Jump to ContentJump to Main Navigation
Proportionate SentencingExploring the Principles$

Andrew von Hirsch and Andrew Ashworth

Print publication date: 2005

Print ISBN-13: 9780199272600

Published to Oxford Scholarship Online: January 2010

DOI: 10.1093/acprof:oso/9780199272600.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 20 January 2017

(p.186) Appendix 3 Gauging Crime Seriousness: A ‘Living Standard’ Conception of Criminal Harm

(p.186) Appendix 3 Gauging Crime Seriousness: A ‘Living Standard’ Conception of Criminal Harm

Proportionate Sentencing
Oxford University Press

(Co-author: Nils Jareborg)


The present topic, of gauging criminal harm, is part of the larger topic of assessing the seriousness of crimes. The question of how to judge crime seriousness has gained importance, with the increased influence of desert-oriented conceptions of sentencing - conceptions which make the severity of punishment depend in large part on the gravity of the offence.

Nevertheless, the jurisprudence of crime seriousness has received comparatively little attention. To someone unfamiliar with the criminal law, this will seem surprising. The gravity of the crime has such obvious relevance to the sanction, he would think, that judges and legal scholars surely must have been theorizing about it for years. Yet that has not been the case. Apart from a number of empirical studies on popular perceptions of the seriousness of various crimes (of which, more later), there has not been much discussion of the subject either in the Anglo-American or the Continental penological literature.

Seriousness of crime has two dimensions: harm and culpability. Harm refers to the injury done or risked by the act; culpability, to the factors of intent, motive and circumstance that determine the extent to which the offender should be held accountable (and censured through the criminal law) for the consequences of his act. Both dimensions affect crime seriousness; to use familiar examples, murder is more serious than aggravated assault because the injury is greater, and it is more serious than negligent homicide because the actor's culpability is greater. The problem is to develop criteria for harmfulness and culpability that are more illuminating than simple intuition.

With respect to culpability, there is some theory and doctrine that can be drawn upon, and it comes from the substantive criminal law. The substantive law already distinguishes intentional (ie, purposive, knowing, (p.187) or reckless) conduct from criminally negligent behaviour. For sentencing doctrine, more refined distinctions could be developed, concerning the degree of purposefulness, indifference to consequences, or carelessness involved in a sentenced offender's conduct. The doctrines of excuse in the substantive law could also be drawn upon in order to develop theories of partial excuse - for example, of partial duress and diminished capacity.1

With regard to harm, however, matters are otherwise. Few legal doctrines have been developed on how the gravity of harms can be compared. The substantive law does not provide much guidance, because the occurrence of harm is not explicitly made a general condition of criminal liability. The legal offence definitions simply describe various kinds of prohibited conduct - which the legislature is assumed to have considered harmful enough to be criminalized. This Appendix is devoted to the subject of gauging harm, because we think it is particularly in need of exploration.


We have limited the scope of our analysis of criminal harm in various respects, and should explain at the outset what those limitations are.

2.1 Victimizing criminal conduct

The present chapter addresses only the harmfulness of criminal conduct, which injures or threatens to injure identifiable victims. We are restricting ourselves to conduct already declared criminal, because our interest in harm arises in the context of sentencing policy. We wish to assess harm, in order to see how serious the offence is; and we wish to determine the offence's seriousness for the purpose of determining the severity of the punishment. There is, of course, another context for considering harm -namely, when deciding whether conduct should be declared criminal at all. Whether conduct is harmful, and how harmful it is, should be an important factor in the legislative consideration about its proscription. However, it is not the only factor: the legislature should also consider how much countervailing social value the conduct has, the extent to which prohibition could be intrusive upon privacy or personal autonomy, the practical feasibility of enforcement, and so forth.2 Thus the fact that a given species of conduct might meet our suggested criteria for harmfulness, and even for grave harmfulness, does not necessarily settle whether it should be proscribed. We will be speaking of conduct that has been prohibited already.

Our focus is also on victimizing behaviour, that is, behaviour the primary apparent evil of which consists in that it sets back the interests of identifiable natural persons. We will thus be speaking chiefly of the (p.188) traditional offences such as theft, burglary or assault in which an individual is the primary sufferer, and the concern is with how much he has been harmed. It is these crimes that are most obviously injurious and constitute a logical starting place for the analysis. In thus focusing on victimizing conduct, we are fully aware of the existence of other forms of criminal harm. Some crimes (eg, certain environmental offences) have harm that is primarily collective or aggregative: no identifiable person has his interest set back by activity of an identifiable actor; but if a sufficiently large number of persons engage in the conduct, the public's interests are adversely affected. In focusing on individual victimizing crimes, we do not wish to suggest these other crimes are necessarily less important: only that their harmfulness is a more complicated matter to analyse. Having a theory that addresses victimizing harms should be a useful first step.

2.2 Standardization

We will be concerned chiefly with the standard harm involved in a given category or subcategory of crime. How hurtful a given intrusion is depends on the situation of the victim, and particular victims' situations may vary greatly. The theft of a car from a rich person may be no more than a temporary inconvenience; a car theft from a poor person in a rural area may leave him bereft of any means of transportation. We shall assume, in rating the standard case of a given species of crime, that injury occurs to someone who is neither especially vulnerable nor resilient.

Why this emphasis on standard harm? Particular criminal acts are too diverse to be rated on an individualized basis. The analysis is aided when one (1) rates the standard case of an offence, and then (2) addresses unusual cases through supplemental principles (eg, of aggravation and mitigation). The first of these tasks will keep us amply occupied here. The second, on aggravation and mitigation, is complex enough to call for separate treatment.

In sentencing doctrine, one judges harm in order to determine how blameworthy the actor is in committing the conduct, not in order to compensate the victim for his loss. This actor-perspective requires one to consider how much, and within what limits, the act's harmful consequences may justly be attributed to the actor. Conceptions of culpability require that, in assessing the seriousness of conduct, unforeseeable consequences should not be taken into account.1470a The standard harm in (p.189) any given type of criminal conduct is, ordinarily, foreseeable. However, when one is talking about atypical harms, foreseeability diminishes. The burglar may be expected to understand the typical consequences of a burglary, but not, ordinarily, Jane Doe's particular situation - for example, the extraordinary personal value the vase taken from her apartment had for her as a gift from a deceased friend. A final point about standardization: we should be aware from the outset of the limits of particularizing ratings of seriousness. One can rate the harm category - and also take certain features of the harmfulness of the particular event into account through the aggravation/mitigation principles. But one cannot reasonably expect to approximate, in legal judgements of seriousness, the particularity of everyday judgements. Precisely how much Ms Doe is injured by the loss of her particular vase would require a delving into her personal situation to a degree which the law cannot possibly undertake, and should not try to undertake. The rehabilitative penal ethic lost its credibility in part because of unrealistic claims it made about the individualization of sentence. A proportionalist penal ethic should not make the same mistake. The criminal law reflects a system of public standards, not an arena for personalized judgements. What the law realistically can accomplish is to assess crime seriousness in standard cases, and then permit deviations from that assessment for specified types of special circumstances.

2.3 Normative focus and the question of cultural variability

Our analysis of harm is normative: we are offering a theory on how harms should be rated. The reason for this normative emphasis is that our account is part of a broader penal theory: one that claims that the severity of punishment should be proportionate to the gravity of the criminal conduct. The grounds for the proportionality principle are those of fairness: that punishment, as a censuring response to criminal behaviour, should reflect the degree of blameworthiness of the criminal conduct (see Chapters 2 and 9). Such an argument stands only if the judgements of blameworthiness of conduct (including those of harm) reflect how much the offender and his conduct ought to be blamed.

Since Thorsten Sellin's and Marvin Wolfgang's 1964 research,3 a number of studies have surveyed popular perceptions of seriousness. The studies provide the persons surveyed with brief descriptions of a variety of crimes, and ask them to rate those crimes' seriousness on a numerical rating scale. Some consensus has been found: people from different walks of life tend, for example, to give similar rank-orderings to crimes.

(p.190) Interesting as such research is, it does not address our question - for it only deals with how various segments of the public in fact rate crimes. Popular ratings do not settle the harmfulness of conduct, for two reasons. One concerns the possibility of factual misjudgement. Respondents have not studied the actual consequences of the crimes which they are asked to assess, and their assessments may be based on erroneous beliefs about those consequences. People may believe that burglary typically involves greater risk of violence than the available empirical evidence suggests it does, or may underestimate the injurious effect of some white-collar crimes. The other reason is that the criteria for judging the known facts have not been reflected on. In judging assaultive crimes, for example, is it only the physical risks that count, or also the impact on the victim's personal dignity? If the latter is considered, why so? Just asking people about their assessments of the harmfulness of assault does not support reasoned judgements about such questions. The advantage of a normative theory of harm is that it addresses the ‘why’ questions.

We are not, however, seeking ratings that are invariant worldwide. How harmful burglary is, for example, depends on its typical impact, and that impact may vary across cultures - depending on both factual and normative differences. The factual differences concern the underlying social arrangements. In our culture, a burglary is an intrusion into someone's home and the home is ordinarily the focus of much of the person's personal life. Were people to sleep in tents, and perform most activities in communal facilities, the impact of a burglary could be much less. Normative differences also matter. Part of the assessment of the harmfulness of burglary depends on the importance accorded to privacy; in a culture where privacy is less valued, the conduct would have less significance. One thus needs to examine the consequences of the offence, given certain assumed social living arrangements, and then evaluate those consequences in the light of certain assumed values (eg, concerning privacy). The role of our proposed theory of harm is to aid in that evaluation. That the theory could yield different harm-ratings in another place, where living arrangements and values are different, is a strength, not a weakness.


In gauging criminal harms, the difficult task is that of comparing the harmfulness of crimes which involve different interests. How is car theft to be compared with burglary, when the former involves a significant property loss, and the latter a smaller financial setback but an invasion of privacy as well? Making such comparisons calls for common criteria, or at least common guiding ideas, for assessing the interests involved.

(p.191) The guiding idea, which we have come to find most natural, is one concerned with the quality of a person's life. The most important interests are those central to personal wellbeing; and, accordingly, the most grievous harms are those which drastically diminish one's standard of wellbeing. Mayhem is so serious because it makes its victims live in misery; burglary seems less serious because it does not create such misery, although it still has a significant impact on the quality of life in its intrusion on the person's privacy.

We required a term for this idea, and came upon Amartya Sen's useful essay, The Standard of Living.4 Economists traditionally have used ‘living standard’ to refer only to degrees of economic affluence or want. Sen, however, suggests a richer use for this term, one that includes not only economic means but other, non-economic capabilities that affect personal wellbeing. We decided to adopt the term ‘living standard’, understood in this broader sense.

3.1 Inadequacies of the ‘welfare-interest’ criterion

Our interest in the living standard grew out of dissatisfaction with the ‘welfare-interest’ criterion which Joel Feinberg proposed in his 1984 book, Harm to Others.5 Previously, one of us (von Hirsch) had relied upon Feinberg's criterion, for the purpose of grading crime seriousness for sentencing.6 Feinberg suggests that the importance of interests infringed by criminal conduct can be gauged according to the degree to which those interests characteristically affect choice.

The most important interests, according to Feinberg, are so-called welfare interests, and hence the most grievous harms are those that intrude substantially upon such interests. Welfare interests are defined by him as the interests a person needs satisfied in order to have any significant capacity to choose and order his way of living. They provide, in his words, ‘the generalised means to the advancement of [the person's] ulterior interests’.7 When those interests are intruded upon or destroyed, the person is foreclosed from a wide variety of other choices. Physical safety, for example, is a welfare interest because the person is impeded from pursuing almost any aims of his choice, when physically incapacitated. Next in importance come so-called security interests, which Feinberg defines as those needed to provide a certain safety margin to the welfare interests. Lowest-ranked would be all other interests, which he entitles ‘accumulative interests.’

How plausible is it to define the importance of interests in the way Feinberg does? There are two ways of interpreting his theory. One is that it describes the logic of assessing interests: people actually judge interests to be of fundamental importance when they provide the generalized means to the advancement of more particular concerns. Sometimes, such an (p.192) account would seem plausible. Most of us would consider it a basic interest to possess a certain modicum of financial resources. Why so? Because that is necessary to pursue our particular life plans, whatever they may be. But counter-examples readily come to mind, where such an account would seem strained. Consider the interest (which almost all of us would deem very important) in avoiding intense and protracted physical pain. Such pain would doubtless interfere with the pursuit of various specific life plans, but does that really account for why we think the interest so important? Is it not, instead, that the avoidance of pain is in itself essential to the quality of a person's life; that a life of physical suffering is an awful life?

Alternately, Feinberg's welfare-interest criterion could be explained more modestly, as a theory of state action. Interests are being ranked in importance to help guide the coercive action of the criminal law. The law's criteria for harm should make allowances for differences in how people value life's goods, and should not elevate one set of focal aims above another because that is for persons themselves to choose. The attraction of the welfare-interest theory is that it helps assure that the criminal law, in its criteria for harm, gives due recognition and scope to individual choice: welfare interests are valued highly because they are a prerequisite to whatever ways of life individuals themselves might choose. The welfare-interest idea thus appears to rest on the tenets of philosophical liberalism.8

Relying thus on notions of choice can produce odd results, however. Much as one may value choice, its diminution is not necessarily the worst thing that can happen to one who is victimized. The person who has been maimed or made destitute obviously suffers grievous harm. But is this so merely because his freedom of choice has been restricted? It would seem more straightforward to say the maimed or destitute person might survive but only barely so, with loss of the most basic human satisfactions. As long as one regards the fate of individual humans to be important, one would consider such harms to be serious, without need for invoking further assumptions about the importance of choice.

The idea of liberalism also seems somewhat misapplied. Liberal notions of autonomy are chiefly concerned with limiting the state's power to compel an actor to adopt one or another set of aims for his own life. It is thought better that a person be permitted to pursue his chosen manner of existence (however unconventional or seemingly irregular) than that he be required to live as state authorities deem wise or virtuous. In the present context, however, we are speaking of harm to others; the standards of criminal harm do not concern the actor's own way of life, and he remains free to make his self-regarding choices. Instead, the actor has chosen to interfere with the interests of others; and the harm criteria are supposed to measure the degree of that intrusion. That harm may not merely be to other persons' capacities (p.193) to choose, but to whatever vital interests they may have - and those other vital interests are not necessarily reducible to a capacity to choose.

Granted, the potential victims of harm have a variety of life aims, and that variety should be recognized. Harm should thus be defined broadly enough so that it is not just the interests of a conventional person that are protected. This means that the harm-criteria need to be scrutinized to see if they give sufficient leeway to differences in life style and life aims of potential victims. We will argue later that a living standard criterion provides this leeway.

Besides these conceptual difficulties, the welfare-interest classification fails also to supply sufficient distinctions among harms. Thus:

  1. (1) The highest category, of welfare-interests, would consist of those concerns essential to a minimum of choice. Basic physical health would qualify, as a person will almost always need that to go about his chosen business. A minimum of economic support would be included, for similar reasons. Perhaps, a few more interests would also qualify. Feinberg notes, however, that his choice criterion does not supply a ranking among the welfare interests, because they are interdependent and each is essential for choice.9 (Good health but destitution debilitates, just as bad health but affluence does.) This is not a problem for his purposes, since he is addressing the decision to criminalize conduct, and is asking whether the conduct is sufficiently injurious to be worth proscribing. It is troublesome for our purposes, however, as for sentencing purposes we require rankings even among the most serious crimes.

  2. (2) The intermediate category - ‘security interests’ - are defined as those needed to ‘cushion’ or provide a safety margin to welfare interests. This conception seems deficient in several ways:

    • It is so imprecise as to be virtually unworkable. How does one tell, even in principle, whether a given interest is or is not needed to ‘cushion’ welfare interests?

    • It gives middle-level interests, inappropriately, a purely instrumental role. Consider the interest in living with a modicum of comfort. We would regard this interest as having inherent value, because it is so important to personal wellbeing. Its importance does not lie merely in the fact that such comfort provides a margin of security against destitution.

    • It allows no distinction in importance among middle-level interests. Some intrusions would put the victim near the subsistence level. Others would merely put him just below a level of wellbeing that seems normal or adequate.

  3. (3) The lowest level of interests in Feinberg's classification - ‘accu-mulative interests’ - strikes us as a catchall. Some of these still would play a significant role in the person's overall wellbeing; others would (p.194) be purely accumulative in the literal sense. Such cases need to be distinguished.

3.2 The idea of the living standard

The term ‘living standard’ has a narrow economic meaning, referring to affluence. It has, however, a broader meaning - the one which Sen emphasizes,10 and which is used also in quality of life studies such as the Stockholm Metropolitan Project.11 This refers to the quality of persons' existence in a sense that includes not only material support and amenity but other non-economic capabilities that affect the quality of a person's life.

The living standard is one of a family of related notions, including wellbeing, that refer to the extent of human flourishing. Wellbeing, however, can be highly personalized. The quality of my life depends upon my particular focal aims. To the person who wants to devote his life to contemplation and prayer, material comfort and social amenities may matter little. Thus to determine a particular person's wellbeing, one ordinarily needs to know much about his life goals and his reasons for adopting them.

The living standard, however, does not focus on actual life quality or goal-achievement, but on the means or capabilities for achieving a certain quality of life.12 It is also standardized, referring to the means and capabilities that would ordinarily help one achieve a good life. Consider the person who is in good health, affluent, and with a wide social network from which to draw friends and acquaintances. If his chief goal in life is to write great novels and he lacks the requisite talent he may still be frustrated and unhappy. Nevertheless, it would be appropriate to say that he has a good standard of living: not merely because of his wealth but because he has the other means that people ordinarily can use to live well -whether or not he chooses to make full use of those means or actually enjoys their fruits. These features are important for our present purpose because, as noted in the previous chapter, we are primarily concerned with standard harms.

A given capability or resource can support achievement of a variety of ends of a person's choosing. One thus can make living standard judgements without having either to know a particular person's focal aims or to specify in detail what the standard person's aims are or ought to be. The judgements, in other words, assume a certain degree of pluralism. One may judge, for example, that a person's undisturbed possession of his home has a certain degree of importance to his living standard. The judgement can hold, notwithstanding the varying uses of the home in people's lives: that one individual may use his home chiefly as the locus of family life, another as a place for study, another as a place for entertaining numerous guests. One does not need to assume which of these ends are (p.195) preferable, or even to assume that a person should choose personal goals that involve spending much time at home. The judgement is merely that the home plays a significant role in a variety of lives people choose to live in our society. This pluralism is important for our purposes, for it will permit us to make living standard judgements without an ambitious specification of what people's ultimate focal aims are or ought to be.

The living standard differs, nevertheless, from welfare interests in that choice is not the criterion: the means for achieving a certain quality of life is. Many of the interests important to people's living standard afford, as we just noted, a considerable range of choice. Still, the ultimate focus is not on the extent of choice but on the life quality which those interests help support.b

While the living standard, thus conceived, includes non-material capabilities, it is restricted to those capabilities that are self-regarding: those that are involved in the quality of a person's own life. Altruistic satisfactions and dissatisfactions would not be taken into account. In Sen's words, ‘one's misery at the sorrow of another… is not a reduction in… the person's living standard’.13

3.3 Why the living standard?

Why rely on the living standard in order to gauge harms? A simple answer is that it appears to fit the way in which we ordinarily judge harms. Why is mayhem more harmful than burglary? It is not merely because the maimed person's choices have been narrowed more; but rather, because the overall quality of his life has been so much more adversely affected.

In adopting a living standard criterion, we are not claiming that goodness or badness of conduct depends, as a general matter, on the conduct's effect on the quality of people's lives.c Our claim is more (p.196) modest: that the living standard provides, not a generalized ethical norm, but a useful standard which the law can use in gauging the harmfulness of criminal acts. We think this is so for several reasons.

Victimizing harms involve intrusion into various personal resources or interests people have. If one asks why the state should seek to protect such interests through the criminal law, the most plausible answer would be that people require those resources to live decent lives. Robbery and theft are prohibited because people need safety and certain possessions to live tolerably. It thus makes sense to gauge the gravity of criminal harms by the importance that the relevant interests have for a person's standard of living.

Much of a modern state's protective concerns for its citizens is aimed at safeguarding interests of theirs that constitute the generalized means for pursuing a satisfactory life. Social welfare programmes concern the provision of a minimum of economic support to enable people to live decently. Health and safety regulations serve to ensure the physical condition people require to have a comfortable existence, and so forth. Our conception sees the criminal law as having a comparable primary role (see 2.6.2).

The living standard approach also has the advantage of a certain modesty; no ‘deep’ theory of preferred life aims or appropriate social roles is presupposed. It recognizes, as noted earlier, a considerable degree of pluralism in the uses and ends to which people put their various interests.

An alternative to the living standard analysis could be the more ambitious one, of specifying certain focal aims or social roles as pre-eminently important or valuable; then, harms could be assessed by their degree of interference with the achievement of these life aims or the performance of these roles. David Braybrooke takes this view in his analysis of needs. The definition and importance of needs, he asserts, should depend upon their function in carrying out certain central social roles, viz., the role of parent, householder, worker, and citizen.14

We find such an approach troublesome for a number of reasons. First, there is a problem of justification: why these particular life-ends or social roles and not others? Braybrooke asserts that his list of social roles is self-evident. In his words, ‘having a life of minimum normal scope consists just in having an opportunity to perform those roles and tasks’.15 The list is not self-evident to us, however. If the claim is that Braybrooke's four roles are normatively of overriding importance, it needs to be explained why so. If the claim is descriptive (that most people do primarily direct their lives to the four social functions he lists) it would need to be supported by empirical evidence. Such evidence would not be so easy to supply - given the various and overlapping character of people's goals. Braybrooke lists ‘householder’ as one of his four functions. Most people in our society establish households, but is this because they see householdership as something valuable in itself, or because (as we think more plausible) a (p.197) household may serve various ulterior ends people have? If so, why not consider these ulterior ends as pre-eminent?

Second, there is the problem of achieving sufficient specificity. A broad enumeration of pre-eminent focal aims or social roles such as Braybrooke's would not take us far in rating criminal harms, especially intermediate or lesser harms. Homicide or mayhem obviously interferes with the social roles of which he speaks, but how much does a purse-snatching interfere with the victim's role as parent, householder, worker, or citizen? Braybrooke could reply that he need not address such questions since he is only trying to define basic needs. But the problem would become acute if one tried to extend his approach, of defining needs in terms of specified social roles, to the assessment of the wide range of interests protected by the criminal law.

Finally, we wonder about the appropriateness of the preferred social roles approach. Criminal harms are being rated to help decide upon the severity of state sanctions. We doubt whether a theory of state sanctions, in a plural society, should attempt to specify the ways of living that merit special respect. Consider Braybrooke's category of parenthood. Much as many persons value family life, others in our society legitimately have different life priorities. Granted, moral paternalism is not the issue here: no one is being compelled to raise a family, since the standards of criminal harm (as noted already) do not seek to regulate the actor's own lifestyle. Still, pluralism is a matter of concern. The standards of harm affect the extent to which the state protects different interests people have, through the sanctions of the criminal law. In a society that respects diversity, those standards should to the maximum feasible extent acknowledge the variability of people's personal aims and purposes. This it cannot do if those standards rest on assumptions that certain life aims or social roles have pre-eminent worth. If a person wishes to devote his life to contemplation, how are his preferences respected if legally protected interests are valued solely by reference to aims (such as child-rearing) not his own?

The living standard criterion is not so vulnerable to such objections. The security of the home is judged to contribute to the living standard to a certain degree, not because householding is deemed an intrinsically valuable social role or one that contributes to another preferred role such as child-rearing, but because the home contributes to various ways a person may choose to live his life (from study through child-rearing to valetudinarianism); that potential contribution makes the home a resource having a certain role in the standard of living. Since the living standard is concerned with standard cases - with the means or capabilities ordinarily needed to achieve a good life - certain non-standard modes of life con-cededly will receive less attention. The valuation given the home does not do much good to the person who opts to live as a travelling mendicant. But (p.198) the grounds for this treatment of non-standard cases are less troubling. It is not that mendicancy is deemed less worthy than (say) child-rearing. It is, rather, a recognition that the criminal law, with its general rules and its emphasis on foreseeable harms, is mainly equipped to deal with standard cases.

In thus emphasizing the pluralism of the living standard criterion, have we re-embraced the earlier-rejected theory of welfare interests? (After all, the supposed raison d'être of that theory was choice.) We think not. A distinction should be drawn between the harm standard itself and the reasons supporting it. One reason, among others, why we favour the living standard criterion is that it recognizes a certain pluralism of legitimate life-ends and social roles. Nevertheless, pluralism should not be the harm criterion's principal basis. In assessing how burglary affects the living standard, one certainly should be mindful of the variety of legitimate uses to which people put their homes. But to account for the injuriousness of burglary, it does not suffice to call attention to the extent to which the occupant's choices have been narrowed; one needs to consider the impact the conduct also has on his comfort and peace.

The living standard idea is a rich one. It rests on something we consider in everyday existence and various arenas of social policy: the elements of a good life. It will be possible to make comparisons: if a typical person's standard of living is at issue, the interests protected by the criminal law can be compared with other interests. We can ask how a beating compares with a minor accident, or how an arson compares with an accidental fire in one's apartment. We can compare these harms by reference to various generic interest dimensions that may be involved: for example, the beating may compare in physical injuriousness with a minor accident, but has an additional component of humiliation that the accident lacks. It will be possible also to make the factual and value judgements more explicit: to separate out (1) the factual question of the various typical consequences of a given type of crime, from (2) the value judgement about how significant those consequences are in reducing a person's standard of living.

The living standard is also rich in another sense, of leaving room for ancillary moral judgements. Freedom, privacy, and a variety of other values may (indeed, should, as we will see shortly) be considered in applying living standard judgements. What the living standard thus provides is not a single, essential standard but a perspective through which a variety of ethical and practical judgements can be made about the impact of criminal acts on the quality of people's lives.

Finally, the living standard analysis would explicitly allow for cultural variation. Different social living arrangements can affect the consequences of a criminal act; and normative differences among cultures can affect the impact of those consequences on how well a victim lives. This sensitivity is (p.199) an advantage of the living standard approach: as noted earlier, we need criteria capable of taking cultural variation into account.

3.4 Factual v normative elements of living standard judgements

How, and from whose perspective, are living standard judgements to be made? The kind of analysis we are proposing presupposes certain social practices of our society and certain assumed values; but, given those practices and values, calls for a reasoned application of the living standard criteria to particular species of criminal harm. Let us explain.

When one examines how much a given species of criminal harm typically intrudes upon someone's living standard, certain factual, but also certain normative judgements are involved. The mix of fact and norm varies as the analysis proceeds. Consider burglary again. This is an intrusion into the home, requiring one to assess the importance of the home in a person's living standard. The assessment might begin with the interest in health and safety. The home has certain functions in providing shelter, warmth and rest that are critical to basic physical well-being. It is fact that in cooler climates people's health deteriorates when they are homeless and not offered substitute shelter (a fact too sadly evident on the streets of many English and American cities). This interest, however, is not ordinarily involved in burglary, as the offence intrudes upon but does not deprive the person of his dwelling-place. Another kind of concern, that of privacy, brings normative judgements more to the fore. A breaking-and-entering of a dwelling invades the privacy of the home, but the significance assigned to privacy is ultimately a moral judgement. That judgement, too, may be rooted in certain ethical and cultural traditions.d How, then, should the evaluator (that is, the agency responsible for gauging the conduct's harmfulness) make and justify these various judgements?

An ambitious approach would be for the evaluator to try to justify the underlying values: for example, to give an account of privacy and derive therefrom the judgement of its importance to the living standard. Such an account - if possible at all - would have to be grounded in a more fully articulated moral theory than we can supply here. The account is also not needed, for reasons that should be evident already: we are not trying to develop an invariant harm-analysis, but instead to derive ratings (p.200) applicable in our society, given certain prevailing social practices and also certain ethical traditions. Nor do we even need to try to describe those practices and values in full, which would be a formidable task in itself. Instead, we need merely incorporate certain assumptions into the analysis: for example, about the role of the home in everyday living, and about the importance of personal privacy. Granted, the making of such assumptions involves the exercise of judgement. The analysis, however, requires merely that such judgements be explicitly spelled out. Other evaluators, who disagree, can simply make their own, adjusted assumptions. Our aim, as noted earlier, is not to settle such differences of opinion, but to pinpoint their location, and thus permit conditional conclusions: if privacy is accorded such-and-such a degree of importance, these are the consequences that follow.

What, however, of the next step, of applying the living standard principles to the valuation of particular species or subspecies of harm? Suppose, for example, we are asking how much a typical home burglary - involving a breaking-and-entering, theft of a TV, but no ransacking - reduces the living standard, considering certain assumptions about the importance of privacy and other interest-dimensions to wellbeing. That judgement has to be made in the analysis itself, by reasoned application of our suggested principles. It cannot be made simply by reference to prevailing opinion. Why so, should be evident already: ours is a normative analysis that begins with certain factual judgements about the consequences of the crime, and then applies certain stated valuation principles to those facts. The value of the exercise -the reason we have opted for the living standard approach rather than simply asking members of the public for their ratings of harm - lies precisely in the opportunity (1) to check the factual basis of such judgements, and (2) to scrutinize how well-reasoned is the application of the assumed valuation principles to those facts.

3.5 Why our conception is not utilitarian

Is reliance on the living standard a reversion to a utilitarian doctrine of punishment? After all, a deterrence theorist might employ the idea of living standards to estimate the social harm wrought by (say) the offence of burglary; and then set a penalty designed to reduce the crime's incidence and hence its social injuriousness.

The answer is (obviously) negative. A utilitarian penal theory is aggregative and consequentialist; it seeks to estimate how much the totality of future harms may be reduced through this or that penal strategy. Our theory of harms, and the use to which we propose to put it, is neither of these things. We are not aggregating. Because a burglar is responsible only for his own conduct, it is the harm that his conduct typically causes or risks that determines the gravity of the offence - and not the (p.201) totality of harm caused by the acts of all burglars, over whom he has no control.

Moreover, we are assuming a desert-oriented criterion for how much to punish burglary, not a consequentialist and preventive one. We look to the living standard to decide the degree of harmfulness of a standard act of burglary; and look to that assessment, in order to determine the seriousness of the crime of conviction. The sanction - according to desert assumptions that have been detailed earlier in this volume (Chapters 2 and 9) - should be proportionate in severity with the gravity of the offence, thus determined. This is wholly different from designing a sanctioning system in order to optimize citizens' future aggregate living standards; such a system, to have the desired preventive effects, might have to have sanctions that are disproportionate to the gravity of the offence.


Having sketched our rationale, we need to explain how criminal harms may be graded through a living standard analysis. Let us, then, sketch the main elements of that analysis.

4.1 The standard person

Estimating the impact of a given intrusion on someone's living standard requires assumptions to be made on what other resources he possesses. The more slender those resources, the more devastating the intrusion. How, then, should one determine the standard impact?

Our living standard analysis is designed to gauge harms affecting a variety of interests, from the most to the least important. It is thus helpful to imagine someone who has various interests, and then ask how deprivation or intrusion into this particular interest would affect that person's quality of life. The hypothetical victim would thus be assumed to have the interests needed for the various living standard levels: certain interests or resources required to subsist; and others required to raise his wellbeing to higher levels. These assumptions provide a perspective for judging the importance of a particular intrusion: it enables one to judge how important the particular interest or resource intruded upon ordinarily is to a good life, compared to the various other interests and resources a person may have.

4.2 Temporal perspective

What temporal perspective should one adopt when judging the impact of a crime on someone's wellbeing? The perspective makes a difference. Consider having one's pocket picked. If the question is ‘How was your (p.202) day?’ it makes perfect sense to answer that it was awful, because one has lost one's wallet with all of one's IDs. However, if the question is how one's year has been, it would be odd to say that it has been awful, merely because one's pocket was picked four months ago.

Living standard judgements are addressed to the quality of someone's life, and this suggests a considerably longer temporal perspective than the highly variable quality of experience from day to day. The appropriate perspective is a middle-term one - something approximating ‘How has your year been?’ or perhaps even a slightly longer duration. An extremely long temporal duration, however, would be too reductive: even awful experiences diminish in importance if the question were: ‘How was your last decade?’ - and indeed, such questions are seldom asked in ordinary life.e

Our suggested temporal perspective is addressed, however, not to the actual duration of the crime and its visible after-effects, but to the importance of the experience from the overall viewpoint of a given time-frame. A physical assault, and its immediate trauma, may soon be over; yet if the experience was painful or humiliating enough, it may still loom large in the evaluation of, say, the quality of a whole year's experience.

4.3 Rating the living standard: The four levels

If harms are to be gauged according to the extent to which they affect someone's living standard, then the living standard needs to be graded. Any grading scheme is bound to be arbitrary, but one should be able to make rough distinctions: for example, between a person's suffering an intrusion that makes him somewhat less comfortable, and one that affects the person's subsistence interests. We propose four living standard levels, which may be formulated as follows:



General Description


Survival, but with maintenance of no more than elementary human capacities to function. No satisfactions presupposed at this level.

Minimal wellbeing

Maintenance of a minimal level of comfort and dignity.

‘Adequate’ wellbeing

Maintenance of an ‘adequate’ level (but no more) of comfort and dignity.

Standard wellbeing

Maintenance of full life quality.

This scale is designed to gauge the degree to which a given intrusion affects the person's living standard. If, for example, Crime X affects interests (p.203) required for subsistence (living standard level 1°), it qualifies as gravely harmful. If Crime Y affects only interests needed to maintain the quality of the person's life at the ‘adequate’ level (level 3°); this renders the harm less grave but still substantial.

Having four grades is by no means inevitable. Since we are dealing with a continuum, a larger or fewer number of notches could have been cut. We propose these four gradations because the differences among them seem reasonably apparent. Having a larger number of living standard gradations would make it easier to rate harms once the gradation affected by the criminal conduct is determined; but it would make the latter determination more difficult, as the differences among the grades would be less easy to discern.

There will also be variations in wellbeing within the four grades. Subsistence, for example, ranges from bare survival to survival but with significant debilitation: both mayhem and aggravated assault affect subsistence, albeit to different degrees. Again, this should not be surprising, if one keeps in mind that we are dealing with a continuum, not neatly demarcated steps.

The function of these four gradations is to measure the extent to which a criminal act typically intrudes into a person's living standard. The (numerically) lower level that is involved, the greater the intrusion. To take an obvious example: an aggravated assault threatens subsistence (1°) and thus is substantially more harmful than a petty theft affecting only level 4°. This explains why we are using a ‘positive’ conception (the living standard) to rate harm (something negative, involving a taking-away). Criminal harm consists in the intrusion into legally protected interests, and its seriousness depends on the importance of the interests involved. We are gauging those interests’ importance in terms of their significance for a person's living standard.

Let us consider the four living standard levels more closely, then, and give a gloss on their meaning.

Subsistence (1°). We have defined subsistence as ‘survival, but with maintenance of no more than elementary human capacities to function.’ This means barely getting by. Included would be preservation of one's major physical and cognitive functions, and preservation of a minimal capacity for social functioning. Being killed obviously destroys subsistence, but being maimed or made destitute would intrude upon subsistence also.

Critical to our definition of subsistence is that no significant satisfactions are presupposed at this level. Comfort thus is not part of subsistence -although avoidance of intense pain would be. Neither is privacy or self-respect included: one can survive or get by without privacy and despite repeated humiliations.

(p.204) Minimal wellbeing (2°). This level we have defined as ‘a minimum level of comfort and dignity.’ It means having more than barely getting by. In addition to surviving, certain elementary human satisfactions are assumed: some comfort, a minimum opportunity for self-respect.

Part of minimal wellbeing is thus a certain level of material support: shelter from inclement weather, nutritious food, etc. Another part is some degree of privacy and personal autonomy. One can subsist without the slightest privacy, but one hardly can be said to have even a barely satisfactory life. Protection against grossly demeaning or insulting treatment falls in this category also: one can subsist despite such treatment, but it impedes the maintenance of any degree of self-respect. Nevertheless, we are speaking only of bare minima of comfort, privacy, etc.: an existence at this level would still be having a substandard quality of life.

‘Adequate’ wellbeing (3°). We have defined this level as ‘maintenance of an adequate level (but no more) of comfort and dignity.’ The term ‘adequate’ is used here not in the sense of being satisfactory, but in the more restricted sense of being not plainly inadequate. Included would be a level of material amenity needed for a minimally comfortable existence, but no more than that. It also includes a degree of privacy and avoidance of demeaning treatment needed to avoid a substantially deprived existence.

Standard wellbeing (4°). This we have defined as ‘maintenance of full life quality.’ It includes those concerns that improve someone's quality of life to the point where no substantial grounds for complaint exist. Included would be some modest luxuries, and being treated by others with reasonable civility.

Marginal impact. Some criminal harms have only a marginal impact on the quality of a person's life. Paradigmatic of harms in this category are petty thefts. The conduct may cause brief inconvenience, but typically has no significant lasting impact on the person's wellbeing.

4.4 The generic interest dimensions

A criminal act can intrude upon a variety of different kinds of interest dimensions.f To cite previously mentioned examples, assault affects both the person's safety and his self-respect; a burglary, a person's material (p.205) comfort and privacy. The crime can affect these various dimensions in different degrees, and with differential impacts on the living standard: the privacy intrusion in a burglary, for example, may affect the standard of living more than the comfort intrusion does. Our analysis thus needs to distinguish various generic interest dimensions. We suggest three main dimensions, namely:

  • Physical integrity

  • Material support and amenity

  • Privacy/freedom from humiliation

We do not claim this list is complete, and have used no deep theory to derive it. Instead, we rely on our impressions of the main kinds of concerns that seem typically involved in victimizing crimes. Adding supplemental interest dimensions would not alter the mode of analysis.

How are these interest dimensions to be used? One would first identify and separate out the interest dimensions involved in an offence. Next, one would apply the living standard criteria to each dimension successively. To illustrate, consider simple residential burglary: the offender, during the residents' absence, breaks into an apartment and steals a television set. First, one identifies the interest dimensions. Here, those chiefly involved appear to be (1) material amenity, and (2) privacy. The material loss consists in the loss of a useful home appliance (a TV), the expense of having locks repaired, etc. The loss of privacy derives from the intrusion into the victim's personal space. Second, one applies the rating criteria to each dimension, successively. With respect to the material amenity dimension, the impact on the living standard is usually quite limited: a workable replacement TV, for example, can be obtained at limited cost. However, the rating may well be higher when the privacy/autonomy dimension is taken into account, as we will see shortly.

Let us examine the three interest dimensions more closely.

Physical integrity. This embraces concerns about health, safety, and avoidance of physical pain. An intrusion into physical integrity can, depending upon its extent, affect any of the living standard gradations: from a homicide that destroys subsistence (living standard level 1°) to a (p.206) jostling causing only momentary discomfort that scarcely disturbs the quality of one's life.

Material support and amenity. This embraces someone's material concerns. These range from the most basic ones needed for subsistence (eg, food, drink and a minimum of shelter); to the various material amenities needed for a life of tolerable comfort; to various luxuries. An intrusion into a person's material interests may thus range from those that are very serious (affecting living standard level 1°) to those that are quite trivial and do not affect the living standard at all.

Privacy/freedom from humiliation. Privacy is important to wellbeing not only because it helps promote self-respect, but also because it helps the person pursue preferences of various kinds. It is affected by various offences, from burglary to illegal wiretapping. They are concerns that are culture specific to a degree: in our culture they affect wellbeing considerably, but another social setting with different moral assumptions may give privacy less weight.

The other interest included in this dimension concerns avoidance of certain kinds of disrespectful treatment.g We considered various broad terms for this interest, such as ‘self-respect’ or ‘dignity’. These terms, however, seemed overbroad - as one's self-respect and dignity depend on one's own self-conceptions as well as on what harm befalls one. The narrower term - ‘freedom from humiliation’ - seems more apt, because it refers to those setbacks to self-respect that derive from degrading treatment. That this is a dimension of quality of life should be evident: one is worse off when the intrusion of one's interests also involve being treated in a degrading fashion. This interest is affected by a variety of criminal acts, especially physical assaults.

The two elements of this dimension - privacy and freedom from humiliation - are conceptually distinct. However, there is in practice enough overlap between these concerns to warrant their being treated as a single interest dimension. If one's apartment is burglarized and ransacked, one suffers both from an invasion of privacy and from having been treated in a degrading fashion.

A corollary of this discussion is that some of the interest dimensions do not run through the entire living standard scale. Physical integrity relates to all four living standard levels, depending on the degree of the intrusion. (p.207) Material support and amenity does also. However, the third dimension (privacy/freedom from humiliation) relates to concerns that arise only at living standard level 2° (minimum wellbeing) or higher. The reason is that level 1o (subsistence) concerns merely ‘getting by’, and no satisfactions are assumed at that level.

4.5 Replacement

What of replaceable items? Something may be important to the quality of one's life and yet readily replaced. Consider the theft of an umbrella in rainy Bergen, Norway. Isn't that serious? After all, the victim will constantly be drenched, cold, and ill without it. Obviously not: the person can obtain a new umbrella cheaply enough so that it makes little difference, ordinarily, to the quality of his existence.

The principle can be stated generally, as follows. Suppose an offence damages or destroys an interest, X. Suppose that X can be replaced for a certain replacement cost, Y. If, in the standard case, making the outlay Y would have significantly less impact on the living standard than the loss of the item X itself, then the harm should be measured by the impact on the living standard of incurring the cost, Y.

In dealing with replacement in this fashion, however, we should keep the generic interest dimensions in mind: an item may be readily replaceable from the perspective of one dimension but not another. Consider our case of burglary and the theft of a television. From a material amenity perspective, the loss is readily replaced. However, the intrusion into privacy is not something which is replaced or made whole by the acquisition of an ersatz TV.

Is this analysis class-biased? After all, some impoverished persons will have great difficulty affording a replacement that most people easily could afford. We are, however, speaking of the hypothetical standard case - so that replaceability should be judged by how affordable the replacement item ordinarily is. One can expect deviations from that norm in both directions: persons who can easily pay even for the most costly replacements, and others who for whom even the most inexpensive could represent a financial sacrifice. These are cases of special resiliency or vulnerability, a matter which principles of aggravation and mitigation may need to address.

4.6 Psychological harm

We have not included ‘psychological harm’ as a distinct dimension in our analysis. The omission is intentional, because the term is too much of a catchall to be of help.

(p.208) Emotions, like beliefs and attitudes, are capable of being supported by valid reasons, or not. When I lose a tennis match I may be furious at my opponent, but have no reason for being so. However, if I lose because the opponent has cheated, then my anger becomes warranted: I am entitled to resent the injury.

A variety of emotional states justifiably flow from being criminally victimized. Attacks on one's safety (quite properly) elicit fear; debasements elicit shame and a sense of humiliation, and so forth. Such distress, however, can be considered part of the intrusion into the interest dimension involved. In assessing conduct that involves threat to personal safety, for example, the fear justifiably elicited as well as the actual injury would be considered in assessing the living standard impact. When examining conduct that involves humiliating treatment - as several of the offences discussed below do - the sense of abasement is integral to what reduces the living standard. In this fashion, we can analyse the various forms of warranted distress that various intrusions produce.h

Other emotional states may flow from criminal victimization, but with less good reason. A routine burglary may make someone fear serious physical harm, without regard to the conduct's actual effects or risks. Some crimes may specially cause anxiety because of aversions felt against those thought to be typical offenders - for example, against their assumed ethnicity or lifestyle. Such responses would be disregarded in our analysis, as they cannot be were all forms of psychological impact lumped together as ‘psychological harm.’

4.7 Some illustrations

Having outlined the elements of a living standard analysis, it might be illuminating to try it out on some hypothetical cases. Here then are some brief offence descriptions. The acts described are meant to be typical instances of various genera of robberies, burglaries, thefts, etc. (The descriptions are not necessarily co-extensive with statutory offence descriptions.)i To control for culpability, we shall assume for the moment that each crime is purposely committed, and that the offender is aware of its usual (p.209) consequences. That way, the differences in ratings will be due to the degree of harmfulness of the conduct involved.

Homicide Our standard case is: A (unlawfully and purposely) kills B. The analysis is simple enough. The conduct destroys subsistence, which is rated as 1°. Hence the conduct has the highest harm rating.

Common assaults The assumed standard case is: A beats up B, but produces no lasting injury. Suppose the beating is quite painful, and results in substantial bruises and some lacerations - but not enough to require hospitalization.

Consider, first, the interest dimension of physical integrity. Let us suppose that the harm involved is no worse than accidentally walking into a solid glass door. It hurts badly when it happens, one has a black eye and a headache for a few days, and then it is over. How serious is that?

  • Is the living standard level involved level 1° (subsistence)? Obviously not. There is no loss of functioning. It is merely uncomfortable, and comfort, physical or material, is not presupposed at this level.

  • Is it level 2° (minimal wellbeing), or 3° (merely ‘adequate’ well-being)? Again, no. We hardly think that the person who walks into a glass door and gets a black eye has his overall living standard reduced to or below such levels of adequacy.

  • Is it level 4° (standard wellbeing)? For a certain period, yes. One's quality of existence has definitely gone down while one is still sore and black-eyed. But the period involved may be too brief to qualify from the middle-term time perspective of which we are speaking.

Let us, however, consider another interest dimension: that of freedom from humiliation. Here, the intrusion into the living standard is potentially more significant: being beaten up is demeaning. This link between being beaten and being humiliated probably holds in most cultures, and may have deeper roots: even a dog cringes if it is beaten and cannot defend itself. In our culture, at any rate, a beating is ordinarily deeply humiliating. What makes it so is not just the physical intrusion and pain, but the being put at someone else's mercy. The person beaten is literally abased -knocked down, abused - and the beater establishes direct physical dominion over him.

So the question becomes - at which level is this deprivation? It is definitely not as high as level 1°, because subsistence does not presuppose self-respect, in our taxonomy. Is the intrusion at least at 3°? We would say so. A reasonable degree of self-respect is surely a part of this level. And the humiliation of a beating strikes us as grave enough to compromise self-respect at the ‘adequate’ level.

Is the intrusion still more serious, that is, at 2°? We would think not. There are humiliations that are so grave as to compromise even a minimal (p.210) level of wellbeing - as we will see when we discuss rape. But we remain unconvinced, for the moment, that the humiliation of such a beating, substantial though it is, would be so devastating to a normal person's self-respect, that it would place the quality of his life at minimal wellbeing, that is, barely above the subsistence level.

Petty assaults Consider the case: A slaps B's face. Here, the material discomfort is trivial. So the question is again one of humiliation. Having one's face slapped is humiliating. But one is not being made helpless, as in the case of beating. One can extricate oneself with dignity - remonstrate, move away, call the authorities, etc. We thus doubt it qualifies as involving level 3°. A reasonable person would not regard himself as having suffered a major deprivation, sufficient to reduce his whole quality of life for an appreciable period to below the ‘adequate’ level. Conceivably, the humiliation involves level 4°, on grounds that there has been some diminution of overall wellbeing. However, even this conclusion is debatable, given a middle-term time perspective.

Armed robbery The assumed standard case is: A robs B at knifepoint, takes his wallet and the small amount of cash in it. The property loss here is quite minor - ordinarily, too small to have more than marginal impact on the living standard. The major element is the threatened bodily intrusion: namely, the threat to kill or gravely to injure someone. This is a threatened intrusion into a level 1° interest: subsistence.j However, we need a discount to reflect that the intrusion is merely threatened and contingent. We therefore say there is a discounted 1°. How such discounts might be made will be discussed later.

Forcible rape The standard case is the classic stranger rape: A rapes B at gunpoint or knifepoint. One interest dimension involved is bodily safety. Here, the analysis would be similar to the preceding case: since a threat to survival is involved, this is a discounted 1°. The other interest dimension is that of freedom from humiliation. We hardly need belabour that forced sex is about the most demeaning imposition that can be imagined - far more humiliating than a beating. Because of the strength, in our culture, of the norm that sexual favours may be granted only with consent, forced sexual intercourse is an extreme form of being subjected to another's dominion. Thus we would rate this at the highest level of intrusions to self-respect, namely, at 2°. This analysis explains why we consider rape at knifepoint worse than the just-cited armed robbery. It supplements the discounted 1o for menace to bodily safety, with the actual 2° for humiliation.

Date rape’ A typical case might run something like this. A female student induced to drink to excess at a male student club and then, with (p.211) her capacity of physical resistance reduced, is made to have sexual intercourse without her consent. The difference between this and the preceding case is that the threat to bodily safety is eliminated: the victim is not threatened with serious injury (although, that conclusion may change, with increasing prevalence of AIDS). However, the interest in sexual integrity is intruded upon much as in the preceding case - with its attendant grave humiliation. For reasons explained already, therefore, this would seem to involve a 2°.

Burglary with ransacking The scenario is all too familiar: a burglar enters a home and, after stealing some of the contents, defaces the walls and pictures, smashes furniture and china, etc.

Let us first consider the interest dimension of material support and amenity. In its extreme forms, destruction of a person's living space can be very serious. If the person is left literally homeless, and cannot afford alternative accommodation, he joins the ranks of the homeless who -at least in our social environments and climates - live at the subsistence level.

The kind of intrusion involved here, however, is less drastic. A ransacking is, in its material consequences, about comparable to having a kitchen fire in one's flat - in which smoke damage and firemen's efforts result in substantial damage to furnishings, clothing, etc. How much are one's interests set back by such an event?

  • Certainly, one is not reduced to the subsistence level (1°). One is not rendered homeless, and still has warmth and shelter.

  • One also is not reduced to the minimum wellbeing level (2°). By obtaining some temporary replacement furnishings and a few items of replacement clothing, one can continue to live in minimal comfort.

  • Would the intrusion involve ‘adequate’ wellbeing (ie, level 3°)? A home furnished with basic amenities does appear to be part of our conception of adequate wellbeing. Here, however, the issue of replacement costs arises. The standard person may well be able to afford to replace those basic amenities without a major impact on his standard of living.

  • In any event, the intrusion would involve level 4° - reduction in a full life quality for a significant period.

Let us turn, then, to the interest dimension of privacy. An important part of privacy is having a home of one's own, in which one may spend a substantial portion of leisure time and which may be the focus of personal, family, and social life. The home may be arranged to suit the taste of the occupant, and may be entered only by invitees. Any residential burglary, therefore, intrudes on this interest.

(p.212) Being deprived altogether of a home of one's own would be a very serious intrusion. Suppose someone were involuntarily removed from his home to collective living quarters with strangers. Even if board, shelter and amenities were adequate, the loss of privacy and autonomy would seem a large enough setback to wellbeing to affect level 2° - a minimal living standard. (Being institutionalized has this quality.) Ransacking of a home in a burglary is not so great an intrusion, for one still keeps a home of one's own. However, the forced entry of a stranger - coupled with the damage to the occupant's chosen physical environment - would seem to compromise an adequate living standard. Thus, from the privacy/autonomy standpoint, the living standard level involved probably should be rated as 3°.

Simple residential burglary Here, the dwelling is entered, and a common item of property is removed. No ransacking occurs.

  1. (a) Material amenity The intrusion on material comfort is relatively minor. A readily replaceable item of property is lost. Otherwise, the liveability of the dwelling is not affected. In material terms the living standard would scarcely be affected.

  2. (b) Privacy/autonomy Here, the entry constitutes a significant intrusion into privacy. However, the occupant's preferred arrangements of his/her home are not significantly disturbed. Our judgement is that this would involve level 4°.

Car theft In car theft, only material amenity is at issue. The car is not a zone of privacy akin to the home, ordinarily. For many people, use of a car is important to leisure or even to livelihood. However, cars are insurable against theft, and a stolen car can be replaced by a functioning vehicle. Provided that theft insurance premiums and/or replacement costs are modest, there would be shor-term inconvenience but the standard impact on a person's wellbeing would be limited.k

4.8 Constructing a harm scale

Constructing and applying the living standard ratings helps, but does not suffice to grade harms. First, for reasons to be explained shortly, there need to be more (at least, two more) harm gradations than the four living standard levels. Second, discounts need to be made for (p.213) risked or threatened harm. A robbery at knifepoint, for example, risks subsistence, so we described it as a ‘discounted’ 1°. But it makes little sense to discount directly onto the living standard scale: if one reduced the armed robbery rating to 2°, for example, that would not be saying the offence risks subsistence, but rather that it actually affects the living standard level of minimal wellbeing - which may well not be an accurate characterization. Third, an offence may give rise to multiple living standard ratings in different dimensions. A routine assault, as just noted, has a relatively low rating in the physical integrity dimension, but a higher one in the freedom from humiliation dimension. We thus need a way of combining these diverse assessments into a net harmfulness grade.

Creating the scale. A variety of techniques for constructing the penalty scale might be imagined. Let us suggest a relatively simple structure, for the sake of illustration.

The harm scale would have six gradations. The highest gradation should be for crimes that do not just intrude upon, but terminate the living standard: when the victim is killed. The lowest gradation should be for crimes that do not significantly affect the living standard at all. The simplest harm scale thus would have six gradations, as follows. Each gradation could be set forth as a band, so as to allow differentiations within the band. The gradations would be the following:

Harm gradation

Living standard level affected

I-very grave

Living standard terminates (death)


Subsistence (living standard level 1°)


Minimal wellbeing (level 2°)


‘Adequate’ wellbeing (level 3°)


Standard wellbeing (level 4°)


Living standard not affected or only marginally so

Notice this mapping rule, elementary as it is, involves value judgements, for we are characterizing the gravity of various levels of intrusions into the living standard. We are, for example, asserting that injuries affecting subsistence involve grave harms, etc. However, these judgements seem straightforward enough (given our preceding discussion) to require no further elaboration.

In this scale, within-category variations can be denoted simply by the location within a given band: no numbers are necessary. For example, the difference between serious wounding and mayhem (permanent disfigurement) - both of which affect subsistence, but to different degrees - by (p.214) locating both crimes within the ‘II-grave’ band, but placing the latter crime above the former within the band. Alternatively, one might be more ambitious - for example, by dividing each of these six categories into an upper and lower range, and thus having 12 gradations of harm.

Discounts for threatened or risked harms. Many crimes only create a threat or risk to an interest, so what matters is not only the importance of the interest but the degree to which it is threatened. The more remote the risk, the greater should be the discount - and hence, the lower the harm rating.

Consider these four offences: aggravated assault, attempted aggravated assault,l armed robbery, and drink-driving. All affect the interest in subsistence, but in different degrees. An aggravated assault directly intrudes upon physical integrity; an attempt involves a significant risk to that interest; robbery threatens that interest, but the threat is contingent (the robber might use the gun only if the victim resists). In drink-driving, the risk is more remote, depending on how much alcohol the offender has consumed and how crowded the highways are.

To deal with these situations, a two-step process seems appropriate. The first is to make a living standard valuation of the completed harm. In aggravated assault, we have done that already: the harm directly affects subsistence and hence is in the ‘grave’ range. The second step is to make an appropriate discount for threat or risk.16 We might, for example, treat certain attempts as constituting a sufficiently high risk to keep the conduct in the ‘grave’ range - albeit at a point in that range below the completed harm. Next, we give the threat that is present in armed robbery a somewhat larger discount: perhaps, one would place that offence in the next (‘serious’) harm category - because (while no assault actually takes place) the risk of grievous injury occurring is still high. As we descend to more contingent risks to physical safety (eg, those typically involved in drink-driving), the discount would be larger, so that the conduct would be placed in a lower harm category.

Combinations. The combination problem arises when an offence (eg, assault, as we just saw) affects two or more interest dimensions. Here, a first step could be to identify the dimension that yields the highest harm rating, and identify that as the ‘primary harm’. Suppose that offence X affects both physical integrity and self-respect, but in different degrees. Suppose the physical integrity intrusion is a substantial threat to subsistence, that is, a discounted 1°. We might, according to the discount (p.215) principles just outlined, rate this harm in the third (‘serious’) category. Suppose the intrusion on self-respect involves an actual (not just risked) humiliation - but only such as to affect living standard level 3° - ‘adequate’ wellbeing. This, according to our just-described mapping principles, would qualify the intrusion at the fourth harm category, of ‘upper-intermediate’. On this analysis, the physical integrity intrusion would have the higher rating, and thus constitute the primary harm.

The next step would be to consider the secondary harms - those in other interest dimensions - as possible exacerbating features. It makes offence X somewhat worse that - besides the primary harm to physical integrity - it also humiliates. How much the exacerbation is depends on how high a rating the secondary harm has. The higher that rating, the more weight it should be given as an exacerbating factor.17


Could our proposed analysis be helpful in actual sentencing law? Our answer is that it could be helpful, provided that a number of other steps are also taken. One step is completion of the tasks discussed here. We have only begun, by describing the living standard analysis. The additional tasks, of constructing a harm scale and devising rules for discounts for risk and for combinations, have only been sketched here. A second condition is that culpability be dealt with. Ultimately, we are seeking a jurisprudence of crime seriousness, which calls not only for an assessment of the conduct's degree of harmfulness but also of the actor's degree of culpability. Failing a full theory for gauging culpability, we at least need a way of controlling for the more obvious differences in culpability among crimes, so that we can assess how much such crimes' gravity is affected by variations in their harmfulness (see below).

Suppose, however, that these conditions have been satisfied. In what way could our approach be useful in making sentencing decisions? The answer to this question will depend on the sentencing structure of the jurisdiction. Let us take a look at two structures: sentencing guidelines such as Minnesota's, and statutory sentencing principles such as those embodied in Sweden's law. Both schemes emphasize the gravity of the offence, but use different decision-makers.

Numerical sentencing guidelines such as Minnesota's rely upon a special rule-making agency (the sentencing commission) to devise the sentencing guidelines.18 The commission's guidelines are embodied in a sentencing grid. The vertical axis of the grid constitutes a crime-seriousness score; the horizontal, a criminal history score. The cells within the grid set forth the presumptive sentences or sentence ranges, which the courts are expected to (p.216) apply except where aggravating or mitigating circumstances are found present.

Such a scheme requires the sentencing commission to take each statutory offence category and assign it a numerical seriousness grade, say from ‘1’ (least serious) to ‘10’ (most serious). Where the offence category is broad and embraces heterogeneous conduct (eg, a category such as ‘burglary’) the commission may establish subcategories and assign these different seriousness grades.

To date, when sentencing commissions undertook this rating task, they relied largely on their members' intuitions. In Minnesota, for example, each commission member was given a list of offences and asked to assign each offence a seriousness rating. When the members agreed, this rating was adopted; when they disagreed, discussion ensued until agreement was reached. This meant that no identifiable theory supported the seriousness judgements; indeed, when there was agreement, no reasons were elicited at all.

Here, a theory for gauging harm would have obvious utility, in providing commission members with a method of reasoning. A simple procedure might be as follows:

  1. (1) The commission would rank the various offence categories in seriousness, using the members' sense of these offences' gravity - as Minnesota did.

  2. (2) The commission would select, from the list of offence categories, a shorter list comprising intentional, victimizing crimes: victimizing, because that is what the present analysis is about; intentional, to try in a simple fashion to control for variations in culpability.

  3. (3) To this subset of offences, the members of the commission or its staff would be invited to apply the analysis we have suggested in this Appendix: make a living standard analysis of the harms involved, and actually rank the harmfulness on a simple scale.

  4. (4) With this harm analysis in mind, the members would be asked to re-rank the offence categories in seriousness. They also would be asked whether the analysis suggests a new or altered sub-categorization: to distinguish, for example, between ordinary residential burglary and burglary with ransacking, on grounds that those acts' impact on the living standard are different. Where disagreements arose, members also would be asked to argue their ratings in terms of this theory - that is, to try and identify whether and how their disagreements are based on divergences of view about the conduct's effect on the living standard. The commission could thus determine whether the analysis helps resolve disagreements, and whether and to what extent their initial (intuitive) ratings require change as a result of the analysis.

(p.217) A procedure of this kind would use our theory the way it should be used: as a guide for judgement, not a formula. Ultimately, the commission would still have to rely upon its best judgement in making the ratings; only now, that judgement would be guided by identifiable, coherent principles.

The theory, of course, would aid the commission only in its first seriousness assessment task: rating the gravity of categories or subcategories of crime. The commission would also need to address aggravation or mitigation: that is, specify what types of special circumstances affect the harmfulness of the conduct or the culpability of the actor, and thus justify a departure from guideline ranges.

The Swedish law takes a somewhat different tack. There is no sentencing commission, and no numerical seriousness score. What the statute prescribes is that the choice of sentence should depend principally on the ‘penal value’ (ie, seriousness) of the conduct, and that penal value ordinarily depends on the harm or risk of harm of the conduct and the culpability of the actor.19

Swedish sentencing doctrine distinguishes between ‘abstract’ and ‘concrete’ penal value. The abstract penal value reflects the general harmfulness and culpability of the type or subtype of offence involved. Against the background of the abstract penal value, the concrete penal value of the particular case is then considered - taking into account the particular injuriousness and culpability of the actor's conduct. To aid in the determination of the concrete penal value, the statute sets out a list of aggravating and mitigating factors that increase or diminish the gravity of the conduct in the particular case.20

In assessing the abstract penal value, the courts are supposed to consult the statutory penalty scales, that is, the maximum and minimum penalties for various offences prescribed in the Swedish Penal Code. Parliament is expected, in setting penalty scales, to consider the gravity of the conduct compared to that of other crimes. This means that two kinds of agencies are involved in the assessment of seriousness: the Parliament and the courts.

This makes the procedure seem more cumbersome than that of a sentencing commission. How can the legislature and the courts, with their large, diverse memberships and numerous other duties, possibly apply a theory of harm such as ours? Two features of Swedish practice, however, may help facilitate matters. First, major law changes - such as those involved in altering penalty scales - are seldom undertaken at Parliament's sole initiative. Ordinarily, a government-appointed study commission, aided by an advisory panel of experts, drafts the legislation. Such a commission or its advisory panel could deliberate on harm in much the same manner as a sentencing commission. Second, sentencing doctrine (as other major legal doctrines) is developed by the Supreme (p.218) Court and the six regional Courts of Appeal. These are more compact bodies, capable of developing doctrines of criminal harm based on the general idea of the living standard, or of utilizing scholarly advice on such doctrines.

Let us emphasize, however: the present discussion is not a recipe for how a sentencing commission, or a Swedish-style panel of experts, should proceed. We are only suggesting that our theory can have practical utility. Our aim has been to suggest a way to think about victimizing criminal harms. The practical applications come later.

Although there remain important topics to discuss (eg, relevant empirical research),m we will conclude our discussion here. We have put forward a general approach - one that would gauge criminal harm through the notion of the living standard. What we are proposing is a way of thinking about criminal harm, not a formula. The living standard categories we sketch are far from precise, and their application to criminal harms leave much to judgement. However, we do supply a systematic conceptual framework for making such judgements - and believe that superior to untutored intuition and guesswork.

(p.219) Notes


(1.) See further Asp & von Hirsch 1999, 171–3.

(2.) Feinberg 1984, 190–3.

(a) In crimes of intent, the defining elements of the crime must actually be known to the actor. What, however, of the further harmful consequences? How much, for example, need the defendant have known of the humiliating character of an assault in order for humiliation to be considered in gauging a crime' seriousness? A full answer would require an excursus into culpability theory, which we will not undertake here. However, it should be clear that those ulterior consequences should at least be foreseeable: unforeseeable consequences have no place in the assessment of the gravity of the conduct.

(3.) Sellin and Wolfgang 1964.

(4.) Sen 1987.

(5.) Feinberg 1984.

(6.) Von Hirsch 1985, 66–74.

(7.) Feinberg 1984, 42.

(8.) von Hirsch 1986, 702–706.

(9.) Feinberg 1984, 57–8.

(10.) Sen 1987.

(11.) Erikson and åberg 1987, ch. 1.

(12.) Sen 1987, 30–1, 36–7.

(b) The distinction between (1) wellbeing understood as possessing certain means and capabilities that permit a degree of freedom of choice, and (2) freedom of choice in itself, is usefully elucidated in Amartya Sen's 1984 Dewey Lectures (Sen 1984, 206–8). He supplies the example of the person eating his sandwich on the riverbank, who sees a man drowning in the river. The event adds to the person's choices–because now he has the extra option of saving the victim or not – but those do not add (and may actually detract) from his wellbeing. As Sen notes: ‘The additional opportunity of saving the drowning person did not give you a better (or even as good a) way of pursuing your own wellbeing. You were no longer free to eat your sandwich without anxiety, and … there was a genuine loss of opportunity to pursue your own wellbeing.’ (at 207).

(13.) Sen 1987, 27.

(c) Harm consists of being made worse off. One can gauge how much one is harmed (ie how much worse off one is made), by the impact of the conduct on the quality of one's life. Other evils, however, do not necessarily involve harm as part of their defining characteristics. A possible example is offence. According to the view of one of the authors (von Hirsch), being treated offensively does not necessarily involve being made worse off in the sense of having one's personal resources diminished; its evil resides simply in being dealt with without consideration or respect. To gauge the gravity of the offensive conduct, one thus may require a standard other than one that refers to reduction in the offended person's living standard. For an analysis of offence, and how it differs from harm, see Simester and von Hirsch 2002.

(14.) Braybrooke 1987, 48.

(15.) Id., 49.

(d) A philosopher might ask whether such norms are rooted in the ethical traditions of our own culture, or have more universal applicability. We do not have to address the question, however, as we are not at the moment attempting to apply our living standard criteria in other cultural settings. Our own view would be that respect for other person's interests is essential to any valid system of ethics, but that particular moral principles through which that respect is expressed – eg principles regarding privacy – may legitimately reflect a culture's ethical traditions. For a somewhat comparable view, see Hampshire 1989, ch. 2.

(e) Of course, the very worst intrusions would remain significant even from such a long-term perspective – a serious assault being an example.

(f) Are the interest dimensions themselves interests? That one would depend on how broadly one defines ‘interest’. Were interests defined as resources, then some of the listed dimensions – for example, physical integrity and material support and amenity would be interests; but other dimensions, such as privacy, would not necessarily be – as one might have privacy concerns that would extend beyond one's resources; see Simester and von Hirsch 2002. In that event, however, privacy would still be helpful as a way of evaluating the harmfulness of intrusions into one's resources. One's living space is a resource, for example. When that resource is intruded upon through a burglary, one of the harmful aspects of that intrusion is the invasion of privacy involved. That privacy intrusion can be treated as an interest dimension in evaluating the harm, even if one does not consider privacy in itself to be an interest.

(g) This dimension creates a degree of overlap between harm and culpability. Humiliation normally presupposes intent (or, at least, apparent intent). I am being humiliated by being beaten – but not normally, by being injured through someone's criminally negligent handling of a vehicle.

(h) A comparable issue is that of offensive conduct. Should its offensiveness be assessed by the degree of subjective distress the conduct causes, or by how much the conduct violates norms of respectful and considerate treatment of others? One of the present authors would favour the latter view; see Simester and von Hirsch 2002.

(i) Many jurisdictions have broad statutory offence descriptions. The American Law Institute's Model Penal Code (Art. 221), for example, defines burglary as any entry into a building with intent to commit a crime; and has special provisions only for nighttime residential burglary. Such broad definitions are insufficient for rating the harmfulness of the conduct, as they disregard critical factors typically affecting the living standard impact – in burglary, the amount of vandalization, if any, wrought to the burglarized premises. Hence, in the above hypothetical cases, we may consider separate subspecies of a statutory crime category, viz.: residential burglary with ransacking, and ordinary residential burglary.

(j) There may also be an element of humiliation in robbery, but for the sake of simplicity we do not address this.

(k) Thefts become more difficult to analyse, however, when they involve sums of money, rather than standard possessions such as TVs or cars. The difficulty relates to the fungibility of money, and the uneven distribution of financial resources. How a theft of x Pounds Sterling affects someone's living standard depends critically on what other financial resources he is assumed to have. Whether our usual suggested techniques suffices – of judging the impact by a hypothetical standard case, and dealing with deviating cases through aggravation/mitigation – is a question deserving of further discussion.

(l) We are assuming here that attempts should be treated as less serious than completed crimes, see Duff 1996; but see Ashworth 1993.

(16.) See also Robinson 1987b, 44–45.

(17.) These steps have been outlined more fully in von Hirsch and Jareborg 1991, 28–32.

(18.) For an analysis of these guideline's aims and structure, see von Hirsch, Knapp and Tonry 1987, ch. 5; von Hirsch 1995.

(19.) Swedish Penal Code, ch. 29, s. 1.

(20.) Swedish Penal Code, ch. 29, ss. 2 and 3.

(m) To date, two kinds of empirical research are available, neither especially helpful. One is survey research on public perceptions of crime seriousness, such as Sellin and Wolfgang's (1964). The limitations of such research have been suggested already; see 2.3 of this Appendix. The other kind of research has been living standard surveys such as Stockholm's (Erikson and åberg 1987). The researchers construct an index that measures area residents' living standard in terms of various conditions assumed to bear on their wellbeing. Housing conditions, for example, are assessed in terms of number of rooms of the person's residence relative to family size. While such surveys are helpful for their intended purpose, of measuring changes in a population's living conditions over time, the measures used (eg, housing space) were based on researcher's common-sense judgements about the amenities important to a good life, rather than on any general theoretical perspective of how such amenities relate to levels of wellbeing.

  • Two kinds of research, however, might prove more profitable. One was proposed by the late Richard F. Sparks: empirical study of the type of injury resulting from various offences; see von Hirsch 1985, 65–6. Sparks proposed that the assessment of the harm element should rely not on opinion surveys such as Sellin's and Wolfgang's, but on empirical enquiry into the type of injury actually involved in various types of offences. (Traditional victimization studies have been more concerned with measuring frequency and distribution of criminal acts, than with detailing injurious consequences.) Sparks suggested kind of research would inquire systematically into the short- and long-term consequences of various criminal acts. It could provide a better factual basis for judgements about various crime's typical consequences – which could then be used when applying our suggested living standard rating categories.

  • Another species of research would be simulated application of our proposed rating criteria. The researcher might, for example, organize a small group as a simulated sentencing commission, and ask it to rate various crimes. The first step might be to ask the members to rate crime's gravity intuitively, in the manner that Minnesota did. Next, the members would be asked to rate crimes using the reasoning process suggested in this Appendix. The researcher would identify the extent to which this process produced ratings differing from the original ‘intuitive’ ones. The group would then be asked to what extent it wished to abide by those original ratings, and to what extent it preferred the revised, theory-based ratings – and why. Such research might suggest how helpful the proposed rating living standard approach is in decision-making situations.