The Journal of Ethics, 2005, 475-99.
© The Journal of Ethics
Feel free to read or link to these essays. But please do not place copies of them on another server.
The moral roads on which we travel are slippery. Our individual and collective actions inevitably affect others, ourselves, and our institutions. They shape the people we become and the kind of world we inhabit. They increase or decrease the likelihood, however slight, that certain futures will occur. Sometimes those consequences are positive, a giant leap for moral mankind. Other times they are detrimental or morally regressive. We should not try to avoid slippery terrain. That is not an option. Rather we should seek to understand and successfully navigate it.
What, then, is the function of slippery slope arguments in moral debate? Do they just point out these obvious facts? No. If that was all they did, then it seems they would be part of every moral assessment. They aren= t. That is because their principal use is to defend the status quo by making us fear change. Change, of course, is sometimes bad. But not inevitably. Change is also the engine of progress, moral and otherwise. What is inevitable is that we, our relationships, and our institutions change. So fearing change is irrational.
That is why I claim that although (a) life is slippery, and (b) we can sometimes develop cogent slippery slope arguments (Walton 1992; Little, Groarke et al. 1989; Govier 1982; Volokh 2003; Mayo 1990-91; Lode 1999; Williams, B. 1995), given their most common usage, (c) it would be prudent to avoid them in moral and political debate. They are often fallacious and have often been used for ill. I recognize that this proposal seems hasty. After all, all argument forms are sometimes offered with false premises and are sometimes used for ill. Yet that doesn= t lead us to jettison modus ponens from our argumentative arsenal. Why slippery slope arguments? Three explanations. One, slippery slope arguments are especially prone to be vague and ill-formed. Two, people are easily swayed by themC more easily than by faulty instances of modus ponens. They sound suggestive even when argumentative details are vague or absent. Three, even when they are cogent, we can always find alternate, usually preferable, arguments that capture their insights without carrying their argumentative baggage. These problem do not plague the use of modus ponens.
I am concerned here with causal slippery slope arguments. I will not address the rich literature on logical slippery slopes. First, I am inclined to think that they are neither as common nor as rhetorically powerful. Second, even if we had a cogent response to all logical slippery slope argumentsC say, by drawing the line (Williams, B. 1995: 219-20; Schauer 1985: 380) or showing that their logical structure is flawed (Shapiro 1998)C people may nonetheless be psychologically or socially or politically or judicially inclined to slide from one side of the conceptual divide to the other. It may be, as Lode expresses it, A humans arguably have a tendency to psychologically [emphases mine] assimilate closely related cases@ even if they are logically distinguishable (1999: 1513). That is why logical versions of the slippery slope argument, even when flawed, may causally move people or institutions. Whether I am right about this, I think causal arguments are sufficiently common, interesting, and important, to focus on them.
The Structure of Slippery Slope Arguments
The philosophical and legal literature is replete with competing, and sometimes wholly incompatible, accounts of slippery slope arguments. Those who regularly use these arguments may employ several of them. This often makes their positions difficult to critique since if one objects to one formulation, they may slide to another. I will not try to canvas them all. Rather I will briefly outline one prominent alternative and then contrast it to my own. Throughout the first section I will explain why I think my account is preferable to alternatives. I will then evaluate the use of slippery slope arguments.
Eugene Volokh recently offered a statement of slippery slope arguments in the Harvard Law Review. Although his account explicitly concerns only social policies, his description captures the nub of all causal slippery slope arguments. A You think A might be a fairly good idea on its own, or at least not a very bad one. But you're afraid that A might eventually lead other legislators, voters, or judges to implement policy B, which you strongly oppose.@ So you oppose A (2003: 1028).
There is clearly something right about his account. Slippery slope arguments do claim that we should reject some proposed behaviors or policies because their likely consequences will be bad. However, his definition is too broad: it describes all negative consequentialist arguments, only some of which are slippery slope arguments. We must isolate what distinguishes slippery slope arguments from other negative consequentialist arguments. Most often that is the mechanism that leads from what he dubs > A= to > B= .
I propose that slippery slope arguments have the following general structure:
Let me say something about each of these elements. First, although there are philosophical disputes about the precise meaning of A prima facie,@ these differences have no bearing on the current project. We need only acknowledge that those advancing slippery slope arguments claim (or grant for purposes of argument) that action x is prima facie permissible in some sense. To use Volokh's language, x must be A a fairly good idea on its own, or at least not a very bad one.@ Therefore if we could stop after taking the first (and perhaps a few additional) step(s), then we would have done nothing (very) wrong. However, given the kinds of creatures we are (psychological version), the nature of institutions we inhabit (political version), the kinds of categories we use (logical version), or the types of laws we employ (legal version), we are unlikely to stop after the first step. That is why the first step (action x) is immoral. It is not immoral in itself; it is immoral because it probably leads to consequences that are. Second, the mechanism of slippery slope arguments is a series of small analogous steps which presumably lead us from an action that is prima facie permissible to one that is not. Finally, all assume that the latter action or circumstances are, in fact, immoral.
Differentiating Slippery Slopes from Related Argument Forms
To understand and evaluate these arguments, we should differentiate them from other argument forms with which they are often confused or conflated. Some of these forms are quite plausible and thereby lead us to think slippery slope arguments are more forceful than they are.
Those Clearly Distinguishable from Slippery Slopes
The first two forms are, I think, clearly not slippery slope arguments, even though they are sometimes confused with the them.
Consistency arguments are schematized as follows:
C1: We should do x for reason R.
C2: Reason R justifies doing y.
C3: Y is immoral. Therefore,
C4: Doing x is not justified by reason R.
Some people treat these as slippery slope arguments (Volokh 2003: 1037). However, since their second premise is critically different, we should distinguish them (Schauer 1985: 363, 369). Slippery slope arguments claim that x leads to y by means of some series of small analogous steps. Consistency arguments claim that the reasons that justify doing x straightforwardly justify doing y. We challenge consistency arguments by demonstrating that x and y are relevantly different, and therefore, that although R will justify doing x, it does not justify doing y. We challenge slippery slope arguments by denying that action x will probably lead to y.
[make part of note] It is interesting that we might be able to construe the logical version of the slippery slope argument as a series of small consistency arguments: once one takes each small step, she is thereby warranted in taking the next step, etc. However, I would say that if each step genuinely does warrant taking the next step, then it is really just a consistency argument, as schematized above. Breaking down the argument into small steps is just a rhetorical device to help others see it. In other cases, though, such arguments are not really consistency arguments: each step does not wholly warrant the next one; rather, people merely assume than if A warrants B, and B warrants C, then C must warrant D. The question, of course, is why does each step warrant the next one. If the mechanism is a small set of analogous steps, then it is a slippery slope argument.[end note]
Arguments from Cumulative Effects
Some arguments exploit the fact that even when a single act-type does not have noticeably harmful effects, the collection of many such acts might. If one person walks on the grass, she will not harm it, while if 10,000 people do, they will. If one person discharges a small volume of mild pollutants into the air or water, she may not create a serious health risk, while if 10,000 do, they will. People use this fact to mount an impressive argument:
Although this is a plausible argument if the premises are true, it is not a slippery slope argument. Arguments from cumulative effects concern the ways many innocent actions, taken collectively, can be detrimental. Slippery slope arguments concern the ways that a seemingly permissible action can lead through small analogous steps to detrimental consequences (Schauer 1985: 361-62). The reasons for thinking the former are very different from reasons for thinking the latter. The second premises of these respective arguments must be defended and challenged differently.
Those More Related to Slippery Slopes
The following two argument forms are more closely related to slippery slope arguments, and, therefore might sometimes be difficult to distinguish from them. What I shall show, though, is that for each we face a dilemma. If someone has a strong inductive generalization or causal argument, then she would not recast it as a slippery slope argument. That would be argumentatively anemic. On the other hand, those who cast their argument as slippery slopes, even if, in other ways, they resemble one of these forms, do so precisely because their evidence supporting the generalization or causal claim is weak.
Straightforward Inductive Generalizations
Suppose someone proposes that we raise the speed limit on Interstate Highways to 90 mph. I would argue that the death rate from automobile accidents will skyrocket; therefore, we should resist the proposal. But this is not a slippery slope. It is a simple inductive generalization (Salman 1989: 90). We have ample empirical data about how changes in speed limit impact death rates from automobile accidents. We saw what happened when we increased speed limits from 55 mph to 65 mph. The same evidence suggests death rates would increase if we raised speed limits to 90 mph.
The reasoning employed in slippery slope arguments differs. Each step down the slope differs from, but is analogous to, the previous step. It is not a straightforward generalization. Of course one might use a slippery slope argument to oppose increases in speed limits. But someone would do so only if she lacked solid empirical evidence of the effects of this change in policy. If she had the evidence, she would use an inductive generalization to support her position, not a slippery slope.
Straightforward Causal Arguments
Although slippery slope arguments are distinct from A straightforward causal arguments,@ they will, in some cases, bleed into them. But as with inductive generalizations, someone would use a slippery slope argument only if she lacked the empirical evidence to support a straightforward causal argument. To explain why, consider the following example. Frank intentionally drops a Ming vase from six feet above a bare concrete floor. The vase breaks. It would have been silly to have mounted a slippery slope argument against his dropping the vase since dropping the vase, barring something or someone to cushion the fall, just is to break the vase. Increasing the temporal gap between x and y does not alter the facts: my detonating strategically placed explosives atop a Swiss mountain is not the first step down a slippery slope to killing people at the bottom. Rather, barring some freakish intervention, I kill villagers below by means of an avalanche. Adding a month-long timer does not relevantly change matters, although it does slightly increase the probability that something or someone might intervene, thereby making the consequences a bit less certain, albeit still clearly causal. In each case x starts the causal chain that standardly leads to y. As Lode puts it, such chains are A more reminiscent of a cliff or a wall than a slope@ (1999: 1477).
They are more like a cliff than a slope because the mechanisms of change in paradigm slippery slope arguments differ from paradigm cases of straightforward causal arguments. Some causes are direct (e.g., the water from the leaking gutter erodes the foundation of the house) while others are probabilistic (e.g., smoking causes cancer). But in each case there is a clear causal chain from x to y. If I get lung cancer from smoke, it is because I myself smoked or regularly inhaled second-hand smoke. In contrast, slippery slope arguments hold that x leads to y by means of small, usually barely indistinguishable, analogous steps. For instance, those opposed to physician-assisted suicide may claim that even (seemingly) justifiable instances of PAS would ultimately lead some other physicians to take their patients= lives inappropriately (Bok 1998). How? Presumably successive physicians will make analogous (slightly different albeit similar) exceptions to the A rules@ against taking patients= lives; these changes in agents= perspectives will A accumulate@ over time, diminishing doctors= psychological repugnance to killing. Eventually some doctors will kill some patients unjustifiably (Wright 2000: 82-84). The earlier changes causally pave the way for later ones. However, he causal connection is not causal in the ordinary sense; rather, the change results from a series of small analogous steps. The doctors who will purportedly kill later patients are not the same ones who helped the first patients end their lives, nor did the former doctors make the latter doctors kill their patients. The probability that y will occur is also far less than one.
Nonetheless, people offering slippery slope arguments rarely conclude that doing x is probably immoral. They conclude or imply that doing x is immoral. Perhaps this omission is rhetorical since acknowledging it would diminish their arguments= ability to sway public opinion. It could also be that they think it is immoral to do x even if x only probably leads to the immoral y. If the probabilities were high enough, that might be plausible. Nonetheless, this claim should be clearly stated and defended. Additionally, I would think that if we object to x only because of these deleterious consequences, then our moral disdain for doing x would be less than if we had independent reasons for thinking it is immoral. If nothing else, we should regret we cannot do xC after all, x is prima facie morally permissible, apparently desirable, and only probably leads to immoral consequences. Yet many who employ slippery slope arguments have the same disdain for x as they have for actions they deem intrinsically immoral. That is the first suggestion that the common use of slippery slope arguments is rhetorical.
Evaluating Slippery Slopes
To evaluate causal slippery slope arguments, I begin indirectly, by examining several cases in which slippery slope arguments have been or might be used, and contrast them with a clear case where such arguments would never be used. These will help us better identify the nature, function, and reliability of these arguments.
Looking at Some Cases
Cases Where Slippery Slopes Look Plausible
If a parent wants to convince her child to be honest with her friend Susie, she might use a slippery slope sounding argument. A I understand why you want to tell Susie a `white lie= about why you cannot attend her party. That seems like a good idea right now. But be careful. By telling this small lie now, you will be more likely to later lie about more important matters,@ the parent might say. A With each lie you will become less inclined to tell the truth, and more prone to lie about more serious matters. If you do not want to become a liar, you should resist the urge to lie to Susie@ (an argument suggested by Bok's insightful discussion of lying, 1978: 57-72).
Or suppose Bob, an alcoholic who has not had a drink for two years, asks his counselor whether he could have a drink at an office party the coming weekend. The counselor will almost surely say A No.@ A Although you might think it would be acceptable to take a drink just this once, under these unusual circumstances,@ she might say, A even if this first use will not make you drunk, you will become more likely to drink again later. After all, you think, `I took a drink that time and didn't get drunk.' Each time you drink again, you will tend to increase both the frequency and amount that you drink. Before long, you will regularly be getting drunk. So don= t drinkC not even once@ (for different examples see Mayo 1990-91: 91; Lode 1999: 1504) Even if oversimplified, the counselor and the parents offer sound advice. Both claim that a relatively harmless and plausibly permissible action may increase the propensity of acting badly later. This propensity makes doing the initial actions immoral. These kinds of cases lend credence to slippery slope arguments.
Cases Where Slippery Slopes Are Implausible
Not all slippery slopes are so plausible. Some are ludicrous. You and your spouse are devoted parents. You are rarely away from your children. But you want an evening alone, without interruption from the kids. You go out for dinner together and leave your children with a sitter. Your children would prefer that you be home, but the sitter is adequate. You also spend money on yourselves, money you could have spent on the children.
Suppose someone offers a slippery slope argument against your going out. They argue that by going out this evening with your spouse rather than doing something with and for your kids you have started down a treacherous slide. Your action will probably lead you to do something more extravagant for yourself. Before long you may mortgage your children= s college education and even their health so you and your spouse can take a five-star around-the-world cruise. These consequences are so horrendous that you should not start down this road. That is why it would be wrong of you to go out to dinner with your spouse, even this once.
This argument, unlike the first two, is absurd because the projected consequences are so clearly improbable, and because few people, if anyone, seriously believe it would be wrong of you to go out for dinner occasionally. Barring some special knowledge about you, we have no reason to think these disastrous consequences are more than the faintest of faint possibilities. In fact, given our background knowledge of what often happens to parents of young children, we have far more reason to think that failing to go out for dinner in these circumstances will create or reinforce a habit of neglecting your spouse to spend time with your children. That failure would likely have serious consequences for your marriage.
Cases Where Slippery Slopes Are Morally Disastrous
The history of moral debate is littered with slippery slope arguments used to defend morally horrific behavior. Such arguments were regularly used to resist abolition. For example, a prominent Protestant preacher claimed that we should not grant A colored men@ freedom because of the A terrible consequences@ to which that would lead:
Then a colored man might be the next governor; and colored men might constitute their Legislature, and set on the bench as judges in their courts. Thus the entire administration of the government in those States would be placed in the hands of degraded men, wholly ignorant of the principles of law and government (Rice 1846: 33).
These arguments did not end with the Civil War or with the turn of the Twentieth Century. Growing up in Nashville, I regularly heard slippery slope arguments against granting equal rights to blacks. In my town blacks were required to ride on the back of the bus, to drink at separate water fountains, and to use different toilets. Proposals to change these practices were met by racists who claimed that even small changes to these rules would ultimately lead to more fundamental (and A clearly immoral@ ) changes: before long blacks might want to marry our daughters or our sisters!
Slippery slope arguments were also used to resist granting full rights to women. Thomas Taylor wrote the Vindication of the Rights of Brutes (Taylor 1966/1792) as a spoof of Mary Wollstonecraft's Vindication of the Rights of Woman (Wollstonecraft 1986/1792). He assumed that we should not acknowledge the rights of women since, if we did, we would, through some series of small analogous steps, ultimately embrace the A ludicrous@ claim that animals had rights. Here is a vivid case where the advocate did not believe that x was morally permissible. Rather he tried to prove that granting rights to women was immoral by showing that it led to absurd consequences (reduction ad absurdum). However, such arguments are convincing only if the reader (or listener) is unwaveringly committed to the third premise. Most people then thought that granting rights to non-human animals was ludicrous, so they likely found his argument convincing. Today few people think the idea is ludicrous, even if they think it is wrong. Hence, they would not be convinced by Taylor= s argument.
What These Cases Show
By reflecting on these cases, by understanding when slippery slope arguments areC and are notC used, we can isolate what is both insightful and worrisome about them.
The Importance of Habit
The arguments against single instances of lying and drinking gain their plausibility by exploiting a significant psychological and moral truth: our previous choices, actions, and deliberations inevitably shape our current behavior, while current choices and actions shape future behavior. Yet by using slippery slope arguments to evaluate only some behaviors, we imply that this is only an episodic feature of human life. It is not. It is the heart of human life. We call it A learning.@ We consciously learn words and syntax so we are able to speak and think. We consciously attend to what is around us so that we will be A spontaneously@ attentive in the future. We consciously reflect on our action so that we become predictably self-critical. Although the precise ways that we use language, attend to our surroundings, and reflect on our behavior and beliefs are sometimes conscious, the character of these conscious deliberations is likewise shaped by earlier actions, choices, and behavior. The philosophically inclined just think about problems differently than do most people. That is the kind of people our training has made us.
To use Dewey= s language, we are habitual creatures (Dewey 1988/1922). By A habit@ Dewey does not mean some set of repetitive (and usually negative) behaviors, but behavior that (1) is influenced by prior activity, especially our interactions with others, (2) organizes a person= s action, (3) is typically exhibited in overt behavior, and, (4) is operative, even when not exhibited in standard ways (LaFollette 2000). For instance, when I learn a new word, (1) I learn the word because of my interactions with others, (2) it empowers me to speak, (3) I may use the word on future occasions, and (4) even when I do not, it shapes my dispositions for future behavior by enabling me to understand the word when others use it, and it empowers me to think new ideas.
On this way of understanding human action, it is misleading to say that a propensity to be dishonest is a mere consequence of lying to Suzie. That implies a false separation between the later event and the earlier behavior. It is like saying that coughing without covering your mouth has, as a consequence, releasing germs into the air. Not so. Coughing without covering one= s mouth just is, in this world, to release germs into the air. Of course there is a temporal gap between your daughter's initial lie to Suzie and your daughter's becoming a liar (and between my coughing and germs being released into the air). That means another factor (another person, circumstances beyond one= s control, or the person's other habits) might intervene so that the later behavior does not occur. However, when the connection between an action and what follows is sufficiently tight, we do not ordinarily distinguish the action and its consequence. Put differently, human action is temporally thick: It is not something we do once, in some narrow slice of time. An action is what it is in important measure because of the ways it typically extends into the future.
A prominent way in which past actions extend into the future is by changing our individual and collective propensities for future action. That may sound like a truism. But if it is, it is a truism we have oft forgotten when we think about ethics. Slippery slope arguments gain much of their credence by exploiting this phenomenon, but they do so in ways that mask habits' central role in human behavior.
The use of such arguments also implies that all habits are negativeC that all slopes lead downward. They thereby create an A undifferentiated risk aversion@ (Schauer 1985: 376), an irrational fear of change. They make us slopeaphobic. But that is to be life phobic since all actions occur on a slope. That is the kind of creatures we are and the kind of world we inhabit. We are changing creatures living in a world of change in which each choice affects the direction and character of that change. Sure, some slopes do lead downward, but others lead upward (we call them A learning curves@ ). Whether we tend to move up or down the slope depends, in part, on how we view change, and whether we have experience traveling on slippery terrain. We can learn to better traverse downward slopesC to slip and occasionally slide, without sliding all the way down to the bottom. We can also learn how to ascend after having slipped on a downward slope: we can learn from a bad experience.
Omissions, as Well as Actions, Can Create Habits
Our habitsC our propensities for future actionC are created not only by what we do, but also by what we fail to do. Habits emerging from omissions differ from those shaped by actions. They diminish, rather than increase, the propensities for particular future actions. Every day I fail to do my scheduled exercise or practice the piano (Mayo 1990-91: 91), I initiate or reinforce a habit of missing exercise or practice. That does not mean that I will become a couch potato, but it does make it less likely, albeit slightly, that I will make or sustain a successful regimen of exercise or practice. Every time I am indifferent to the needs of a friend, I initiate or reinforce a habit of being indifferent. That does not mean that I will become a selfish pig, it just makes that more likely, albeit slightly. This is a phenomenon of which most of us ordinary mortals are well aware.
Once we recognize that both actions and omissions shape my propensities for future behavior, it is apparent that slopes cannot be avoided. Rather we should learn how to navigate them successfully.
The Importance of Empirical Data
The counselor= s argument against Bob= s (the alcoholic) taking a drink is plausible not because of some vague causal connection presumably tracked by slippery slope arguments, but because she has strong empirical evidence of Bob= s inability to handle alcohol. A slippery slope argument not only does not add anything, it detracts from the counselor= s argument. It is far more powerful for her to present Bob with the clear empirical evidence: his history of alcoholism, his past attempts to A drink just once,@ and how even a single drink repeatedly led to his resuming his alcoholic behavior. The counselor might use slippery slope sounding language, but if she does, she does so to present the evidence, not as a substitute for it. Absent such evidence, there is no good reason to tell Bob not to drink.
Once we step back and understand these arguments= function, we see that they persuade (or fail to persuade) people based almost entirely on the listener= s current beliefs about what is right and wrong. When people are predisposed to think that the initial behavior (x) is acceptable, then they are rarely swayed by slippery slope arguments. For instance, since most parents want to go out for dinner with their spouses, they are not afraid of what will happen if they do. That is why the second argument has no bite. Conversely, if the listener is already inclined to believe that x is wrong, then they will be receptive to slippery slope arguments and will not be inclined to notice the absence of empirical evidence in support of premise two. Those already opposed to euthanasia will likely think that slippery slope arguments against it are telling. This is the second reason for thinking that such arguments' primary use is rhetorical.
Their Function Is Conservative
Taylor and the racists were right: small changes not only can, but sometimes do lead, via small analogous steps, to more substantial changes. About this, those who use slippery slope arguments are right. We are creatures who learn and adapt to new environments; our previous actions change propensities for future action. Where they went wrong was in implying that all changes are morally objectionable. Many people would not have recognized that racism and sexism were fundamentally wrong until they first took those moral baby steps. That is why those small changes were not immoral. They dislodged people from their immoral views and initiated positive moral change. Change is not, as the slippery argument suggests, inevitably downward. Change can lead upward as when we learn from experience. Yet the standard use of slippery slope arguments ignores this by presupposing the moral status quo. If premise three is false, then the conclusion is not supported. This is worrisome since, as the historical examples reveal, the moral status quo is always debatable, is not infrequently inappropriate, and is sometimes seriously unjust. Yet these are precisely the circumstances in which slippery slope arguments are normally brandished: to defend assaults on the moral status quo. These are the same conditions under which such arguments are unacceptable. Once someone has mounted a critique against the status quo, we cannot defend the status quo by simply reasserting it. Yet, as Glanville Williams put it:
it is the trump card of the traditionalist, because no proposal for reform, however strong the argument in its favor, is immune from the wedge objection. In fact, the stronger the argument in favor of reform, the more likely it is that the traditionalist will take the wedge objectionC it is then the only one he has (Williams, G. 1986/1958: 426).
In such cases, their real use is rhetorical.
Their Real Use Is Rhetorical
Let= s rehearse some of our findings. Causal slippery slope arguments are plausible only if the second premise is true, yet we have no reason to believe the second premise is true unless we have evidence of the causal link between x and y. The counselor= s advice to the alcoholic makes good sense only if she has specific evidence of the alcoholic= s past; without that evidence, the advice is unduly cautious. Many people drink without becoming alcoholics. However, if we do have ths evidence, we do not need slippery slope arguments. So why do people use them? They use them as rhetorical tools. This rhetorical use may not, in itself, always be objectionable. A presenter might have the required empirical evidence to support the second premise but does not forward it because she believes the recipient does not have and cannot understand that evidence. As a heuristic device (Volokh 2003: 1125), this may be sensible. If the listener cannot comprehend the evidence, then that may be the best she can do. Even so, this use of the slippery slope is defensible only if the speaker has the necessary evidence. If pressed in philosophical debate, she should be able to produce those goods. If she can, then within that debate, the slippery slope argument adds nothing. If, however, she cannot produce the evidence, then the slippery slope argument is a rhetorical device that plays on the listeners= fears or prejudices.
Social-political Versions of the Slippery Slope
The habitual nature of humans largely explains how social mores evolve and how past political decisions shape future choices. Moreover, a central aim of social institutions and political decisions is to enable some options and to foreclose others. This combination of institutional aims and our psychological natures, seems to support the moves exploited by the second premise of slippery slope arguments in the political arena. There are five examples of these arguments.
In A On Killing the Innocent with a Clear Conscience@ Sabini and Silver convincingly argue that it is easier to get people to do morally outrageous actions by first getting them to do mildly immoral ones (Sabini and Silver 1982). Stanley Milgram exposed this tendency in his research on obedience to authority (that research is described in detail by Sabini and Silver). The Nazi= s exploited this tendency in getting German= s support for their program to exterminate the Jews. The Nazis did not initially advocate genocide. Instead they incrementally increased their mistreatment of Jews and slowly garnered wide support for their genocidal policies. This tendency, Sabini and Silver claim, is an inevitable feature of large institutions and amorphous groups, a feature for which we should be on guard.
Second, political theorists, legal scholars, and judges sometimes employ slippery slope arguments to defend free speech. Although free speech is fundamentally important to the flourishing of individuals and the state, there are instances where each of us would like to curtail some speech. We might even think we would be justified in doing so in select cases. However, if we forbade speech in these presumably justified cases, we would lessen the political and legal barriers to more frequent and substantial limitations on speech, and thereby increase the likelihood that the state will squelch speech that we need. Even when that is not a likely consequence, limitations on speech will have a A chilling effect@ on desirable speech. Citizens will increasingly be afraid to air their views in public, even if their speech would have passed constitutional muster. Rather than opening the possibility that the government will limit important speech (and even engage in wholehearted censorship), we should permit forms of speech we find grossly objectionable. We must stick by our general principles; otherwise we start down the slippery slope (Arthur 1997).
Third, slippery slope arguments are commonly used to criticize physician-assisted suicide (PAS). These arguments predict that there will be untoward consequences of legalizing physician-assisted suicide, even in cases where we might be sympathetic to the patient who wants to die:
Practices may be extended to groups of patients beyond the original few who fit the strict requirements; and distinctions may be blurred so that patients may have to die without having requested euthanasia, perhaps quite against their wishes (Bok 1998: 112-3).
These purported changes will not occur all at once, but will accumulate from a series of smaller analogous steps. To avoid these immoral results, we should refuse to take the first step.
Fourth, people occasionally offer what van der Burg calls the A apocalyptic slippery slope@ (van der Burg 1991: 43). In these cases the proponents claim not that y is especially likely, but rather that y is so terrible that the mere risk of its happening is sufficient to justify refraining from doing x. This form of the argument has been used to condemn the nuclear arms race, extensive reliance on nuclear power, recombinant DNA research, and cloning.
Fifth, people sometimes critique proposals simply because of who supports themC what Volokh calls the A ad hominem heuristic@ (2003: 1075). Members of an identifiableC and by your lights, distastefulC group offer a proposal that you think acceptable, or perhaps just a bit misguided. Nonetheless, you fear that if the group is given a political inch, then they will, over time, gain more power, and begin to institute significant and severely negative changes.
The Problems with These Arguments
It is not difficult to see why each of these examples is rhetorically persuasive. However, I would contend that they, like the personal versions canvassed before, are either (a) not slippery slope arguments, (b) that they are flawed, or (c) that their insights, however valuable, can be accommodated as well, if not better, in other ways. Let me explain this claim by returning to look at each example.
(1) Sabini and Silver claim that doing something wrong (harassing Jews) might lead us to do something horrible (killing Jews). Although that may well be true, this is not a slippery slope argument since the proponents acknowledge that the initial act is wrong, not permissible. However, even if we were to extend what we mean by a A slippery slope argument,@ it would relevantly differ from standard ones. Thinking that bad behavior will probably cause worse behavior is not surprising: it is precisely what one would expect given our habitual natures. However, it is difficult to imagine why, barring specific empirical evidence, we should generally think that doing a morally permissible action will cause us to do evil in the future.
This example does suggest one plausible rhetorical use of the argument. Suppose I have independent evidence that x is bad, I offer that evidence and convince others. However, those whom I convince are not moved to stop x, perhaps because they do not think x is so bad as to warrant strenuous effort (e.g., it is not worth the effort to support a political candidate who is only mildly better than the opposition). I want to impel them to act. So I offer a quasi-slippery slope argument to show them that once x occurs (the inferior candidate gets elected), the morally terrible y is likely to happen.
Although this is a sensible argumentative strategy, I am inclined to think that even here it would be best to forego vague talk of a slippery slope and focus on the specific empirical evidence of how and why the mildly bad x will lead to the terrible y (why electing the inferior candidate will have consequences more serious than we first thought).
(2) Although the free speech argument is plausible, in its strongest form it is not really a slippery slope argument. Proponents claim that if we prohibit Nazis from speaking then we thereby license the majority to prohibit any speech they deem immoral. That is true, however, only if the reason we prohibit the Nazis= speech is that the majority objects to it. If so, then this is a claim about what our reasons commit us to: it is an argument from consistency, not a slippery slope.
Suppose, though, that we justify prohibiting Nazi speech not because the speech offends the majority, but because we judge that the speech is especially harmful to Jews. That rationale would not straightforwardly justify restricting all unpopular speech. Under those circumstances, this free speech argument might be a slippery slope. It would be a slippery slope inasmuch as it claimed that forbidding Nazi speech for one set of reasons might lead us (via small analogous steps) to prohibit desirable speech for different reasons. The claim that the action would have these consequences is plausible only if based on sound empirical evidence, an inductive generalization employing the demonstrated propensities of people in social and political institutions. Without such evidence the argument would not be plausible. Slippery slope arguments, as they are ordinarily used, are too dull to do precise philosophical carving.
(3) Slippery slope arguments have played a central role in the debate over physician-assisted suicide (PAS). These arguments take several forms and are often offered in concert. One common move is to argue that even advocates of PAS must recognize that A the logic of justification for active euthanasia is identical to that of PAS@ (Arras 1999: 276). As stated, however, this is not a slippery slope argument but a consistency argument. It claims that since PAS is relevantly similar to active euthanasia, then if we permit one, we must permit the other. It does not claim that permitting one will lead us, via small analogous steps, to permitting the other. Furthermore, this argument assumes that active euthanasia is morally objectionable. If it were not morally objectionable, the argument has no bite. Yet Arras does not defend that claim.
Even if Arras were to mount such an argument, this general strategy, oft employed in the euthanasia debate, drives home the earlier point that slippery slope arguments are the preferred weapons against social changeC including some changes that we now regard as significant moral progress. Not only were these arguments used to battle equal rights for blacks and women, they were also used to challenge public education, the forty-hour work week, government supported retirement and medical care, etc. This does not show that current moral wisdom is always flawed; it does mean, however, that when someone challenges that wisdom with a plausible argument, then we (a) need to defend that wisdom, and (b) we cannot defend it merely by reasserting it. In arguments about social institutions as in arguments about persons, the third premise can be false. And all social institutions, like persons, are on a slope. Incremental change can be bad, but it can also be the engine of improvement. Given the creatures we are and the institutions we inhabit, we would not have decided overnight that blacks are equal to whites or that women should have the right to vote. We reached these desirable moral ends only by first taking small steps on the slippery slope of life.
Arras and other critics of PAS, however, rarely rely on a single slippery slope argument. Arras argues that permitting even seemingly permissible cases of PAS will likely lead to abuse: (a) physicians may euthanize patients even when their A decisions@ are not A sufficiently voluntary@ ; (b) the practice will have more detrimental effects on A the poor and members of minority groups@ ; (c) physicians= failures to A adequately respond to pain and suffering@ will lead some ill people to prematurely choose death; and (d) we will not establish a reporting system that A would adequately monitor these practices@ (1999: 277).
Arras has isolated some serious worries, ones we would be ill-advised to ignore. These should give us pause before permitting active euthanasia. If we proceed, we should seek ways to lessen the probability of those detrimental effects, and proceed only if the gains are worth the costs. Nonetheless, I fail to see that his points vindicate the use of the slippery slope. First, PAS did not create the problems Arras mentions. Doctors and philosophers disagree now about what constitutes a A sufficiently voluntary@ action. Doctors now fail to give adequate pain relief, the current U.S. health system is often unfair to the poor and minorities, and medical reporting in that system is shoddy (Ibid.).
Second, to whatever extent that these worries are legitimate, it is not because seemingly permissible actions will be transmuted via small analogous steps into morally objectionable ones. Rather, we can note the significant failings of our current health care system, and straightforwardly predict what will happen if we permit physician-assisted suicide unless we take due care. That is, we can inductively generalize, as we might in speculating about the likely consequences of electing an incompetent president or of raising the speed limit to 90 mph. But such speculations have nothing to do with slippery slope arguments. To use a slippery slope argument to make any of these points would be argumentatively weak.
Third, we must not forget that forbidding PAS will also have demonstrable costs, that not permitting physician-assisted suicide may be A the callous abandonment of patients to their pain and suffering@ (1999: 279). I think the conclusion we should draw is that whether we legalize PAS, we should make significant changes in our medical system. Once we make these changes, we can conduct a careful risk analysis of the benefits of permitting and forbidding PAS.
The importance of careful risk analysis is most easily seen when evaluating the apocalyptic versions of the slippery slope. Those who employ this version claim that since x might lead to some supremely terrible y, then we should refuse to do x, no matter how appealing. Such arguments have been used to criticize cloning, certain forms of genetic engineering, and our reliance on nuclear power. Consider, e.g., the claim that widespread use of nuclear power could lead to two different, but related, supremely terrible results: (a) a nuclear A accident,@ more serious than that at Chernobyl, and (b) long-term contamination of the earth from disposal of radioactive wastes.
But to see why this does not vindicate the use of slippery slope arguments, let us compare it with two structurally similar, but wildly implausible, slippery slope arguments. In the first, the same x (using nuclear power) is claimed to lead to a different but still terrible y (the moral collapse of the country). In the second, a different x (educating the poor) is claimed to lead to the same terrible y (a nuclear meltdown). Unlike the original case, these arguments are laughable. Why? Because we have no evidence that either x will have these terrible consequences. Without such evidence, the mere terribleness of y gives us no reason to refrain from x. After all, any action could lead to terrible consequences. The original argument about nuclear power, on the other hand, is plausible precisely because we can see a possible causal connection between x and y.
Once again we see that we canC and mustC assess this claim without employing slippery slope arguments. We should make an informed judgement of risk. We must determine the seriousness and likelihood of the risk and compare it with the importance and likelihood of the benefits. As the likelihood and seriousness of harm increase, we have increased reason to refrain from acting, while as the likelihood and importance of the benefits increase, we have increased reasons to act. The questions are: (a) just how risky is using nuclear power, and (b) how beneficial is it? To the extent that we have real evidence for thinking that it might have these disastrous consequences, then that should give us some pause in relying on nuclear power. Minimally it should compel us to make serious efforts at ensuring safety. Of course, that is precisely what we do. We make stringent safety demands of nuclear power plants, and we do so because we have empirical evidence that a meltdown could occur. We also know about the dangers of storing radioactive materials.
Of course knowing these risks of using nuclear power, even if substantial, does not solve the issue. For, omissions, as well as actions, have consequences. The failure to use nuclear power plants would arguably make power exorbitantly expensive, and that could lead to our country= s financial demise. Minimally it could make us too dependent on fossil fuels. These consequences are also terrible, and someone might argue that these risks, although perhaps less terrible, are far more likely than the consequences of a meltdown. I cannot here defend either argumentC or the range of other possibilities. What I do know is that slippery slope arguments, as they are regularly used, are poor substitutes for a careful assessment of risk.
In saying this, I do not wish to suggest that cost-benefit analysis is a cure-all. It, too, is beset with problems. We typically lack the knowledge to make precise predictions about the outcomes of complex social policies. However, skepticism about cost-benefit analysis does not require us to embrace slippery slope arguments. Rather we might think about how to behave in cases were we cannot accurately predict the outcomes of available actions.
This leads to the last form of the social political versionC the A Give > em an inch@ version. Such arguments are used by both sides of the political spectrum: Some people use these to critique any changes in abortion laws. They fear that if they permit so-called A right to life@ groups to win on any point, no mater how small, that the groups will be emboldened and empowered to seek more serious restrictions on abortion rights. Others may use this argument to resist any gun registration law, no matter how minor, on the grounds that if these laws are adopted, gun control groups will be emboldened to seek to confiscate guns.
I understand the appeal of these arguments. They are far from crazy. However, they are not slippery slope arguments. The issue in these cases is not, strictly speaking, whether slightly limiting abortions (or having minimal gun registration) will transform, via small analogous steps, into more significant restrictions on abortion or guns. The issue is whether giving a group you dislike a political victory, however small, empowers them to make more substantial and unwanted changes (Volokh 2003: 1075-80). These are plausible claims, but only inasmuch as they are sound inductive generalizations, grounded in knowledge of the group in question and our appreciation of the temporal thickness of action.
Living on a Slippery Slope
I have not argued that all slippery slope arguments are faulty, although many are. I have not claimed that slippery arguments never isolate morally relevant features of action, for many do. What I have argued is that given the way they function in moral debate, we should avoid them. They do not add anything and often do more harm than good. I will briefly reiterate the arguments, and then explain why it is better not to try to avoid slopes, but rather to understand that all life is in some sense a slippery slope. If so, we must learn how to successfully navigate slopes.
(1) When offering slippery slope arguments, advocates suggest they are isolating a psychological, social, or political feature that is relatively unique: the tendency of current behavior to have morally relevant consequences. This morally significant tendency is not unique, but ubiquitous. All choices occur on a slope, and any slope can be slippery, especially in the pouring rain and especially if one is wearing the wrong conceptual boots.
(2) Slippery slope arguments as standardly used not only mask or ignore the pervasiveness of change, even when they acknowledge it, they often misdescribe or misunderstand it. They claim that once we take the first step (do x), we have no ability to stop the slide to y, although y may not, for some independent reasons, occur. However, this overestimates our predictive powers while underestimating our control. We can rarely predict the long-term outcomes of a single action. On the other hand, we do have some control over which track our lives take. By acting in certain ways now we shape the people we are to become. By establishing institutions and laws now, we shape the kind of society we will become. We do not always know what the future holds, but we can make ourselves the kind of people better able to cope with whatever future we find.
(3) Even when slippery slope arguments are logically impeccable, there are equally good, and usually better, ways of capturing their insights. They are plausible since they acknowledge that our actions now shape the kinds of people we are to become and the kind of world and institutions we are to inhabit. But these phenomena are better captured by emphasizing the habitual nature of humans, by offering sound inductive generalizations, by citing specific empirical evidence about the causal relationship between x and y, or by a simple consistency argument, than by relying on some vague causal connection supposedly captured by slippery slope arguments.
(4) Standard uses of causal slippery slopes arguments make us fear change. However, life is change. And why assume the first step will take us down a slope? After all, some slopes are ascendingC they empower us to learn, grow, and flourish.
(5) To believe otherwise is to blindly embrace the moral status quo. The status quo is where we begin moral deliberation. If it is where we always end, we will sometimes perpetuate grave injustices. We can make moral advances only if we are willing to deviate from current moral norms.
(6) By making us fear slopes, these arguments make us more likely to slide on the slopes we must traverse. Consider the following analogy: people who must walk on slippery surfaces might not know that the slopes are slippery. Others might fear them. Still others might know the surfaces are slippery but are prepared to navigate them. Who can best move on slippery surfaces? The first person, being unaware of the nature of the surface, is most likely to slip. The second person is so afraid of slipping that she doesn= t venture out, while the third person will have the surest footing.
This resembles living on the slippery slope of life. Those who do not understand the propensities of current action to shape future behavior (for example, young children), are more likely to make mistakes. Those who are unduly afraid of slippery surfacesC who unduly fear changeC will stay crouched in their moral corners, afraid to do anything new, different, or innovative, because any new action might lead to perdition. While those who understand that all life is on a slopeC those sensitive to the ways in which current choices and actions have morally relevant consequencesC will be better prepared to navigate those slopes. They will see the ways that personal actions and social changes can have detrimental effects, and will be on guard against them, and, when feasible, find ways of insuring the detrimental effects do not occur. These people will have the conceptual boots to give them a relatively firm grip and the experience of walking on slippery surfaces that gives them more secure footing.
In short, the knowledge that actions occur on a slope should neither incapacitate us nor make us unduly fearful. If we did not change, then we could not learn, grow, improve, and progress. What we thought was a descending slope might turn out to be ascending. In other cases we may discover, what every hiker knows, that a partial descent down one slope is often required to climb to a higher neighboring peak (Depew and Weber 1996: 282-4). Finally, even when we are on a descending slope, we can often descend part of the way without sliding to the bottom.
Whether we can and do depends, in large measure on our recognition of the moral terrain on which we travel, and from our experience in traversing such terrain. Of course change is not always for the good. It must be watched, scrutinized, and evaluated. However, that is just to say that we should reflect on what we do. We will then be more likely to intelligently guide our conduct: to act when we should, to refrain from acting when appropriate, and the wisdom to discern the difference.1
1. I wish to thank David Copp, John Hardwig, Todd Lekan, Michael Pritchard, Hillel Steiner, and especially Eva LaFollette and Brook Sadler for helpful comments and suggestions on earlier drafts of this paper.
Arras, J. D. (1999) "Physician-Assisted Suicide: A Tragic View." In Arras, J. D. and Steinbock, B. (eds) Ethical Issues in Modern Medicine (5th edn.). Palo Alto, CA: Mayfield Pub. Co., 274-80.
Arthur, J. (1997) "Sticks and Stones." In LaFollette, H. (ed.) Ethics in Practice: An Anthology (2nd edn.). Oxford: Blackwell Publishers, 364-75.
Bok, S. (1998) "Part Two." In Dworkin, G., Frey, R. G., and Bok, S. (eds) Euthanasia and Physician-Assisted Suicide: For and Against. Cambridge: Cambridge University Press, 83-139.
C C (1978) Lying: Moral Choice in Public and Private Life. New York: Pantheon Books.
Depew, D. J. and Weber, B. (1996) Darwinism Evolving. Cambridge, MA: MIT Press.
Dewey, J. (1988/1922) Human Nature and Conduct. Carbondale, IL: Southern Illinois University Press.
Govier, T. (1982) "What's Wrong with Slippery Slope Arguments." Canadian Journal of Philosophy 12, 303-16.
LaFollette, H. (2000) "Pragmatic Ethics." In LaFollette, H. (ed.) The Blackwell Guide to Ethical Theory. Oxford: Blackwell Publishers, 400-19.
Little, J. F., Groarke, L. A., and Tindale, C. W. (1989) Good Reasoning Matters! Toronto: McClellan and Stewart.
Lode, E. (1999) "Slippery Slope Arguments and Legal Reasoning." California Law Review 87 (6), 1469-544.
Mayo, D. J. (1990-91) "The Role of Slippery Slope Arguments in Public Policy Debates." Philosophic Exchange 20-21, 81-97.
Rice, N. L. (1846) A Debate on Slavery. New York: Wm. H. Moore & Co.
Sabini, J. and Silver, M. (1982) Moralities of Everyday Life. New York: Oxford University Press.
Salman, M. H. (1989) Logic and Critical Thinking. San Diego, CA: Harcourt Brace Jovanovich.
Schauer, F. (1985) "Slippery Slopes." Harvard Law Review 99, 361-83.
Shapiro, S. C. (1998) "A Procedural Solution to the Unexpected Hanging and Sorites Paradoxes." Mind 107, 751-61.
Taylor, T. (1966/1792) A Vindication of the Rights of Brutes. Gainesville, FL: Scholars' Facsimiles & Reprints.
van der Burg, W. (1991) "The Slippery Slope Argument." Ethics 102 (1), 42-65.
Volokh, E. (2003) "The Mechanisms of the Slippery Slope." Harvard Law Review 116, 1026-134.
Walton, D. (1992) Slippery Slope Arguments. Oxford: Clarendon Press.
Williams, B. (1995) "Which Slopes Are Slippery?" In Making Sense of Humanity and Other Philosophical Papers, 1982-1993. Cambridge: Cambridge University Press, 213-23.
Williams, G. (1986/1958) "Euthanasia Legislation: A Rejoinder to the Nonreligious Objections." In Mappes, T. A. and Zembaty, J. S. (eds) Biomedical Ethics. New York: McGraw-Hill.
Wollstonecraft, M. (1986/1792) A Vindication of the Rights of Woman. London: Dent.
Wright, W. (2000) "Historical Analogies, Slippery Slopes, and the Question of Euthanasia." Journal of Law, Medicine, and Ethics 28 (2), 176-86.