#falsifiable

"I see science in general as a form of arrogance control, in the sense that it's one of the most organized methods we have of forcing us to put our beliefs to the test, and forcing us to face dissonance if the test does not confirm what we believe."

"A theory is scientific if it has the property where a person can articulate an instance in which it would be wrong, in which it can be falsified.

If I tell you 'if the light is blue, x is true, and if the light is red, x is true,' that's not a scientific statement because there's no way that it will not be true at the end. I have to be able to propose an instance in which you might prove me wrong.

This is a very appealing idea, it has a bravado and a wagering quality to it, that a scientist must risk being wrong in order to be a scientist.

In 1919, eclipse expedition which confirmed Albert Einstein's theory of general relativity, and what most impressed him about it wasn't the ambition of the theory or the fact that it was confirmed. What he liked about it was Einstein said, 'if starlight isn't measured to bend around the sun during this eclipse then my theory is incorrect.'

He liked the wagering on being wrong. And he thinks that's the basic status of what a scientific theory is."

This is why religious ideas not just aren't scientific, but cannot be scientific. "Faith" insulates people from figuring out they're wrong.

Reality is under no obligation to be convenient and comforting to you.

What would disprove evolution? - an excerpt from Jerry A. Coyne's book 'Why Evolution is True'.

If evolution is a scientific theory worth its salt, then there must be some conceivable observations that could show it to be wrong. I just wanted to put down, for the record, what some of those observations might be. First, let’s reprise what I see as the major components of the theory of evolution.

Evolution occurs, that is, there is gene frequency change in populations over generations.

Significant evolution takes time—that is, it usually (though not always) requires hundreds to thousands of generations to occur. It is not instantaneous, and it is the population and species rather than the individual that evolves.

Lineages of organisms split, or speciate, so that the single lineage that gave rise to life 3.5 billion years ago has undergone numerous splitting events to produce the millions of species alive today (and also the even more millions that went extinct).

The converse of #3: any pair of living species has a common ancestral species some time in the past.  That is, if you trace any pair of twigs on the tree of life, you will find a node where the line from the trunk bifurcates to produce them.

The process producing the appearance of design in organisms is blind, purposeless natural selection. (There are, of course, evolutionary forces other than selection, including genetic drift, but they don’t produce the marvelous complexity that was once seen as the prime evidence for the hand of God.)

These parts of the theory don’t all stand together. For example, you could have evolution without branching: that would mean that only one long-evolved species would be alive today. Or you could have evolution, but not occurring by natural selection. (The complete absence of natural selection is hard to imagine, though, given that organisms replicate their genetic material, and some genes will replicate better than others in different environments. Lamarckism and teleologically-driven evolution, however, were once seen as the main drivers of adaptation.)  Finally, you could have evolution but not gradualism: every population could experience great saltational leaps in one generation.

In my general talk on the evidence for evolution, I give a list of seven observations that, if repeated and confirmed, would disprove parts of the theory of evolution described above. This shows that it is a scientific theory in the Popperian sense of being falsifiable.  Here are some of those conceivable observations:

Fossils in the wrong place (e.g., mammals in the Devonian). If the fossil record were all out of order like this (a single anomalous fossil might not overturn everything, of course, since it could be in the wrong place for other reasons), we’d have to seriously question the occurrence of evolution.

Adaptations in one species good only for a second species.  There are plenty of adaptations in species that are good for other species, but also help members of the first species: these are the basis of mutualisms. (Cleaner fish, for example, remove parasites and dead tissue from other marine fish, but thereby gain a meal.) But we don’t expect to see—and don’t see—adaptations in one species that evolved solely for the benefit of another species.

A general lack of genetic variation in species.  Evolution depends on genetic variation. If most species had none, they couldn’t evolve.  However, the universal efficacy of artificial selection (I’m aware of only three lab experiments that failed to show a response to such breeding experiments), shows that genetic variation is ubiquitous in nearly all species.

Adaptations that could not have evolved by a step-by-step process of ever-increasing fitness. This is of course the contention of advocates of Intelligent Design like Michael Behe. But adaptations like the flagellum, which Behe and other IDers cite as features that couldn’t have arisen by a step-by-step process of increasing adaptation, have been shown to plausibly arise by just that process.  We don’t need to completely reconstruct the evolution of things like flagella, but simply show that their evolution by a stepwise adaptive process was plausible.

The observation that most adaptations of individuals are inimical for individuals or their genes but good for populations/species.  Such adaptations aren’t expected to evolve often because they would require the inefficient process of group or species selection rather than genic, individual, or kin selection.  And indeed, we see very few features of organisms that seem inimical to organisms or their genes but useful for the population or species. One possible exception is sexual reproduction.

Evolved “true” altruistic behavior among non-relatives in non-social animals. What I mean by “true” altruistic behavior is the observation of an individual sacrificing its reproductive output for the benefit of individuals to which it is either unrelated or from whom it does not expect to receive return benefits.  In this “true” altruism your genes give benefits to others and get nothing back, and this shouldn’t evolve under natural selection. And, indeed, we don’t see such altruism in nature. There are reports that vampire bats regurgitate blood to other individuals in the colony to whom they’re unrelated, but those need confirmation, and there may also be reciprocal altruism, so that individuals regurgitate blood to those from whom, one day, they expect a return meal. Such cooperation can evolve by normal natural selection.

Complete discordance between phylogenies based on morphology/fossils and on DNA. While individual genes can show discordance by lateral transfer—rotifers, for example, have incorporated into their genome from DNA from very unrelated organisms, and this is also common for bacteria. But lateral transfer of genes, as opposed to their direct descent from parent to offspring, is relatively uncommon.  So, for example, if we sequenced the genome of a blue whale and found that on the whole the species was more closely related to fish than to mammals, we’d have a serious problem for the theory of evolution.

We don’t see any of these anomalies, and so the theory of evolution is on solid ground. As I say in my book, “Despite a million chances to be wrong, evolution always comes up right. That’s as close to a scientific truth as we can get.”

About the Author

Fun fact: There’s no evidence that falsifiability isn’t un-falsifiable.  ¯\_(ツ)_/¯