Thursday, 19 September 2013

More on Speeding

     A video is making the rounds, supposedly pulling the rug out from under the "Speed Kills" campaigns of the insurance industry, police, and the media. Essentially the argument is that speed limits are too low in many places, and that it's perfectly safe to drive at higher speeds and the only reason police enforce these limits the way they do is to generate revenue, not to promote safety. The insurance companies, it is alleged, are also making money from this because they get to raise premiums when someone gets a speeding ticket, while media is just lazily repeating the "speed kills" rhetoric because it makes a better story.

     Where to begin? Well, let's start with the basic claim, that speed doesn't kill. No, of course it doesn't. It's an abrupt change in speed that kills, such as the sort that arises when two objects of greatly differing vectors collide. And as we know from our basic high school physics, kinetic energy is proportion to the square of the velocity, which means if you double the speed, you quadruple the energy. It takes energy to crumple fenders and bones, so the greater the speed, the greater the capacity to do damage. This much is obvious, but that kinetic energy formula is important in another way, which I'll get to in a moment. The point here is that actually, yeah, higher speeds are more dangerous, and in a non-linear way: going 10% faster is more than 10% more dangerous.
     To be fair, that's not really the point of the video, which argues that higher speeds (on highways or other roads that can handle higher speeds) do not necessarily increase the likelihood of a collision. Well,  maybe not, but even if the chance of a collision remains constant (and it would be bizarre to claim that higher speed limits reduce the chance), the fact remains that higher speeds do more damage.

     This, by the way, answers part of another one of the claims of the video, namely that insurance companies love for people to get speeding tickets because it gives them an excuse to raise premiums. Actually, and actuarially, insurance premiums are supposed to be calculated to reflect the expected cost of insurance. Insurance companies have limited information upon which to make accurate estimates of the likelihood that a given customer will be responsible for a claim, and how much of a claim. Knowing that a driver has earned a speeding ticket reveals a little bit of data about their driving habits which allows for somewhat more precise estimates. Now, I'm the last person to trust in the selfless integrity of the insurance industry, and I haven't actually researched the actuarial data here, but it seems likely to me that at least part of the reason your insurance rates go up when you get a speeding ticket is because statistically speaking, you are a bigger risk.

     The video makes a similar argument about police departments making money from speeding tickets, and while I agree that it's problematic to introduce a profit motive to law enforcement, it's worth noting that the police are still not actually a for-profit organization; the money governments raise from traffic enforcement doesn't exactly go to line the pockets of shareholders.
     And yes, police use stealth traps to catch speeders, when the visibility of a police car is much more effective at getting people not to speed. So what? Is this supposed to show that speed traps are aimed at making money rather than promoting safe driving? Nonsense. It's easy to get people to obey the law when there's a cop around. The point of speed traps is to encourage drivers to obey speed limits when they don't see a cop. Sorry, I shouldn't need to explain that, but apparently I do.

     But my main beef with this video is that it misses the main point of speed limits, which as I've written before, aren't just (or even mostly) about the state deciding how fast you can be trusted to stay in control of your vehicle. Cruising along in your private vehicle, it's easy to feel free and independent and lose sight of the fact that our public roads and highways are part of a complete system of transportation, and that they are a shared public resource. We hire traffic engineers to design and manage this system for us, and we want them to make it serve our needs as efficiently as possible. So they do things like place stop signs and traffic lights and turning lanes and set speed limits and so on with the goal of minimizing the time it takes to make a typical trip. In other words, speed limits are meant to speed up traffic. Let me say that again in a separate line all by itself for emphasis.

     Speed limits are meant to speed up traffic.

     That sounds counterintuitive, I know, but it's true. If you thought the purpose of speed limits was to reduce accidents, well, sure, that's nice, but considering that even very minor accidents create huge delays and reduce average speeds by a ridiculous amount (and accidents where there are injuries or deaths even more so), it's enough to say minimizing accidents is just a good way for system engineers to minimize delay.

     Time for a simple chart.



     Obviously, if the enforced speed limit is zero, traffic's not going to be moving very efficiently. As we increase the enforced limit, the effective speed of traffic increases, but only up to a point. At that point, higher speed limits can actually decrease the effective speed of traffic. Why?
     In an ideal world, where everyone is zipping along at a uniform speed, it doesn't really matter what the maximum is so long as everyone is going close to the same speed. Yes, we need to worry about stationary obstacles, but presumably an ideally-built highway will be free of obstacles.
     But this ideal world is sort of like the ideal universe of basic high school physics, where friction and air resistance are ignored. In the real world of driving, cars do not simply travel along at uniform speeds in happy obedience to Newton's First Law; they are generally moving because the people in them are travelling from point A (where they presumably were at rest for a while) to point B (where they presumably intend to stop for a bit). So unavoidably, there's going to be some acceleration involved somewhere along the way. Moreover, since not everyone has the same points A and B, there's going to be some merging of traffic along the way also, which means that not everyone is going to be travelling at the same speed. And sometimes people are going to need to stop in places they didn't intend to stop, such as by the side of the highway.
    So in the real world, there will be a range of different speeds we have to accommodate, and unfortunately one of those speeds will be zero. Now, traffic can cope with a range of speeds, especially if there are multiple lanes. But changing from a slower lane to a faster one involves acceleration, and remember that whole kinetic energy equation? Among other things, it means that accelerating from 90 km/h to 100 km/h takes almost twice as much energy as accelerating from 50 km/h to 60 km/h. So lane changes and passing are going to be more demanding, and you'll need a bigger break in traffic to be able to execute them safely. That means that if you happen to be stuck in the slower lane, you don't actually benefit from the fact that the faster lane is really really really fast.

     To put it another way, road space is a scarce public resource, and you use more of it when you go faster, leaving less available for others to use. Speed limits are an attempt to distribute that resource equitably, so everyone can enjoy the benefits of a public roadway. People who drive faster may not recognize it, because the delays they cause to other people are barely visible in their rear-view mirror, but they are slowing everyone else down, even if they never actually cause an accident. Oddly enough, a lot of the time they complain about being stuck behind some inexplicable delay in traffic, it's actually due to the ripple effects of some other selfishly impatient driver.

     Are some speed limits too low? Possibly. But not for the reasons given in the video. Want to get where you're going faster? Drive at a reasonable speed, leave a reasonable gap ahead of yourself for other vehicles to change into and out of your lane, and be a patient, strategic driver. Let the speed limits do their job.


Thursday, 5 September 2013

Bring on the Jurassic Salad!

     I used to talk about how to me, the scariest part of Jurassic Park had nothing to do with the dinosaurs. I mean, let's be honest here: Homo sapiens is the scariest killing machine on the planet. We don't have big teeth or claws, but we are pretty clever with the use of tools and terrain, and we have language and can use it to plan and execute coordinated operations with terrifying effectiveness. This deceptively wimpy bipedal primate hunted mammoths, aurochs and moas to extinction. We went after sperm whales, fercryinoutloud! A tyrannosaur gets loose from a zoo? Bah. It might eat a few people before we can mobilize an attack helicopter to take it out, but we probably wouldn't even need to do that; a resourceful bunch of humans on the ground would probably figure something out by the time the  pilot finished his preflight checks. (Yes, I know the velociraptors were supposed to be the really scary, smart ones, but seriously, they'd be hunted to extinction in a matter of months, and more people would die in car accidents driving to the store to buy anti-velociraptor gadgets than would be eaten by velociraptors.) Sure, scary to the people on the ground actually being chased by dinosaurs, but not especially more so than most of the countless other ways each and every one of us will face death.

     No, to me, the scariest part was the scene just before we actually saw the dinosaurs, when the paleobotanist played by Laura Dern is marvelling over a fern she identifies as having been extinct for millions of years. I thought that fern was the most dangerous thing in the whole movie.
    Why? Because over time, invasive plant species can do way more harm to a lot more people than a couple of hungry carnivores, and while we humans are unchallenged apex predators who excel at killing other predators to the point that we have to rein ourselves in so we can still have things to kill, we totally suck at controlling little things like weeds. The economic damage caused by kudzu, water hyacinth, and "harmless" little critters like zebra mussels and carp and pine beetles adds up year by year by year, and there's very little we seem to be able to do about it.
     So the Jurassic Park geneticists revive a fern that's been extinct for millions of years, and whose natural predators are, presumably, still extinct. Oh crap. What's going to eat it? What's going to keep it from growing everywhere, including places where we're trying to grow corn or rice or potatoes?

     But then I happened to see this video, "Why aren't all plants poisonous?" in which the answer to the question is actually that all plants are poisonous, but we creatures who feed on plants keep evolving counters to the toxins the plants come up with.  I agree with the analysis, subject to some qualifications: some plants benefit from the action of hungry animals, and thus make delicious fruit or nectar to encourage them. Also, some plants, like thistles and cacti, may rely more on physical deterrents against predation than chemical ones.  (I always suspected that thistles must be tasty, or they wouldn't devote so much metabolic effort to growing those prickles. I first checked a couple of books to confirm they weren't toxic, and then picked some thistle from my weedy backyard, and trimmed the prickles from the leaves. Turns out, they're actually quite palatable!)

     Anyway, it got me to thinking: this arms race of plant toxins vs. evolved herbivore resistance to those toxins has kept on going ever since that fictional fern in Jurassic Park went extinct. That fern may have had state-of-the-art chemical defenses against the herbivores large and small of the late Cretaceous, but the art has evolved considerably since then. It may well be that many of the genetic tricks that herbivores developed to feed on the extinct plant have been retained and refined in the genome of modern animals (including us), particularly if that plant's surviving relatives have continued to use and refine those defenses.

     So maybe, I was wrong, and rather than deathweed juggernauts blacking out the sun, the reanimation of extinct plant species would instead lead to delicious and healthy "new" foods for us to eat. Either that, or they'll be completely ungrowable because every agricultural pest and microbe will eagerly devour them before we can.

Monday, 2 September 2013

Consciousness and Free Will

Some time ago, I read a book called My Brain Made Me Do It, in which the author Eliezer Sternberg (at that time an undergraduate in philosophy and pre-med) talked about some implications of neuroscience he found profoundly disturbing. He seemed most upset by the results of some interesting work using functional MRI to track brain activity while test subjects were asked to perform certain tasks.

Subjects were asked to perform a very simple task: press a button when they felt like doing so. The experiment was intended to piece together the sequence in which various mental events take place, and so the subjects were also asked to make a note of exactly when they decided to press the button, by noting the position of a dot moving regularly in a circle on a screen in front of them.

Now, it turns out that the moment the subjects thought they were deciding to press the button was actually quite a few milliseconds later than the fMRI showed activity in the part of the brain where the choice was actually made. In other words, the conscious experience of choosing was considerably later than the choice itself. The author of this book found that very disturbing with respect to free will, because he felt that a choice made some circuit in one’s brain that lies outside of the actual seat of consciousness was not genuinely a free choice. We do not hold people responsible for epileptic fits or other brain phenomena, after all; doesn’t the discovery that none of our choices are conscious completely do away with the notion of personal responsibility?

I wanted to like this book, because it was engagingly written and earnest, but I kept wanting to shout at it because it was based on an unfounded assumption about the role of consciousness, that it’s the seat of autonomy and choice, the place from which the body is controlled, where “we” as individuals ultimately live. That’s not how I understand consciousness at all, based in part upon my reading of books like Marvin Minsky’s Society of Mind and Daniel Dennett’s Consciousness Explained.

We know now that the brain is made up of a whole lot of specialized agents, little programs that are good at doing one thing and one thing only. There’s a bit of your brain optimized for recognizing the letter “Q”, and one for using chopsticks, and another for remembering that gasoline is flammable. Most of these little components are wired up to other components that perform related tasks, so they can share information as needed; your “Q”-recognizer talks to your word readers that recognize words with “Q” in them, letting them know if there’s a “Q” in the visual field, and keeping quiet if there isn’t.

Now, these agents don’t usually have direct connections to other agents if there’s no frequent reason for them to talk to each other. The “Q”-recognizer needs to be linked to your visual cortex, but rarely if ever needs to share information with the part of your brain that knows how to sing the tune to “Happy Birthday”. That’s where consciousness comes in.

Consciousness is just another part of the brain, but its specialized job is to serve as a kind of public bulletin board to which all the other parts of the brain have access, allowing them to share information they wouldn’t normally be able to share. You can see how this works by catching yourself when it doesn’t work. Example: I had just got out of the ground-floor shower (I usually use the shower in our basement), and was thinking about something else (that is, the bulletin board of my consciousness was busy sharing some other kind of information) when I suddenly realized I had just kicked my dirtly clothes into the closet. A habitual behaviour (kicking the dirty clothes down the laundry chute) that usually operates without the need for conscious oversight had encountered a problem, and posted the error message on my bulletin board: “Something went wrong trying to kick dirty laundry down chute.” The instant this was posted, and thus made available to all parts of my brain, two other brain elements immediately made relevant reports: “The laundry chute is located on the ground floor of your parents’ house”, and “You are currently in your own house”. (Yet another brain element added: “Dumbass!”)

You have probably had similar experiences, where you just did something on autopilot that, if you had stopped to think about it consciously, you would have immediately recognized that you were making a mistake. You probably could have avoided the error if you had been consciously paying attention to what you were doing, but consciousness is a limited resource. If your consciousness is busy sharing information about which house you’re in and what dirty-laundry-protocol to follow, it’s probably not sharing information between the part of  your brain that remembers what’s in the fridge and the part that’s trying to plan what to cook for dinner. But all these parts of your brain are still part of you, even if they aren’t always (or ever, even) starring in the spotlight of consciousness.

Yet there is something to the moral intuition that we don’t hold an epileptic morally accountable for the actions triggered by brain events during a seizure. So how do we draw a meaningful moral distinction between those unconscious brain events and the ones for which we are responsible?

I would suggest that the answer lies not in whether or not the decision is a conscious one (because no decisions are truly conscious in themselves, even if they are immediately reported through consciousness), but whether or not the decision could be influenced by the information content of consciousness. Let me demonstrate with two examples.

First, take a deep breath. Actually, it doesn’t matter if you take a deep breath or not. The decision to do so is yours, and you may decide not to, but at a minimum, the idea of taking a deep breath is now posted in your consciousness bulletin board for all the other bits of your brain to see and work on. There is probably a part resisting: “Wait, why should we take a breath just because someone told us to?” Another part is saying, “Breathing deeply once in a while is good for you anyway! Do it!” Lots of other little bits are piping up with various arguments for or against, while your poor motor control brain segment has been on alert for the decision to proceed or abort with Operation Deep Breath since you first saw the sentence. Whatever your decision element decides, you will become conscious of your decision mere milliseconds later as it is posted to the bulletin board.

Now, you can and do breath unconsciously, most of the time in fact. The part of your brain that controls breathing usually doesn’t bother consciousness with it, and doesn’t require any special information from consciousness to keep doing its job. But it can modify its behaviour in response to such information, such as “Don’t inhale now! You’re under water!” The fact that it can be affected by conscious information is what qualifies taking a deep breath as our act, even though it may not always (and usually isn’t) done under conscious control.

Second example: Sneeze, right now. Now, again, it doesn’t matter if you did just happen to sneeze as you were reading this, because odds are that if you did it’s because you already needed to sneeze anyway. Sneezing is, for most people at least, a reflex action that is driven entirely by physiological stimuli, over which the contents of your consciousness have almost no direct influence. Yes, you can take deliberate actions that will indirectly lead to your sneezing (such as inhaling ground pepper or snuff), but the act of will there is “inhale pepper”; the sneeze is at best a consequence of your volitional act, rather than your own act. And yet sneezing is certainly something that takes place using muscles and organs you generally control, and even involves brain events in the very brain whose decisions you are said to be responsible. 

It’s possible, of course, that you’ve trained yourself somehow to establish some neural link between your consciousness and the sneeze reflex, and maybe you can sneeze at will, in which case I’ve just chosen a bad example for you. For me, at least, the sneeze reflex pays no attention to the content of consciousness, and so it would be inappropriate to attach moral blame or praise to me for my “decision” to sneeze, while it might be perfectly appropriate to hold be responsible for a decision to take or not to take a deep breath. 

To me, then, the fact that decision-making may take place pre-consciously does not in any way raise problems for practical free will. It is a mistake to identify ourselves too closely with only the conscious part of our brains. Our minds are the emergent phenomena of complex networks of distinct brain subsystems, not merely some tiny bit that lives in one little corner, pulling the strings.