I've been meaning to follow up on some more thoughts about Unscientific America. I suppose what bothers me about the book is that there is no distinction between positive statements--the way things are or will be--and normative statements--the way things ought to be. As an example, consider yesterday's post about type I diabetes. When someone (often a child) is diagnosed with type I diabetes, we have several options:
- Provide the resources to treat the disease and enable the diabetic to live as best as one can with the disease.
- Let him sink or swim on his own--if he (or his parents) have the financial resources to treat the disease, great. If not, well, then it sucks to be him.
- Stake him to the ground outside, and let him be devoured by wild animals.
To help bring those of you in the slow ethics group up to speed, options #2 and #3 are fucking evil. But back to positive and normative statements. Positive statements could include, but aren't limited to, the following:
- Said person is a type I diabetic.
- Said diabetic should follow a certain regimen to maintain blood glucose levels at a healthy level.
- It will cost some, quantifiable amount of money to treat this person's diabetes.
Note that positive statements don't have to be agreed upon statements of obvious fact: we could disagree on how much it could cost to treat diabetes. But the important point is that none of these statements are about how things should be or what we should do.
(An aside: Economists are particularly bad at confusing positive and normative statements. Often, stated results, such as policy X will maximize productivity are implicitly assumed to be desirable, when maximizing productivity above all else might not be the 'best' outcome. For instance, one might be more concerned with controlling the release of a toxic compound into the environment, even if that doesn't allow the maximization of productivity.)
Normative statements (e.g., we should stake diabetics to the ground to be devoured by wild animals) are different. As a scientist, I can tell you if someone has diabetes. As a scientist, I can tell you what the possible outcomes of the disease are. As a scientist, I can tell you what the best regimen for controlling blood sugar levels is. These data can and should inform what we ought to do. But ultimately, as a scientist, I have no special prerogative to generate normative statements--what we should do. In other words, when it comes to how we respond to my scientific observations (which I am more qualified to make as a trained scientist), I, like everyone else, am just another asshole with a blog.
Which brings me to Unscientific America (no, I am not implying that Chris and Sheril are assholes. Although they do have a blog). To the extent scientists can contribute to the debate by clearly, convincingly, and compellingly making positive statments (e.g., global warming is real, and has a significant human component), I agree with the arguments in Unscientific America. And, sadly, since most of the Republican party has decided to headbutt the crazy train, it is absolutely necessary.
It has also led to unprecedented unity among scientists since we don't have to discuss what we ought to do; we are busy defending reality from the theopolitical Uruk-hai right, the corporate useful idiots, and narcissistic contrarians. But figuring what we ought to do about global warming involves more than positive statements. Yet, my reading of the book (which is also, no doubt, colored by my regular reading of their blog) suggests that they want more than just positive statements--that scientists should use some kind of scientific authority to defend normative statements.
I have no problem making normative statements: Sweet Baby Intelligent Designer, just scroll through the blog. But I make those statements based on my being a citizen and a human being (on my good days), not based on being a scientist.
- Log in to post comments
Mike,
So what's wrong with you, as a sceintist, making mnormative statements, especially about scientific endeavours that are within your disciplinary pervue? You have the expertise to determine, statistically, the probability of certain policy outcomes meeting their stated goals. From the perspective, you are in a far better place to make normative statements about which policy should be pursued.
And, leaving that aside, if scientists aren't making normative statements about the outcomes o fthe policies that are being promulgated based on (or the rejection of) our science, then lobbyists will. And they, afterall, are SOOOOOOOO god at getting the science right that they can easily speak for us.
That example is just dishonest economics, in that it sweeps cost externalization under the rug. There's an implicit "net" on "productivity" as people hear it, so at best the speaker is relying on misdirection to convey an impression other than the truth.
There's an additional element in a normative statement: values. It's one thing for a scientist to say, "in order to eliminate all human life from Island X at minimum cost, we must dust it with Chemical Y," and saying "we must dust Island X with Chemical Y." The latter contains a presumption that eliminating all human life from Island X is desirable.
It's possible that eliminating all human life from Island X is desirable -- but that's not "speaking as a scientist" any more.
This kind of dishonesty is well-known, and it even has a mathematical proof that it is dishonest: Lipsey-Lancaster's theorem of the "second best".
There is a also a philosophical confusion economists practice. Doctor Dumbo, who has a beautiful model that nicely describes mass behavior as a statistical summary of ideal agents, goes on to criticize individuals who behave differently from his ideal agents. Really! Doctor Dumbo, if he had an ounce of scientific integrity about him, would realize his goal is to fit his model to people, not tell people to behave like his agents.
But there is a deeper kind of dishonesty practiced in the profession, one rarely addressed at all. The ultimate question is just who rates as an agent, and this is outside all models. The economics of a slave nation look very different when all humans are agents versus when the only agents are the free populace.
You will find pretty much all professional economists today understand this, and of course, would never make this error. But do they really? I do not believe in animal rights--well, I could sympathize, but in honor of PETA I never will--but perhaps in 50 or 100 years, animals will take what everybody acknowledges as their rightful place as economic agents on par with humans.
Rather more likely, what happens when AI gets us to the point where corporations have the actual (as distinct from legal fictitious) autonomy to act as agents in their own rights distinct from their stockholders?
Last time I looked, economists were still looking the other way on the subject of management interests vs. corporate interests.
I wish I could find an essay from about ten years ago by (IIRC) Charles Schwab on the change from an ownership system to a management system for equities. Looking back that was seriously prescient.
TrackBack uses a REST model, where requests are made through standard HTTP calls. To send a TrackBack ping, the client makes a standard HTTP request to the server, and receives a response in a simple XML format.
Actually, economists have been studying this question for decades.
Absorbing post at that point, and iâve bookmarked this blog tooâ¦keep up the high work. I guess that it will be great if some students use your material in custom research paper situation.