Science - Building Models, Not Facts

I'm going to harp on a subject that really annoys me. The use of the word "fact" - this is a totally useless concept and is the source of a lot of confusion out in the general public and amongst science journalists.

Science does not build facts. It builds models, ideas concepts and theories - tools for our mind. Not facts. The idea of a "fact" leads to confusion and misunderstanding. This notion subconsciously promotes the comparison of the products of scientific endeavor to religious dogma. Facts = immutable truths.

Is gravity a "fact"? Does evolution "exist"? Is calculus "real"? To me these questions are the equivalent of asking whether a bank loan is a solid or a liquid. The question doesn't compute.

Our brain (or mind) is a prediction device. We take in information from our sensory organs and we reconstruct the world our minds in order to make sense of it. The ideas and concepts that float in our mind are not objects in the same way that an apple is, they are just tools that help our brain to better react, and predict how the world around us works. There is a real world out there, with real objects that do exist, but we can only access these through our sensory organs. We can only understand and predict how all these objects operate by performing simulations in our mind. But we must not forget that these are nothing but simulations. All of our ideas, models and concepts - the building blocks of these simulations - are nothing but tools. They are not reality. They do not exist in the same way that an apple does.

So evolution is not a fact, but it is not because we are still debating this idea. No scientific concept is a fact. Many who have been exposed to this line of reasoning ("evolution is a theory") immediate jump to the conclusion that if facts do not "exist" then every idea is equivalent. This is false as well.

Science is the process of building and evaluating tools that we call "models", "concepts" and "ideas". Scientific battles and debates revolve around what observations can say, how to interpret experimental results, and which models have the best predictive power. In the process, science builds and then evaluates all of these tools, whether you call them "ideas", "models", "concepts" or "theories". These tools have value in that they can be used by our minds not only to explain the makeup and history of the world around us, but also to predict how the seemingly chaotic universe works. These tools have utility in that they generate insight. Our scientifically generated models give us simple rules that can also predict what else is out there, and how the world around us works. Simplification and predictability.

Ideas such as "intelligent design" are useless because they have no predictive power, and their way of "simplifying" the world is the mental equivalent of representing unknown processes (such as "how did the eye come to be") by uninformative ideas ("the intelligent designer made it so") - there is no insight, no general theme that can unify disparate observations and no predictive power. All that is accomplished is the equivalent of a mental suicide.

So next time you are tempted to use the idea of "fact", please remember that this tool is deeply flawed in its ability to predict and to give insight into the inner workings of the world and of the human mind.

More like this

I woke up yesterday, made myself a cup of coffee and sat down with the New York Times, and a left over piece of corn bread from Thanksgiving. It was a beautiful morning and I was at peace. Then I read this (I will try to be polite) by Paul Davies. Apparently scientists operate on faith. Faith that…
Sorry, I've been busy these past few days. I want to respond to a comment posted by Dan and take this opportunity to broaden the discussion about how we use language to construct models. Dan's concerns about information and life have been echoed by many out there, for example by John Wilkins. Can…
Michael Nielsen is planning to attend an "unconference" and is considering possible topics. He quotes one from Eva Amsen: My idea: find 4 or 5 volunteers from different backgrounds to sit on a 20 minute panel and (with audience feedback) make a list of Ten Things Everyone Should Know About Science…
Our state science standards are being patched up right now, and while they're mostly just fine, one sneaky provision is still on the books. "The student will be able to explain how scientific and technological innovations as well as new evidence can challenge portions of or entire accepted theories…

So, what is your definition of "fact"? for me a fact is something we observe using our senses (for instance, an apple falls from a tree) while gravity is the theory that "explains" this. you do an experiment to test an hypothesis and your results are facts.

I would argue that science is actually the only means to acquire truth (facts), through reason and logical deduction. Humans invent theories and models which are validated with facts obtained via the scientific method. This is an important distinction.

In your example of the theory of evolution (as the current origin of species), the facts are the fossil record, genetic mutation, etc. These facts are interpreted by scientists to support the theory of evolution.

"So, what is your definition of "fact"? for me a fact is something we observe using our senses (for instance, an apple falls from a tree)..."

First, my point is that science doesn't generate what you would call facts, such as an apple falling from a tree, it produces and evaluates theories - that's the main point of this entry.

Second, I was very careful to call what we gather from our senses as observations - a much more precise term than fact. Observations are influenced not only by the real world that is being measured, but also by our instrumentation and our interpretations. Now you can label the "apple falls" or some fossil as a fact, but you know as well as I do that observations are the subject of much scrutiny. To pick a example that you could relate to, it was once believed that the MTOC would reorient in a polarizing cell. Although "MTOC reorientation" was the observation, was it a fact? Well later on it was seen that the MTOC didn't *really* reorient towards the front of a migrating cell, instead the nucleus moved away from the site of polarity and the MTOC stayed right where it was through out the whole process - it only "reoriented" as a byproduct. In the future, we may find a better way of describing this observation, but by calling our current interpretation a "fact" you trap yourself into saying that this description of the measurement is definitive - but in practice this is rarely the case.

Don't forget that the concepts of "fact" and "observation" are themselves tools - the first is more flawed than the second.

I also would add that not all observations are created equal. Some are closer to reality than others. A 3D crystal structure is closer to reality (or a "fact") than say an immunofluorescence micrograph. But does the structure really mimic exactly what happens in vivo? In fact if the structre has an R factor which is is a function of how close the 3D model is to the data. It may be very close, but we can never say that it is 100% exact.

Glad to see you blogging again, Alex! Could you email me your new email address? The one I have keeps getting bounced. Thanks a bunch!

Uhm, that "MTOC reorientation" story sounds familiar..."MTOC reorientation" was the observation, or even better, the observation was "the MTOC is between the leading edge and the nucleus" and this was named "MTOC reorientation" because the theory was that the MTOC moved. the problem here was the name of the observation that suggested a theory. It was a biased observation name!

According to google (i.e, the source of all the truth), fact is "a piece of information about circumstances that exist or events that have occurred" while observation is "the act of making and recording a measurement" which goes in accordance with your suggesting of using "observation" instead of fact. 
Interesting the 3rd entry for fact is "a concept whose truth can be proved; scientific hypotheses are not facts". 

By evil gomez (not verified) on 28 Sep 2009 #permalink

Science does not build facts. It builds models, ideas concepts and theories - tools for our mind. Not facts. The idea of a "fact" leads to confusion and misunderstanding.

Only to a confused mind or a mind that's set up to create a confusion on purpose. There is nothing close to the demagoguery than pretending that common universally understood terms are not defined. You sound Clintonian (in a sense that next you might be asking of what exactlly "is" is).

You can muddle the water all you want but the truth is that the fact is something that's overwhelmingly accepted as existing and real, and a non-fact is the opposite of it. So, for example, it is a fact that I=V/R. One can go on endlessly about universes with alternative laws, about potential relativity of mathematics, and about how that equation is only approximation of what can actually observed by us, etc, etc - none of that will ever change the fact that Ohm's law is a fact (particularly solid facts are called laws when it is useful to set them apart).

"Model" used to be a respectable term to designate an evolved hypothesis. With enormous amount of the bullshit unreliable observations sold by cell biologists as "facts" on which the models are built, the word now has a dirty aftertaste to it. Finally, I am starting to notice many scientists trying to avoid it.

When everyone repeatedly observes lysozyme structure to be nearly the same under huge variety of conditions and crystal forms (and, in truth, that has little to do with the R factors observed) - the structure of lysozyme is a fact. Once in a while - quite rarely, really - the known fact(s) will turn out to be non-facts. Big deal! That's the nature of continuum. Anyone who denies the usefulness of "fact" should logically deny the existence of color "green".

To summarize, science strives to build facts.
And, by and large, it succeeds.

fact is something that's overwhelmingly accepted

The way you are using the word "fact" is not the same as the typical usage by a layperson. Any prevailing model is accepted by the scientific community because it is the most useful tool we have for representing reality.

The way you are using the word "fact" is not the same as the typical usage by a layperson.

Actually, it is. It is exactly how it is used by lay people. Some of them may not realize it but it does not change the fact that is it.

Newton's laws of motion were overwhelmingly accepted for 200 years, but are these facts?

Well we now know that Newton's laws of motion do not predict how objects behave when they accelerate at high speeds. Einstein's relativity is better. But is it a fact? Well if you look at small objects relativity breaksdown. So just like Newton's laws, our current models do not exactly represent reality. But both Newton's and Einstein's work are useful because they do predict much of how the world works.

Newton's laws of motion were overwhelmingly accepted for 200 years, but are these facts?

Yes, laws of motion are facts. Of course. Indisputably. Unquestionably and without a doubt. And, by the way, Newton's laws of motion are overwhelmingly accepted today.

If you believe that Newton's laws of motion are facts then I cannot help you. Newton's laws of motion are off - Einstein demonstrated that. Of course the errors are negligible when a body is not accelerating at high speeds.

Instead of thinking of these theories as facts, think of them as tools. Some tools are better than others. Science is the act of sorting out what is the best available tool.

good entry. i've been talking a lot about this with my partner in crime lately, and you put very elegantly and sucinctly what i have been trying to say! glad to see you back on teh internets. i may resurrect my blog one of these days, too, if i have any time to write after i start my post-doc.

p.s. your daily snark: you also can't help someone who writes "muddle the water". and if i have to explain why that is, i definitely can't help you.

If you believe that Newton's laws of motion are facts then I cannot help you. Newton's laws of motion are off - - Einstein demonstrated that.

Oh, thanks you for letting me know. Who is Einstein? Sheesh, for someone with such a profound of misunderstanding of scientific method, you seem to be too full of it. For example, you could have paid attention to what I said earlier: "One can go on endlessly about ... how that equation is only approximation of what can actually observed by us ... none of that will ever change the fact". And that's exactly what you are doing - invoking approximation argument.

Hint: every law has or potentially has its applicability and every fact is/can be observed only under a given set of conditions.