This world is filled with crazy stuff. I get tempted to write posts about a lot of things. I try to resist most of the time. I figure if I can resist writing about something, it wasn't worth writing about in the first place. Today's topic is one I managed to resist writing about for a while, but I misspoke about it on Twitter so now I feel obliged to. It began with this tweet of mine:
Which was regarding this blog post by William Briggs I saw highlighted in my Twitter feed. The post as a whole bugged me. It criticized a blog post at RealClimate, portraying it as completely incompetent. Briggs was very uninterested in what I had to say, blocking me after just a couple tweets. Presumably that explains why he didn't respond when I specifically pointed out what his misquotation was:
The RealClimate post had said:
They are not confidence intervals on whether a warming has taken place – it certainly has.
But despite quoting that sentence, Briggs wrote:
Second, the author says that a trend “certainly was” in the data.
The difference between "certainly has" and "certainly was" is rather minor. It was just the proverbial straw. Briggs's post contained misrepresentation after misrepresentation, so when he flat-out misquoted someone, I gave up. I have little respect for people who consistently misrepresent what people say in order to belittle it. For instance, when Rahmstorf said:
Rather, these confidence intervals refer to the confidence with which you can reject the null hypothesis that the observed warming trend is just due to random variability
Briggs responded by saying:
He claimed his test was needed to rule out whether the data was caused by (or was “due to”) “random variability“.
But this is an misrepresentation. Rahmstorf clearly referred to "the observed warming trend" not "the data" when he discussed random variability. Briggs uses this obvious misrepresentation to then say:
This term is nonsensical. It quite literally has no meaning.
I don't think Briggs is a person who should be listened to in determining what "is nonsensical" if he's going to misquote the person he criticizes and misrepresent what the guy is talking about in obvious ways. This is especially true when much of Briggs's post is based upon nothing but rhetoric. The next part is:
Randomness, as I’ve said thousands of times, cannot cause anything. Instead, something caused each and every temperature datum to take the value it did.
Again, Briggs is talking about the data, not the trend. That's not the important part though. The important part is: What in the world is he smoking?!
Technically speaking, nothing is random.* Everything has a cause. When you flip a coin, you do not have a "random" chance of it landing on heads or tails. The coin will land on heads or land on tails, depending upon a variety of factors like the height from which you flip it, the force your thumb applies to the coin, the angle that force is applied at, the amount and direction of airflow in the area the coin passes through, the amount of air pressure, the temperature of the area, the elevation you are standing at (affects gravity), any surface the coins hits, and hundreds of other factors.
But who in the world would think that's a meaningful criticism of what Stefan Rahmstorf said? If I'm playing a game with a friend and say, "Let's decide who goes first randomly" and pull out a coin, who would respond with a tirade about how nothing is random? Nobody. Everybody would understand what I meant, just like everybody would understand what Rahmstorf meant.
I tried to highlight this issue on Twitter, but I was using a phone, and I did a terrible job of explaining the nuance in 140 characters:
That tweet is horribly unclear. That's why I felt obligated to write this post. The point I was trying to make is Rahmstorf began the ridiculous sidetrack by saying "caused by (or was 'due to')." These phrases are generally interchangeable. However, switching from "due to" to "caused by" allowed Briggs to pretend it made sense to say, "Randomness... cannot cause anything."
I'll demonstrate. I ran out of milk this weekend. Due to that, I went to the grocery store today. One could fairly say, "My trip to the grocery store was caused by my lack of milk."
Only, Briggs would then come along and say, "No it wasn't. Your desire could only cause you to think about how much you wanted milk, which could make you think about the time it would take to go to the store, which could make you..." and continue on and on for a few hours explaining exactly what chain of causality existed.
The switch from "due to" to "caused by" allowed Briggs to be hyper-literal and pretend Rahmstorf was saying something ridiculous in order to dismiss it. That let him then write things like:
Don’t skimp your thinking on this. Prove to yourself how bizarre the author’s notion is.
Which one can only do by intentionally ignoring the obvious meaning the author had in mind.
I don't know that somebody writing such ridiculous things merits a blog post. If so, there are probably other parts of William Briggs's post I should have focused on instead. He makes a number of claims about statistics and mathematics which are just as wrong. There might at least might be some value in clarifying what's wrong about them since they don't rely upon obvious semantic tricks anyone should be able to spot. Still, I tried and failed to explain this issue so I feel it's the one I ought to focus on.
*This statement isn't provably true. There are parts of our universe where we cannot know there is a particular cause to an observed effect. Quantum mechanics is a prime example, as is the human consciousness. There could be randomness in either, or there could just be some deeper pattern we are unaware of. Because of uncertainty like that, in fields like cryptography, "random" is understood to not be literally random. It merely means "random enough we cannot discern a pattern." That, of course, fits perfectly with how people commonly view "random" in regard to things like coin flips. It also rebukes Briggs's argument by showing "randomness" is often used to merely refer to a set of factors so complicated we can't discern whatever pattern there is.