VoijaRisa
(?)Community Member
- Posted: Tue, 28 Aug 2007 05:40:03 +0000
As a veteran of the Evolution “debate” (I put debate in quotes because there’s really no debate in the scientific field, but rather a contrived political/religious controversy), one of the frequent claims that I’ve seen is that much of what science has revealed about our universe isn’t really science, because we “can’t test it.” After all, the claim goes, we can’t recreate the entire history of the Earth to make sure things evolved the way science says they did, or repeat the Big Bang in a beaker.
As a scientist, I can say this is absolute nonsense. The only thing such comments reveal is a profound misunderstanding of what is meant by evidence and testing in the scientific field. Thus, the purpose of this post is to explain how hypotheses are actually tested, especially how the hypotheses of Evolution and the Big Bang gained the highest status a hypothesis can become in science: a theory.
Before I go any further, let’s first define these terms and how they’re used in science. In every day usage, “theory” just means “guess” or “conjecture”. However, in science, “theory” has a very different connotation. What people generally consider a theory is actually a “hypothesis”: a proposed statement of how something works that has not undergone rigorous testing. If a hypothesis does undergo rigorous testing and every test confirms the predictions it makes, then it is eventually upgraded to the status of “theory”. Again, this means that it’s exceptionally well supported. This is the definition I will be using for the rest of this thread, unless otherwise noted.
But what about all those other important terms that we always hear people tossing around when discussion science? What about laws, and facts, and proof?
Laws are, for the most part, very similar to theories. The main difference is how much they cover; Laws are extremely narrow and don’t seek to explain much. They’re usually associated with a mathematical expression. For example, Kepler’s laws are very short (Planets travel in ellipses, p^2 = a^3, ….). Meanwhile, theories are far more general. Evolution seeks to explain the diversity of life, provide a mechanism, explains progressions in fossils, inheritable traits…. Both are exceptionally well tested. It’s just that theories are much broader. But just like theories, laws are not necessarily “proven”. In fact, many have been shown to be incorrect. For example, Newton’s laws of gravity did not work properly for Mercury’s orbit. It was eventually realized that this was because of relativistic effects (ie, close to the speed of light) and Newton’s laws were supplanted by the theory of relativity.
Facts, as big and important as they sound, are actually quite useless in science. Since for all you know, gravity could turn off tomorrow and we’d all float away, we can’t say for certain that gravity or anything else in science is 100% fact. Theory is the best it gets. So what are facts? Facts are the bits of evidence. They’re things like “at time t I observed something at position y”. In and of itself, a fact is useless. Combine it with a bunch more and explain it all with a nice theory, and now you’re actually getting somewhere. The point is, that contrary to popular belief, facts are not better than theories.
Similarly, there’s no such thing as proof. The best we can do is to say that every observation has confirmed predictions made by a hypothesis; but since we can’t be sure what the next observation will show, there’s no proof. Only exceptionally well supported hypotheses called theories.
Now that we’ve defined our terminology, how do we use all these terms to understand how science works?
First, let’s start off with something general that no one should really have any disputes about. Let’s take an example of a ball falling. The scientific method starts with an observation. In this case, the observation is that things fall (profound, I know). This is a fact.
From there, a question is posed. For a ball falling, we can ask how it falls. A prediction is made: It falls due to the force of gravity which, on Earth causes an acceleration of 9.8 m/s^2. Going through a bit of math, we can use that to predict where a falling body should be at any given time and then check it against reality. This gives us our hypothesis. Hypotheses are generally mathematical or at the very least, qualitative models of an event.
So we then rig up an experiment. We’ll take a ball, drop it and measure it’s position every second. We can then plot its position on a graph and compare it with the prediction. If it matches closely, then we say that our prediction is good and if we test it a whole bunch more and it continues to do well, it eventually becomes a theory. If not, then the hypothesis needs to be reformulated or scrapped all together.
But let’s imagine what a graph of this should look like. On the x-axis we’d have time in seconds. On the y-axis would be how far the ball had fallen. Here’s a quick sketch of that.
IMG 1
The dots are the actual measurements, and the solid line is the position predicted by your hypothesis.
But OH NOES! The points don’t all lie on the line! Does this mean that they gravity hypothesis is wrong?!
Of course not. To understand why, let’s take a slightly more in depth look. One of the things this graph doesn’t show is what scientists call “error”. This doesn’t mean anything was done wrong, or that the measurements were wrong, but rather, that there is some inherent limitation to how accurately we can measure things. Basically, there’s some random error associated with measuring anything. One of the strong points of science (albeit frustrating for those partaking in it) is that we do what’s known as error analysis, in which we take these errors into account and see if the data is still close enough to match our prediction from our hypothesis.
This is all done through the power of statistics. From this, we actually know (in terms of a percentage) how well our data is matching our prediction. Again, if it fits pretty well, we keep it (perhaps with a bit of tinkering to get the prediction a bit closer), and if it sucks, we scrap it.
However, just because our data doesn't match, doesn't mean that the hypothesis is completely flawed. If the error intrinsic to the measurement is large, then we need to take that into account and find a way to reduce that uncertainty. This can be done either by making more accurate experiments, or simply by taking the average of a lot of observations. The more observations, the smaller the error gets. This is why we keep testing things in science. Incidentally, because the first time an experiment is run no earlier tests exist, this means that the initial hypothesis is likely to be quite different than reality. It requires tinkering and more observations to reduce errors. In this light, when people say that some of what Darwin said was wrong, it should be considered beside the point. We never expected it to be dead on. There's details that obviously wouldn't match. But as we've learned more, we've corrected the details he got wrong. The theory of evolution today, is not the same as the one Darwin posited over 100 years ago.
Another thing that’s important to realize here, is that we don’t have to have a data point at every point along the line. In other words, we don’t need to see the height of the ball every nanosecond. Every second or two is good enough to get a good idea of how well the prediction matches up with reality. Keep this in mind. It will become important later.
So if we take into account the inherent limitations of measuring anything, it turns out that even though those data points don’t perfectly match up with the prediction made by our gravitational hypothesis, it still matches extremely well.
Now, what does any of this have to do with evolution?
Well, let’s walk through the same process. First, we make the observation that Darwin did: That there are a number of species that while similar, are uniquely adapted for their environments. Darwin then posited that these similarities were indication of a common ancestor, and changed due to environmental pressures, which drew out beneficial mutations. The prediction is, that when we look at the fossil record, we should observe the gradual changes. In other words, we should see the branching of the species tree (aka the phylogenic tree). If that’s not familiar enough for you, let’s take a look at what Darwin predicted:
IMG 2
Can we find this in nature? Absolutely. To save time though, I’m only going to look at individual splits instead of the whole three though. Basically what this means is that we’ll zoom in on one fork.
To demonstrate this, let’s look at an ancient species known as Rhizosolenia. It’s a genus of single celled organism that have very distinctive silicate cell walls. Fossils of their remains from 1.7 million years ago show two separate species. One of the defining characteristics of is the area of these shells. If we plot these areas vs. time, we get a graph that looks like this:
IMG 3
Again, with this graph we have a nicely defined trend with several data points. We don’t have a size at every point in time. There’s lots of “gaps” or “missing links”, but it’s still enough to clearly show what’s going on: One species split into two. This is just one example among many. Is this proof of evolution? Absolutely not.
As I stated before, it’s impossible to “prove” everything in science. All scientists can do is prove something beyond a reasonable doubt. In other words, the case of Rhizosolenia just begins to lend credibility to evolution. But we need even larger changes to really make evolution credible. This is clearly demonstrated by the evolution of whales from terrestrial ancestors.
So it’s clearly established that Darwinian evolution can account for changes between phyla, one of the largest differences in animal groups. Again, this doesn’t prove evolution, but it makes it an extremely powerful candidate since it is able to predict how species branch that matches closely with what is observed.
One of the most contentious instances of evolution, however, is the evolution of humans from an ape-like ancestor (often mischaracterized as humans from monkeys). One of the main characteristics that defines a human is the size of the skull which has gifted us with our larger brain capacities. If evolution is true, then we should be able to trace the size backwards in time until it is similar to the species from which we diverged.
Despite the claim of creationists, there are a good number of intermediary forms from which we can study the brain capacity. So let’s take a look at this:
IMG 4
From this, we can see a clear trend of increasing brain size, from ~400 cc to the roughly ~1400 cc we have today. So what we have is a very nice trail in accordance with evolutionary predictions. Are there “gaps”? Sure, but we still have a very nice trend, even without having a fossil for every step along the way. Thus, the “missing link” argument creationists frequently tout is rather pointless. We just don’t need links at every step. The trend is clear without them.
But this instance is only looking at one feature along one branch. From this next graph, we can see the trend and branch going the other way, even more distinctly:
IMG 5
Near the branching point, we have the species most similar to the common ancestor with a small brain, large teeth, and use facultative bipedalism (ie, the wear on fossils show they occasionally walked upright for short periods of time). That group then split to become the green branch which continued to diversify into modern humans, while another leg of the branch went off to blue group and eventually the orange, modern apes.
Here's another graph with essentially just the left branch of that graph:
IMG 6
The "missing links" that are so often referred to would be the "?"'s between branches. While it's true we don't have every piece, what should be understood is that this is not necessary. What we see is that the overall prediction is correct. Note the similarities in the shape of this graph to the earlier one of the Rhizosolenia. Both show clear divergences, even without having every step explicitly defined.
There may be some minor holes, but these aren't all that important. Because we've seen that the process works so well in every case we can observe, there's no good reason to suspect it doesn't where we haven't gathered enough data yet. This would be the equivalent of saying that, although I can trace my genealogy back several generations, when I can't figure out who our great, great, great grandmother was, that it meant that I didn't have one and my lineage suddenly popped into existence. Such an answer is horribly unparsimonious! In other words, it invokes unnecessary and untestable mechanisms.
Thus, we see that the concept of “gaps” and “missing links” is really an inconsequential bit of rhetorical nonsense.
But what about other tests for humans being inexorably linked to the rest of the great apes family? All great apes besides humans have 48 chromosomes. Humans only have 46. Since the loss of two whole chromosomes would have absolutely been fatal, this would imply that one set of our chromosomes came from a merger from the common ancestor. We can look for this by looking for the “stop” instructions in the DNA, known as telomeres.
It turns out that in the middle of chromosome #2, there is a deactivated telomere, which is the result of a gene fusion. If this were not observed, common descent from the great ape family would be falsified. No telomere; No common ancestor.
Can other hypotheses account for this? Creationists often posit that similarities in DNA could very well be the result of the creator using similar methods and thus, is “evidence” for creationism under that interpretation. However, the difference is that if the DNA weren’t similar, evolution’s prediction would be falsified and evolution would be in trouble. Meanwhile, if that were the case, creationists could just claim that it’s evidence that the designer were really creative. Either way, they can claim victory for their designer. In other words, it can’t be falsified. Thus, if there can’t be evidence against, you can’t really claim to have evidence for either.
Evidence only comes when you pass a test that could have falsified your hypothesis.
Another often argued part of evolution by creationists is that of the "Cambrian explosion" (which is universally misrepresented by creationists). They often moan about how there's no transitional fossils and that species formed "as if overnight". But yet again, we see the distinct overlapping and branching pattern predicted by evolution:
IMG 7
But let's go back to the concept of "error" I discussed earlier and explore how it works a bit more. It should be clear that data at every single point along the way isn't necessary. But as I mentioned, the more times you test something the more your error goes down, and the greater your confidence in your hypothesis becomes. I've just presented a handful of tests that are extremely conclusive. In science, having just a handful of data points that are so well constrained and with such low errors, already gives you extremely high confidence.
However, like I said, having lots of measurements, even without really small errors on each individual ones, can be taken collectively to still give extremely high confidence. Here, evolution shines again. Consider this: It would only take one fossil to disprove all of evolution. This is often referred to as the proverbial "fossil rabbit in the Precambrian." Effectively, what it says, is that if we even find one fossil out of place chronologically, then all of evolution is shot. So if evolution were wrong, it should be pretty easy to find fossils out of place that didn't match with evolutionary predictions. In this light, every fossil dated is a test of evolution. This gives thousands of tests for evolution, of which it has passed every single one. Like I said earlier, the more tests you have, the greater the confidence. In this regard, the confidence level of evolutionary theory is almost unprecedented.
Another way we often look at things in science is to check what's known as the "bounds". In terms of evolution, this would just to do a simple check to make sure that the fossil record doesn't show things evolving faster than mutations can account for. If species were changing faster than that, mutations couldn't account for this change and evolution would be falsified.
To test this requires two values: One for how fast things change, and another telling us what the maximum rate mutations can compile is. To measure this, we use units called "darwins". It turns out that the upper limit for how fast mutations could compile (under very harsh forced selection criteria) is somewhere around 10000 darwins. Typically, the fossil record only shows changes compiling at the rate of about .05 darwins, so it can be well accounted for by a factor of a hundred thousand. Even the most fast paced evolution observed to take place in nature is only on the order of a few hundred darwins. Still easily accounted for by Darwinian evolution.
Thus, again, where evolution could have been potentially falsified, it shows that it can hold its own.
I also promised to discuss how this all fits with the Big Bang too. If you’re following things well, you can probably tell where this is going: Creationists claim that the Big Bang isn’t testable because we can’t recreate it. But we don’t need to. We just need to make sure that our hypothesis makes predictions that can either be confirmed or refuted.
It all starts with an observation. For the Big Bang, the observation is that almost all galaxies in the universe are moving away from us (only nearby ones being influenced by our gravity aren’t). The further they are, the faster they’re moving.
The hypothesis is that, if you rewound things, everything would instead be coming together. At some point in time, it’d all meet. Play it back again, and that’s the Big Bang. But what sorts of predictions can we make from that?
One of the main ones comes from the fact that, if you shove a lot of energy and/or mass really close together, it gets very hot. When things get hot, they also start giving off light.
The trouble is that when you start squishing all that mass together, it also gets somewhat “foggy”. There’s too much matter in the way for the light to travel anywhere. Thus, there’s a very special balance in which the density is just low enough that we should be able to see that light.
Combine that with the fact that as something moves away from us (due to the expansion of the universe), it “stretches” the light to longer wavelengths, and we can put this all together to make a very specific prediction about how we should observe that leftover light from the moment that the universe was becoming dense enough to let the light escape. It turns out that we should end up actually seeing it as radio waves.
This was first predicted in 1964. Later that same year, radio astronomers Penzias and Wilson stumbled across it when they were performing a completely unrelated experiment. Either way, they’d confirmed a major prediction of the Big Bang theory and were awarded a Nobel Prize in 1978.
If you’ve hung around Evolution or Big Bang threads very much, you’ve probably seen this image at some point (with the caption: SCIENCE: It works BITCHES):
IMG 8
From here
If it didn’t make sense before, it should now. That’s what’s known as a blackbody curve. Basically, it’s a prediction of how the temperatures of a gas should be distributed described by the equation in the upper right corner. The line on there is the prediction.
Those boxes on the line are the actual data points. Notice where they lie: exactly on the predicted line!
So what does this mean? It means that the prediction made by the Big Bang theory, one that no other hypothesis can even account for, was exactly right. As always, this doesn’t prove the theory, but it does make it a damn well supported one.
But as if that wasn’t enough, Big Bang theory makes another stunning prediction. If everything were completely evenly distributed after the Big Bang, there would be no places that were more dense than others around which galaxies could start forming. Thus, for the Big Bang theory to be right, there’d better be some very slight variations in the initial distribution. In fact, to get a universe like the one we see today, we can predict how large those variations would have to be.
Enter the WMAP probe which went looking for these variations. Again, predictions were made:
IMG 9
And the data matches it extremely well. This stunning verification earned the Big Bang another Nobel Prize.
As I keep saying, this isn’t proof, since proof can never truly be obtained in science, but it does lend strong credibility to our theory. No other hypothesis can even come close to the power of these predictions. These two tests show that our Big Bang model matches observations of reality to more than 99% accuracy.
These aren’t the only observations that the Big Bang theory makes, but they’re enough to illustrate the point I was trying to get across. For more information on the Big Bang, I’d recommend reading my blog entry on the 4 common misconceptions of the Big Bang.
So let’s recap.
In science we don’t have to be able to recreate the entirety of an event to make observations (ie, to test) of that which nature has laid out before us. The creationist claim that we fill in every single step along the way and take care of those pesky “missing links” and “gaps” sounds nice rhetorically, but in reality is both impractical (we can’t expect every single step to have fossilized) and unnecessary. Even sporadic fossils can clearly paint the picture we need to confirm our predictions, just as periodic observations of the height of a falling ball can confirm the predictions of gravitational theory.
Statistics can tell us just how closely those sporadic data points match our predictions. If they match closely, then we consider the hypothesis to be good, and after additional testing, a hypothesis gains the most highly accepted title in science: Theory. Gravity, Evolution, and the Big Bang all do exceptionally well and have gained this title.
This doesn’t prove it, because for all we know, tomorrow we might find a fossilized rabbit in the Precambrian, which would completely falsify evolution. But with the fossils we do have, we can paint enough of the picture to be confident in what’s going on: Life evolves.
Similarly, going through the steps of the scientific method for the Big Bang, we get amazingly well confirmed predictions, which verify that theory as well. Again, we can’t recreate the entire history of it in a beaker, but our models can be checked in several different ways (analogous to several data points) to ensure accuracy.
Thus, despite not having a complete picture, we do have a pretty damned good one. Every time we confirm a prediction through observation, it adds evidence to our already well-supported theories. Creationists meanwhile, since they cannot make any testable predictions, have yet to provide any evidence to support their conjectures.
Special thanks to my proof readers:
Poindextra heart
Westenblum
Maxdom
Chrisoya
Cuddle-fish
mitoguard
Jaaten Delost-Syric
Further reading:
The Failure of Irreducible Complexity
Evolution, Entropy, the Big Bang, and the Second Law
Is Intelligent Design Different that Creationism?
Why are ID Proponents/Creationists intellectually dishonest?
Expelled: No Intelligence Allowed (An ID Propaganda Film)
Evolution - What it is and isn’t
Evolution and its Compatibility with Creationism
Updates
8/31/07 - Expanded discussion on reduction of errors.
9/7/07 - Added brief discussion of "Cambrian Explosion".
9/14/07 - Added additional hominid phylogenic tree and discussion.
9/15/07 - Added discussion on rate of change and "darwin" units.
9/22/07 - Added link to The Failure of Irreducible Complexity thread.
9/22/07 - Added Discussion of Probabilities and the Second Law to Optional Reading section.
12/26/07 - Added new link to further reading section: Why are ID Proponents/Creationists intellectually dishonest?
4/15/08 - Added new link to Expelled: No Intelligence Allowed (An ID Propaganda Film).
As a scientist, I can say this is absolute nonsense. The only thing such comments reveal is a profound misunderstanding of what is meant by evidence and testing in the scientific field. Thus, the purpose of this post is to explain how hypotheses are actually tested, especially how the hypotheses of Evolution and the Big Bang gained the highest status a hypothesis can become in science: a theory.
Before I go any further, let’s first define these terms and how they’re used in science. In every day usage, “theory” just means “guess” or “conjecture”. However, in science, “theory” has a very different connotation. What people generally consider a theory is actually a “hypothesis”: a proposed statement of how something works that has not undergone rigorous testing. If a hypothesis does undergo rigorous testing and every test confirms the predictions it makes, then it is eventually upgraded to the status of “theory”. Again, this means that it’s exceptionally well supported. This is the definition I will be using for the rest of this thread, unless otherwise noted.
But what about all those other important terms that we always hear people tossing around when discussion science? What about laws, and facts, and proof?
Laws are, for the most part, very similar to theories. The main difference is how much they cover; Laws are extremely narrow and don’t seek to explain much. They’re usually associated with a mathematical expression. For example, Kepler’s laws are very short (Planets travel in ellipses, p^2 = a^3, ….). Meanwhile, theories are far more general. Evolution seeks to explain the diversity of life, provide a mechanism, explains progressions in fossils, inheritable traits…. Both are exceptionally well tested. It’s just that theories are much broader. But just like theories, laws are not necessarily “proven”. In fact, many have been shown to be incorrect. For example, Newton’s laws of gravity did not work properly for Mercury’s orbit. It was eventually realized that this was because of relativistic effects (ie, close to the speed of light) and Newton’s laws were supplanted by the theory of relativity.
Facts, as big and important as they sound, are actually quite useless in science. Since for all you know, gravity could turn off tomorrow and we’d all float away, we can’t say for certain that gravity or anything else in science is 100% fact. Theory is the best it gets. So what are facts? Facts are the bits of evidence. They’re things like “at time t I observed something at position y”. In and of itself, a fact is useless. Combine it with a bunch more and explain it all with a nice theory, and now you’re actually getting somewhere. The point is, that contrary to popular belief, facts are not better than theories.
Similarly, there’s no such thing as proof. The best we can do is to say that every observation has confirmed predictions made by a hypothesis; but since we can’t be sure what the next observation will show, there’s no proof. Only exceptionally well supported hypotheses called theories.
Now that we’ve defined our terminology, how do we use all these terms to understand how science works?
First, let’s start off with something general that no one should really have any disputes about. Let’s take an example of a ball falling. The scientific method starts with an observation. In this case, the observation is that things fall (profound, I know). This is a fact.
From there, a question is posed. For a ball falling, we can ask how it falls. A prediction is made: It falls due to the force of gravity which, on Earth causes an acceleration of 9.8 m/s^2. Going through a bit of math, we can use that to predict where a falling body should be at any given time and then check it against reality. This gives us our hypothesis. Hypotheses are generally mathematical or at the very least, qualitative models of an event.
So we then rig up an experiment. We’ll take a ball, drop it and measure it’s position every second. We can then plot its position on a graph and compare it with the prediction. If it matches closely, then we say that our prediction is good and if we test it a whole bunch more and it continues to do well, it eventually becomes a theory. If not, then the hypothesis needs to be reformulated or scrapped all together.
But let’s imagine what a graph of this should look like. On the x-axis we’d have time in seconds. On the y-axis would be how far the ball had fallen. Here’s a quick sketch of that.
IMG 1
The dots are the actual measurements, and the solid line is the position predicted by your hypothesis.
But OH NOES! The points don’t all lie on the line! Does this mean that they gravity hypothesis is wrong?!
Of course not. To understand why, let’s take a slightly more in depth look. One of the things this graph doesn’t show is what scientists call “error”. This doesn’t mean anything was done wrong, or that the measurements were wrong, but rather, that there is some inherent limitation to how accurately we can measure things. Basically, there’s some random error associated with measuring anything. One of the strong points of science (albeit frustrating for those partaking in it) is that we do what’s known as error analysis, in which we take these errors into account and see if the data is still close enough to match our prediction from our hypothesis.
This is all done through the power of statistics. From this, we actually know (in terms of a percentage) how well our data is matching our prediction. Again, if it fits pretty well, we keep it (perhaps with a bit of tinkering to get the prediction a bit closer), and if it sucks, we scrap it.
However, just because our data doesn't match, doesn't mean that the hypothesis is completely flawed. If the error intrinsic to the measurement is large, then we need to take that into account and find a way to reduce that uncertainty. This can be done either by making more accurate experiments, or simply by taking the average of a lot of observations. The more observations, the smaller the error gets. This is why we keep testing things in science. Incidentally, because the first time an experiment is run no earlier tests exist, this means that the initial hypothesis is likely to be quite different than reality. It requires tinkering and more observations to reduce errors. In this light, when people say that some of what Darwin said was wrong, it should be considered beside the point. We never expected it to be dead on. There's details that obviously wouldn't match. But as we've learned more, we've corrected the details he got wrong. The theory of evolution today, is not the same as the one Darwin posited over 100 years ago.
Another thing that’s important to realize here, is that we don’t have to have a data point at every point along the line. In other words, we don’t need to see the height of the ball every nanosecond. Every second or two is good enough to get a good idea of how well the prediction matches up with reality. Keep this in mind. It will become important later.
So if we take into account the inherent limitations of measuring anything, it turns out that even though those data points don’t perfectly match up with the prediction made by our gravitational hypothesis, it still matches extremely well.
Now, what does any of this have to do with evolution?
Well, let’s walk through the same process. First, we make the observation that Darwin did: That there are a number of species that while similar, are uniquely adapted for their environments. Darwin then posited that these similarities were indication of a common ancestor, and changed due to environmental pressures, which drew out beneficial mutations. The prediction is, that when we look at the fossil record, we should observe the gradual changes. In other words, we should see the branching of the species tree (aka the phylogenic tree). If that’s not familiar enough for you, let’s take a look at what Darwin predicted:
IMG 2
Can we find this in nature? Absolutely. To save time though, I’m only going to look at individual splits instead of the whole three though. Basically what this means is that we’ll zoom in on one fork.
To demonstrate this, let’s look at an ancient species known as Rhizosolenia. It’s a genus of single celled organism that have very distinctive silicate cell walls. Fossils of their remains from 1.7 million years ago show two separate species. One of the defining characteristics of is the area of these shells. If we plot these areas vs. time, we get a graph that looks like this:
IMG 3
Again, with this graph we have a nicely defined trend with several data points. We don’t have a size at every point in time. There’s lots of “gaps” or “missing links”, but it’s still enough to clearly show what’s going on: One species split into two. This is just one example among many. Is this proof of evolution? Absolutely not.
As I stated before, it’s impossible to “prove” everything in science. All scientists can do is prove something beyond a reasonable doubt. In other words, the case of Rhizosolenia just begins to lend credibility to evolution. But we need even larger changes to really make evolution credible. This is clearly demonstrated by the evolution of whales from terrestrial ancestors.
So it’s clearly established that Darwinian evolution can account for changes between phyla, one of the largest differences in animal groups. Again, this doesn’t prove evolution, but it makes it an extremely powerful candidate since it is able to predict how species branch that matches closely with what is observed.
One of the most contentious instances of evolution, however, is the evolution of humans from an ape-like ancestor (often mischaracterized as humans from monkeys). One of the main characteristics that defines a human is the size of the skull which has gifted us with our larger brain capacities. If evolution is true, then we should be able to trace the size backwards in time until it is similar to the species from which we diverged.
Despite the claim of creationists, there are a good number of intermediary forms from which we can study the brain capacity. So let’s take a look at this:
IMG 4
From this, we can see a clear trend of increasing brain size, from ~400 cc to the roughly ~1400 cc we have today. So what we have is a very nice trail in accordance with evolutionary predictions. Are there “gaps”? Sure, but we still have a very nice trend, even without having a fossil for every step along the way. Thus, the “missing link” argument creationists frequently tout is rather pointless. We just don’t need links at every step. The trend is clear without them.
But this instance is only looking at one feature along one branch. From this next graph, we can see the trend and branch going the other way, even more distinctly:
IMG 5
Near the branching point, we have the species most similar to the common ancestor with a small brain, large teeth, and use facultative bipedalism (ie, the wear on fossils show they occasionally walked upright for short periods of time). That group then split to become the green branch which continued to diversify into modern humans, while another leg of the branch went off to blue group and eventually the orange, modern apes.
Here's another graph with essentially just the left branch of that graph:
IMG 6
The "missing links" that are so often referred to would be the "?"'s between branches. While it's true we don't have every piece, what should be understood is that this is not necessary. What we see is that the overall prediction is correct. Note the similarities in the shape of this graph to the earlier one of the Rhizosolenia. Both show clear divergences, even without having every step explicitly defined.
There may be some minor holes, but these aren't all that important. Because we've seen that the process works so well in every case we can observe, there's no good reason to suspect it doesn't where we haven't gathered enough data yet. This would be the equivalent of saying that, although I can trace my genealogy back several generations, when I can't figure out who our great, great, great grandmother was, that it meant that I didn't have one and my lineage suddenly popped into existence. Such an answer is horribly unparsimonious! In other words, it invokes unnecessary and untestable mechanisms.
Thus, we see that the concept of “gaps” and “missing links” is really an inconsequential bit of rhetorical nonsense.
But what about other tests for humans being inexorably linked to the rest of the great apes family? All great apes besides humans have 48 chromosomes. Humans only have 46. Since the loss of two whole chromosomes would have absolutely been fatal, this would imply that one set of our chromosomes came from a merger from the common ancestor. We can look for this by looking for the “stop” instructions in the DNA, known as telomeres.
It turns out that in the middle of chromosome #2, there is a deactivated telomere, which is the result of a gene fusion. If this were not observed, common descent from the great ape family would be falsified. No telomere; No common ancestor.
Can other hypotheses account for this? Creationists often posit that similarities in DNA could very well be the result of the creator using similar methods and thus, is “evidence” for creationism under that interpretation. However, the difference is that if the DNA weren’t similar, evolution’s prediction would be falsified and evolution would be in trouble. Meanwhile, if that were the case, creationists could just claim that it’s evidence that the designer were really creative. Either way, they can claim victory for their designer. In other words, it can’t be falsified. Thus, if there can’t be evidence against, you can’t really claim to have evidence for either.
Evidence only comes when you pass a test that could have falsified your hypothesis.
Another often argued part of evolution by creationists is that of the "Cambrian explosion" (which is universally misrepresented by creationists). They often moan about how there's no transitional fossils and that species formed "as if overnight". But yet again, we see the distinct overlapping and branching pattern predicted by evolution:
IMG 7
But let's go back to the concept of "error" I discussed earlier and explore how it works a bit more. It should be clear that data at every single point along the way isn't necessary. But as I mentioned, the more times you test something the more your error goes down, and the greater your confidence in your hypothesis becomes. I've just presented a handful of tests that are extremely conclusive. In science, having just a handful of data points that are so well constrained and with such low errors, already gives you extremely high confidence.
However, like I said, having lots of measurements, even without really small errors on each individual ones, can be taken collectively to still give extremely high confidence. Here, evolution shines again. Consider this: It would only take one fossil to disprove all of evolution. This is often referred to as the proverbial "fossil rabbit in the Precambrian." Effectively, what it says, is that if we even find one fossil out of place chronologically, then all of evolution is shot. So if evolution were wrong, it should be pretty easy to find fossils out of place that didn't match with evolutionary predictions. In this light, every fossil dated is a test of evolution. This gives thousands of tests for evolution, of which it has passed every single one. Like I said earlier, the more tests you have, the greater the confidence. In this regard, the confidence level of evolutionary theory is almost unprecedented.
Another way we often look at things in science is to check what's known as the "bounds". In terms of evolution, this would just to do a simple check to make sure that the fossil record doesn't show things evolving faster than mutations can account for. If species were changing faster than that, mutations couldn't account for this change and evolution would be falsified.
To test this requires two values: One for how fast things change, and another telling us what the maximum rate mutations can compile is. To measure this, we use units called "darwins". It turns out that the upper limit for how fast mutations could compile (under very harsh forced selection criteria) is somewhere around 10000 darwins. Typically, the fossil record only shows changes compiling at the rate of about .05 darwins, so it can be well accounted for by a factor of a hundred thousand. Even the most fast paced evolution observed to take place in nature is only on the order of a few hundred darwins. Still easily accounted for by Darwinian evolution.
Thus, again, where evolution could have been potentially falsified, it shows that it can hold its own.
I also promised to discuss how this all fits with the Big Bang too. If you’re following things well, you can probably tell where this is going: Creationists claim that the Big Bang isn’t testable because we can’t recreate it. But we don’t need to. We just need to make sure that our hypothesis makes predictions that can either be confirmed or refuted.
It all starts with an observation. For the Big Bang, the observation is that almost all galaxies in the universe are moving away from us (only nearby ones being influenced by our gravity aren’t). The further they are, the faster they’re moving.
The hypothesis is that, if you rewound things, everything would instead be coming together. At some point in time, it’d all meet. Play it back again, and that’s the Big Bang. But what sorts of predictions can we make from that?
One of the main ones comes from the fact that, if you shove a lot of energy and/or mass really close together, it gets very hot. When things get hot, they also start giving off light.
The trouble is that when you start squishing all that mass together, it also gets somewhat “foggy”. There’s too much matter in the way for the light to travel anywhere. Thus, there’s a very special balance in which the density is just low enough that we should be able to see that light.
Combine that with the fact that as something moves away from us (due to the expansion of the universe), it “stretches” the light to longer wavelengths, and we can put this all together to make a very specific prediction about how we should observe that leftover light from the moment that the universe was becoming dense enough to let the light escape. It turns out that we should end up actually seeing it as radio waves.
This was first predicted in 1964. Later that same year, radio astronomers Penzias and Wilson stumbled across it when they were performing a completely unrelated experiment. Either way, they’d confirmed a major prediction of the Big Bang theory and were awarded a Nobel Prize in 1978.
If you’ve hung around Evolution or Big Bang threads very much, you’ve probably seen this image at some point (with the caption: SCIENCE: It works BITCHES):
IMG 8
From here
If it didn’t make sense before, it should now. That’s what’s known as a blackbody curve. Basically, it’s a prediction of how the temperatures of a gas should be distributed described by the equation in the upper right corner. The line on there is the prediction.
Those boxes on the line are the actual data points. Notice where they lie: exactly on the predicted line!
So what does this mean? It means that the prediction made by the Big Bang theory, one that no other hypothesis can even account for, was exactly right. As always, this doesn’t prove the theory, but it does make it a damn well supported one.
But as if that wasn’t enough, Big Bang theory makes another stunning prediction. If everything were completely evenly distributed after the Big Bang, there would be no places that were more dense than others around which galaxies could start forming. Thus, for the Big Bang theory to be right, there’d better be some very slight variations in the initial distribution. In fact, to get a universe like the one we see today, we can predict how large those variations would have to be.
Enter the WMAP probe which went looking for these variations. Again, predictions were made:
IMG 9
And the data matches it extremely well. This stunning verification earned the Big Bang another Nobel Prize.
As I keep saying, this isn’t proof, since proof can never truly be obtained in science, but it does lend strong credibility to our theory. No other hypothesis can even come close to the power of these predictions. These two tests show that our Big Bang model matches observations of reality to more than 99% accuracy.
These aren’t the only observations that the Big Bang theory makes, but they’re enough to illustrate the point I was trying to get across. For more information on the Big Bang, I’d recommend reading my blog entry on the 4 common misconceptions of the Big Bang.
So let’s recap.
In science we don’t have to be able to recreate the entirety of an event to make observations (ie, to test) of that which nature has laid out before us. The creationist claim that we fill in every single step along the way and take care of those pesky “missing links” and “gaps” sounds nice rhetorically, but in reality is both impractical (we can’t expect every single step to have fossilized) and unnecessary. Even sporadic fossils can clearly paint the picture we need to confirm our predictions, just as periodic observations of the height of a falling ball can confirm the predictions of gravitational theory.
Statistics can tell us just how closely those sporadic data points match our predictions. If they match closely, then we consider the hypothesis to be good, and after additional testing, a hypothesis gains the most highly accepted title in science: Theory. Gravity, Evolution, and the Big Bang all do exceptionally well and have gained this title.
This doesn’t prove it, because for all we know, tomorrow we might find a fossilized rabbit in the Precambrian, which would completely falsify evolution. But with the fossils we do have, we can paint enough of the picture to be confident in what’s going on: Life evolves.
Similarly, going through the steps of the scientific method for the Big Bang, we get amazingly well confirmed predictions, which verify that theory as well. Again, we can’t recreate the entire history of it in a beaker, but our models can be checked in several different ways (analogous to several data points) to ensure accuracy.
Thus, despite not having a complete picture, we do have a pretty damned good one. Every time we confirm a prediction through observation, it adds evidence to our already well-supported theories. Creationists meanwhile, since they cannot make any testable predictions, have yet to provide any evidence to support their conjectures.
Special thanks to my proof readers:
Poindextra heart
Westenblum
Maxdom
Chrisoya
Cuddle-fish
mitoguard
Jaaten Delost-Syric
Further reading:
The Failure of Irreducible Complexity
Evolution, Entropy, the Big Bang, and the Second Law
Is Intelligent Design Different that Creationism?
Why are ID Proponents/Creationists intellectually dishonest?
Expelled: No Intelligence Allowed (An ID Propaganda Film)
Evolution - What it is and isn’t
Evolution and its Compatibility with Creationism
Updates
8/31/07 - Expanded discussion on reduction of errors.
9/7/07 - Added brief discussion of "Cambrian Explosion".
9/14/07 - Added additional hominid phylogenic tree and discussion.
9/15/07 - Added discussion on rate of change and "darwin" units.
9/22/07 - Added link to The Failure of Irreducible Complexity thread.
9/22/07 - Added Discussion of Probabilities and the Second Law to Optional Reading section.
12/26/07 - Added new link to further reading section: Why are ID Proponents/Creationists intellectually dishonest?
4/15/08 - Added new link to Expelled: No Intelligence Allowed (An ID Propaganda Film).