Violence in Proportion is a very good, very densely argued book about the proportionality condition on the justified infliction of harm. More specifically, it is a book about what Tomlin calls ‘preventive limiting proportionality’: the version of the condition that limits the permissible magnitude of harms inflicted in service of preventing wrongs – in other words, the kind of proportionality condition most often employed by ethicists of defense and war. In its simplest form, this condition requires that permissible harms not be disproportionate to the goods they achieve. But specifying and applying this view can often be, even in relatively simple cases, a complex and difficult task. One of Tomlin's principal aims (and accomplishments) in this book is to lead the reader to ‘appreciate the complexity of proportionality’ and to show that ‘the number of ways in which we can develop this seemingly simple idea is dizzying’ (p. 3). It is important to say at the outset that this book does not build up and defend a unified account of proportionality. It is, rather, an ‘exploration’ of a ‘tangled web of philosophically knotty problems’ (p. 2). In practice, this means that the book discusses a series of puzzles about how to specify preventive limiting proportionality in various circumstances, the solutions to which are ‘by and large … independent of one another’ (p. 3). (Tomlin says: ‘My own account of proportionality is built up of the different positions I take on these issues, but it does not come as a take-it-or-leave-it offer’; p. 3.) The book is thus best read as an invitation to think, alongside Tomlin, carefully through the puzzles he raises. It is also an invitation to philosophers of war and defense to think more deeply about what is (I think it is fair to say) the least thoroughly studied of the three canonical conditions on permissible defensive harm. It is also worth noting at the outset that Tomlin says little – excepting an excellent first chapter about the conceptual terrain of proportionality and a later chapter on aggregating harms that argues for the priority of proportionality over necessity – about what proportionality basically is, or about the fundamental relationship between the basic principles governing defensive harm. Insofar as the book reaches conclusions bearing on these questions (which it sometimes does – more on which below), it is because they are consequences of Tomlin's solutions to his puzzles, not because they are themselves direct objects of investigation. This is not that kind of inquiry. Whether that is a virtue or a vice will depend on the reader's own philosophical dispositions. What, then, are the puzzles? There are (broadly) four. The first puzzle concerns the counterfactual baseline we should use when assessing what defensive harms achieve – since, recall, preventive limiting proportionality compares the harms defensively imposed to the goods of prevention thereby achieved, and what goods count as achieved is relative to what one might otherwise have done but did not. The second puzzle concerns disparities between judgments of the proportionality of acts versus the proportionality of courses of conduct. For example, could a war whose battles are all disproportionate be itself proportionate? The third puzzle concerns uncertainty, since we can ask questions about proportionality from a subjective or objective point of view – and since we can later gain evidence that might affect how or whether we continue what we previously believed to be a proportionate course of conduct. The fourth puzzle concerns the aggregation of harms: could enough headaches inflicted upon innocents make an act (‘widely’) disproportionate? Could (‘narrow’) proportionality license killing an unlimited number of individually liable threats? I will not reveal Tomlin's answers to these questions, since the beauty of this work is at least as much in how he arrives at them. As this list reveals, some of the knottiness of the puzzles Tomlin presents is attributable to their interaction with background issues of notorious knottiness in moral theory generally: counterfactual evaluation, uncertainty, and aggregation. (Doing versus allowing also takes a central role in some chapters; the non-identity problem makes an appearance in another.) But this book does more than bring a series of well-known, difficult puzzles to bear on a single subject matter. Tomlin makes a convincing case that these classic puzzles give rise to new puzzles of distinctive knottiness and practical importance for questions about proportionate preventive harm itself, which is to say regarding whom we may harm, to what extent, in service of what ends. (These puzzles are indeed new ones; only two of the book's 11 chapters are based on previously published material.). Since the book does not advance a single overall argument, nor develop a unified account of proportionality, summarizing its argument(s) is a more involved task than can be usefully accomplished in a review. Instead, let me discuss one of the puzzles and solutions the book presents in order to give the reader a flavor of the kinds of views developed and defended within. In the chapters on counterfactual baselines, Tomlin's claims are that (1) we do indeed need to engage in counterfactual analysis in order to make proportionality judgments (contra notable claims to the contrary in the literature); (2) the baseline of comparison required must be a moralized one (rather than, say, the simple non-moralized baseline of non-action); and (3) all moralized baselines face objections, so we must choose the least objectionable. We then set out to find this baseline. A baseline is the option against which defensive achievements are to be measured and defensive harms to be weighed. This matters because if a defensive option does not achieve enough, the harm wielded will be disproportionate and so impermissible. In simple cases – e.g. where there are only two options: defend or don't defend – the relevant baseline is set by the option don't defend. Only if defend achieves enough good compared to don't defend, and compared to the harm imposed by defend, is defense proportionate. But in even slightly more complex cases – when there are multiple ways of not defending, when defense is costly to the victim, or when harms to third parties are involved – it can become tricky to settle on the correct baseline. The available option which does not cost the defensive agent (or involves only trivial costs), does not harm anyone, and leaves the victim as well off as is permissible. (p. 80) Egghead: Attacker is about to punch Victim. Victim has a fragile skull on one side, such that if she were punched there, she would die. She must choose whether to turn her head to the fragile or the robust side before the punch lands. She can also choose whether to avert the attack by killing Attacker, or to permit the attack by suffering a punch. (cf. p. 86) Now consider Tomlin's baseline conditions. The world in which Victim turns her head to the robust side is the one that leaves Victim best off. Both options – turn the head this way or that – are antecedently morally permissible, involve only trivial costs to Victim, and harm no one else. It is thus that baseline against which the proportionality of Victim's response is to be weighed, according to Tomlin's view. Since killing Attacker would be clearly disproportionate to suffering a punch, proportionality prohibits Victim from lethally defending herself. This is the correct result. Beneficence: An extravagantly rich Ruler has (permissibly and irrevocably) set aside great riches for Subject, to be given over at the stroke of midnight. Ruler has a flair for the dramatic: the riches are perched above lava, and will at midnight, if not for Ruler's intervention, fall in and be destroyed. Midnight arrives, but alas, so does Attacker, who threatens to break Subject's leg. Now, Ruler must choose whether to give Subject the riches and allow Subject's leg to be broken or to prevent her from suffering a broken leg by harming Attacker and allow the riches to be destroyed. Suppose Subject would prefer the riches, at the cost of enduring a broken leg. (This case is significantly condensed from the original; cf. pp. 89–92.) In response, Tomlin correctly points out that the case requires a number of highly artificial conditions to hold in order to yield this counterintuitive verdict: Ruler must be genuinely permitted, but also not required, to make Subject rich; it must be genuinely non-costly to Ruler to part with the riches; doing so must also be genuinely non-harmful to third parties; and Ruler must face an exclusive choice between making Subject rich and defending her leg. In light of all of these conditions, Tomlin is happy to bite the bullet. (Tomlin points out that if we defend Subject's leg, we are ‘choosing to make Subject worse off than we could, and choosing to harm the attacker in order to achieve this’ (p. 90) – and it would indeed be odd, in its own way, if this were permissible.) Making Subject rich is, Tomlin believes, the baseline against which harmful courses of action (like defending her leg) are to be judged. To defend her leg, and so to deprive her of riches, is thus to make Subject worse off than she should be, and so to impose a harm on Attacker that is disproportionate and therefore wrongful. Importantly, the defensive wrong here therefore is suffered by Attacker, who faces the imposition of a disproportionate harm. It follows that Attacker may rightly counter-defend himself against Ruler's intervention – whether or not this results in Subject gaining the riches after all. Tomlin is aware that this may seem like a bitter pill to swallow, but consoles us with the observation that ‘our intuitions are not clearly trustworthy here’ (p. 90). I am not so sure (and Tomlin is of course content to appeal to our intuitions in other highly constrained and artificial cases). But even accepting Tomlin's verdict on the suitably specified case, notice that the result is not just about what is permissible for Ruler to do. It is also the result that the riches are relevant, in the suitably-specified case, to what counts as permissible defense. This strikes me as a more difficult consequence to accept. It may be that faced with the choice between making Subject rich and defending her leg, Ruler should, all things considered, make her rich. But as a matter of defense, surely (I think) Ruler may prevent the attack on Subject's leg. Since Tomlin's view is that Ruler may not, Attacker suffers a wrong and may engage in counter-defense. But this odd result appears only if we do take the matter of the riches to be relevant to the proportionality of the option of defending Victim's leg. A better view, I think, would succeed in rendering the option of the riches irrelevant to the question of permissible defense. Of course, Tomlin's actual view cannot be seriously objected to on the basis of the mere idea of a better one. But this mere idea returns us to the question of method. For a book of this length, relatively little space is devoted to what could be called fundamental questions about the nature of proportionality and the prevention or infliction of harm. Here is an example of what I mean that is pertinent to the concern raised just above. Two chapters, concerning aggregation, take as their goal the acceptance or rejection of various versions of a relevance rule: a principle that tells us how to handle questions of whom to harm and whom to save when the options involve harms (or preventions of harms) of very different magnitudes. To accomplish this, we begin with simple versions of a relevance rule that quickly become untenable when confronted with slightly more complex cases, and so refine the relevance rules. The refined rules continue to be confronted with increasingly complex cases, and become increasingly refined themselves, until we arrive at one – ‘Relevance Rule 4**’, in this case – that seems, in Tomlin's judgment, to best satisfy all of the cases considered. But while we use the term ‘relevance’, and we appeal to intuitions about what should be relevant or irrelevant to choices between options, we never ask the question, in some more general sense, of what relevance is in this domain, what makes something relevant or irrelevant, and why. ‘Relevance’ is an essentially technical term here, furnishing the name for principles that allow or disallow harms of varying magnitudes from counting against others. The method of cases is, of course, an established and indeed distinguished one, which Tomlin employs with great skill. But one cannot help wondering whether more attention paid directly to basic questions might nevertheless have helped – say, to furnish deeper grounds for treating (for instance) the possibility of riches as irrelevant to the matter of proportionate harm prevention. These worries aside, what is worth highlighting is the philosophical skill and creativity with which Tomlin develops and pursues the puzzles that make up the book. He has given philosophers working on the ethics of defense and war a rich new set of issues to take up – and a formidable set of solutions to contend with.
Edmund Tweedy Flanigan (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: