Philosophy of error

I like technological breakthroughs, and the announcement in this article struck me as good news:

Nuclear power plants smaller than a garden shed and able to power 20,000 homes will be on sale within five years, say scientists at Los Alamos, the US government laboratory which developed the first atomic bomb.

The miniature reactors will be factory-sealed, contain no weapons-grade material, have no moving parts and will be nearly impossible to steal because they will be encased in concrete and buried underground.

Cool. However, reading on, the more I stared at the cost figures provided by the CEO of Hyperion (the company licensed to sell these things in the US), the more confused I became:
The US government has licensed the technology to Hyperion, a New Mexico-based company which said last week that it has taken its first firm orders and plans to start mass production within five years. 'Our goal is to generate electricity for 10 cents a watt anywhere in the world,' said John Deal, chief executive of Hyperion. 'They will cost approximately $25m [£13m] each. For a community with 10,000 households, that is a very affordable $250 per home.'
That just didn't look right, but there are a lot of, you know, digits. So I opened the calculator program on this computer, divided 25 million by 10,000, and got a figure of $2500.

It hardly inspires confidence that the CEO of a company which makes nukes would make such a blatant error in math.

I guess it's possible that the error wasn't his, but was instead that of the writer. You know, a zero got dropped. Does this indicate an "error"? I'm very forgiving of obvious errors, especially typos, but the problem with numerical errors is that the intended meanings aren't always clear the way they are when, say, the word "the" appears as "teh."

The question of whether a number is an error is compounded if we look at the possible biases that might be behind it. In that context, what is an "error"? That a number is simply wrong does not end the inquiry. If I am "just a digit off" when I tell the IRS that I made $25,000 when the number is really $250,000, it is certainly relevant whether I deliberately gave them an incorrect number as opposed to having made a mistake in math or in transcription. Yet the number is "wrong," and in that sense an "error" regardless of my intent. At that point, though, the question of "what is an error" becomes a philosophical one. In the moral and legal sense, if I deliberately gave the IRS an incorrect number, I made no error, for my act was deliberate. Lies are not errors. Yet in math, an error is an error is an error. Suppose that for some pathological reason you take a math exam, and you deliberately gave the wrong answer knowing it was wrong. It is still wrong, and good or bad faith do not affect its inherent wrongness. A wrong answer is no more wrong because it was sincerely believed to be right either. Interestingly, if by your calculations you got the wrong answer, then made a transcription error which made it right, it would still be right. Morality and intent are irrelevant.

Notice that from the facts given in the current example, it is not clear which number is in error. There are three figures:

  • $25 million cost per plan
  • 10,000 customers
  • $250 cost per customer.
  • Because the last number is wrong, we tend to focus on the bad math that must have "led" to it. But if the problem is an error in transcription, if we assume the CEO provided the guardian with correct figures, the 10,000 customers might have been 100,000. But I doubt it, because the use of the word "community" implies smaller, and 100,000 is not a community but a city. So maybe the $25 million is wrong. Maybe it's $2.5 million. Possible, but that seems pretty cheap -- even for a small nuclear plant.

    I can't be sure, but I think the $250 should be $2500. Still affordable. Anyway, the lower the price, the better the news, right?

    It's a minor point, but the philosophical aspects of errors fascinate me, and in an amazing coincidence, earlier this morning I read an article dealing with precisely that subject, titled "Math students find success with philosophical route to the right answers":

    It seems weird at first: Math teachers who don't care if the answer is correct.
    OK, that was the first paragraph. It is supposed to be startling to the readers. However, it didn't startle me because it isn't the first time I've read about these new methods of teaching math. I've blogged about them before, and I'm skeptical.

    However, it is claimed (by the proponents) that these methods work better to improve test scores:

    But Wayne State University professors who teach college math at Detroit Public Schools believe that triumph is not in the answer, it's in the struggle to get there.

    Now, their philosophy -- one that test scores suggest works with surprising success -- has found a bigger and more efficient home on WSU's main campus.

    The Center for Excellence and Equity in Mathematics is about "creating an environment in which kids are intellectually fearless, where they're not so worried that their answers are incorrect or correct," said math Professor Leonard Boehm.

    Boehm is a man in constant motion -- pacing and punching the air in the classroom, shaking students' hands and cracking really bad jokes. Four times a week, he brings college math to fifth-graders at Thurgood Marshall Elementary School, just one of the programs now served by CEEM.

    On a recent morning, the wooden floor boards creaked beneath his stride and 26 students were rapt.

    The lesson: If 2 raised to the second power is 4, and 2 raised to the third power is 8, what is 2 raised to the zero power?

    Hands shot in the air. "My esteemed colleagues," each respondent began their answer. (The answers in Boehm's class always start this way as a sign of respect.)

    The students agreed: The answer must be zero, right?

    No, wait. 1? Their certainty wavered. 2?

    Boehm grinned and stopped. His voice dropped.

    "How come a mathematician," he asked, "might say the answer doesn't exist?"

    It was Desean Washington-Jones, 11, who eventually raised his hand through the quiet.

    The students began gesturing -- like a basketball referee motions for a traveling offense -- to show their support as Desean worked through the answer.

    "My esteemed colleagues," he began. "The answer might not exist because there is no factor form."

    Hands flew in the air and began to shake -- the sign they agreed.

    "Can something come from nothing?" Boehm asked. "If a factor form doesn't exist, can the answer come from something that doesn't exist?"

    And so it went -- onto existential quantifiers, the fallacy of induction, and lessons that will lead to the concepts of a limit as well as infinite geometric and harmonic series.

    It's the kind of college-prep training not always available to or expected of inner-city kids, said Steve Kahn, WSU professor and director of CEEM.

    "The kids in Detroit are getting screwed," said Kahn, who began his career teaching Detroit's high school dropouts.

    I don't doubt they're getting screwed, and it is to be sincerely hoped that the philosophical approach to math will cause them to be proficient in basic everyday survival type math skills.

    The numbers cited are impressive:

    The CEEM approach seems to be working. Before Math Corps, students answered about 30% of the answers correctly on grade-appropriate tests. By the end of the camp, they were averaging 90%, according to CEEM stats.

    "The math is good and simple. ... But the real key is changing a kid's attitude about her or his expectations," said Robert Thomas, dean of WSU's College of Science.

    That, and matching students with peer mentors "turns out to be magical," he said.

    A jump in scores from 30% to 90% really does seem magical to me.

    Assuming that the numbers are correct, I wondered if anything might explain this other than magic, so I went to the Math Corps web site, and saw this:

    In reviewing applications, staff seeks indications of a serious desire to succeed in mathematics as well as evidence that the student is willing to work hard to achieve that success.
    That means they've pre-selected for motivation. That may be more important than magic. Also, I notice that there's little information provided about CEEM's tests. CEEM stands for the Wayne State University Center for Excellence and Equity in Mathematics, which has no website, and appears to be the creation of one professor. Thus, there is no way to know exactly what is on the test that CEEM appears to have designed it. However, since it teaches only motivated students and uses its own test for the before and after, I am not surprised that there would be marked improvement.

    Of course, if the goal is "math equity" (a complex issue described here) what might be the implications for the test? If the goal is to improve scores to promote equity, that might not translate into real world proficiency. What I'd like to know is how they would do on a standardized test, and whether they end up being able to do things like long division. It's one thing to teach that correct math answers don't matter as an initial stage, but I'd hate to think that they never really matter. Otherwise, how is anyone supposed to make sense out of anything numerical?

    But once again, maybe my hangup about accuracy reveals my own educational deficiencies. I was bad at math, and one of the reasons was that there might have been too much emphasis on finding accurate numbers. Here's how I put it before:

    ...appreciating how many dead bodies there might [as in victims of Hurricane Katrina] be is a highly personal process. To one person, there might be hundreds. To others, there might be thousands, and depending on social skills and psychological considerations, still others might see the answer as millions.

    Aren't higher numbers more relevant to what's going on in the world? If the goal of math is to make things relevant, then the numbers have to be higher, because otherwise, people might not care as much.

    What this means is that the hangup that bloggers like me have with finding accurate numbers reveals an educational deficiency which is being remedied.

    I should be glad. Because it means my being bad at math really wasn't any shortcoming on my part. And my hangup about it only reflects the wrong social attitudes of the times in which I grew up.

    This is all changing.

    We should be glad.

    There's still a lingering question, though. If numerical unaccountability is more fair and equitable than numerical accountability, then what does that suggest about the basis for fairness?

    If there is no accountability, then why are only some are off the hook?

    posted by Eric on 11.09.08 at 12:35 PM





    TrackBack

    TrackBack URL for this entry:
    http://classicalvalues.com/cgi-bin/pings.cgi/7630






    Comments

    I would have loved being in that class.

    Donna B.   ·  November 9, 2008 12:59 PM

    Yet more evidence explaining why Detroit is such a basket case.

    First, regarding the cost of personal nukes, I would presume the error was from the reporter. If the reporter had been math-competent, the reporter would've caught the error. Unless you have an audio file of the CEO saying $250, I would presume the reporter dropped the digit and was clueless it happened. In other words, the journalist made an error in not catching the obvious math error. It is reasonable to presume that was because he/she -- and not a nuclear engineer -- made the math error.

    Second, as for how math is being taught in the Detroit schools, this sounds like more feces-flinging from the "self-esteem is more important than actual knowledge" school of education.

    Rhodium Heart   ·  November 9, 2008 01:13 PM

    As Euclid stated there is no royal road to geometry, similarly there is no royal road to making inner city kids math-competent.

    However, I can speak for myself regarding "philosophical" approaches to math helping computational skills. Years ago my high school math featured one of the then so-called "New Math" ( UICSM, in my case) programs. In 9th grade math we spent a lot of time on proofs of the various number properties ( distributive, associative etc). I was a competent multiplier and adder before I took "New Math." IMHO, my exposure to number properties via many proofs in 9th grade helped my estimation capabilities. Since 69= 70- 1, one can easily estimate 69X 7. Etc. While one has calculators today, estimation helps one to determine if the answer makes sense, is ballpark.

    At the same time, many students of lesser ability did not like the New Math. But it did work for me.

    Gringo   ·  November 9, 2008 06:15 PM

    The amazing things about the description of that math class falls into two categories: 1) the enforced recognition of esteem of colleagues and 2) the sheer excitement of discovery.

    It certainly may not be the answer to improved math ability in inner city schools, but I don't think that proves it is useless.

    Donna B.   ·  November 9, 2008 07:09 PM

    You know, in High School I was a wizard at solving calculus and physics problems. I knew exactly what to do. Yet I often made numerical errors. Which got marked off a little. The most important thing is getting the concepts right. You can recheck your numbers in a more leisurely setting than a timed test. i.e. the real world.

    M. Simon   ·  November 9, 2008 09:32 PM

    Exactly, M. Simon. I sucked at math classes in college, but excelled in chemistry. The math in chemistry made sense to me even when I did get the arithmetic wrong.

    The more I view mathematics as a language, or philosophy if you wish, the more sense it makes and the more use it is.

    Donna B.   ·  November 9, 2008 10:11 PM

    I think that the reason for the $250 was that the reactor needs refueling every 7-10 years as stated in the body of the story. This would, in typical sales speak, "As low as $250 dollars per household per year!!!!111!eleventy!!" which, if you think about it is a better way to report costs. if its ongoing costs are $250 a year per home, plus 10 cents a KW/Hr (not per watt!). now if one was to read their website, it actually states that the "battery" is for 20,000 houses for 5 years. same answer, $250 per house, 10 cents a KW/Hr. this price is right in line with some of the more expensive states average electricity prices. my guess is that the math made sense, the reporters, being reporters, didn't really understand what they were talking about and left out the important bits that made it all make sense.

    Sean Sorrentino   ·  November 10, 2008 07:21 AM

    It's long past time we fired all the teachers and started from scratch. Yes, we can!

    Brett   ·  November 10, 2008 08:13 AM

    These are great comments, every one of them. Sean, thanks for apparently figuring this out. (It never would have occurred to me, but I guess where it comes to interpreting the news, two heads are better than one.)

    Eric Scheie   ·  November 10, 2008 07:13 PM

    Math is the quantative system to describe reality. So it is a language that many find difficult to understand.

    I can describe a circle in mathematical terms or just draw a picture. Most find it easier to understand the visual than the abstract.

    But to be able to replicate exactly the curved form af a chair back the abstract language is better.

    RAH   ·  November 11, 2008 08:16 AM

    Post a comment

    You may use basic HTML for formatting.





    Remember Me?

    (you may use HTML tags for style)



    November 2008
    Sun Mon Tue Wed Thu Fri Sat
                1
    2 3 4 5 6 7 8
    9 10 11 12 13 14 15
    16 17 18 19 20 21 22
    23 24 25 26 27 28 29
    30            

    ANCIENT (AND MODERN)
    WORLD-WIDE CALENDAR


    Search the Site


    E-mail




    Classics To Go

    Classical Values PDA Link



    Archives




    Recent Entries



    Links



    Site Credits