Take the 2-minute tour ×
Philosophy Stack Exchange is a question and answer site for those interested in logical reasoning. It's 100% free, no registration required.

When you look at the world, I think it's a rather non-controversial statement that a good percentage, if not a majority, of social problems are caused by people making choices based solely on short-term outcomes--"if X then Y, and Y is desirable, therefore I will do X"--without taking long term consequences into account. (ie. Y leads to Z, which is more undesirable than Y is desirable, therefore you should not do X.)

But in a recent discussion with a friend that got into this territory, when I tried to point out undesirable long-term consequences, he accused me of employing the Slippery Slope Fallacy and declared my reasoning to be therefore invalid, and that was the end of that. And it made me wonder.

Why is it formally considered a fallacy to try to reason beyond a single degree of cause and effect, and why does this explicitly only apply when reaching a negative consequence? (I distinctly remember, when we learned about fallacies in high school, that "X leads to Y, which leads to Z, and Z is desirable, therefore you should do X" is not considered an example of the fallacy.)

It seems just a little bit Orwellian to see the name of Logic invoked to formally dissuade people from one of the highest and most practical forms of logical reasoning, namely reasoning about the long-term consequences of one's own actions. So what am I missing? Why is this considered a fallacy?

share|improve this question
1  
Fallacies of the form "X leads to Y, which lead ... to Z", where Z is desirable but where the chain of consequences is tenuous, would typically be called wishful thinking. –  Niel de Beaudrap Oct 24 '12 at 22:02
    
Personally, I think your friend is limiting his thinkig capabilities by a story about possible fallacies in reasoning... In my opinion, short term thinking isn't a choice, its more a unconscious act. If your intention is right, you do not have to think in long or short terms, it will always turn out right and you only have to think in the moment: "Is my intention right?" Good intentions always yield good results. What you (or your ego) want is not always what you need.... –  Mike de Klerk May 16 at 13:02
    
@Mike: Good intentions always yield good results? Try telling that to all the millions of people who had their homes foreclosed as the end result of well-intentioned politicians deregulating the mortgage industry to make it easier for people to get a house. Just to name one obvious example. –  Mason Wheeler May 16 at 13:12
    
Politicians do not always act from genuine good intentions. And btw, if a workaholic father loses his home, he might spent more time with his family. Though initiate perforce, it might wake him up to understand that family time is more important than money. So losing your home doesn't necessarily has to be a bad thing. Everything is relative and how you deal with things is what matters, and you need good intentions for that to yield good results. –  Mike de Klerk May 16 at 13:30
2  
@MikedeKlerk "if a workaholic father loses his home, he might spent more time with his family" shouldn't it be "if a workaholic father loses his home, he might spent more time with his family under the bridge"? –  LIUFA May 16 at 14:12
show 2 more comments

5 Answers 5

The problem with some instances of the kind of reasoning you are deploying is that, in a train of thought of the kind If A, then probably B; if B, then probably C; if C..., then probably Z, although every of the single steps might be likely and compelling, the transition from A to Z might not be.

Suppose that each transition has a probability of 0.99 -- that is, P(B|A)=P(C|B)=...=P(Z|Y)=0.99. Then, for all the argument tells us, the probability of Z given A is the much lower 0.99^28 = 0.75

This doesn't mean that, in general, reasoning from causes to consequences is fallacious. As you rightly point out, it is not.


Now that we are at it, it is probably useful to note that hardly any of the traditional fallacies single out reasonings that are never compelling. Take, for example, post hoc, propter hoc: while, indeed, in general it is not true that if an event follows another they are both related causally, noting a strong temporal correlation between events of one type and events of the other can provide very good evidence of the existence of a causal process. Ignoring it on the grounds that post hoc, propter hoc reasonings are fallacious would be very silly.

share|improve this answer
    
if I'm reading you correctly there - a fallacy states "This logic is not always true" - not "this logic is always false" (at least in these examples). Just out of interest, is there a term for "this logic is always false" specifically? –  Ryno Oct 24 '12 at 8:53
    
@Ryno you are probably thinking of (in)valid arguments –  Schiphol Oct 24 '12 at 14:30
1  
+1 for the nice example using conditional probabilities! –  DBK Oct 29 '12 at 0:19
add comment

According to the Fallacy Files, a slippery slope argument is an example of an informal fallacy:

An informal fallacy is one that is not formal, that is, it is a type of fallacy in which the content of the argument is relevant to its fallaciousness, or which is fallacious for epistemological, dialectical, or pragmatic reasons. Typically, informal fallacies occur in non-deductive reasoning, which relies on content as well as form for cogency. Also, because content is important in informal fallacies, there are arguments with the form of the fallacy which are cogent. For this reason, when forms for informal fallacies are given, this is for identification purposes only, that is, one cannot tell from the form alone that an instance is fallacious. Rather, the forms will help to differentiate between distinct types of informal fallacy.

To put it another way, just because someone uses an argument in the form of a slippery slope does not mean that their argument is invalid or fallacious. It's perfectly possible to construct a slippery slope argument that works and is persuasive:

Don't light that cigarette or the next thing you know you'll have lung cancer!

It works because smoking is addictive (so once you start it's difficult to stop) and does cause various types of cancer. Obviously not all people who take up smoking will get lung cancer and some people with lung cancer have never smoked. But that's not what the argument is claiming. Rather it's attempting to persuade the hearer that making a certain choice (taking up smoking) could have unintended consequences outside of their control (getting cancer).

A similar argument is fallacious:

Don't go to that dance or the next thing you know you'll be pregnant!

This doesn't work (even if the hearer is female) because there are many things a person can do to avoid getting pregnant at a dance.1 A dancer might say "No" to a potential liaison or use contraception, for instance. The argument doesn't hold up if there are subsequent decisions that can be made to prevent an a person from slipping to the (presumably) undesired bottom of the slope.


There's a closely related argument that's sometimes called the semantic slippery slope. Rather than being of the form "X leads to Y, which leads to Z", it goes "X is like Y, which is similar to Z". This version depends on the vagueness of the lines between X and Y and Z. We see this sort of reasoning a lot in the abortion debate:

Killing a baby after birth is wrong, therefore killing a fetus at 9 months is wrong, therefore killing a fetus at 8 months wrong, ...

And:

A fetus is just a mass of cells in a woman's body so removing them has the same ethical complications as removing [some other mass of cells].

Neither of these arguments are cogent on their own since reasonable people can draw clear lines between one situation and another. Is a fetus a part of the woman's body or is it an independent human? The lines are fuzzy, so it sort of depends on where you start. Neither argument is particularly persuasive for people who make different assumptions or have different ethical priorities.

A semantic slippery slope argument could be cogent, but it's also a red flag that there might be something fishy going on under the surface.


Footnote:

1. The old joke: A Baptist minister once warned young people against sex because it leads to dancing.

share|improve this answer
add comment

The problem with slippery slope arguments is that they strictly speaking only allow extreme positions, because every move away from one of the extremes is a move towards the other extreme, and if the other extreme is identified as something bad, then the slippery slope argument will say that a slight move towards that extreme is bad because you can add another small move, and another one, and eventually you'll get so close to the other extreme that it is clearly bad.

For example, a slippery slope argument would say that you should always reduce your food intake as much as possible, because if you eat a bit more, the next time you might eat again a little bit more, and again a little bit more, until you eat so much that you get obese. Of course obesity is a real danger, as the large number of obese people shows. But the principle of eating as little as possible isn't healthy either because it leads to anorexia.

In most cases, the optimum is not at one extreme, but somewhere in the middle (you should not eat too much, but you also should not eat too little; you should not waste your money, but you also should not be stingy). Of course if you are somewhere in the middle, there's always the danger of failing to remain there and moving to far to one side. But what the slippery slope argument fails to take into account is that this danger exists for both directions.

The problem is that extreme positions are very easy to maintain because it is very easy to tell what's "wrong": Everything which moves away from the extreme. If you try to keep a healthy middle, things are not so easy: For each single move, it may not be always obvious if you already get too far away into one of the directions, or maybe are even correcting a previous move too far in the other direction. That is, if you try to maintain a middle position, you always need to think about whether something already goes to far. Moreover, different people will not agree where exactly is the optimal point, and it's not that easy to explain why you consider your point the optimal one, or even what this point is (and sometimes you'll even have to adjust your position about what the optimal point is, as you learn something new, or as the situation changes). A maintainer of an extreme position has it easier. He has a clearly defined position, and every move away from it can be argued against with the slippery slope argument. That's what makes extreme positions and slippery slope arguments so attractive: You don't have to think about whether this specific move is good or bad. You just have to look at the general direction of the move to determine if you should reject it.

share|improve this answer
    
But isn't this simply the Golden Mean Fallacy: the correct position is generally somewhere in between the two extremes? Its incorrectness can be trivially demonstrated with the following argument: "Bob takes the extreme position that murder is never acceptable, whereas Frank takes the equally extreme position that everyone he does not like should be murdered. We should work out a reasonable compromise; maybe only people who have caused real harm to you should be murdered." –  Mason Wheeler Oct 27 '12 at 12:38
    
A middle-ground position is just as slippery, because it can be moved simply by moving one of the extremes. See the concept of the Overton Window, for example--the theory that this exact phenomenon not only happens, but is consciously manipulated by those who understand it, in politics. –  Mason Wheeler Oct 27 '12 at 12:39
    
@MasonWheeler: First, generally is not the same as always. Also note that "murder" is by the very definition "unjustified killing of humans" and therefore by definition cannot be justified; the extreme position is not "murder is never acceptable" but "killing humans is never acceptable". And in this form, it is indeed a disputable claim. For example, most people would consider killing in self-defense acceptable in certain situations. Also, a position in between the extremes is not necessarily (and indeed almost never) a position exactly in the middle (indeed, in most cases, ... –  celtschk Oct 27 '12 at 16:28
    
... you would be hard-pressed to even define that middle). And finally, if you say "a middle-ground position is just as slippery" then you are certainly not arguing about what I wrote. The extreme position is exactly the position which is not slippery; my point is exactly that not being slippery does not mean that you are right; indeed, often (but, again, not always) it is a clear sign of being wrong. –  celtschk Oct 27 '12 at 16:31
add comment

In the course of debating many political issues, it is very common for both sides to claim that various things will have various effects, without either side being able to definitively prove its case. Over the years, I have found that in many cases when one side claims that one political action will cause some particular chain of effects and the other is arguing that such a chain of effects will not occur, the latter party in practice ends up deliberately instigating the very chain of events they claimed would not occur.

In formal argument where there is fundamental agreement upon premises, the credibility of one's opponent doesn't matter. In politics, however, when arguments are predicated upon unproven claims, it is necessary to judge the credibility of the participants' claims, and such judgements may often be very reasonably based upon judgments of the participants themselves. The fact that a person is of bad character would not imply that they are incapable of making a valid argument, but should cause one to regard with extreme suspicion the claims upon which their argument would rest.

share|improve this answer
add comment

I make an argument that in a situation A, we should do X. X isn't entirely positive, but in situation A the benefits are hugely outweighing the negatives. You then say that in situation A' I would also suggest doing X, even though in situation A' the benefits are a tiny bit less. And then you make the "slippery slope" argument that eventually in situation B, which we get by applying 100 tiny changes to A, I would also argue to do X, even though in situation B doing X would be entirely negative. And because of that "slippery slope", we are not supposed to do X in situation A, where it is clearly beneficial.

If the long-term consequences that Mason claimed are that we might in the future do X in inappropriate situations if we do it today in an appropriate situation, that would be the "slippery slope" fallacy. But in most situations, looking at long-term consequences is nothing like that.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.