Critical Thinking

In today’s Britain, the ability to think logically appears to be under serious threat. This is most apparent when browsing sections of the blogoball* and when listening to our politicians.

Here’s a non-specific example of a particularly common error. Let’s say that A is a defined characteristic or experience and that X is a particular act. We do a study and discover that every single person who commits act X has first conformed to characteristic A. Can we conclude that characteristic A causes act X?

No, we bloody well can’t.

Let’s try a real example. Characteristic A will be defined as those who have smoked cannabis. Act X will be injecting heroin. In that example, studies have shown that almost all injectors of heroin were cannabis smokers first. I’ve seen any number of politicians say that this proves that cannabis is a “gateway drug” but it is a conclusion which cannot be drawn.

Here’s another example. Characteristic A will be defined as those who have smoked cigarettes. Act X will be injecting heroin. In that example, studies have shown that almost all injectors of heroin were cigarette smokers first. Now if you’re a smoker and you think I’m wrong in the previous example, you might be starting to feel unease here. Doesn’t this prove that tobacco is also a “gateway drug”? No, it bloody well doesn’t.

Here’s a third example. Characteristic A will be defined as those who have eaten bread. Act X will be injecting heroin. In that example, studies have shown that almost all injectors of heroin were eaters of bread first. To suggest a causal linkage there would be patently absurd; bread is not a “gateway” to hard drugs. But almost every heroin user was a bread eater first.

I do love a good reductio ad absurdum (not least because it’s just about the only Latin I know). The point is that plugging in a stupid example demonstrates that the original assumption, which still holds true in the ridiculous example, is false. And in this example, the reason is fairly obvious. The vast majority of bread eaters do not go on to inject heroin.

The fact that all people who commit act X first exhibited characteristic A says almost nothing about any causal relationship which might exist between them. In the first example, the “proof” of a causal relationship seems plausible to some people because it confirms their already held view. In the third, we’ll all instantly dismiss it because we all know it’s ridiculous but the same logical “proof” is used in both cases. If one conclusion is valid then the other must be too.

Here’s a final example. Characteristic A will be defined as those who believe in a version of Islam. Act X will be suicide bombing. In that example, studies have shown that almost all suicide bombers believed in a version of Islam first. (Let’s just put aside whether we believe this claim is true or not and assume that it is for the purposes of this post.) Would that prove that a belief in a version of Islam causes suicide bombing?

No, it bloody well wouldn’t.

* I’m dissatisfied with that other word. This is only in improvement in that it isn’t that other word.

  1. Wolfie said:

    You noticed!?

  2. ian said:

    I prefer ‘blogistan’

  3. Robert said:

    Of course I agree with you about the cannabis/heroin., and the Muslim/Bomber. These are variations on Hume and his problem of induction.

    However, there is a difference between iron, irrefutable logic establishing facts, on the one hand; and the degree of empirical evidence we accept as ‘fact’ in order to go about our daily lives, and in order to make decisions to govern.

    Standing in the middle of the road does not mean you’ll be hit by a bus. But nevertheless, you discourage your kids from doing it. So, the employment lollipop ladies is a political decision based on empirical ‘facts’, not absolute facts.

    In the same way, the anti-cannabis lobby is not trying to establish that it always leads to hard drug use. They just point out that (to continue your bread anology above) you don’t buy heroin from the baker.

    In fact, you probably buy it from a dope dealer, and I’ll wager there would be a lot less cannabis-to-heroin cross-over if you could buy a spliff from the newsagents. Yes – Cannabis is a gateway drug, and selling it and taxing legally will change this. The pro-cannabis position rests on inductive reasoning too.

  4. Garry said:

    Robert, I agree.

    What I really wanted to stress was the problem with causation. Empirically, we can be fairly certain that any near-future British suicide bombers will believe in some version of Islam. Many people then conclude that this indicates that belief in some version of Islam is a significant cause of suicide bombing.

    I should perhaps add that the reason I chose the cannabis example was that I saw a politician make the case exactly as I’ve described it here. They were convinced that this was an absolute proof of a causal relationship, that it proved that the consumption of cannabis caused people to then use heroin. Can’t remember who it was, unfortunately.

  5. Garry said:

    Blast, I accidentally posted the previous comment before it was finished.

    This should go before the last paragraph:

    What worries me is that other factors are not being properly considered because so many people think they already know the primary cause.

    Like in my bread and heroin example, the vast majority of the world’s 1.3bn Muslims will not become suicide bombers. Empirically, this does not suggest a strong causal link.

    I believe more time needs to be spent studying the psychology and personality of suicide bombers. Many, it seems to me, are likely to suffer from the same psychological disorders (probably one also shared by perpetrators of school shootings). This won’t produce “proof” either of course but I suspect it will produce useful empirical evidence of the significant causes.

  6. Thought said:

    For starters, I’d say it’s far more salient that most of them appear to be common or garden narcissists. We should treat them like the little pricks they are instead of elevating them to the level of global importance.

  7. I always get a bit confused with the taxonomy of logical fallacies, partly ‘cos they seem to overlap so much depend on whose classifications you’re using, but I think the one you identify here is a combination of the fallacy of the undistributed middle (cats have tails and whiskers; Rover has a tail and whiskers, so Rover must be a cat) and ‘cum hoc, ergo propter hoc’ (with this, therefore because of this).

    Anyway, should anyone be interested I’ve written about it at more length at my place.

  8. G. Tingey said:

    More to the point, they are a specific type of religious believer (islam in this case)

    And all religions are forms of organised blackmail, with, usually morder and torture bolted on at some point.

    At present, 99.9% of terrorists are muslims.
    But, they are less than 0.01% (or an even smaller proportion) of the “muslim population” ……

    Erm ….

  9. Robert said:

    At present, 99.9% of terrorists are muslims

    Is that true though? Think of FARC and the other rebel movements in South America. Think of the Tamil Tigers. Think of the Hindus who blew up a muslim cemetary. And the IRA is still a proscribed organisation, isn’t it?

    That’s before we get into a semantic debate about ‘state terrorism’ which also involves a lot of narcissts and non-muslims…

  10. Monjo said:

    Well Garry your argument is a little silly. There’s the extreme feminist view that because all rapists have a penis, that all people with a penis (men) are rapists. Heck the extreme feminist view was that all sex was rape.

    Going back to the drugs issue. The use of bread has also been used in such silly comparisons with causing deaths. (As has water). Yet even though 100 per cent of people who eat bread will die – they generally live a little longer than those who don’t eat bread.

    So yes hard drug users may go straight from bread to heroin. But most will progress through other drugs – smoking cigarettes at 16; trying cannabis at University… These are clear documented patterns that are more than incidental.

    You are right that critical thinking is under extreme threat in this country. You have none of it. Instead you wish to spout some nonsense about causation without fully understanding the issues.

    Your article is about “Logical Fallacies”, but you succumbed to the ultimate illogical fallacy – that just because something does not hold to be 100 per cent true, it does not make it untrue. So yes a lot of cannabis smokers will not go onto heroin, but as a percentage a much higher percentage of cannabis smokers than non-smokers will make the leap. Hence the notion that cannabis is a gateway drug.

  11. chris said:

    Those corrolations are not the only evidence we have. You can also ask the bombers what they consider their motivations via the often produced statements of responsibility.

    This will almost always be something described in terms of a national liberation struggle; be it for a Tamil homeland, Palestine, Iraq, or Chechnia. The bombers may only have the most tangential relationship to this struggle (such as the British Muslim suicide bombers of Pakistani decent in Isreal) but there is almost always something like that.

    There is also often (but not always) references to Islam. Especially when their connection to whatever liberation movement they are supporting is rather obscure. Such as the British Muslim suicide bombers of Pakistani decent in Isreal to link themselves with the Palestinians. Or the British Muslim suicide bombers in Britian to link themselves with just about every Muslim fight worldwide.

  12. Garry said:

    Monjo, thanks for that interesting view but I think you may have missed the point.

    Consider it this way; what causes someone to take cannabis? Perhaps it is dissatisfaction with their life or an inability to come to terms with the world they live in. Perhaps this same dissatisfaction (probably on a greater scale) is also what causes people to take heroin . Perhaps this is the reason for the relationship between cannabis and heroin consumption. If this is true, the relationship between cannabis and heroin use is not causal; the use of cannabis does not cause people to use heroin. Rather, the same underlying causes are present in both cases. What appears obvious, that the use of cannabis causes some people to then use heroin, is actually obscuring the underlying truth.

  13. Garry said:

    Chris, those are good points. I should again stress that I’m not denying that belief in a version of Islam plays a part in this.

    I’d put it this way. What you’ve described is the end result of a process of radicalisation. What I’m saying is that we should be paying more attention to the underlying causes of that process. To say that Islam is the primary cause, given the fact that the overwhelming majority of Muslims do not become suicide bombers, seems to me to be unsatisfactory.

    BTW, a minor point but I believe the Tamils are mostly Hindus.

  14. chris said:

    The Tamils are mostly Hindu, I included them as the Tamil Tigers have committed by far the most suicide attacks of all the groups mentioned, in what they call the cause of their homeland. However the vast majority (I believe there was some use of suicide by Christian groups during the Lebanese Civil War) of the suicide attacks that are likely to affect anybody outside the Tigers area of operation, such as here in Britain, are by Muslims who will claim to be doing them in part because of Islam.

  15. Garry said:

    Sorry chris, having read that back, I realise that my bit about the Tamils sounded really sarcastic and that I’d missed your point. That wasn’t my intention. (Got a nasty cold and having trouble communicating this week.)

    I see what you’re saying and again, I’m not disputing that these bombers claim to be acting in the cause of Islam. What I’m saying is that we should spend more time studying what causes people take up these causes (if you see what I mean) in this particular way .

  16. Thomas said:

    congratulations on the digg. It was a great post and I’m glad it got all the attention it did.