2. The Argument in General

We are here dealing with the slippery slope argument (SSA), as often scorned by philosophers, though often with such unhelpful qualifying phrases as ‘if the likelihood of such trouble is exaggerated’Footnote ³ as it has been invoked by juristsFootnote ⁴ and political advocates. Many critics have dismissed the SSA as the last resort of the traditionalists bereft of better arguments.Footnote ⁵

Textbook discussions of this alleged fallacy are systematically inadequate, in a way that reflects the authors’ sympathy for proposals against which SSA's have been deployed. So far as I can see, their chief complaint is that SSAs are examples of the fallacious argument from fear, with some reference to the fallacy of the heap as well.

Thus the premier American logic textbookFootnote ⁶ cites a ‘keen critic’ for the following response to the SSA against physician assisted suicide.

At the same time, it endorses the following SSA against punishing hate crimes more severely than ordinary crimes.

But the first of these quotations ignores the line, salient on any view of which I am aware, between acts that intentionally take human life and acts that do not. And the second fails to specify bad effects arising from motive-defined criminal laws comparable to a doctor's deliberately killing his unprofitable patients or pressuring them to kill themselves. (Like most bad arguments it can be improved.)

I here argue that SSA's, though they do not provide a deductive guarantee, sometimes give us strong reasons to reject an initially attractive proposal. Such a proposal need not be acceptable in itself.Footnote ⁹ But unless someone finds it appealing, there is no need for an SSA. If an otherwise seemingly acceptable A (in my initial example, legal marijuana without advertisement) threatens to lead to an unacceptable B (dealers aggressively marketing marijuana to children), this is a reason, though not always a decisive reason, for rejecting A. That the argument is non-deductive, and sometimes leads us astray, is no argument against my thesis. The same could be said of induction.

Slippery slopes happen, even if sometimes we welcome their results. They are always contextual: if we have hit bottom (say we are living in Germany under the Nazis)Footnote ¹⁰ they cease to be relevant. And different countries will be in different places on various slippery slopes. Hence I use mainly American examples, though my audience is all philosophers interested in applied logic. Legal scholars do not need instruction, since lawyers (and political advocates) invariably use or respond to SSA's as their advocacy requires. If they reject the argument as ‘pernicious nonsense’Footnote ¹¹ or ‘obscurantist flim-flam’,Footnote ¹² the suspicion that they are in fact trying to push us down the slope will be acute.

If a white Southerner argued in 1860 that abolishing slavery would lead to legal intermarriage and an African-American President, his historical prophesies would have been correct. We might not think that the bottom of the slope was a bad place, but that – my hypothetical pro-slavery writer would say – is exactly what we should expect. And those who promoted this development appealed to benign slippery slopes: Frederick Douglass, urging his fellow African-Americans to enlist in the Union Army maintained, correctly, that the limited emancipation proclaimed by President Lincoln made general emancipation inevitable.Footnote ¹³

Anneli Jefferson has observed:

But in fact the values of individuals, as well as of groups, do change, sometimes in ways that support SSAs. To give an old-fashioned example, someone brought up by socially conservative parents may come to accept heavy drinking, marijuana use, sex outside of marriage, abortion, ‘open’ relationships, bisexuality, and orgies. Whether we regard such a person as debauched or liberated, the phenomenon is real. And at each stage it may seem clear to him or her that the next one is plainly wrong, but habits and associations operate powerfully to undermine such perceptions. And so at the end of the slope the person will accept things that, at the beginning, he or she would have regarded as utterly unacceptable. And each step down the slope will involve a shift of moral perceptions, while seeming too small to worry about. The person at the beginning of the slope will view the person at the end with horror; the person at the end will look back to the person at the beginning with contempt, perhaps tinged with nostalgia.

Moreover, the outcome of some slippery slopes is bad by any standard. The erosion of constitutional government in Weimar Germany proceeded from more and more extensive use of emergency decrees to circumvent legislative deadlock, through the overthrow of the Prussian state government, through the summary killing of Hitler's rivals in the Nazi Party in Night of the Long Knives, through the euthanasia campaign, and lastly to the Holocaust. The doctrines of the jurist Carl Schmitt facilitated this development.Footnote ¹⁵

We may also observe the slippery slope at work in areas of contemporary controversy. The Supreme Court of Vermont, by way of establishing same-sex (quasi-) marriage in the state, turned the sour conservative maxim, Give some people an inch and they will take a mile, in to an imperative of constitutional jurisprudence. The court reasoned,

And the slope has gone on to the full recognition of same-sex marriages. This particular slippery slope has been welcomed, or at least accepted, by many of my readers. But it is still worthwhile to observe the slippery slope at work. And whatever their views on marriage, my readers may be troubled by the story of euthanasia in the Netherlands.

To spell the matter out, we may be confronted with an audience favorably disposed toward assisted suicide by the terminally ill, but horrified by the suicide of emotionally troubled adolescents. If all forms of assisted suicide are illegal, then such people will be disposed to change the status quo. The question then is, how likely it is that change in favor of some suicides will lead to the general licensing of suicide. One crucial element in our reasoning will be the vague expression unbearable suffering, which in any event cannot be limited to physical pain (or at any rate is not so limited in practice).

In any event, the central questions raised by SSAs are those of evaluation and of real-world effects. If the status quo before the proposed change is bad enough, or the condition at the bottom of the slope is relatively acceptable, then the SSA will fail. To recur to my original example, someone who finds the present ‘War on Drugs’ intolerable, or views with aplomb the aggressive marketing of recreational drugs to children, will not be troubled by the difficulty of maintaining a ban on marijuana advertising in practice. Likewise, if the supposed causal relations do not obtain – and they are likely to be highly controversial – the SSA will fail.

Bernard Williams is wrong to distinguish ‘horrible result’ from ‘arbitrary result’ SSAs.Footnote ¹⁸ Slippery slopes trouble us only because their results can be horrible (though some possible results, such as endless litigation, are horrible only relatively speaking). In no case do we require horrors of the Nazi variety; there are many unpleasant states of society that do not go that far.

The question of historical causation from the top to the bottom of the slope is complicated. Ideas have consequences, but they are not the only thing that has consequences. Nor are the consequences of ideas in practice always what one might expect them to be in the abstract. Bad arguments can be historically influential. In practice, this means that when an acceptable change A is followed by an unacceptable result B, it is often uncertain that that it was A, rather than some other development within the society in question that is to blame for B. Which battles to fight, and which are the most important, is an inescapably prudential question. But the counsel to avoid bad results by opposing bad beginnings remains sound, however complicated its application in practice.

Even when we speak of actions rather than justifications, the SSA concerns, not bare behavior (if there is such a thing), but behavior understood and legitimated in certain terms – as the exercise of a constitutional right for example. Hence some legal or legal-like formulary is always in the background.

Some facts about language are therefore crucial. With one narrow class of exceptions, all language is open to interpretation. The exception is the rules of a game such as chess, in which the number of possible positions is both finite and precise, however enormous. Aside from such cases, legal formularies fall upon a continuum with respect to the extent to which they invite dispute about their application. That each State of the American Union is entitled to two Senators raises serious interpretive issues only under conditions of plague or civil war, where the issue is whether the relevant entity is entitled to act as a State. On the other end of the continuum, the notorious ‘mystery’ doctrine (‘At the heart of liberty is the right to define one's own concept of existence, of the meaning of the universe, and of the mystery of human life’)Footnote ¹⁹ is as indeterminate as a formulary can be without losing all content. For there is no conduct – whether suicide bombing or pursuing wealth in ways dangerous to the nation's economic stability – that could not be defended as a way of responding to the mystery of human existence.

Williams is correct to see that slippery slopes are inherent in some of our language, for example the word person as commonly used by bioethicists. In his own words,

This slope has in fact been exploited by some writers in bioethics.Footnote ²²

As Eric Lode has pointed out, ‘People frequently remain steady in their application of vague terms.’Footnote ²³ But this is true only when the background cultural and political understandings that govern their use are stable. Our legal system and our customary morality, however, have their home in a world where many different groups strive to push them in a direction they desire.

Vague doctrines cover up clashing agendas, in a way that can smooth a slippery slope. As Mario J. Rizzo and Douglas Glen Whitman point out,

In many cases the demand for legal or policy consistency generates slippery slopes. If two elements of the law are in tension with another, then the inconsistency can be resolved in more than one way. But each of them will be regarded, by some parties to the dispute, as involving a slippery slope in an undesirable direction.Footnote ²⁵ For these reasons, SSAs in real-life practical discourse are not ‘sorties SSAs’, which exploit the inevitable open texture of language to oppose the extension of a term even part of the way into a contested area.Footnote ²⁶ For example, someone might argue against severe taxation of the top one percent that there is no clear line between rich and poor, and that everyone would then be at risk of expropriation. Or we should not exempt low income people from the income tax, lest we end up without any revenue. Such SSAs would bring practical reasoning to an end. In real SSAs, there is always some principle or social force pushing us down the slope; the issue is its intellectual or political power.

Logic is one thing, and social psychology another. As Lode puts it, ‘Effective distinctions may fail to be reasonable. … Similarly, reasonable distinctions might not be effective.’Footnote ²⁷ Again, bad arguments can be politically effective. The reason for this phenomenon is, as Rizzo and Whitman observe,

Thus arguments and principles employed by people with one agenda may be captured by others whose agenda is very different and possibly even abhorrent to those who make the original argument.

There are different versions of the understandings of the connective, If … we will. Most important is the question, whether the relation here is one of logic, however broadly construed, or only of social psychology. Or in terms of the more formal version, the question is whether the property P, which the person making the SSA asserts obtains throughout the series, is understood in consequenatialist or deontological terms – as bad in its social effects or offensive in itself. Sometimes the evaluation will switch from deontology (at the bottom of the slope) to consequentialist (at the top).

There are three different kinds of argument that go by the name slippery slope, depending on how these questions are resolved. One is an argument by analogy, the second looks to the nature of the justifications offered, and the third relies, not on intellectual considerations, but on the dynamics of social and political life.

3. The Analogical Slippery Slope

Arguments by analogy pervade discussion of practical issues, both in the law and in more informal contexts. Here are two influential examples:

Neither (A) nor (B) ends the argument. Those who reject their conclusions argue that the cases are relevantly different; an argument by analogy can always be answered with a distinction. But neither are they are irrelevant: they present a case their opponents are obliged to answer. As Douglas Walton has pointed out,

Simple arguments by analogy like the above are dyadic; an analogical SSA involves a repeated use of analogy. Thus for example:

Writers hostile to SSA's take (C) as their paradigm of the alleged fallacy. But there is nothing inherently wrong with it: it differs from (A) and (B) only in that it reiterates the analogy. And like them, while not decisive, it presents an argument its opponents are obliged to answer.

What (C) calls for is what is somewhat brutally called a ‘cutoff point’ – an intellectually and politically defensible line between a human being or person and pre-human organic matter. Its proponents also need such a cut-off point, in order to rebut the suggestion that they are committed to the absurd result that any abstention from reproduction is murderous. But, since 1827, we have known of at least one clearly defensible cut-off point: the moment of conception.Footnote ³¹ In all cases we need to know at what point there is a distinction that stops the series of analogies.

But why not draw an arbitrary line, as we do when we designate a speed limit?Footnote ³² The answer lies in the gravity of the issue at stake. The ‘mistakes are fatal’ argument against capital punishment is very popular, though it has its problems.Footnote ³³ But a parallel argument against punishing minor traffic offenses would be grotesque. We can be unconcerned about arbitrariness in the voting age, since a sixteen-year old need only wait two years for full citizenship. And a similar point holds for jury size, since there is no reason to believe that six-person juries are especially likely to perpetrate miscarriages of justice. But if we were to establish a maximum voting age, whether that age were seventy-five or eighty would be a more sensitive issue. The issue would be more sensitive still, if we accepted the proposal that people beyond a certain age were to be killed or denied life-saving medical care. Likewise, the possibility that we define the line between persons and non-persons in such a way as to wrongly exclude some entities from personhood ought to give us serious concern.

In response, we can adopt a ‘precising definition’ of person, designed to remove or reduce its vagueness or open texture.Footnote ³⁴ At this point analogical and sorites understandings of the SSA merge, since each step in the sorties is like its predecessor, and the question in either understanding is where to draw the line. But these definitions commonly have considerable open texture as well.

Another approach is to embrace the open texture of the relevant concepts, and adopt, for example, a ‘gradualist’ approach to the fetus question. The problem lies in establishing an intellectually and politically defensible correlation between stages of development and ‘indications’. We lack the necessary conceptual equipment for this task, in part because the moral metaphysics of the West treats personhood as an on-off category, and affirms that all persons (however much inequality we may tolerate at the practical level) as inherently equal. A third approach is to brave the charge of ‘speciesism’ and formulate the relevant rules in terms of membership in the human species. Members of extraterrestrial species, or even chimpanzees and dolphins, might also end up getting recognized as, as was said in the sixteenth century of American Indians/Native Americans, as ‘true men’.Footnote ³⁵ But such recognition would not be automatic, and would require both deliberation and experience.

4. The Argumentative Slippery Slope

The analogical slippery slope focuses on the objective characteristics of the pertinent actions, to the limited extent that these can be distinguished from the agent's motives, intentions, and justifications. The argumentative slippery slope links forms of behavior that may look very different, but are linked by the justifications offered in their behalf. The form of the argument is as follows:

1. (1) All the arguments for A are also arguments for B.

2. (2) Therefore, those who accept A also should accept B.

Where such an argument is valid, refusing to accept it is to commit the ‘taxicab fallacy’. Just as someone who has taken a taxi to the airport pays and tips the driver, and wishes him or her a good day, so one who commits the taxicab fallacy dismisses the argument when it has served its purpose. He behaves like a lawyer who is content to win the case and hand and feels himself entitled to neglect the possible wider consequences of a favorable ruling.

Here are two examples:

These arguments both seem good, at least in the absence of persuasive considerations to the contrary. In defense of (E), it is true that those people who oppose Roe v. Wade for women demand a greater sacrifice of people who have brought about a pregnancy than do those who oppose Roe v. Wade for men. But the proponents of Roe v. Wade for women also demand more for women than their counterparts do for men – not only to have no part in the child's upbringing, but also to do something that prevents the child being brought up at all. Any kind of argument can be involved in an argumentative slippery slope, and the criteria of validity (or goodness) vary from argument form to argument form. What is essential to a valid argumentative slippery slope is that the critic uses the same kind of argument as did the original arguer: if a deductive argument, then a deductive argument, if an argument by analogy then an argument by analogy, and so on. If the argument used by the critic is weaker than the original argument, then the argumentative slippery slope is fallacious.

The strength of an argumentative slippery slope depends on the arguments advanced for starting the slope. The more proponents of voluntary euthanasia and assisted suicide emphasize the burden sick and disabled people create for society, the slipperier the slope will be.

When gay marriage is defended as the triumph of love over ‘archaic’ custom, then the slope toward incest and bestiality is very slippery. For we expect parents to love their children, and pet owners to love their companion animals. And since we are talking about marriage, the tradition that opposes love to marriage is irrelevant here. A more sophisticated argument for same-sex marriage is that marriage has no essence or nature, but designates whatever sort of relationship we choose to define as such (a ‘social construction’, if you will). The inevitable issue is why a polygamous family, a bisexual, polyamorous commune, two brothers living together as Platonic roommates, or a woman who has decided to wed herself,Footnote ³⁷ should not be defined as married.

The strength of an SSA also depends on our reasons for rejecting the bottom. Consider a slope ending in incest. If our concern is defective offspring, then gay incest is no different from homosexual activity of other sorts. If our concern is the disruption of the patterns that provide the grammar of social relations, then we have considerable reason for alarm. If there is nothing wrong with incest, there is nothing wrong with a slippery slope with incest at the bottom. That there is at present no significant incest liberation movement pertains to the next sort of SSA.

5. The Prudential Slippery Slope

Prudential SSAs are of the following form:

1. (1) Policy P will (or may) lead to result R.

2. (2) Result R is bad.

3. (3) Therefore, we should not adopt policy P.

Two prudential SSAs, each of which can be taken as a sample of a far larger family are

The same sort of argument applies to the assassination of alleged terrorist leaders.

Both (F) and (G) are consequentialist arguments, though they are both used – especially (F) – by writers whose approach to ethics is non-consequentialist. Consequentialist arguments cannot be avoided in ethics,Footnote ³⁸ though to rely on consequences alone would produce a morality that we should reject on consequentialist grounds. Both also turn on the consequences of policies: not individual acts, but acts carried out under public authority more or less by the light of day. The Nazi euthanasia campaign was somewhat clandestine, and arguably contributed to the, also somewhat clandestine, Holocaust. But if clandestine operations have untoward implications, we are entitled to expect that public policy announced by the White House, the Congress, or the Supreme Court will have such implications as well (and even more so).

Whether a prudential slope is slippery depends on a host of socio-psychological mechanisms).Footnote ³⁹ I here hit the high points only. The status quo in all societies outside the gates of Eden is the result of a compromise among a number of groups each having both their own distinctive material interests and their own distinctive understandings of the good society and the way to it. Some such groups are acutely unhappy with the status quo, and have a program for change. Sometimes accommodation or even alliance is possible but sometimes these programs clash deeply. The saying of Mao Tse-tung, ‘We shall support whatever the enemy opposes and oppose whatever the enemy supports’, has adherents these days all over the political spectrum.

Those who use the slippery slope defensively sometimes hope to exploit it as well. Disaffected groups would like to change things all at once, and some of their adherents have moral objections to incrementalism.Footnote ⁴⁰ But political effectiveness requires a willingness to proceed ‘step by step’, if only to plant premises in the public mind that make possible more radical policies in the future.Footnote ⁴¹ And this means trying to push the larger society down a slippery slope.

As a result, a range of possibly attractive moderate positions becomes harder to defend. These range from moments of silence at the beginning of the school day, through proposals defuse the same-sex marriage issue by privatizing marriage,Footnote ⁴² measures of gun control short of total prohibition of private possession, artificially induced coma as an alternative to euthanasia and assisted suicide, and attempts to find a middle ground between open borders and draconian immigration policies, to states’ rights approaches to the abortion issue. At the constitutional level, they include the American compromise about religion: all religions are granted freedom, but none allowed dominance. Some of these proposals might be the best that can be done in an imperfect world. But we are constantly made conscious of the possibility that advocates of a moment of silence might, to put their opponents’ suspicions as charitably as possible, have a program for using state power to pressure children to pray.

In view of the possibly pernicious uses of the SSA, we might mount an SSA against it. Once we start using SSAs, or so the argument goes, we will not stop until we renounce all attempts at rationally defended reform.

But, first, the SSA is directed against innovations, and the use of SSAs is part of the practice of advocates today.Footnote ⁴³ Opponents of the SSA are proposing a reform of our argumentative practices. Second, the SSA gets its force from skepticism about our ability to predict and control the unintended consequences of our policy choices. If this skepticism were total, reasoning about questions of policy would come to an end, but it need not be so: all our decisions – including the decision to use or accept an SSA – must be made with appropriate humility, but this need not be paralysis.

Another version of this objection runs as follows. The SSA turns on the difficulty we may have in making distinctions (e.g. between voluntary and involuntary euthanasia, or between extrajudicial execution of fellow citizens suspected of being terrorists and rubbing out political opponents). But the SSA itself supposes that we are able to make the distinction between the (at least relatively) acceptable behavior at the top and the unacceptable behavior at the bottom. Hence the SSA undermines itself.Footnote ⁴⁴ In reply: we are neither totally unable to make distinctions (say between killing an animal and killing a human being) nor perfectly good at making them (especially under conditions of stress, as in end-of-life decisions).Footnote ⁴⁵ And we will be less good at making decisions when the distinction is not one embodied in existing practice. SSA is a conservative argument, though not every cause in which it is used is standardly conservative. Until just recently, civil liberties were considered a liberal cause, and SSAs are routine in their defense.Footnote ⁴⁶ Everything depends on the contours of the particular issue.