4 ms·

The author links to a much better critique of Pinker by Nassim Taleb. The basic idea is that Pinker uses data about very thin-tailed processes, like street crim

by lbarrow 12y ago

The author links to a much better critique of Pinker by Nassim Taleb. The basic idea is that Pinker uses data about very thin-tailed processes, like street crime, to make inferences about very fat-tailed processes, like casualties in war.

a8da6b0c91d 12y ago
Some have successfully picked apart Pinker's numbers even with regard to street crime. Murder rates have fallen because of amazingly improved trauma care, not a decline in assault and battery. Beatings and stabbings that were lethal 50 years ago are now routinely survived.
- leroy_masochist 12y ago
  That's an important point. For a more detailed discussion of it, I recommend Dave Grossman's On Killing. He makes a convincing argument that the long-term downward trend in the U.S. murder rate is a function of faster trips to the hospital (due to better first responder protocols), and better care once in hospital. An examination of the combined total of reported aggravated assaults and murders, according to Grossman, implies that we're getting more violent over time.
  - schoen 12y ago
    So Grossman's theory is that it's just plain harder to kill people nowadays, not any particular lack of trying to do so?
    - leroy_masochist 12y ago
      His theory is that many attempted murders are prosecuted as aggravated assaults, and that advances in various medical best practices (care delivered on ambulance, emergency surgery, etc) have increased the survival rate of people who have been victims of assault, thus making the total number of people who die from murder decrease as a percentage of the population....but the actual rate of people trying to kill each other has increased.
  - kahirsch 12y ago
    > He makes a convincing argument that the long-term downward trend in the U.S. murder rate is a function of faster trips to the hospital (due to better first responder protocols), and better care once in hospital. There was a disconnect between murder rate and aggravated assault rate from the mid-1970s to about 1990, but before then and after then, the two rates were highly correlated. Graph: http://i.imgur.com/bVG4RHm.png http://i.imgur.com/bVG4RHm.png
  - njharman 12y ago
    When your stats are reported rates. You can't make claims about actual rates. could be we are just reporting more violence. Cause better record keeping.less lynchings. Which knows.
  - Gibbon1 12y ago
    Yeah but that doesn't explain the decline in forcible rape in the US from 1970 to present.
- brownbat 12y ago
  The "improving trauma care" can't explain why there have been declines in categories of crime across the board,[1] or why murder rates spiked almost as steeply from the 50s to the 90s (while trauma care was also improving) as they declined afterwards.[2] Vehicle theft rates are less than a fourth what they were in 1990. Burglary less than half. Aggravated assault is down 40%. Vandalism has been cut in half. Fraud is less than a fourth of what it was in 1995. Most notably, homicide rates track really closely with violent crime rates overall.[2] You can say BJS is biased and collecting terrible stats, but this would have to be a new bias, since we're comparing their stats to earlier results. And they at least try to cross check trends across different methodologies, and publicize their methods for critics to scrutinize. Their approach would make it likely they'd notice some obvious trend (that would only affect death rates) that was skewing their stats.[3] Also, other data sources seem to confirm their findings, including those from local police.[4] It could be that many factors are conspiring to distort these statistics, each one covering for the other. The amount of other factors that would have to be messing with the statistics simultaneously though make the trauma care hypothesis seem a likely target for occam's razor. Crime rates absolutely did fall. Every criminologist and even several economists are scrambling to solve the puzzle, and no one has definitively succeeded. It's a real trend driving everyone to study it, and there's no "gotcha" explanation for it, it's going to be a product of several complex factors. [1] http://www.bjs.gov/index.cfm?ty=datool&surl=/arrests/index.cfm http://www.bjs.gov/index.cfm?ty=datool&surl=/arrests/index.c... [2] http://www.krusekronicle.com/kruse_kronicle/2012/12/a-historical-perspective-on-violence-in-the-united-states.html http://www.krusekronicle.com/kruse_kronicle/2012/12/a-histor... [3] https://www2.fbi.gov/ucr/cius2009/about/crime_measures.html https://www2.fbi.gov/ucr/cius2009/about/crime_measures.html [4] http://www.nyc.gov/html/nypd/html/crime_prevention/crime_statistics.shtml http://www.nyc.gov/html/nypd/html/crime_prevention/crime_sta...
  - tim333 12y ago
    >Crime rates absolutely did fall. Every criminologist and even several economists are scrambling to solve the puzzle I blame the internet. They're all sitting about on facebook and twitter when they should be out there stealing.
drited 12y ago
Pinker's comments on Taleb's critique are worth a read, available here: http://stevenpinker.com/files/pinker/files/comments_on_taleb_by_s_pinker.pdf http://stevenpinker.com/files/pinker/files/comments_on_taleb...
- oskarth 12y ago
  ...Which Taleb in turn responded to in the (updated) version of the rebuttal: Pinker’s Rebuttal of This Note: Pinker has written a rebuttal (ad hominem blather, if he had a point he would have written something 1/3 of this, not 3 x the words). He still does not understand the difference between probability and expectation (drop in observed volatility/fluctualtion ≠ drop in risk) or the incompatibility of his claims with his acceptance of fat tails (he does not understand asymmetries-- from his posts on FB and private correspondence). Yet it was Pinker who said “what is the actual risk for any individual? It is approaching zero”.
  - mjfl 12y ago
    Nassim Taleb has a weak habit of ascribing every argument against him to an ad hominem attack. Taleb does have a point, war could very well become less frequent but more deadly. If the current nuclear stockpile stays where it's at, the next major war could kill all of humanity. However I think Pinker's argument is, barring that, life does seem to be getting better, and I think his argument there is valid. In his chapter "The Long Peace", he goes over how there has never been a period in history where the 500 lb gorillas of the world have been without war in a long time.
    - curun1r 12y ago
      > life does seem to be getting better, and I think his argument there is valid. I think anyone who's protested or supports the #BlackLivesMatter movement might quibble with that conclusion. Making conclusions like that based off crime data is specious if violence is in anyway systemic. Violence doesn't have to be murders, assaults and robberies...incarceration is its own form of violence and incarceration rates are as high as ever. Poverty is its own form of violence and income inequality is as high as ever. And civil asset forfeiture is basically legalized robbery, so if you included those statistics along with reported robbery statistics, you'd probably find that robberies have actually increased. What I see happening is that some (rich) people have realized that nationalistic bonds no longer tie us together as strongly as they used to. They've been replaced by economic and religious bonds. So yes, the leaders of the largest countries in the world have largely avoided war against each other. But the top stratum of these countries has been very active in colluding against the bottom strata (G8, WTO, TPP, etc). They've largely waged war against the lower classes in all countries for many years now without most people realizing it. Just because the violence is perpetrated under the guise of law enforcement does not change the fact that it's still violence. And we still have "official" (not declared, but acknowledged) wars against religious extremist groups (Al Qaeda, Isis, etc), so there's that too. In short, the nature of violence has changed, not the prevalence of it. That you don't realize it only reflects that you're not (yet) a target.
ArkyBeagle 12y ago
War had this massive technological increase in violence from about 1850 to the lats 20th century. That explains pretty much it all to my satisfaction. The thing that is missed by these critiques of Pinker seem to be that Determinist genetic theories, which caused much of the "Hemoclysm" held sway for a long time, really until the advent of DNA.So once a better tool came on line, the justification for the violence faded. Of course, this has been uneven; so we have Rwanda. To me that's sort of the exception that proves the rule. It's tribal and unmodern violence. I think all that's required to see how far we've come it so read Kipling - especially "My Boy Jack". There's also a stage play and a "movie" of the stage play. The horror of the end implication of the Edwardian sense of manly military honor is brought into sharp relief.

by lbarrow 12y ago

a8da6b0c91d 12y ago

Some have successfully picked apart Pinker's numbers even with regard to street crime. Murder rates have fallen because of amazingly improved trauma care, not a decline in assault and battery. Beatings and stabbings that were lethal 50 years ago are now routinely survived.

leroy_masochist 12y ago
That's an important point. For a more detailed discussion of it, I recommend Dave Grossman's On Killing. He makes a convincing argument that the long-term downward trend in the U.S. murder rate is a function of faster trips to the hospital (due to better first responder protocols), and better care once in hospital. An examination of the combined total of reported aggravated assaults and murders, according to Grossman, implies that we're getting more violent over time.
- schoen 12y ago
  So Grossman's theory is that it's just plain harder to kill people nowadays, not any particular lack of trying to do so?
  - leroy_masochist 12y ago
    His theory is that many attempted murders are prosecuted as aggravated assaults, and that advances in various medical best practices (care delivered on ambulance, emergency surgery, etc) have increased the survival rate of people who have been victims of assault, thus making the total number of people who die from murder decrease as a percentage of the population....but the actual rate of people trying to kill each other has increased.
- kahirsch 12y ago
  > He makes a convincing argument that the long-term downward trend in the U.S. murder rate is a function of faster trips to the hospital (due to better first responder protocols), and better care once in hospital. There was a disconnect between murder rate and aggravated assault rate from the mid-1970s to about 1990, but before then and after then, the two rates were highly correlated. Graph: http://i.imgur.com/bVG4RHm.png http://i.imgur.com/bVG4RHm.png
- njharman 12y ago
  When your stats are reported rates. You can't make claims about actual rates. could be we are just reporting more violence. Cause better record keeping.less lynchings. Which knows.
- Gibbon1 12y ago
  Yeah but that doesn't explain the decline in forcible rape in the US from 1970 to present.
brownbat 12y ago
The "improving trauma care" can't explain why there have been declines in categories of crime across the board,[1] or why murder rates spiked almost as steeply from the 50s to the 90s (while trauma care was also improving) as they declined afterwards.[2] Vehicle theft rates are less than a fourth what they were in 1990. Burglary less than half. Aggravated assault is down 40%. Vandalism has been cut in half. Fraud is less than a fourth of what it was in 1995. Most notably, homicide rates track really closely with violent crime rates overall.[2] You can say BJS is biased and collecting terrible stats, but this would have to be a new bias, since we're comparing their stats to earlier results. And they at least try to cross check trends across different methodologies, and publicize their methods for critics to scrutinize. Their approach would make it likely they'd notice some obvious trend (that would only affect death rates) that was skewing their stats.[3] Also, other data sources seem to confirm their findings, including those from local police.[4] It could be that many factors are conspiring to distort these statistics, each one covering for the other. The amount of other factors that would have to be messing with the statistics simultaneously though make the trauma care hypothesis seem a likely target for occam's razor. Crime rates absolutely did fall. Every criminologist and even several economists are scrambling to solve the puzzle, and no one has definitively succeeded. It's a real trend driving everyone to study it, and there's no "gotcha" explanation for it, it's going to be a product of several complex factors. [1] http://www.bjs.gov/index.cfm?ty=datool&surl=/arrests/index.cfm http://www.bjs.gov/index.cfm?ty=datool&surl=/arrests/index.c... [2] http://www.krusekronicle.com/kruse_kronicle/2012/12/a-historical-perspective-on-violence-in-the-united-states.html http://www.krusekronicle.com/kruse_kronicle/2012/12/a-histor... [3] https://www2.fbi.gov/ucr/cius2009/about/crime_measures.html https://www2.fbi.gov/ucr/cius2009/about/crime_measures.html [4] http://www.nyc.gov/html/nypd/html/crime_prevention/crime_statistics.shtml http://www.nyc.gov/html/nypd/html/crime_prevention/crime_sta...
- tim333 12y ago
  >Crime rates absolutely did fall. Every criminologist and even several economists are scrambling to solve the puzzle I blame the internet. They're all sitting about on facebook and twitter when they should be out there stealing.

drited 12y ago

Pinker's comments on Taleb's critique are worth a read, available here: http://stevenpinker.com/files/pinker/files/comments_on_taleb_by_s_pinker.pdf http://stevenpinker.com/files/pinker/files/comments_on_taleb...

oskarth 12y ago
...Which Taleb in turn responded to in the (updated) version of the rebuttal: Pinker’s Rebuttal of This Note: Pinker has written a rebuttal (ad hominem blather, if he had a point he would have written something 1/3 of this, not 3 x the words). He still does not understand the difference between probability and expectation (drop in observed volatility/fluctualtion ≠ drop in risk) or the incompatibility of his claims with his acceptance of fat tails (he does not understand asymmetries-- from his posts on FB and private correspondence). Yet it was Pinker who said “what is the actual risk for any individual? It is approaching zero”.
- mjfl 12y ago
  Nassim Taleb has a weak habit of ascribing every argument against him to an ad hominem attack. Taleb does have a point, war could very well become less frequent but more deadly. If the current nuclear stockpile stays where it's at, the next major war could kill all of humanity. However I think Pinker's argument is, barring that, life does seem to be getting better, and I think his argument there is valid. In his chapter "The Long Peace", he goes over how there has never been a period in history where the 500 lb gorillas of the world have been without war in a long time.
  - curun1r 12y ago
    > life does seem to be getting better, and I think his argument there is valid. I think anyone who's protested or supports the #BlackLivesMatter movement might quibble with that conclusion. Making conclusions like that based off crime data is specious if violence is in anyway systemic. Violence doesn't have to be murders, assaults and robberies...incarceration is its own form of violence and incarceration rates are as high as ever. Poverty is its own form of violence and income inequality is as high as ever. And civil asset forfeiture is basically legalized robbery, so if you included those statistics along with reported robbery statistics, you'd probably find that robberies have actually increased. What I see happening is that some (rich) people have realized that nationalistic bonds no longer tie us together as strongly as they used to. They've been replaced by economic and religious bonds. So yes, the leaders of the largest countries in the world have largely avoided war against each other. But the top stratum of these countries has been very active in colluding against the bottom strata (G8, WTO, TPP, etc). They've largely waged war against the lower classes in all countries for many years now without most people realizing it. Just because the violence is perpetrated under the guise of law enforcement does not change the fact that it's still violence. And we still have "official" (not declared, but acknowledged) wars against religious extremist groups (Al Qaeda, Isis, etc), so there's that too. In short, the nature of violence has changed, not the prevalence of it. That you don't realize it only reflects that you're not (yet) a target.

ArkyBeagle 12y ago

War had this massive technological increase in violence from about 1850 to the lats 20th century. That explains pretty much it all to my satisfaction. The thing that is missed by these critiques of Pinker seem to be that Determinist genetic theories, which caused much of the "Hemoclysm" held sway for a long time, really until the advent of DNA.So once a better tool came on line, the justification for the violence faded. Of course, this has been uneven; so we have Rwanda. To me that's sort of the exception that proves the rule. It's tribal and unmodern violence. I think all that's required to see how far we've come it so read Kipling - especially "My Boy Jack". There's also a stage play and a "movie" of the stage play. The horror of the end implication of the Edwardian sense of manly military honor is brought into sharp relief.