h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

What is more important in a scientific theory (or any theory or description of the world): that is actually describe the true nature of existance or that it make accurate predictions in as many circumstances as possible?

Ooooooh. What a splendid question. OK. Don't know much about scientific theory, but my two interests are History and Language (shading into Linguistics) and in both of those disciplines ir is much more important (and much more interesting) to describe the true naure of existance rather than predict.

In science (though that's a broad church, isn't it), I'd imagine that a lot of it is, well, both. I mean Biology and Astronomy are more about diescribing the nature of... whereas, what, physics? or at least chemistry is very interested in coming up with formulas to not exactly predict the future, but if you do X then Y _will_ happen. But then I suppose you could say the same about evolution theories, except it would happen really really slowly.

So I's say: either or both, depending on the focus, and I'm more interested in the description. But I will be fascinated to see how this descusion develops.

Ultimately it's the first, with some reservations. The second is just a guide (if your predictions are always wrong, then you have some aspect of "reality" wrong). That's why scientists and other thinkers are not necessarily happy with theories that work, but contain ad hoc or inconsistent elements.

Noggin

Primary reservation

"We have every reason to believe that the best theories in physics are true, but they present us with a picture of reality that makes no sense to the intuitions about space, time and matter that evolved in the brains of middle sized primates....It's impossible to stop thinking thoughts that are literally incoherent, such as 'What was it like before the big bang?' or 'What lies beyond the edge of the universe?' or 'How does the damn particle manage to pass through two slits at the same time?'"
Steven Pinker


Noggin

How can you tell the difference between a theory which accurately describes something and one that makes predictions? Surely to know whether a theory describes something accurately you compare its descriptions with what you see (or hear, etc.), and this is a sort of prediction. For example, the following would probably be called a description: "The human body has a heart, liver and lungs, amongst other things." Whereas this next sentence would probably be called a prediction: "Momentum is conserved." But really, what the former is saying is: "If you open up a human body, you'll find a heart, liver and lungs", which is similar to what the latter is saying, i.e. "if you measure the momentum of a system, and then again a while later, it will be the same". If you'd said "The human face has freckles" it would be an inaccurate description because the sentence "If you look at a human face you'll see freckles" is an inaccurate prediction.

If you mean to ask whether it's important that there is some sort of correspondence between the basic elements of a theory and what is 'really out there', I don't think that's even a well defined question, and it's certainly irrelevant. If you can't even know whether or not the basic elements of your theory correspond to something in the world or not (and you can't, you can only know if your theory makes accurate predictions or not, and that only in a very limited way), how could it be important that it do so?

"If you mean to ask whether it's important that there is some sort of correspondence between the basic elements of a theory and what is 'really out there', I don't think that's even a well defined question, and it's certainly irrelevant. If you can't even know whether or not the basic elements of your theory correspond to something in the world or not (and you can't, you can only know if your theory makes accurate predictions or not, and that only in a very limited way), how could it be important that it do so?"

That's the question I was trying to ask, and I think I agree with your answer to it.

Right, I think that it's a good answer to the question, although it doesn't seem to be as widely held as you might think. Most philosophers seem to believe in the correspondence theory of truth which is sort of the opposite of what I was saying.

"Most philosophers seem to believe in the correspondence theory of truth which is sort of the opposite of what I was saying."

Perhaps that is an essential difference between philosophy and science. Certainly not the only one, but a significant one. Philosophy perports do define the "Truth", science merely attempts to understand and predict what happens.

I think you're probably right to a certain extent, but it's more a difference of attitude than anything else. There are philosophical theories and philosophers who are dubious about 'Truth' (like Wittgenstein, or the pragmatists: James, Peirce and Dewey), and equally most scientists tend to speak about their work as if it were something that could be true or false, for example you might find a physicist saying "It's an open question whether or not there are gravitons (a form of fundamental particle)". Despite that, philosophers, even the pragmatists, seem to still be very interested in resolving this issue, whereas scientists just don't worry about it and get on with things.

Anyway, how about this as a modified question: Despite what I said earlier, there does seem to be a difference between the theories of biology and physics which seems to best expressed by saying the former is descriptive, the latter predictive. Can we clarify what we mean by saying this sort of thing? and perhaps if we can then we can answer a modified form of your original question, whether it's more important for science to describe or predict.

*phew*

Have nothing to add right now, just observing

"Despite what I said earlier, there does seem to be a difference between the theories of biology and physics which seems to best expressed by saying the former is descriptive, the latter predictive. Can we clarify what we mean by saying this sort of thing? and perhaps if we can then we can answer a modified form of your original question, whether it's more important for science to describe or predict."

I think that that difference is illusiary. I think that really the difference it that, because biology is still reletively more primative, it has fewer true theories and more of it is merely attempts to describe things. But the true theories are intended to predict. Physisists just know more about their area of study and can thus usually predict mathematically, which is more accurate. Biology will probably eventually end up looking like physics--mathematical predictions.

okay I am gamely gonna have a go at this one..

Niether reality or predictions are erroneous, part of our nature is to examine, evaluate and come to conclusions, and even have fun and brain storms whilst doing it, it help us simians evole somewhat and alter our environment. For a prediction right at that precise moment in time by an 'experiment' is I suppose a correct prediction for reality at that time...

I also think that if you are going to conduct an experiment to make a 'prediction' about something, its very hard in 'reality' to do, because of a million different factors ie: chaos theory involved, before, during and after said experiment.

Is that where you meant <> RDO?

Think my brain just went 'pop'

So do you think that mathematical predictions should be the goal of all theories? What about the theory of evolution? It doesn't and probably never will predict anything, but it still seems to be a useful and important theory.

That last post was for Daneel (a fellow Asimov fan I think). Vixen, I think you're right about the chaos theory (and so on) thing. This is why I think it might be important to have theories which aren't predictive, because there are some things that can't be predicted mathematically because of chaos theory (and the like).

While I broadly agree with Dogster, I think what I was trying to say earlier is that scientific theory is not "just" about prediction, but also elegance, economy of explanation, and fruitfulness (suggesting new lines of enquiry).

I think it could also be said that while we can never encounter reality-as-it-is-in-itself the assumption that there *is* such a thing is absolutely necessary for any enquiry of a meaningful kind.

Noggin

"So do you think that mathematical predictions should be the goal of all theories? What about the theory of evolution? It doesn't and probably never will predict anything, but it still seems to be a useful and important theory."

I think it should be the goal to be quantitative rathher than qualitative. A qualitative theory is still useful, but it is preferable to make it qualitative if possible.

And yes, I am an Asimov fan. I am the same person as the researchers Susan Calvin, Hari Seldon, R. Giskard Reventlov, The Last Imperial, and some other non-Asimovian ones.

"I also think that if you are going to conduct an experiment to make a 'prediction' about something, its very hard in 'reality' to do, because of a million different factors ie: chaos theory involved, before, during and after said experiment."

It depends how specific the prediction is going to be. I can be very certain that if I drop a basketball on cement, it will fall and bounce. How high it will bounce and in what direction are more chaotic and harder to predict.


"Is that where you meant <> RDO?"

No, I was talking about the idea that physics is predictive but biology is descriptive.

<>

Okay,

Is also an Isaac A, fan even tho I don't understand everything I read.

Hi Everyone,

Great discussion, right up my alley

I think this question comes from a common misunderstanding, that a scientific theory has to predict something to be valid. I could be wrong, however.

All of science (which is a type of philosophy, so no 'science vs philosphy' comparisons are really valid) starts with a prediction: the hypothesis. It basically states that with all of the available information we think this will happen here, and we can test it. So the prediction is being *used* to test reality. The scientific method can only infer what's real or not through the testing of hypotheses. In the eyes of science, reality only unveils itself under the scrutiny of a well-designed and experiment.

As you may notice, this is quite limited if you look at it from a non-science viewpoint. I don't think any scientist will say anything they discover represents true, certain reality. There is no absolutes in science, and that follows for the last three words of the previous sentence. The question posed is restrictive because it forces you into one of two camps that are not mutually exclusive nor necessarily valid.

However, to address the question directly:
"What is more important in a scientific theory ...: that is actually describe the true nature of existance or that it make accurate predictions in as many circumstances as possible?"
I think that any theory that closely describes the nature of reality will hold up to scrutiny by it's very definition. Therefore it should also be able to extend itself into the future, making predictions. There aren't two types of theory. Just theories, which are the end product of the scientific method.

Regarding evolution: The theory can and does make tremendous predictions, and a lot of it *is* expressed in mathematical equations. For examples, see molecular biology, genomics and population genetics, including the Hardy-Weinberg equilibrium, modelling of genetic drift, and game theory. Biology is predictive and descriptive. Evolution is said to be happening when alleles (different forms of a gene) change in frequency in a population. That's it. And that's what population genetics involves. Using statistics we can make prediction and models based on what has been observed in nature. Because of the random nature of facets of evolution nothing can be predicted with certainty, which in itself reflects the maturity of the field (interestingly, the modern synthesis of evolutionary theory was decided upon over 50 years ago, before the discovery of DNA by Watson and Crick, it has held up and been augmented ever since).

The main difference between biology and physics is complexity. Biological systems are unimaginably complex things. Reducing them down to their components often doesn't do anything (hence holistic fields like genomics). Physics is highly reductionist and simple in comparison. Old school biology (in the 1800s and well into the 1900s) was divided up into Zoology, Botany, etc. It was basically stamp-colecting all of God's 'creations' to shown His Infinite Wisdom. It aint like that anymore. Biology is an advanced, multidisciplinary science. It's no accident that it's at the forefront of science.



Ste (genomics guy)

"I think this question comes from a common misunderstanding, that a scientific theory has to predict something to be valid. I could be wrong, however."

While arguing with a biology PhD student about this will probably result in me looking like a moron, I would say that a scientific theory must make predictions to be valid. If it makes no predictions, how can it be falsified?


Actually, the question is the result of a Steven Hawking quote. I was reading the text of a debate between him and Roger Penrose and he said that one of the main differences between them was that Penrose considered how acurately a theory described reality important, but Hawking, as a posivist, thought that only the predictions it makes are important.