Hacker News Reader

Comment by maskedinvader | Hacker News Reader

The biggest misconception in most people (at least from my perspective from conversations involving evolution) is what constitutes a scientific theory, By definition from [1] , "A scientific theory is a series of statements about the causal elements for observed phenomena. A critical component of a scientific theory is that it provides explanations and predictions that can be tested. ".

People I speak to think its just a theory, hence a guess, a hunch, and not a fact, not proven etc, where as its actually the opposite of that, i.e we can test the predictions it makes, its as close to being proven as anything can be. If people can simply recognize the definition I bet more people can learn and accept Evolution. [2] explains this better.

[1]: http://rationalwiki.org/wiki/Scientific_theory [2]: http://www.notjustatheory.com/

View on HN

More to the point:

Theories don't become laws, theories generate laws; testing those laws is a test of the theory. Laws are mathematical equations which relate observable quantities such that we get quantitative predictions.

The Ideal Gas Law, PV = nRT, is one example: It relates Pressure (P), Volume (V), number of moles of gas (n), a physical constant (R), and Temperature (T) such that we can predict the values of certain traits, or the relationships between traits, given some known or hypothesized values for the other traits. The theory it's part of is the one that says gasses are made out of tiny particles called atoms or molecules, which fly around freely and bounce off each other and solid surfaces with a speed proportional to their temperature. Every time the Ideal Gas Law makes a good prediction, that theory gets validated.

I'm going to have to disagree. There's no meaningful distinction between "laws" and "theories". One famous consequence of the theory of special relativity is the mass-energy equation - but we do not call E = mc^2 a "law". We just say it's a part of the theory of special relativity.

That, however, does not mean the theory of relativity and E = mc^2 has a lower status than Newton's law of universal gravitation and F = (G m1 m2 )/r. That we call one a "theory" and the other a "law" is just an accident of speech. It's not even a difference of semantics - in this context, "law" and "theory" have the same meaning. We just continue to say "law" for some things for historical reasons.

We call it the law of conservation of mass.

Or well its a limit case, but yeah we absolutely do call it a law.

Theory is the explanation, law is the model.

The distinction you propose is demonstrably not true in modern theories. (Where "modern" means a bit over 100 years old.) We call such things a law, but there is no difference between a theory and a law. Such "laws" just predate our current philosophy of science, and we continue using names like "the Ideal Gas Law" because we always have. If someone came up with an equivalent thing now, we would not use the term "law" - which we can observe with physics from the 20th century.

Perhaps you're right. Mark this one down to schools teaching outdated things as fact, I suppose.

Still, my idea is better than the "hypotheses become theories, which become laws" nonsense, which is still taught at the elementary school level.

This is literally semantics.

The vast majority of people know what an "accepted fact" is, and they also know what a "proposed model of reality" is (even if they don't use such phrases).

The opposition to evolutionary theory as a model of reality is a topic that principally obsesses only a specific population: American Protestants.

In most of the Western world, including in the Catholic Church with its over 1 billion adherents, evolution is an "accepted fact" (or "scientific theory" however that may be rendered semantically).

Some propose using the word theorem to differentiate between these two meanings.

I wrote a novel based in part around the idea that one could prove something like Boltzmann's H-theorem (which has problems of its own) for evolution: http://www.amazon.com/Darwins-Theorem-TJ-Radcliffe-ebook/dp/...

Darwin's Theorem would start with the facts about genetic material--the probability distributions for replication--and use further distributions of features, consequences for reproductive advantage, and so on to demonstrate that it is mathematically necessary for evolution to give access to a more-or-less unbounded adaptive landscape.

The book is actually mostly about sex, violence, politics, religion, regret and redemption, so it's not as dry as that probably makes it sound!

I have absolutely heard people say, "Evolution is just a theory"

It's not just semantics. Maybe some times it is, but not always, and I even think, not usually. Many people don't actually understand the basics of scientific theories, evidence and falsification. It requires going beyond learning just the "scientific process", and going into the actual philosophy of science. In my experience, that is not commonly known.

Stressing the term "theory" by creationists attacking evolution is a rhetorical device intended to imply that "even scientists don't know".

Obviously, scientists, and other educated people, know that this attack is misguided.

However, recognizing that such arguments are indeed rhetorical, rather than logical, is a step towards not having to go through them again and again...

I'm making a different point. People who make that attack have a fundamental misunderstanding of how science works. They don't understand that it is not possible to prove a theory correct, only falsify it. They don't understand that quantity and quality of supporting evidence matter. They don't understand the difference between "proof" and "evidence". They don't understand that their arguments could be levied against any scientific theory, including ones that they accept. That they can do this means they don't understand how science works.

But, nor do a lot of people who don't make such rhetorical attacks. I think that the philosophy of science is an easy thing to miss - it's possible to make it through a Bachelor's program in science without really being forced to think about it.

>They don't understand that their arguments could be levied against any scientific theory, including ones that they accept. That they can do this means they don't understand how science works.

Exactly, which is why epistemological arguments won't win the evolution versus creation debate.

The fact that people don't understand the word "theory" is worrying because it's a flag that they don't understand scientific method. People don't know that a "hypothesis" is the guessing[1] part and that a hypothesis supported by experimental evidence becomes a theory - it becomes less of a guess.

Popular science programmes or reporting fail to explain this and reporting often uses the word theory to mean "has a hunch".

[1] supported by principles I guess.

Sometimes "hypothesis" persists well past the guessing phase (e.g. "The Hygiene Hypothesis" is pretty well supported, even if it's not a complete explanation of autoimmunity and allergies)

I don't think your explanation is quite right, either. Typically, a "hypothesis" is in the form of "we think x will happen in y circumstance". Then they run some experiments to see if that's the case. If it is, then they may try to come up with a theory to explain why, then run (many) additional experiments to test the theory.

In other words, I usually use the word "hypothesis" to mean a simple prediction of outcome, without any explanation of how. A "theory" needs some explanation in it, not just a prediction of correlation.

I keep telling these people that we have things like number theory as well that's as hard science as it can get :) - and it's still a theory. I'm not sure how 'strong' is the word 'theory' in English though, but in my mother tongue it definitely has the feeling of being a 'guess'.

In informal English communication "theory" is a synonym for "hypothesis". It's also far more commonly used in everyday communication than hypothesis, which looks/sounds more formal or technical.

This is mostly the problem, I think. In this sense, both 'theory' and 'hypothesis' are antonyms of 'fact'. Leads directly to uninformed people saying that evolution hasn't been established as a fact, or that it's not "true".

Mathematics is not a science, and number 'theory' is not theory as meant in the sciences.

Unlike in science, where enough evidence for a particular hypothesis makes it generally believed to be true, in math empirical evidence is not and will never be enough to classify a statement as true.

Yeah, in English I'd say the word 'theory' has three distinct meanings. A guess, as in 'I have a theory about this;' a scientific theory, i.e. set of testable hypotheses, etc.; and a set of related propositions that form a cohesive subject (like number theory).

Thanks. Is there a resource or a book on amazon that focuses on the testability of inter-species evolution ? I've been a bit frustrated by this as most people just allude to the fact that it took thousands or millions of years so hard to test it - but we should extra-polate some observed lab principles at a much tinier and then scale the intuition by a billion times to get inter species evolution.

Specifically are there any:

1. Simulations (i.e. code up the actors in the system and let it proceed it a accelerated time scale similar to what we do for space simulations for heavenly bodies)

2. Emulations (not computational but demonstrating using a real life model of some kind)

3. Lab demonstrations of this happening

4. Reproductions of the whole set of mechanisms that allow a species to evolve into something new that we would classify as a different species

that clearly show the evolution principles in effect for a reasonably complex inter-species case.

I have been frustrated from Richard Dawkins's accounts too. I once read his essay on how the eye might have evolved. His accounts tend to be so speculative and full of "should"s and "must"s, its very off putting.

Edit: This book in fact is the best writing on evolution of eyes: http://www.amazon.com/Evolutions-Witness-How-Eyes-Evolved/dp...

But it completely skips the actual mechanism or testability of anything it says. So basically doesn't address the how at all. A resource that actually addresses the HOW and not just conjecture on the WHAT would be really cool.

I think the fundamental problem is that "species" is a way for human to divide up organisms, rather than a fundamental function of biology or evolution. Every organism has its own DNA but "species" is a way of trying to group those into patterns that make sense to humans.

That is, it's easy enough to demonstrate that evolution drives changes in DNA. That's objective. Whether or not you call it a species change is subjective.

Let me phrase the question a different way - since species seems to be overloaded. What stable state of organism led to the evolution of Giraffe ? Lets call that organism X. (I actually don't know what X is, is it a horse or a zebra or a donkey?)

Is there a simulation that would demonstrate X evolving to a Giraffe or test it ?

Or put another way, if we were to take some snapshots at times where we could consider the organism to be different enough according to the modern "species" classification say A -> B -> C -> D -> E, is there a way to test D -> E or B-> C using the stated principles/theory ?

There is no such thing as a "stable state of organism" in this context. Remove all your goalposts because they are confusing you. Evolution is a continuum.

Then what is evolution trying to explain ? I thought it was explaining how primates evolved to humans, not just micro evolution at the DNA level ? For all practical purposes these are considered stable states in that they are different enough (visually, internally, intellectually) so as not to be considered continuous i.e. just differing by some epsilon.

Are you saying evolution does not claim to answer this and this is a question not worth asking?

I have a feeling the gist of what I am asking is clear enough but we are getting lost in words. I just asked for a link on Amazon that talks about testing/simulating/demonstrating evolution for snapshots enough far apart in time where we would say yes these organisms are pretty different, something like the snapshots demonstrated here: http://www.bot1320.nicerweb.com/Locked/media/ch11/taxonomy.h...

Are you just asking about macroevolution? https://en.wikipedia.org/wiki/Macroevolution

"micro evolution" You are being confused by terms that were invented by a group of people (Discovery Institute) that deny evolution because of their religion not because of lack of evidence. The snapshot evidence you need is found in the genome (genomes that developed over billions of years of "micro evolution") of every organism on the planet not just primates. You can't just ignore that and ask why we can't observe lizards turning into birds overnight.

Edit: I wasn't aware of this until now. We actually have observed one species ("species" being defined as scott_s mentioned up-thread) evolve into another.

http://blogs.scientificamerican.com/science-sushi/2011/12/18...

One of several problems is that for many of the organisms with a short enough generation time to run an experiment like this (see Lenski's experiments with E. coli: http://en.wikipedia.org/wiki/E._coli_long-term_evolution_exp... ) the notion of a "steady state species" in terms of your thinking (Baboon vs. Human vs. Gorilla) is entirely false.

Not entirely. We can define species as groups of organisms that can't reproduce with each other. (For species that have sexual reproduction, of course.) See: http://bioteaching.com/modes-of-speciation/

Why do you think we should be able to extrapolate millions of years of random mutation into the future and be able to test it? If you want evidence there are plenty of other places to look.

I get what you're asking. You want a testable account of how any given species went from what we would call species X to what we would call species Y. There is probably never going to be such a level of resolution in our analysis that would satisfy even the staunchest naysayer. To really comprehend evolution requires an imagination capable of extrapolating a given set of principles into a framework that entails the very minuscule available data we do have under one encompassing theory. That theory is evolution. It is not something that can be fully understood in isolation.

This doesn't directly answer your question, but this was a key insight that finally made the entire picture click in my mind. Your question of wanting hard evidence of cross-species evolution is a red-herring and a distraction to your understanding. The concept of species is not a biological reality. Any definition we can come up with for a species, we can find many instances that break the definition. What we see as a species is just groups of individuals who have exchanged genes in recent history. There is no demarcation to delineate a transition between one species and another: (in the case of sexual reproduction) at every point a child can mate with its parents, with no exceptions (except for degenerate cases). This is true for every species that sexually reproduces, going all the way back to the origin of sexual reproduction. The reason groups of animals that represent a unified gene pool look so different is that groups of individuals fall into stable states with their environmental niche, while other groups fall into other stable states for other niches. Animals appear to be delineated by clear species simply as a function of our perspective from a single time slice. If we could observe gradual change over millions of years, we would see many splitting branches and merging branches, branches that get further apart and then seed an explosion of many smaller branches (i.e. when say land animals split from birds), etc etc. But in reality, every individual is a single branch on this tree. What we judge as a species are just branches that are 'close together' in this tree such that they look like a single unit.

A good way to think about it is to imagine a prism's dispersion of white light. There is clearly red and clearly orange but it gets tricky finding the place where "red becomes orange". The closer you look, the harder it is to choose a spot.

Thanks - your answer is reasonable. What I was really looking for is a mechanism that demonstrates this:

"Animals appear to be delineated by clear species simply as a function of our perspective from a single time slice. If we could observe gradual change over millions of years, we would see many splitting branches and merging branches, branches that get further apart and then seed an explosion of many smaller branches (i.e. when say land animals split from birds), etc etc."

The idea of gradual change is fine to me - and the term species seems to create too much confusion. But the fact that gradual change over time can create seemingly distinct states (lets not call them species) still boggles me mind when I try to conceive of the mechanism. Is asking that question a red-herring or completely unreasonable ? In fact that is the part I find most fascinating. The fact that the change was gradual doesn't preclude from the fact that this gradual change did lead to states which were observed to be quite distinct. One doesn't preclude the other.

To explain more of where I am coming from: As a person who writes reasonably complex software for a living, I am amazed by the "software" in the brain if you will i.e. its remarkable cognitive capacity for many tasks that we are still in the process of replicating in AI. I find it hard to understand how this could arise as a step by step gradual change purely aided by a favorable set of environmental properties. So something to that effect would be really cool to understand (And even emulate or copy in our software development practices). I am a bit disappointed that there are no simulations that can be done or we don't understand the mechanisms well enough to do this.

For a very specific example, it would be cool to simulate "This is how the eye came into being.". Not just enumerating the commonalities between eyes of different species - I find the "commonality" argument as being "factual evidence" of origin of the eye very weak personally. I know its been accepted as fact by most or everyone else but something that would demonstrate the underlying "gradual" principles at work to evolve an eye for example would be uber-cool and satisfying in my book. I don't understand why this question is stupid or or unscientific.

If you just want to play with something that uses a genetic algo, there is this: http://boxcar2d.com/

You can handcraft a simple starting point and let it gradually evolve in different environments. Note how the different environments bring out different traits even tho the starting point is the same. I don't know how many hrs I lost playing with this. :)

The way you ask that eye question is not stupid nor unscientific.

Sadly eyes are used by creationists to ask a similar question in an attempt to discredit evolution. "The eye is so complex and it doesn't work unless it's all there so how could it gradually develop?"

This is one of those things where a whole lot of redundant societal arguments and misunderstandings could be averted if we just used two different words for the two different concepts.

exactly my thoughts too. If only we can use 2 different words... sigh

The real problem is however a philosophical one, how to drive down the point that there is no absolute truth,just absolute contradiction that disproves a (now disfunct) scientific theory . Unless you are making certain axiomatic assumptions or using the constructs of language for example 'all single men are bachelors', the best we can do is let our best theories make predictions and then conduct experiments to test those predictions.

Well said.

Yeah, I agree.

I also think getting people everywhere to change the meanings of words is an uphill, possibly impossible battle. We can't even teach people about figurative vs literally.

I think Scientists should give up and get rid of the word "theory". Scientists had no problem renaming the brontosaurus so it shouldn't be as hard as it sounds.

Reply on news.ycombinator.com