Comment by lutusp | Hacker News Reader

Comment by lutusp | Hacker News Reader

Parent Full thread

> but you would thereby eliminate all observational sciences - evolutionary biology, astrophysics, cosmology, etc.

No. Not if an observation leads to a testable, falsifiable theory, if that theory makes new predictions, if the predictions are verified, and if the theory in all its manifestations remains potentially falsifiable by new evidence. Like quantum theory, the best scientific theory we have, a theory that makes many other fields look like stamp collecting, and a theory that relies on observations of real-world phenomena.

As to cosmology, there are a number of cosmological theories based on observation that have led to testable theories, and those theories have survived reality-testing and have made predictions that are later verified in observations, like the recent observations of the CMB that support inflationary theory and led to some recent Nobel Prizes.

> I suspect most people would find that a somewhat extreme position.

Science is not a popularity contest, and it is not politics.

> Also, fraud is an issue in all fields of Science, so that seems somewhat an unfair issue to introduce.

Many people don't understand that some pseudoscientific fields are all fraud, in which case arguing that all fields have some fraud misses the point.

The hypothesis that lead exposure affects early childhood development, and that this leads to higher crime rates does lead to testable predictions - it predicts, for example, that different regions that removed lead at different times would see crime rates drop at different times. And indeed, this is exactly what we see. Is this as strong as exposing a test population and comparing it to a control? No. But neither is it meaningless pseudo-science.

This line of argument is actually very similar to the cosmological example - the theory makes a prediction about something you should observe, if the theory is correct. However, this is still much weaker proof than being able to generate a Big Bang in the lab and examine the outcome.

The QM case is an example of how much more powerful it is in fields where we can do controlled experiments, I agree. But not everything worth investigating (and worth knowing) is amenable to that type of study.

> The hypothesis that lead exposure affects early childhood development, and that this leads to higher crime rates does lead to testable predictions ...

Yes, but not falsifiability, required for science. Okay, I see people aren't getting this, so here's a classic example. Let's say I am a doctor who believes he has found a cure for the common cold. My cure is to shake a dried gourd over the cold sufferer until he gets better. The cure might take a week, but it always works. My method is repeatable and perfectly reliable, and I've published my cure in a refereed scientific journal (there are now any number of phony refereed scientific journals). And, because (in this thought experiment) science can get along without defining theories, without falsifiability, I'm under no obligation to try to explain my cure, or consider alternative explanations for my breakthrough — I only have to describe it, just as lead in gasoline is being described as inversely correlated with crime rates in the linked article.

> Is this as strong as exposing a test population and comparing it to a control? No. But neither is it meaningless pseudo-science.

But that is exactly what it is -- meaningless pseudoscience -- on the ground that no one is crafting and then testing theories about the correlation. Until we know why A is correlated with B, until we can demonstrate a cause-effect relationship and make falsifiable predictions based on the theory, it's not science.

> But not everything worth investigating (and worth knowing) is amenable to that type of study.

Yes, I agree, science can't provide answers to questions not amenable to the scientific method (i.e. "that type of study"), and not all important questions have this property.

Well in the case of lead wouldn't the prediction be: a jurisdiction that removes lead from its petroleum products will experience a significant decrease in crime rates about 20 years later?

Not without a testable, falsifiable explanation. Observations (i.e. descriptions) can't be falsified, that's reserved for explanations. If I say, "the night sky is filled with little points of light", that's hardly falsifiable. But if I say, "those points of light are actually thermonuclear furnaces like our sun, but at greater distances", that claim is open to meaningful test and possible falsification.

Here's why a testable, falsifiable theory is required for science. Let's say I'm a doctor and I've created a revolutionary cure for the common cold. My cure is to shake a dried gourd over the cold sufferer until he gets better. The cure might take a week, but it always works. My method is repeatable and perfectly reliable, and I've published my cure in a refereed scientific journal (there are now any number of phony refereed scientific journals). And, because (in this thought experiment) science can get along without defining, falsifiable theories, I'm under no obligation to try to explain my cure, or consider alternative explanations for my breakthrough — I only have to describe it, just as the linked article describes the correlation between lead in gasoline and the crime rate.

Because I've cured the common cold, and because I've met all the requirements that social psychology recognizes for science, I deserve a Nobel Prize. Yes or no?

Reply on news.ycombinator.com