Wednesday, March 21, 2007

The robotic ape: Morality and the ventromedial prefrontal cortex

Persons with damage to the ventromedial prefrontal cortext behave more like Mills and less like Kant (emphases mine):
Study Finds Brain Injury Changes Moral Judgment - New York Times

... findings, published online by the journal Nature, confirm the central role of the damaged region — the ventromedial prefrontal cortex, which is thought to generate social emotions, like compassion.

Previous studies showed that this region was active during moral decision-making, and that damage to it and neighboring areas from severe dementia affected moral judgments. The new study seals the case by demonstrating that a very specific kind of emotion-based judgment is altered when the region is offline. In extreme circumstances, people with the injury will even endorse suffocating an infant if that would save more lives.

“I think it’s very convincing now that there are at least two systems working when we make moral judgments,” said Joshua Greene, a psychologist at Harvard who was not involved in the study. “There’s an emotional system that depends on this specific part of the brain, and another system that performs more utilitarian cost-benefit analyses which in these people is clearly intact.”...

...The new study focused on six patients who had suffered very specific damage to the ventromedial area from an aneurysm or a tumor. ...The area in adults is about the size of a child’s fist.

People with this injury can be lucid, easygoing, talkative and intelligent, but blind to subtle social cues, making them socially awkward. They also have some of the same moral instincts that others do.

... All three groups also strongly rejected doing harm to others in situations that were not a matter of trading one certain death for another. They would not send a daughter to work in the pornography industry to fend off crushing poverty, or kill an infant they felt they could not care for.

But a large difference in the participants’ decisions emerged when there was no switch to flip — when they had to choose between taking direct action to kill or harm someone (pushing him in front of the runaway boxcar, for example) and serving a greater good.

Those with ventromedial injuries were about twice as likely as the other participants to say they would push someone in front of the train (if that was the only option), or to poison someone with AIDS who was bent on infecting others, or suffocate a baby whose crying would reveal to enemy soldiers where the subject and family and friends were hiding...

...The ventromedial area is a primitive part of the cortex that appears to have evolved to help humans and other mammals navigate social interactions. The area has connections to deeper, unconscious regions like the brain stem, which transmit physical sensations of attraction or discomfort; and the amygdala, a gumdrop of neural tissue that registers threats, social and otherwise. The ventromedial area integrates these signals with others from the cortex, including emotional memories, to help generate familiar social reactions.

... This tension between cost-benefit calculations and instinctive emotion in part reflects the brain’s continuing adjustment to the vast social changes that have occurred since the ventromedial area first took shape. The ventromedial area most likely adapted to assist the brain in making snap moral decisions in small kin groups— to spare a valuable group member’s life after a fight, for instance. As human communities became larger and increasingly complex, so did the cortical structures involved in parsing ethical dilemmas. But the more primitive ventromedial area continued to anchor it with emotional insistence an ancient principle: respect for the life of another human being.
It's hard not to wonder what a similar study would find find on adult genetic relatives of children with autism. The study is too small to be persuasive on its own, but it's just another in a flurry of recent research that hammers home the reality that we are our brains, and that we have more in common with science fiction robots than we once fancied. Pull out our emotion chip and we switch to a relatively "cold-blooded" judgment system. Disable that, and you probably get a sociopath. I would wager that a successful sniper has a relatively inactive VPC.

Eventually we'll discard the illogical concept of 'individual responsibility'. I wonder if I'll live long enough to see what will replace it ...

Science and faith in the senate

Emphases mine.
BBC NEWS | Americas | Al Gore makes global warming plea:

...Representative Joe Barton, the leading Republican on the House Energy and Commerce Committee, challenged global warming science as "uneven and evolving".

"You're not just off a little, you're totally wrong," he said of Mr Gore's conclusions that carbon dioxide emissions contribute to global warming....
My understanding is that there's probably a 1/15-1/20 chance that we'll eventually discover that carbon dioxide emissions have been a relatively small contributor to global climate change to date. Even if the 20:1 odds paid off, Mr. Barton, of course, would still be operating on faith, not reason. There is nothing in the data to support his confidence, ergo it must arise from some sort of faith.

Had history played out a little differently, he would be chairman of the House Energy and Commerce Committee. Chairman. The thought chills the soul.

How can any reasoning person vote for the modern GOP? The party needs to be rebuilt from top to bottom.

Tuesday, March 20, 2007

The Homo ergaster Turing test: a mind set in stone

The 8 yo and I were browsing the book companion to the excellent BBC series 'Walking with Cavemen' the other night. I was struck by a commentary on the peculiarity of the eternal stone tool. Modern humans appear to have gone from stone tools to quantum computers in about 30,000 years, but ancient hominids spent millions of years doing exactly the same thing. Even Neandertal seemed to create the same tools year in and year out -- though perhaps with a mechanical precision that modern humans cannot equal.

Paleontologists have tried to image what a mind would be like that showed such limited creativity. They speculate about a highly compartmentalized "intelligence", with nothing like the "general" intelligence we're alleged to have. Something, in other words, like a modern computer.

We can't yet build a non-organic sentient computer (we build millions of the organic variety every day), but we might be able to build one that would "pass" for Homo ergaster. That is, it would past the Turing-egaster test. All it needs then is a bit of evolution ...

PS. The new book costs about $35 or so. At Amazon.com, including shipping, the cost is about $9. Amazon's used book service must be having some impact on new book sales and the used book marketplace. I've not read anything about this though. We ordered a used copy.

Dementia: is it really a disease?

So is Alzheimer's-type dementia really a disease?
Prevalence of Alzheimer’s Rises 10% in 5 Years - New York Times:

... The updated estimates, based on the rising occurrences of the disease with age, not new disease research, were released yesterday by the Alzheimer’s Association, along with a compilation of other information about a progressive brain disease that afflicts 13 percent, or one in eight people 65 and over, and 42 percent of those past 85...

...Mary Mittelman, an Alzheimer’s researcher at New York University, had mixed feelings about disproportionate attention to early onset Alzheimer’s disease. On the one hand, Dr. Mittelman said, these cases are such a small minority that she fears will take focus and resources “from the majority who are much older.” On the other, she said, 'because of the ageism of this society” far too many people still believe dementia to be part of normal aging and attention to this younger group will clarify that it is a 'real disease'.”
When 42% of people over 85 have advanced features of a prolonged process, is this a "real disease" or an aging process? The boundary here is pretty fuzzy. Nobody has the brain or body at 85 that they had at 45, much less 25.

My personal bias is that the "Alzheimer's process" is a "normal" part of aging, right up there with getting weak, flabby, and shorter. What varies, of course, is how fast it progresses, and whether one dies before or after the process becomes disabling (disability also being a fuzzy term -- many persons who we consider "able" at 85 are unable to do the work they did at 45).

The distinction between an aging process and a disease has a practical implication. In general, diseases are treatable, but aging is much harder to stop. If we accpet the model of an Alzheimer's "Process" rather than Alzheimer's Disease, then we can better judge how great our challenge is. We need to do more than arrest a disease, we need to slow the aging process of the human brain. The bright side is that there's been a lot of research lately that suggests that might be possible.

Pet poisoning: 1/6 die in testing?!

Whatever the Menu Foods contaminant is, it is so lethal that it killed 7/45 animals in testing:
San Jose Mercury News - Pet food recall has owners searching for answers

... Menu Foods told the Food and Drug Administration it received the first complaints of kidney failure and deaths from pet owners Feb. 20. It began new tests Feb. 27 in which 40 to 50 dogs and cats were fed its product. During those tests seven animals died, according to the FDA's top veterinarian, Stephen F. Sundlof. The contamination appeared more deadly to cats than to dogs, he said, although the mix of species tests was not immediately known....
Other articles have added a bit more detail. The animals became ill about 5-6 days post-exposure and the chief suspect is contamination of a gluten based thickening agent at a Kansas plan..
FDA has sent inspectors to company plants in New Jersey and Kansas. Most complaints stem from products made at the latter factory, though both received shipments of wheat gluten, identified as a likely source of contamination, from the same supplier, said Stephen F. Sundlof, the FDA's chief veterinarian. The ingredient is a protein source used to thicken the pet food gravy. The FDA is screening pet food samples for substances known to be toxic to the kidneys, like toxins produced by molds.
It's impressive how little we are hearing about the testing. We're going to beef and rice at our house until we know more -- and our dog is on a dry fish based Gluten free diet (yeah, she has canine sprue - happily gluten free diets are easy for dogs). In theory we should be pretty safe, but to put it mildly we have zero trust at the moment.

Update: A Kansas source has more information. If the source is the Kansas plant, the Kansas City Star may be the best information source. This article says the contamination was detected incidentally in taste tests of an experimental product, the company wasn't looking for problems ...
... Menu Foods, according to one veterinary toxicologist, was working on some experimental products when it did some testing, and found that the food sickened or killed some animals. The company reportedly traced the problem back several weeks to when they started using a new wheat component in their foods. The company immediately stopped using that new wheat source, which it used in high concentration in the experimental food. To be safe they recalled every other brand that might contain the new wheat, even in tiny quantities.
I wonder if it will be aflatoxin again ...

Monday, March 19, 2007

The surgeon gene

FuturePundit: Sleep Gene Determines Performance Of Sleep Deprived. I bet most surgeons have the short version of the gene.

Flatland and the 248 dimensions of time

In the 19th century two dimensional world of Flatland the hero infers the existence of dimensions beyond those of his god-like 3 dimensional correspondent. He would be willing to image a 248 dimensional creature:
Is this the fabric of the universe? | Science News | Connected | Telegraph

Mathematicians have successfully scaled their equivalent of Mount Everest. Today they unveil the answer to a problem that, if written out in tiny print, would cover an area the size of Manhattan.

At the most basic level, the calculation is an arcane investigation of symmetry – in this case of an object that is 57 dimensional...

... What makes this group of symmetries so exciting is that Nature also seems to have embedded it at the heart of many bits of physics. One interpretation of why we have such a quirky list of fundamental particles is because they all result from different facets of the strange symmetries of E8. I find it rather extraordinary that of all the symmetries that mathematician’s have discovered, it is this exotic exceptional object that Nature has used to build the fabric of the universe...

... Today’s feat rests on the drive by mathematicians to study symmetries in higher dimensions. E8 is the symmetries of a geometric object that is 57-dimensional. E8 itself is 248-dimensional...

....The ways that E8 manifests itself as a symmetry group are called representations. The goal is to describe all the possible representations of E8. These representations are extremely complicated, but mathematicians describe them in terms of basic building blocks. The new result is a complete list of these building blocks for the representations of E8, and a precise description of the relations between them, all encoded in a matrix, or grid, with 453,060 rows and columns. There are 205,263,363,600 entries in all, each a mathematical expression called a polynomial.

..."This is an impressive achievement," said Hermann Nicolai, Director of the Albert Einstein Institute in Potsdam, Germany. "While mathematicians have known for a long time about the beauty and the uniqueness of E8, we physicists have come to appreciate its exceptional role only more recently - yet, in our attempts to unify gravity with the other fundamental forces into a consistent theory of quantum gravity, we now encounter it at almost every corner...
I'd love to hear an explanation of how E8 might relate to non-local action and the seemingly bizarre nature of time. Are meaning-free instantaneous communications between entangled photons instances of higher dimensional structures protruding into 4 space? Is there really only one dimension of time, or is it simply that meaning can't exist in the others?

Oh, and did those last two sentences of mine make any sense at all?

It's a thin line between being an amateur commentator on math and physics and being a complete loon! :-)

Update: Another article explained that E8 has merely 57 dimensions, but 248 symmetric rotations. Whatever that means. Mathematics left the scope of mere mortals well over a hundred years ago, it's not surprising that non-specialists should struggle with this announcement.

Industrial food and pet poisoning - lessons?

A year ago it was aflatoxin contamination killing dogs (Diamond pet food is still in business btw, the web site proclaims their lack of connection to this episode). Now it's some poison still to be named...
Pet Food Is Recalled After Link to Animal Deaths - New York Times

More than 60 million cans and pouches of dog and cat food sold under dozens of brand names were recalled on Saturday after being linked to the deaths of 10 animals.

The food was manufactured by Menu Foods, of Streetsville, Ontario, which makes wet food sold as store brands for companies like Wal-Mart, Kroger and Safeway.

The company also makes food on behalf of many brand-name pet food makers. Menu Foods said it had recalled some food made for the Iams unit of Procter & Gamble. Two other pet food companies — Nestlé Purina PetCare and Hills Pet Nutrition, the unit of Colgate-Palmolive that makes the Science Diet brand — recalled some of their products that were made by Menu Foods.

Menu Foods is recalling only certain gravy-style pet food in cans and pouches it made from Dec. 3 to March 6.

The company said in a statement that tests of its food had “failed to identify any issues with the products in question.” But it did associate the timing of the reported deaths with its use of a new supplier for wheat gluten, a source of protein. Sarah Tuite, a spokeswoman for Menu Foods, declined to name the supplier.

The reported deaths of cats and dogs have been from kidney disease, Ms. Tuite said. Symptoms vary but can include lethargy, jaundice and vomiting, she said...
The recall list on their web site is very long (there's another list for cats):
1. Americas Choice, Preferred Pets
2. Authority
3. Award
4. Best Choice
5. Big Bet
6. Big Red
7. Bloom
8. Wegmans Bruiser
9. Cadillac
10. Companion
11. Demoulas Market Basket
12. Eukanuba (ouch, that hit us)
13. Food Lion
14. Giant Companion
15. Great Choice
16. Hannaford
17. Hill Country Fare
18. Hy-Vee
19. Iams
20. Laura Lynn
21. Loving Meals
22. Meijers Main Choice
23. Mighty Dog Pouch
24. Mixables
25. Nutriplan
26. Nutro Max
27. Nutro Natural Choice
28. Nutro Ultra
29. Nutro
30. Ol'Roy Canada
31. Ol'Roy US
32. Paws
33. Pet Essentials
34. Pet Pride - Good n Meaty
35. Presidents Choice
36. Price Chopper
37. Priority Canada
38. Priority US
39. Publix
40. Roche Brothers
41. Save-A-Lot
42. Schnucks
43. Shep Dog
44. Springsfield Prize
45. Sprout
46. Stater Brothers
47. Weis Total Pet
48. Western Family US
49. White Rose
50. Winn Dixie
51. Your Pet
The pet care world is fundamentally Libertarian. So is there a lesson here about the limits of Libertarianism? Or is there a lesson about the risks of industrial food? It seems one plant fed a large portion of North American dogs and cats, so a major problem there will have a huge impact.

These episodes have further lowered my opinions of both Libertarian marketplaces and industrial food. Our family may need to look at the ridiculously expensive "organic" pet food market.

Sunday, March 18, 2007

Vanguard has the security question sickness - big time

Vanguard has taken the "security question" sickness to a new level of absurdity.

If you log into Vanguard from an "unfamiliar computer" (meaning you clear your cookie cache) you have to answer a security question. This means I really have to enter critical information that "only I know".

Except these security questions are the same everywhere. If any site I answer them at is hacked, the hackers know my security question answer everywhere where they can establish my identity.

Since these questions are treated as though they were "secret", the fact that they cannot be "secret" means that they reduce rather than enhance security.

I'm tempted (only slightly) to post the answers to my "secure questions" on my public web site. I could file it under "identity theft made easy" and include my SSN, my birth date, and my favorite passwords ...

I bet Vanguard paid quite a bit for this "feature", probably from a big name consulting firm. Only smart people could possibly be so stupid.

Saturday, March 17, 2007

Life in Web 2.0: just don't ask for reliability

I'd imagined that the most reliable service in the Web 2.0 world would be Google. This morning I tried to do some work with Google's industrial application deployment and domain registration service, mediated by their premier eCommerce solution (Google Checkout).

Of course, this is what I got:
Google Apps - Server error
Help | Sign out
Server error
We are unable to process your request at this time, please try again later.
Much has been made of the millennial generation's lack of concern about privacy. I think another remarkable generational difference is an expectation of unreliability, and a probabilistic strategy for managing unreliable systems.

Geezers like me, who remember when most performance was highly predictable, have more trouble with things that don't work -- especially when we're short of time.

Update: Later the process goes one step further along ...
Oops! An error occurred while processing your request.

Friday, March 16, 2007

TimesSelect: Free for a .edu address

Arrgghh. I just paid about $50 for TimeSelect:
TimesSelect University E-mail Verification

...We are pleased to offer a complimentary subscription to TimesSelect. You must be a student or faculty member with a valid college or university e-mail address to be eligible for this offer. You no longer need an access code to activate your TimesSelect University Subscription.

To start the process, please provide your school e-mail address below. If it is accepted, you will receive an e-mail from NYTimes.com with a link to continue your sign up process...
How will they distinguish academics from those with alumni accounts? I qualify as faculty, so now I'm just grumpy. I wonder if the NYT prorates upon cancelation?

Foreign Policy: latest addition to my blogroll

FP Passport is the blog of Foreign Policy, a political science journal. It's excellent. I've added it to my blogroll. (I'm still using Bloglines, but if they keep updating feeds unpredictably I'll try Google Reader.)

Thursday, March 15, 2007

The Measure of Things

A colleague forwarded me one of those "interesting emails" which always seem suspicious. This anonymous message was interesting, though mysterious. It consisted of a series of images beginning with the solar system and ending with the star Antares. Once I cracked the mysterious file format. I started looking around to see if I could validate the content.

My search took me to a 2006 web page, but a very interesting faculty page (worth a visit) was probably the origin for the last 3 images (they appear relatively "home grown"). I suspect the email is based on the faculty page. A reference there pointed a physics simulation site as the original source for the first 3 images. I've attached all the images below, you can click on them to see larger images.

First, some order of magnitudes examples to warm up with:
Planck Length: 10**-35 m
String Length: 10**-35m (I thought this would be a bit bigger than the Planck length actually)
Neutrino: 10**-24m (Much smaller than a quark!)
Electron and Quarks: ? (10**-15 to 10**-18 to 0 -- the concept of "diameter" may not apply)
Proton: 10**-15 m
Hydrogen Atom: 10**-10 m
Water molecule: 10**-9m
Bacteria: 10**-7 m
Red cell: 10**-6 m
White cell: 10**-5 m
Limit of human vision: 10**-4m
Human: 1 meter
Moon and Mercury: 10**6 m
Earth: 10**7 m
Sun: 10**9 m
Arcturus: 10**10 m
Betelgeuse: 10**11 m (varies)
Antares: 10**12 m
Solar system: 10**13 m
1 Light Year: 10**16 m
Milky way Galaxy: 10**21 m
Quasar: 10**26 m (distance at time light was first emitted from furthest detectable quasar)
Visible Universe: 10**26 m
Universe/Multiverse: Unknown. Around 2007 some said 10**27, but more recent speculations are vastly larger (or infinite).
The span from the Planck Length (smallest possible distance) to the Visible Universe is a factor of 10 **61. On a log scale the mid-point is somewhere around the limit of human vision. In other words, there are as many 'scales of 10' between the smallest thing you can see and the smallest thing that can exist as there are between that wee thing and the universe. As Feynman said, there's a lot of room at the bottom.

So if we treat Earth as the measure of all things, let's take a look at some of the stellar objects in the mid-range. First, though, we'll take a look at the 'string' to 'hand' scale from (Seed magazine 2007)




Earth, Venus, Mars, Mercury and the Moon
Earth is only a bit larger than Venus, but Mars is a relative pipsqueak. It's roughly correct to say that Earth is to Mars as Mars is to the utterly barren Moon.


Jupiter, Saturn, Uranus Neptune and the rocky runts of the system.


The Sun and its leftovers


Sun, Sirius, Pollux and Arcturus
Arcturus is about 25 solar diameters across.


Arcturus, Rigel, Aldebaran, Betelgeuse, Antares
Ten Antares, laid side by side, would span our solar system. An individual human is to Antares as Antares is to our galaxy.


Update 2/4/11: The universe is much bigger than when I wrote this.

Update 5/3/12: Now what we used to call the Universe is merely the Visible Universe, which is some infinitesimal fraction of the multiverse. This Scale of the Universe app (Flash, damnit) is fun to play with. There's a lot of nothing between neutrino's and the quantum foam. I had made some egregious arithmetic errors in my original post and I just corrected them.

The mystery of the apple keyboard

I took the kids to the Apple Store tonight, and did some email on a beautiful 30" display attached to a very powerful desktop machine -- and a completely miserable keyboard.

Why are Apple's desktop keyboards so lousy? Their laptop keyboards are reasonable, it's only the desktop kbs that combine great looks with miserable performance.

Oh, yes, I know, keyboards are personal and Apple's kb gets four stars on Amazon. Bah. Show me just one geek who likes that kb (excepting Andrew)!

If Apple were to combine the cosmetics of their cheapo kb with the functionality of $9 PC kb, they'd dramatically increase the sales of their desktop machines. Trust me on this. I'm sure large numbers of XP users walk into an Apple store, type, and figure everything about Apple must be as bad as the keyboard.

Maybe it's all a plot to sell more laptops ...

Another longtime reader abandons the Economist: Crooked Timer

I abandoned The Economist last year because it seemed to been infected by the WSJ Editorial Page brainworm. It wasn't completely demented, but it was getting there. So this CT post caught my attention. I'd love to know what Fallows thinks of today's Economist. He was writing during their glory years! (Emphases mine)
Crooked Timber: Cringe and whinge

I came across James Fallows’ 1991 piece on The Economist (to which my subscription has just lapsed), The Economics of the Colonial Cringe, and thought it pretty interesting. On the one hand, this seems a little dated:

In functional terms, The Economist is more like the Wall Street Journal than like any other American publication. In each there’s a kind of war going on between the news articles and the editorial pages. The news articles are not overly biased and try to convey the complex reality of, well, the news. Meanwhile, the editorials and “leaders” push a consistent line, often at odds with the facts reported on the news pages of the same issue.

If there’s any marked difference these days between the line touted in the editorial pages line and the perspective of the news articles, I can’t detect it. The WSJ seems to still have a firewall between the two (although in fairness its editorial pages are also far loopier than those of the Economist)...

I agree. The WSJ editorial pages read like meth fueled ravings of brains debauched by mad cow disease, but their news pages are excellent. The Economist editorial pages are only moderately prion infested, but the disease has spread throughout the "newspaper" (their term).