural and genetic evolution, see R Boyd and P Richerson, Culture and the Evolutionary Process (1985); S Pinker, The Language Instinct: The New Science of Language and Mind (1995); and Dennett, above n 54.

[70] J Kagan and S Lamb’s The Emergence of Morality in Young Children (1987) is a useful work for its discussions of the growth of morality in children’s early years. That morality is an evolutionary adaptation is a theme that Hayek stressed more in his later writings, such as Hayek (1973, 1976, 1979), above n 30, and The Fatal Conceit (1988).

[71] For arguments of this kind, see, for example, F A Hayek (1979), above n 30, 153-157; and Hayek, above n 70. For an argument framed more generally, see Millikan, above n 8, though it should be noted that the latter directs most of her attention to the general argument that reason and rationality are biological adaptations, and does not focus to such an extent on more specific analyses of the methods of transmission of these adaptations (be they cultural or genetic or, as above, an essential interaction and interplay of the two forms). One might also be pointed to Margaret Donaldson’s Children’s Minds (1978), which stresses the importance of environmental context in reasoning, and highlights the substantial development of the reasoning abilities of children’s minds in their early years. It is probably important to note that although the arguments developed here are evolutionary and selectionist, this does not imply that an argument is being made that all human cognitive faculties are well-adapted; nor does it imply that the vast majority of such processes are well-adapted. For more on the debate in this area, see Dennett, above n 4 and Stich (1990), above n 11 for an overview of the issues.

[72] It should be obvious, but perhaps bears mentioning, that the processes of ‘socialization’ and ‘enculturation’ presuppose that it is possible to embed social practices into the practice of another and hence are in no way incompatible with the model outlined here.

[73] It should be pointed out that this also applies to sub-systems of mind, that is, they embed a weighted map of their environments, some of which might be other activities of mind..

[74] This stress on the intensity aspects of mind receives what is perhaps its most extensive philosophical development by Friedrich Nietzsche; see his works in The Portable Nietzsche (1954); Basic Writings of Nietzsche (1968); The Will to Power, trans W Kaufmann and R Hollingdale (1968); and The Gay Science, trans W Kaufmann (1974). I would argue that it is not a coincidence that, in common with connectionists, Nietzsche also stressed the physiological aspects of mind, a point often overlooked, or dismissed as irrelevant, by his numerous commentators. For an introductory overview of Nietzsche’s thought, see the comprehensive, if flawed, treatment by W Kaufmann, Nietzsche: Philosopher, Psychologist, Antichrist (4th ed, 1974). As an interesting counter-point to Kaufmann, and for an examination of some of the more interesting aspects of Nietzsche’s thought, see the influential A Nehamas, Nietzsche: Life as Literature (1986); while for an intriguing analysis of the intensity aspects of Nietzsche’s own text which are sometimes overlooked by both Kaufmann and Nehamas, see the superb work by H Staten, Nietzsche’s Voice (1990).

[75] This discussion utilizes a characterization of closures which is familiar to many works in artificial intelligence and cognitive science, this being a distinction between the all-or-nothing and the matter-of-degree. This is in turn related to a distinction made in the theory of logic between bivalent and multivalent logics. In specific, the distinction between logic based on ‘crisp’ sets and those based on ‘fuzzy’ sets is similar to the distinction made in this article between all-or-nothing and matter-of-degree closures. For an introduction to crisp set theory and logic, see P Suppes, Axiomatic Set Theory (1960), and P Suppes, Introduction to Logic (1957), while for an introduction to fuzzy set theory and logic, see H Zimmermann, Fuzzy Set Theory: And its Applications (1985), or C Lin and C S George Lee, Neural Fuzzy Systems: A Neuro-Fuzzy Synergism to Intelligent Systems (1996). On this topic, it is perhaps worth mentioning that there are significant consequences to basing one’s reasoning (or one’s model of reasoning) predominantly on one view of closures or the other. That is, it can be argued that different notions of closure tend to be associated with different general notions of reasoning, and in particular with different conceptions of what reasoning does. For a discussion of these different approaches to reasoning see, for example, the discussion in N Stillings et al, Cognitive Science: An Introduction (2nd ed, 1995) 116-135, and S Russell and P Norvig, Artificial Intelligence: A Modern Approach (1995) 92-520. Finally, it is of some importance to note that these different conceptions of reasoning have a contemporaneous manifestation in the different approaches taken by two alternative (or perhaps complementary) approaches to artificial intelligence and cognitive science: the symbolic (classical) paradigm and the connectionist alternative. For a introduction to the symbolic perspective in cognitive psychology, see J Anderson, Cognitive Psychology and its Implications (1995). For more on the operationalization of reasoning within AI, and for an excellent introduction to its symbolic paradigm, see Russell and Norvig, above, while for an overview of a connectionist view, see Lin and Lee, above. Finally, for a comparison of these two approaches, see Stillings et al, above 15-83.

[76] This is similar to Smith’s point that it can be difficult for us to simulate certain forms of mental life, above n 41.

[77] Ibid.

[78] Although even these remain modifiable to a degree, given adequate ‘training’; for some examples, see the discussion in Hebb, above n 56, Hayek, above n 56 and Rosenzweig, Leiman and Breedlove, above n 53.

[79] It should be obvious, but perhaps bears repeating, that each of these factors tends to operate as a matter of degree rather than as an all-or-nothing effect. This means that each factor tends to operate to make impartiality more or less difficult to achieve, rather than in a more extreme fashion. Impartiality, in the final analysis, remains a matter of degree.

[80] The phrase is Lon Fuller’s, in Fuller, above n 30, 125.

[81] One can also argue in another direction. Hayek, for example, argues that articulating the criteria of inter-personal validity can lead to a decrease in precision of these principles simply by making them more contestable and hence open to question, doubt and hence difficulties in application. For the argument, see Hayek’s Law, Legislation and Liberty: Rules and Order (1973) 60-61.

[82] Quoting from Fuller, above n 30, 64.

[83] For one observer’s impression that this seems to be the case in Italy see B Gunnell, ‘A nation strives to be normal’ (1998) 127 New Statesman 22.

[84] I put forward a variety of arguments for this point in my unpublished University of Edinburgh PhD thesis The Foundations of Conduct Regularity (1997).

[85] One of the most comprehensive sources is the discussion in Hayek’s Law, Legislation and Liberty: The Mirage of Social Justice (1976) and references therein. This is a point that has been hotly disputed; see the discussion in Haakonssen, above n 34, 83-87, and the references cited therein.

[86] A point stressed by Haakonssen, above n 34, 83-87. This idea is widely repeated in the literature of legal theory; see, for example, the discussion in HLA Hart, The Concept of Law (1961) 130; Fuller (1969), above n 30, 42, as well as the references assembled by Hayek, above n 85.

[87] Smith, above n 41, 44-45.

[88] Ibid 43-44.

[89] Although not without exceptions, as a reviewer of this article helpfully pointed out. Orgasms are one example of pleasurable feelings of great strength, although typically not of long duration, that might be argued to a high survival value. The essential issues here, I think, are (a) whether there is any reason to believe that matters of ‘evolutionary importance’ might be to a greater degree focused on harm and pain than on pleasure and (b) whether there is any significant difference in precision, relevant to the purposes of conduct governance and obligational structure, that hinges on the difference between harm and pleasure. As to the former, I believe the answer to be yes, but the evidence to support this would require a much more detailed exposition than I can undertake here. As to the latter, one aspect of this difference is explored in the discussion of the differences between prohibitions and performance-demanding obligations; for further discussion, see below n 98.

[90] This statement was criticized by one of the referees of this article on the grounds that ‘pain doesn’t deter’. Now, while it is not clear to me which forms of pain the referee had in mind, in its general form as a blanket dismissal it must be rejected as contrary to the facts of the matter. I would have thought it obvious, by the depressingly widespread prevalence of the practice of torture, undertaken with the explicit aim of deterring individuals from undertaking certain types of conduct, that it is widely known that certain types of pain can in fact act as a deterrent. While it might be the case that certain forms of pain don’t deter, but rather titillate (which seems to be a theme in some postmodern writing discussing ‘S&M’ practice, for example), it is a fundamental mistake to believe this implies that all forms of pain do not deter. In my view, to fail to differentiate between the consensual infliction of pain and the excruciating, continual, enduring and non-consensual pain of torture would be to evidence the most profound intellectual confusion.

[91] The argument is in Smith’s Lectures on Rhetoric and Belle Lettres (first published 1762-63, 1971) 80-81. Similarly, see Smith, above n 41, 85-91.

[92] The shift towards rules in the argument comes from Smith’s argument that general rules emerge from his mutual sympathy process, discussed above, through the stripping away of person-specific information and knowledge in the aspiration for objective grounds for impartial judgment that underlies the impartial spectator construct. See Smith’s discussion of this point; Smith, above n 41, 174-176. It is important to note that the same process that leads to the generation of rules (that is, the process of abstraction) also leads the generation of general rules of morality which are in large part dealing with necessary, as opposed to sufficient, conditions, for social interaction. Thus, in Smith’s model, abstraction filters out both particularity and content of lesser importance. It might be of some interest to note that a similar point is made by HLA Hart in his discussion of the precision of rules; Hart, above n 86, 130.

[93] The same argument would hold true, of course, for rules formulated around benefit versus rules formulated around harm.

[94] This agreement is noted in Fuller (1969), above n 30, 197-200.

[95] Now, one might, I suppose, try to escape this refutation of the positivist presupposition by accepting that there is such an empirical relationship, but by denying that such a relationship is a ‘logical’ one. Hence, it might be argued, it would make sense to claim that there is no ‘necessary’ (in the sense of logically necessary) relationship between moral merit, precision of content and the amenability of content to conforming with the principles of legality. The argument here, crudely put, would be that it is simply a matter of empirical or historical contingency that this relationship holds (assuming, for the moment, that the relationship is granted to exist).

This is not a very compelling argument. The fact that a relationship is an empirical or historical one does not imply that such a relationship is arbitrary or has emerged by chance (as sometimes seemed to be assumed in the misleading dichotomization of the world into the necessary and the contingent). Nor does it imply that there are no good reasons for why such a relationship might continue to hold, or for why the existence of such a relationship might be of importance to us. So why bother forcing arguments into these two categories, particularly when it is the category of functional necessity that straddles them both and which is essential to investigations of the potential operation of conduct governance mechanisms? To argue that by doing so one achieves an objectivity that one lacks when investigating merely contingent issues is not all that compelling, particularly when one considers the nebulous and all too plastic meaning of ‘logical necessity’, particularly when one turns to actual issues in conduct governance. For more on this, and further critique, see J Touchie ‘Exit Fantasyland: On the Ontological Status of “Pure” Logic and “Pure” Semantics’ (1998) XI International Journal for the Semiotics of Law 193; and J Touchie, ‘Jackson on the “Decisions” Underlying the Application of Rules’ (1997) X International Journal for the Semiotics of Law 317.

[96] This is clearly an empirical question. But it is not, I think, ‘merely’ or ‘solely’ an empirical question, as I discuss above n 96.

[97] The framework for analysis of such issues is discussed at some length in Touchie (1997), above n 84.

[98] I introduce a much more detailed analysis in my unpublished working article ‘The Jurisprudence of Prohibitions’ (Griffith University Law School, 2000).

[99] A term borrowed from F Lyotard, The Postmodern Condition: A Report on Knowledge, trans G Bennington and B Massumi (1984) xxiv.

[100] In political theory, see for example R Bellamy, Liberalism and Modern Society (1992); for further references, and for discussions in moral theory, see the references, above 32.

[101] One widely influential reference is, of course, Rorty, above n 63.

[102] An excellent overview of the issues is found in J O’Neill’s The Market: Ethics, Knowledge and Politics (1999) passim; of particular interest is the discussion at 16-52.

[103] The comments made by Hart, above n 86, 110-111 to this effect are typical fare. Assumptions concerning the division of labour and the consequent fragmentation of social life date back to the beginnings of discussions in political and economic theory; see, for example, O’Neill’s discussion of Aristotle’s value pluralism at n 102, 21-34. For a useful historical perspective on value pluralism in liberal theory, see Bellamy, above 100; for an interesting discussion of its place in economic theory, see O’Neill, ibid.

[104] Consider, for example, one of Hans Kelsen’s objections to natural law theories, sketched out in H Kelsen, ‘The Pure Theory of Law and Analytical Jurisprudence’ (1941) 55 Harvard Law Review 44. One of his chief criticisms of natural law’s notion of justice was that its content was so variable across individuals. From this Kelsen claims ‘[t]he usual assertion that there is indeed such a thing as justice, but that it cannot be clearly defined, is in itself a contradiction’: 48-49. Here, Kelsen seems to be presuming that knowing something implies that one is able to describe it clearly at all its different levels of abstraction. This is obviously false. If it were true, it would be impossible to say, as one often does, that one knows something at a general level of abstraction but cannot describe its particular details. It is obviously possible for one to know and describe, in general, an area of mathematics but not know, or to be able to clearly define, some of the detailed theorems of that particular branch. It might be noted that the nonsense in Kelsen’s view is also connected to what seems to be his implicit belief that things need to be described with precision in order to be precise. This, in effect, conflates linguistic precision with precision, tout court.

[105] H Kelsen, What is Justice? (1957) 21.

[106] Presumably, this is what Neil MacCormick means when he states in his Legal Right and Social Democracy (1982) 54 that ‘I reject the natural law view because in my opinion it makes unsustainable ontological assumptions about the existence of objective and rationally discoverable principles of right. Such principles exist in, not beyond, the reasoned convictions and will of human beings’. Note carefully that the issue is framed as one of ontology, rather than empirical fact; note as well that MacCormick’s distinction is nonsensical unless we assume that he is contrasting principles of right that are determinable by the conviction and will of human beings versus those that are not. That the wrongfulness of murder and torture might derive from something other than the ‘reasoned convictions’ or ‘will of human beings’, and that the wrongfulness of such conduct would continue regardless of whether particular humans or particular societies were convinced of it or desired it (‘willed’ it), is in fact a central theme of one of the arguments of this article, ie. that such wrongfulness stems from a commonly-shared, objectively existing minimalist normativity that is ground upon functional and contextual necessities (‘natural necessity’, as Hart put it, above n 86. The argument here is that such normativity can indeed be in conflict with both the convictions and willing of particular individuals and groups, contrary to MacCormick’s assertion.

[107] Kelsen, above n 105, 21.

[108] I have in mind here some of the proponents of Critical Legal Studies, although this comment might well apply more generally.

[109] Those being based on relative comparisons of need and circumstance, and demanding the identification of a (relatively) specific ‘other’.

[110] ‘Transfer’ in that satisfaction of the obligation is dependent upon some individual or group being the recipient of the perform ance of another. Complexity arises in that a specific ‘other’ must be determined, as opposed to obligations running towards others generally.