Science vs. Truth
Some people follow ambulances and fire engines to the scene of disaster, but my favourite blood sport is watching online confrontations between Intelligent Designers (Creationists) and those who believe in natural selection and the theory of common descent.
It's not terribly interesting when the ID side just sits there repeating "because the Bible says so", but every now and then up pops an IDer who passed zir high-school science classes and is game to try to prove the science behind Intelligent Design. These encounters typically end in one of three ways:
The third outcome, however, is much more interesting because the IDer is in fact correct, and the defenders of science do themselves and science no favours when, having seen their soft, white underbelly exposed, they throw up their own barricades of dogma by insisting that science deals with "physical truth" or, worst of all, that the IDer is simply too stupid to understand.
Mathematics aside, science is not about truth, it's about predictability. In science we make observations of a physical process then construct models which, given the same starting conditions, produce the same results as that process. The value of this model lies in its ability to predict future outcomes. If the model is a faithful reflection of the physical process then we can use it to tell us what outcome to expect for a given set of starting conditions before those starting conditions have occurred. Science is our oracle, seer and prophetess - better in fact because unlike the oracles of ancient times the predictions of science are usually precise and unambiguous. "If you drop the rock from two meters above ground, it will fall to the ground, striking it at a speed of 6.26 meters per second."
However, no matter how accurate a model is, it still remains just that: a model. That model may reflect the real underlying truth of the physical process, but we have no way of knowing that. We only know that every time we test it, it seems to work. An excellent example of this is Newton's theory of motion. For centuries it stood up to every test we could throw at it until we finally started calling his equations "Laws". Then along came Einstein showing that in fact Newton's theory only held true in a non-accelerating frame of reference.
To give a simple example, let's say that we have a function called ƒ which takes a single value ϰ and returns a result:
result ← ƒ(ϰ)
We try using a few different values for ϰ and get the following results:
ϰƒ(ϰ)
o0
11
24
714
3366
-6-12
105210
In each case we see that the result is twice the value of ϰ. Based on that we develop a hypothesis, a model, in which the function looks like this:
ƒ(ϰ) ← ϰ × 2
That is, the function returns the value of ϰ multiplied by two.
We publish our hypothesis and other researchers see if they can reproduce our results or find a counter-example which disproves the hypothesis. They try the function using different values of their own choosing and find that the model holds up. Eventually the model is accepted as valid theory and computer programmers feel safe to start using the function whenever they want to double a value because the model gives them a predictable result.
But does the model represent reality? Does it embody the physical truth of the function?
In this case it does not, because the function is really as follows:
ƒ(ϰ) ← ϰ + ϰ
That is, the function returns the value of ϰ added to itself, not multiplied by 2.
The mathematically inclined may roll their eyes and snort that this is functionally identical to the model, but that only serves to underscore the point, which is two-fold:
That doesn't matter, though. Even if true our Intelligent Design model would not be science because it fails one of the fundamental criteria of science: it does not provide for testable predictions. Will separated populations of the same species eventually diverge into different species? The ID model doesn't tell us. Maybe The Designer built that into Its creatures, maybe not. Will species adapt to changes in their environment or not? The ID model doesn't say. Is it possible for an asteroid impact to kill off 90% of the species on Earth? The ID model doesn't say. Evidence of such occurrances in the past may have been faked by The Designer.
Given natural selection on one hand and a suitably tweaked Intelligent Design model on the other, both explaining all the observable phenomena of speciation without conflicting with any observations made in other fields of study, science does not ask "Which is true?" but rather "Which is useful? Which one can we use to most accurately predict future and past behaviour?"
When advocates of Intelligent Design insist on having their views taught in science classes as an alternative to the theory of natural selection, we shouldn't allow ourselves to get caught up in philosophical discussions of what "truth" is. Science isn't about truth, it's about verifiable predicatability. Intelligent Design provides no verifiable predictions, so it isn't science.
It's not terribly interesting when the ID side just sits there repeating "because the Bible says so", but every now and then up pops an IDer who passed zir high-school science classes and is game to try to prove the science behind Intelligent Design. These encounters typically end in one of three ways:
- Having had every logical argument disproven, the IDer falls back on "the Bible says so".
- The IDer denies those objective, documented facts which would otherwise disprove Intelligent Design.
- The IDer points out that science cannot be "proven" either and is as much of an act of faith as Intelligent Design.
The third outcome, however, is much more interesting because the IDer is in fact correct, and the defenders of science do themselves and science no favours when, having seen their soft, white underbelly exposed, they throw up their own barricades of dogma by insisting that science deals with "physical truth" or, worst of all, that the IDer is simply too stupid to understand.
Mathematics aside, science is not about truth, it's about predictability. In science we make observations of a physical process then construct models which, given the same starting conditions, produce the same results as that process. The value of this model lies in its ability to predict future outcomes. If the model is a faithful reflection of the physical process then we can use it to tell us what outcome to expect for a given set of starting conditions before those starting conditions have occurred. Science is our oracle, seer and prophetess - better in fact because unlike the oracles of ancient times the predictions of science are usually precise and unambiguous. "If you drop the rock from two meters above ground, it will fall to the ground, striking it at a speed of 6.26 meters per second."
However, no matter how accurate a model is, it still remains just that: a model. That model may reflect the real underlying truth of the physical process, but we have no way of knowing that. We only know that every time we test it, it seems to work. An excellent example of this is Newton's theory of motion. For centuries it stood up to every test we could throw at it until we finally started calling his equations "Laws". Then along came Einstein showing that in fact Newton's theory only held true in a non-accelerating frame of reference.
To give a simple example, let's say that we have a function called ƒ which takes a single value ϰ and returns a result:
result ← ƒ(ϰ)
We try using a few different values for ϰ and get the following results:
ϰƒ(ϰ)
o0
11
24
714
3366
-6-12
105210
In each case we see that the result is twice the value of ϰ. Based on that we develop a hypothesis, a model, in which the function looks like this:
ƒ(ϰ) ← ϰ × 2
That is, the function returns the value of ϰ multiplied by two.
We publish our hypothesis and other researchers see if they can reproduce our results or find a counter-example which disproves the hypothesis. They try the function using different values of their own choosing and find that the model holds up. Eventually the model is accepted as valid theory and computer programmers feel safe to start using the function whenever they want to double a value because the model gives them a predictable result.
But does the model represent reality? Does it embody the physical truth of the function?
In this case it does not, because the function is really as follows:
ƒ(ϰ) ← ϰ + ϰ
That is, the function returns the value of ϰ added to itself, not multiplied by 2.
The mathematically inclined may roll their eyes and snort that this is functionally identical to the model, but that only serves to underscore the point, which is two-fold:
- It is impossible to distinguish between two internally consistent models which produce identical results from identical inputs. Either one could represent physical truth and it may be that neither does.
- It doesn't matter whether a model represents physical truth. What matters is that it produces the correct results.
That doesn't matter, though. Even if true our Intelligent Design model would not be science because it fails one of the fundamental criteria of science: it does not provide for testable predictions. Will separated populations of the same species eventually diverge into different species? The ID model doesn't tell us. Maybe The Designer built that into Its creatures, maybe not. Will species adapt to changes in their environment or not? The ID model doesn't say. Is it possible for an asteroid impact to kill off 90% of the species on Earth? The ID model doesn't say. Evidence of such occurrances in the past may have been faked by The Designer.
Given natural selection on one hand and a suitably tweaked Intelligent Design model on the other, both explaining all the observable phenomena of speciation without conflicting with any observations made in other fields of study, science does not ask "Which is true?" but rather "Which is useful? Which one can we use to most accurately predict future and past behaviour?"
When advocates of Intelligent Design insist on having their views taught in science classes as an alternative to the theory of natural selection, we shouldn't allow ourselves to get caught up in philosophical discussions of what "truth" is. Science isn't about truth, it's about verifiable predicatability. Intelligent Design provides no verifiable predictions, so it isn't science.