The Relativity of Wrong by Isaac Asimov (1989)
chem.tufts.edu
chem.tufts.edu
A spherical earth model is useful and a flat earth model is not. Also, as Asimov explains here, wrong is relative and a spherical model of the earth is much less wrong than a flat earth model.
Not if you're going to say the projection is flat. If the projection is flat, and the projection is the model, then I should be able to lay a flat ruler, appropriately scaled, between any two points on the flat map and read off the actual distance between those two points. But I can't do that; and the model knows I can't do that. The model tells me to take the coordinates on the map of two points and plug them into a complicated mathematical formula if I want to get the actual distance between them. So the model is not flat in the sense that matters--in the predictions it makes for distances between points.
Is a constant distance metric part of the definition of flat? Why don't the grid lines on the map count as an appropriately scaled ruler?
Yes. Otherwise "flat" would have no meaning; we can always make a flat map of any surface, if we are willing to allow the distance metric to be distorted to any extent we like, so every possible surface would be "flat".
So all models are wrong but some are indeed useful.
Darwins theory (and later modernizations) are indeed wrong (as in not true) but it's the most useful model we have.
Not true in what sense?
Sure, but that doesn't mean they don't match reality. It just means they don't match all of reality; they leave information out. Of course all of our models have to leave information out, I agree with that. But that's a separate question from the question of whether the information a given model includes matches the portion of reality it is intended to model--or, better, to how good an approximation a given model matches the portion of reality it is intended to model. That's the whole point of Asimov's essay.
Otherwise no model can be representing reality in any objective way they can per definition only be interpretations (The map is not the territory).
Reality don't care about true or false models. Reality has no evolution or planets or black holes. These are all interpretations at a very specific scale from a very specific perspective.
But models that allow us to make predictions does not mean that they match reality it just means that it matches the predictions we can interpret.
Thats a very different thing than reality.
I don't see what "interpretation" is involved in comparing a specific prediction with a specific measurement result. You're just comparing two numbers (with error bars, etc.). And the comparison doesn't care what kind of stuff you put in your model; that can be anything you like. All the comparison cares about is the numbers your model outputs as its predictions.
> models that allow us to make predictions does not mean that they match reality it just means that it matches the predictions we can interpret.
No, it means the model's predictions match the actual measurements we make.
> Thats a very different thing than reality.
Perhaps you are using a different definition of "reality" than I am. In which case I don't see how whatever definition you are using of "reality" is relevant to Asimov's article, which is what we are discussing.
In other words you cannot not interpret. It's not reality. In reality that measurement result doesn't exist.
You are interpreting the world through your very very very limited senses and only seeing it the way your senses allow you to see it.
This is very different from the world as it is.
You are talking about reality within the confines of what humans can interpret. I am talking about reality as it is in it's entirety without any specific vantage point.
Hmm, ok, so we are also giving a different meaning to the term "measurement result". I'm afraid our ontologies are too different to have a meaningful discussion.
Which is a more useful way of looking at it. Theories have a predictive horizon. The point of science is to increase the predictive horizon.
But the other side of that is the ability to create new classes of technology.
Electromagnetic theory, relativity and quantum theory are lot more than a tweak on the Newtonian world view. They make it possible to build completely new kinds of technology that are impossible if all you have is Newton.
Suggesting that everything is a process of gradual refinement to an increasing number of decimal places is a wrong theory about theories.
There's a lot more to science snd technology than the orbit of Mercury.
>Darwins theory (and later modernizations) are indeed wrong (as in not true) but it's the most useful model we have.
That may baffle people, but the problem with the Darwinian view is that it's based on the principle that species evolve - when in fact ecosystem evolve, in the sense that flows of chemicals, information, and energy self-organise into more or less stable configurations.
Species evolution is an interesting epiphenomenon on top of that. But assuming it's all there is to know about natural history is like trying to understand meteorology by classifying icebergs. You're going to catch some of the more obvious patterns, but more complex relationships are going to elude you.
Modelling ecosystems is a very useful skill - especially for long space flight - and it turns out we're still incredibly bad at it, precisely because naive Darwinism doesn't have the predictive horizon to deal with the problem in a useful way.
Without a good theory, systems tend to diverge into instability and everything dies, no matter how fast adaptation happens - which isn't usually the outcome you want.
I'd highly recommend reading the article again a little more carefully.
In say 2000 BC, a flat-earth and spherical-earth theory were probably both about the same level of effectiveness in explaining observations and making predictions about new observations. In the year 1800, Newton's mechanics were extremely effective in making correct predictions. As technology advanced and new observations were possible, some gaps were revealed. But even today, Newton's mechanics are extremely effective and accurate for explaining and predicting many observations in our world; just not all of it. Einstein's relativity does a much better job, but also not perfect.
Now again, you can say that there is a huge fundamental difference between the flat-earth and round-earth theories, or between Newton and Einstein, and I agree, but here I think Asimov is just arguing from this "correctness of model in explaining/predicting" perspective.
A more opposing viewpoint could be that the point of science is to develop these models that explain why things happen, not just how they happen. So conservation of momentum is interpreted not just as a description of what happens when bodies move or collide, but also as a reason why they behave as they do. This is a frightening viewpoint because I tend to agree with you that we are very far from really understanding what's going on in physics; for example, imagining that some variant of string theory is true, this could completely change what we think "mass" means, even if the predictions about how objects behave is not very different.
Whatever misconceptions our English major has about how science works, this strikes me as incredibly unlikely. If you believe in the law of accelerating returns, then it's almost infinitely unlikely — that the number of scientific revolutions yet to come is vastly larger than of those already past. So it's certainly not correct to say that we had "got it wrong" before, as if each attempt is like rolling the dice all over. It's also wrong (though, as he says, not AS wrong) to say that we've "finally" got it "straight".
Asimov is criticizing the banality of the English student thinking, if he were supporting his claim in a more serious way they would had discussed the subject further.
This is an interesting statement, because it means he had some idea of how far away the sun was from the Earth. If it was, say, a thousand miles up, you'd see huge differences in shadow length depending on where you were standing on a flat earth. Since it's tens of millions of miles away, the difference is negligible.
He didn't really have any idea of the actual distance to the Sun. He just knew it was far enough away that the Sun's rays could be assumed to be parallel.
But then his successor Aristarchus under took to measure the sun-earth distance as follows:
He looked at the angle the earth and moon form when the moon is exactly half full. This would give the sun distance as a multiple of the moon's distance using trigonometry.
He observed the earth's shadow during lunar eclipses. This gave the moon's size in terms of earth size.
He knew the earth's size from measuring shadows at different latitudes on earth and the distance between those sites.
Knowing the moons size vs apparent size, simple trigonometry gave him the distance to the moon. And therefor also the distance to the sun.
Unfortunately his angle measuring techniques were primitive and he decided the sun was 20x further from earth than the moon (rather than about 400x). But his other figures are quite passable.
"Susan Haack compares knowledge to a crossword puzzle. Old answers interweave with old one. They all reinforce one another. The clues are the question we ask. And the way the answers fall into a predetermined grid...well, that's our confidence that we are on the right track."
That's... ahem... wrong.
How could there possibly be an infinite number of things "we" (humanity, I presume?) are wrong about, given that there have only been a finite number of human beings with (as far as I know) finite life spans?
For example: exact prime factorization of all natural numbers, etc...
Of course this is true only if you accept induction. So you can still back off by declaring yourself finitist* ;)