Theory, predicability and truth

[Magical Realist]

Is the ability to make accurate predictions a good criteria for deciding the truth of a theory? Suppose for example I believe in the creationist account of Noah's flood and that the fossil record will always show a progression of lifeforms from the aquatic to the semi-aquatic to land-based because that is simply the order in which they were killed by the flood. Fish, then frogs, then reptiles, then dinosaurs, then mammals and birds, because that is the order in which they were drowned by the flooding of the earth by fresh water. And lo and behold that prediction holds true. But that doesn't mean the creationist account of the flood is necessarily true.

Take another example. There was a time when people believed the night air to be noxious and full of vapors that can make you sick. So people would avoid going out at night to keep from getting sick. Turns out they WOULD tend to avoid sickness, but not for the reason they thought. What was really happening is their body temp was not being lowered thus making them less susceptible to illness. The theory allowed for accurate predictions, and yet was totally fallacious.

A theory therefore MAY allow for accurate predictions, but NOT be true. Is predictability necessarily correlative to truth?

 

[sideshowbob]

Suppose for example I believe in the creationist account of Noah's flood and that the fossil record will always show a progression of lifeforms from the aquatic to the semi-aquatic to land-based because that is simply the order in which they were killed by the flood. Fish, then frogs, then reptiles, then dinosaurs, then mammals and birds, because that is the order in which they were drowned by the flooding of the earth by fresh water.

That prediction fails. Fish would not be the first to drown, yet they are near the bottom of the fossil record. Birds would not be less likely to drown than dinosaurs. In fact there is very little correlation between the fossil order and the most likely order of drowning.

 

[Magical Realist]

That prediction fails. Fish would not be the first to drown, yet they are near the bottom of the fossil record. Birds would not be less likely to drown than dinosaurs. In fact there is very little correlation between the fossil order and the most likely order of drowning.

True enough. You won't find grass-eating deer next to grass-eating triceratops either. It's just an example to illustrate my point. Can you think of a theory or folklore that provided predictability yet turned out to be false? What about the Ptolemaic system of astronomy? And otoh, are there examples of true theories, or at least ones with great explanatory power, that provide no predictions? I raise this point to propose a difference between predictability and truth. So often predictability in theoretic science is lauded as proof of its truth. But I suspect it might not always be the case.

 

[Dywyddyr]

Is the ability to make accurate predictions a good criteria for deciding the truth of a theory?

Science doesn't deal in "truth" [sup]1[/sup].
What is required is that it's workable.
If a theory gives predictions that turn out to match subsequent observation then the theory is accepted.
But it can always be replaced or superseded by something that works better without any kerfuffle over whether the prior one was "true" or not.

1 In the not-exactly-trivial sense of "is it seen to match reality" then it does, but certainly not in the truth-with-a-capital-T sense.

 

[C C]

Take another example. There was a time when people believed the night air to be noxious and full of vapors that can make you sick. So people would avoid going out at night to keep from getting sick. Turns out they WOULD tend to avoid sickness, but not for the reason they thought. What was really happening is their body temp was not being lowered thus making them less susceptible to illness. The theory allowed for accurate predictions, and yet was totally fallacious.

Mmmph. Was going to drop this recent yet still belated "Correlation does not equal ___" gospel hymn in that synchronicity thread, but perhaps happenstance has intervened again...
Algorithm Reveals Link Between Sour Cream And Traffic Accidents

"Data suggest: marriages in Alabama are causing deaths by electrocution, divorces in South Carolina are causing bees to produce more honey, and Nicolas Cage movies are saving thousands of lives each year. Tyler Vigen created a program, appropriately titled Spurious Correlations, that finds correlations between random data sets and produces a chart every minute."​

 

[Arne Saknussemm]

Mmmph. Was going to drop this recent yet still belated "Correlation does not equal ___" gospel hymn in that synchronicity thread, but perhaps happenstance has intervened again...
Algorithm Reveals Link Between Sour Cream And Traffic Accidents

"Data suggest: marriages in Alabama are causing deaths by electrocution, divorces in South Carolina are causing bees to produce more honey, and Nicolas Cage movies are saving thousands of lives each year. Tyler Vigen created a program, appropriately titled Spurious Correlations, that finds correlations between random data sets and produces a chart every minute."​

It all goes back to Mark Twain: "There are lies. There are damned lies. And then there are statistics."

 

[Aqueous Id]

I agree with our esteemed member with the unpronounceable and untypeable username.

Theories are explanations of phenomena based on best evidence. They are always instantly true. That is, they are true for the evidence from which they arose. And of course they are always subject to change in the event any new data becomes available which might shift the baseline.

The Ptolemaic system was never quite true. From the outset, the brightest luminaries in the sky, the bright planets, were not observed circling the Earth. This provoked all kinds of inquiry and elaborate schemes to force them to be in orbit around the Earth. Tycho Brahe, before he was burned at the stake, was at least able to take a scientific approach and compile the times and positions of each of the planets, which produced the first repeatable set of ephemerides. Most remarkable is that Tycho's data is good to about three digits, a conclusion which amazed Johannes Kepler when he discovered that the data points described orbital ellipses around the Sun, and from which he derived his three laws of planetary motion.

Repeatability is an essential test for the truth of any explanation. But it is one of many criteria necessary to arrive at the correct answer. We need the scientific method to follow through. That's not just lip service. It takes a lot of actual hard work to collect the data that Ptolemy collected and that Tycho collected. It takes a lot of resilience against the oppressive forces of religion and politics which wage war on any science that would overturn superstitious beliefs or otherwise oppose ignorance. It takes the fluency in math which Kepler and Newton both exhibited in their analysis of the data. And as Newton demonstrated, it takes fluency in physics, and the ability to think outside the box without sacrificing first principles, and prolific creative genius.

Without retreading old tires, suffice it to say that the scientific method is merely a systematic form of intellectual quality control. Wherever it is employed we have the highest probability of finding the truth. Wherever it is discredited is where we have the highest probability of finding the lies.

 

[paddoboy]

Without retreading old tires, suffice it to say that the scientific method is merely a systematic form of intellectual quality control. Wherever it is employed we have the highest probability of finding the truth. Wherever it is discredited is where we have the highest probability of finding the lies.

Keep retreading Aid, keep retreading!
That should be literally shoved down the throats of all our "would be's if they could be's" and those that expect to come to a forum such as this to overturn 20th/21st century cosmology.

 

[sideshowbob]

Can you think of a theory or folklore that provided predictability yet turned out to be false?

How else do you decide a theory is false? A hypothesis is a prediction. I hypothesize that my keys are in my other pants. I predict that I will find them there. If the test fails, the prediction fails and the hypothesis fails.

What about the Ptolemaic system of astronomy?

The Ptolemaic system worked in a kindofa-sortofa way but it required ever more complicated ad hoc modifications, like epicycles, to work. Notice also that it didn't predict epicycles.

And otoh, are there examples of true theories, or at least ones with great explanatory power, that provide no predictions?

I don't see how there could be. A theory must come from the testing of hypotheses. How can you test hypotheses without predicting the outcome? How can you even design an experiment if you don't know what you expect to see?

So often predictability in theoretic science is lauded as proof of its truth.

There is no such thing as "proof of truth" in science.

 

[Yazata]

Is the ability to make accurate predictions a good criteria for deciding the truth of a theory?

Depends on what we think that 'truth' means, I guess.

There's instrumental truth and ontological truth.

Instrumental truth correlates observations, measurements, the results of experiments. It tells us that if we measure X and Y, then we will measure Z too. In physical science that's often quantitative and precise, and that's where the mathematical equations stuff comes in.

Ontological truth tells us about what exists, what's really happening, so as to make the experimental results come out as they do. It asserts the reality of things like atoms and DNA.

A theory therefore MAY allow for accurate predictions, but NOT be true. Is predictability necessarily correlative to truth?

A paradigmatic example of a situation where the instrumental and ontological concepts of truth diverge is in quantum theory. The formalism of quantum mechanics succeeds in correlating experimental observations very well, at least if we are willing to accept its inherently probabilistic nature. But it doesn't really tell us very much about how reality is so as to make the measurements come out the way they do. All the experimental results and the resulting mathematical formalism does is place constraints on the range of possible quantum mechanical interpretations.

But that still leaves us with a huge assortment of mutually inconsistent physical (or perhaps logical) accounts of what might really be happening down there on the microscale.

 

[spidergoat]

If a theory can't say exactly why a phenomenon happens, only that it predicts something will happen, it is an incomplete theory. There are many folk remedies, and some of them actually work. But the practitioners can't explain the active ingredient or why it works.

 

[James R]

Is the ability to make accurate predictions a good criteria for deciding the truth of a theory?

"Truth" is a somewhat slippery concept. Perhaps it would be better to say that the ability of a theory to make accurate predictions is a good criterion for deciding whether the theory is useful.

We should also be a little careful not to confuse predictions with "retrodictions". For example, consider the Noah's flood example mentioned above, and ask yourself what novel predictions that theory makes about the fossil record. That is, what does it predict or what has it predicted that was not already known about the fossil record at the time the "theory" was constructed?

It can be relatively easy to come up a "just so" story that explains the currently-known state of affairs. But the real power of a scientific theory is often seen in its capacity to predict what will be seen in future, or if specific new observations are made.

It should also be said that the power of a scientific theory is not just in its predictive ability. The best theories, for example, tend to be able to account for a very wide range of superficially-distinct phenomena. This is part of the reason why Newtonian mechanics or Evolution by Natural Selection are considered such powerful theories.

 

[wellwisher]

There is applied science and there is pure science. Applied science creates new practical applications from science. Applied science does not need to know how things works, to new create application, although a good working knowledge of the science, can make this easier to predict. For example, a chef does not need to know the chemistry of jello, to be able to make jello gummy bears. However, a food chemists, who is an expert at gels, will be able to predict that jello gummy bears can be made ever without doing a bunch of trials in the kitchen. Applied science can be done with trial error, or with logic. It is extrapolating science in both cases, with only the latter needing a working knowledge of the pure chemistry.

Pure science is more concerned with knowing why something is and/or how it works. This is closer to the truth. With applied science, you don't need the truth to make predictions, since it can be done empirically. For example, statistical experiments can place the truth in a black box and can still get good results. We have no idea how medicines impact the integrated body. We only know about the superficial outputs that appear.

Evolution by natural selection is more empirical version of pure science. It can look at the past, and see important trends, but it can't make exact predictions of the future. It can only make generalizations type predictions, even though it has detailed historical data. It can tell us that humans will continue to evolve, but it can't tell us what change will entail or in what sequence. In that sense, it sort of like saying there will be weather tomorrow, but we don't know when, where or how this will appear. We may have accurate records of the past, and see patterns, but beyond that, it is crap shoot since the science is not that pure.

If the science is pure; close to the truth, it should be a stable platform for accurate extrapolations, via logical applied science. Evolution is not that stable, of a platform, in its present form. It is more like chefs in the kitchen, who experiments with food chemistry; combining and baking, while not knowing about food chemistry at the molecular level. These chefs can still do all types of useful and amazing things, but in an empirical way. This can tell us things but is not a stable platform for accurate predictions of new applications.

The main reason evolution is not pure science, is nothing works in life without hydrating water. That is a truth in life chemistry, that is not given enough weight.

 

[danshawen]

Microorganisms including diseases and parasites don't have a lot of advantage or a lot of survival strategies in terms of mobility for infecting host organisms, or hosts that expose themselves by wider ranging activities, at night or otherwise will be increasing their chances of exposure to disease vectors simply by being active or exposing themselves more frequently to other potential hosts.

The cold night air may or may not have very much to do with it, but it is a scientific fact that moist surfaces, even on host organisms, become much "stickier" at lower temperatures, and more likely to collect and retain the microorganisms responsible for disease for longer periods in colder weather. Static electric charge when weather is dry is also an attractive mechanism that would be difficult or impossible to mitigate completely.

Stick that to anyone who suggests there are no good scientific reasons to shun contact with other humans during the cold and flu season. There is good reason to call them colds. Something to consider before going into education or getting that medical degree also. This idea won't stop most people from doing what they are good at, and Typhoid Mary was no exception.

If you are still alive, infected or not, you take your chances. The more infected you are, the less mobile you will be. At least that mechanism works better for survivors. Evolution will take full advantage of anyone who resists, effectively or otherwise.