There is no accepted definition for what science is. Typically the natural sciences, let’s say physics, are taken as a model. This creates certain tensions to “softer” sciences, like e. g. the social sciences. Science is also a historical phenomenon, so science is what is accepted as science by the respective scientists (e. g. historian or theologian).
The accepted hard definition of science comes from Karl Popper: scientific theories offer hypothetical predictions about reality. The predictions must be in principle be falsifiable, to be scientific. (the more they exclude, the stronger the theory).
In practice science is:
“When you can measure what you are speaking about, and express it in numbers, you know something about it. But when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meager and unsatisfactory kind. It may be the beginning of knowledge, but you have scarcely, in your thoughts, advanced to the stage of science.” – William Thomson, Lord Kelvin, 1824-1907.
“So what, exactly, is science?” – "A science checklist", part of “Understanding Science: how science really works” at UC Berkeley.
The question is, whether hard science is the full answer. A physicist may see it that way. But when his head is aching, he turns to a physician and hopes for his - as scientific as possible - help, although the headache can’t be measured.
(copied from FreeWillAndDeterminism) …
stepping into this with serious reservations, but…
Hans, I think that “reproducibility” is a too severe requirement. For example, in politics you can’t reproduce anything, still one can create a scientific approach. One can e. g. make predictions and check them.
I find my self reluctant to relinquish the need for “reproducibility”.
One reason stems for the “cold fusion discovery” that was announced a while ago. It was a very dramatic example of the need for independent verification of experimental findings.
Another reason I am reluctant stems from my inate reaction to the “politics” example that you use. “Political science” is itself, not one of the “hard” sciences, but rather, its classed as a “social” science. Perhaps my reluctance is just the result of one of my current projects being mired in a debate about the differences between these classes of science. But I think there is likely more to it. In your specific example of making political “predictions”, I think the actual science is the mathematics of “statistics” in which sampling theory is used to predict the expected behavior of a large population, based on a relatively small sample of observations. While the methods used in this are relatively rigorous in their application of mathematics, it has always intrigued me that the fundamental assumptions of a “normal error” curve are applicable “If and only if” all of the “explainable variations” have first been removed from the representative sample. In practice, this is seldom done, which is one reason most political predictions are fundamentally flawed. (I’ve digressed too far from the basic issue of whether or not the “scientific method” requires reproducibility in this last case, but I personally prefer to demand it, since it is an affective defense against much of the “bad science” that is currently being promoted as camouflage for “vested interests”.)
This is interesting. There could be times, I think when one could approach something like politics with a scientific approach as mentioned by Helmut above, and the checking of predictions could be at least investigated to determine whether the phenomenon happened. It may be that the events are not actually reproducible, but that enough people examine the evidence and accept that certain conditions did indeed lead up to an outcome.
Although, many phenomena, once understood, should be reproducible to some degree.
There are also other issues, such as the shifting nature of physical reality in general.
Some one could state that it is a “scientific fact” that the Island of Great Britain is “80,823 sq mi” im area, and we could measure this to reproduce it. But, over time, the exact area of this island is constantly shifting. Because it is surrounded by ocean coastline on all sides, the exact area, down to the mile may never be reproducible from one measurement to another. The same holds true for the “size” of the earth, the solar system, the Universe, etc.
“Facts” have at least an implicit scope. The size of Great Britain, at a point in time, is a fact that can be independently verified within the usual measurement caveats of “accuracy” and “precision”. The reality that such a measurement has a “transient” (as opposed to “permanent” or “dynamic” (with respect to time) range of results does not eliminate the “reproducible” nature of this fact, rather it “qualifies” it appropriately.
That being said, I agree with Sam that it is possible to apply “scientific methods” outside of the usual domains of ‘hard’ science. In many instances, this has even been found to be beneficial.
Actually, for something like the exact size of the UK, it is always shifting, so on a certain scale, there would in fact be no way to accurately verify a previous measurement. But on a larger scale, the “size” would be more stable, and verifiable. So, we as a species have learned to focus on a particular “scale” to make these measurements. So, on the square mile scale, there is still constant shifting if you take into account the entire coastline, but it is a variance of plus or minus a certain amount of area. Taking that variance into account then stabilizes the measurement. But, it also helps to obscure part of the actual nature of what is being measured. It is mostly that it has only been useful for us humans to look at the measurement and take into account the constant variance. This point I am making is an “aside” from the hard/soft science discussion. It a point about “HardScience?” measurements themselves.
Maybe we could answer he question more accurately if we tried to tell what is not science. What are the methods of gaining knowledge and understanding about the surrounding world that are not scientific? Are there any?
I remember that in high school my science teacher made me memorize the 5 steps of “the” scientific method. I assumed these had been handed down since Galileo unchanged. (Or perhaps Francis Bacon ?) Apparently lots of other high school students also learned “the” scientific method. We were surprised to learn that we were given different definitions, some with more or fewer steps.
To answer RadomirDopieralski, yes, people learn tons of things in ways that are not scientific. Some of those things turn out not to be true.
Nevertheless, here are just a few things that I’m pretty sure are true, and yet cannot be deduced using the scientific method:
In other words,
Then there are other things people learn, that are difficult to put into words – how to ride a bike, how to swim, how to write English, how to write software, what spaghetti sauce is supposed to smell/taste like. As far as I can tell, no one can learn these things just by reading a book – they have to go out and actually experience it.
I suppose one could claim that “trying to ride a bike” is a kind of experiment, where one tries out a bunch of ways one things might keep the bike from falling over. So you could claim it is “scientific”, because running experiments is part of science. But most of the other steps of the “scientific method” are missing. And I don’t think we want to call anything that leaves out a bunch of steps “scientific”.
What do we call it then?
Thanks David (and everyone else). You just helped me better articulate what I understand science to be.
Science is a creation of/tool to extend the brain/mind, which is a system of organic tissue and complex activity that is in part trying to model reality.
Science grew from the self-understanding that the brain/mind system cannot always accurately model reality on it’s own. Science, therefore, implements some extra steps that help us assure, and bridge the gap of our biological equipment modeling deficiency in our built in equipment as humans, that what we are modeling is as accurate as possible.
It is really tough, in my mind, to narrow the above down to an extreme, single lasting assumption about science, unless:
The only lasting assumption that I can think of currently, is that science is a tool to extend the human mind’s capability to accurately model the things that are happening around it. I think this will stand the test of time. Perhaps I am missing other lasting assumptions?
I wonder if we are making it hard by cutting too high or too low?
I mean, “scientific method” is a tool, a procedure (or even several different ones, as there are many definitions) that is supposed to keep the “sciencing” at high enough quality. But – although it is in common use today – this tool was not always used. Great scientists of our history had often many troubles because hey didn’t know (or didn’t use for any other reason) this tool. So, although you can probably recognize a scientist nowadays by the tools he uses, just like you can recognize a plumber or a carpenter, it’s not always that obvious when he’s “sciencing” and when he’s just using that tool for something else. So categorizing by the tools used might be cutting too low.
Another way of categorizing activities is by their results. So we list examples of facts and theories and other knowledge, and we say “this one is scientific” and “that one is not”. Then we look for a pattern and try to articulate it. This is easy for some of them, and harder for other. In particular, some results can be achieved in multitude of ways, indistinguishable from each other by just looking at effects, but substantially different. The fact that the Earth is circling the Sun, for example, can be a scientific statement as well as just a statement of faith, a belief. So categorizing by the results might be cutting too high.
I have another idea on how to cut, but tit’s of little value in practical recognition of science. It is by motives: completely unmeasurable and only possible to guess unscientifically. But then again, does science has to be scientific itself? If you do it for money, you are a businessman. If you do it for fame, you are a showman. If you do it to make the world better, you are a politician. If you do it because it’s there, you are a romantic. But if you do it to see what happens – you are a scientist. I believe the most important attribute of a scientist is not cynicism, perfectionism, purism, consequence, etc. – although they are pretty common – but a genuine, unquenched, unstoppable curiosity. Yes, I am aware that this makes a lot of what happens in schools and in laboratories all over he world unscientific. What can I say, this is how I see it.
It’s interesting what you say, that’s for sure. I called it a “tool” above. MarshallMcLuhan called it a “medium”. I think maybe his name is more fitting. Science is a “medium” for extending the human mind (which includes the curiosity of the mind), according to MarshallMcLuhan. I think he was on to something with his “extensions of man”.
Maybe StatusOfKnowledge might be relevant to this discussion.
Hans, when you state the importance of reproducability in physics, I fully agree. This kind of scientific methodology is made by physics and can be taken as a model for some other fields. But still, not even in physics this is everything. What about the “idea of reproducability” as a concept of methodology and the term “reproducability” as an act of language development. Shouldn’t these steps also count as science? Was Einstein only a scientist while he produced reproducable theory, and stopped to be a scientist when he tried something that didn’t work, or when he tried to understand competing and maybe worse theories?
Will a physicist deny the scientific inquiry of headaches because they - and the success of some treatment - can’t be measured?
We appear to have agreed that the “Scientific Method” requires reproducibility (i.e. Physical experiments should produce consistent Observation results that support a current Causal theory).
I’m not sure I adequately understand the issue of reproducibility as an act of language development. The question of whether Einstein was a scientist when he tried something that did not work; seems to be a simple case of “negated” logic. What I am trying to say in this case is that it is the process of using experimental methodologies and the Observations they generate to support Theories, is part of the Scientific Method, without regard to whether or not the “experiment worked”. In other words, even a failed experiment can be part of a scientific method. (Let’s not forget that Einstein specialized in the development of science as a “theoretician”, rather than as an “experimental” practitioner. Science is a very “big” field with many specializations and the “Scientific Method” seems to be an equally large concept with many possible variations.)
Perhaps what we are considering here is simple a consequence of struggling with language and our ability to clearly communicate complex concepts?
(Sorry I can’t go too much further with this just now, because I am already scrambling to complete the tasks I need to finish prior to my imminent departure.)
Hans, I only agree with respect to fields of science that allow mechanical experimentation. Your use of “hard science” and “social science” already suggests, that “science” is something more general. This page doesn’t ask “what is hard science?” but “what is science?”.
Science always was a moving target. People experimented even before they had the word “experiment”. Only in the 18th century the scientific method, as a mechanical modelling of phenomena, became defined. We should not believe that this is the end of the development of scientific methodology. At about 1900 scientists thought physics was at its end … then came Einstein and Schrödinger and physics started anew at a totally different level. The application of statistics in social sciences is a very poor replacement for a scientific methodology that we do not yet have. There is a need for a methodology for unfolding systems - like cities, ecologies, humans - that can’t be subjected to reproducable experiments. The AlexandrianMethod may provide the solution, but is is a starting point, not an end point. It has not reached the consciousness of people.
I like several of the points you’ve made! They will even give me something to ponder while I am traveling, removed from the internet. One of the last things that I am engaged in before departing is a bit if a debate about Scientific Research and Experimental Development, as recognized by Canada’s Taxation authority. I fear their (economic) harness may have bridled my thinking a bit too much. It will be good to get away from civilization and (re)gain a different perspective. Thanks for the “push” into this “broader” mind-set.
As a hypothesis I’d suggest to widen the idea of science from (predictable, reproducable) to (useful/reusable, sharable) knowledge or theory. I think the essence of science is to produce knowledge that is useful and can be shared. The integration of “useful” and “sharing” is imho “reuse”. Reuse in society is the advantage that makes the social process of science happen.
Radomir, I don’t understand the idiom “cutting too high or too low”?
I think that “reproducibility” is a too severe requirement.
Certainly if we want to draw a distinction between “correct thinking” and “incorrect thinking”, cutting out things that are not reproducible is cutting out too much.
Some people limit the term “scientific” to a narrow definition: ideas that are, at least in principle, falsifiable. Other people use the good connotations of the word “scientific” and expand it to a wide definition, covering all kinds of good things. Alas, shifting between these definitions makes it all to easy to (perhaps unintentionally) lead to equivocation (Wiki:EquivocationFallacy).
As I’ve said at MeatBall:WhatIsScience, I’ve been experimenting recently with avoiding the word “science”.
What do we call it then?
Use other, more descriptive words. For example:
What about the “idea of reproducibility” as a concept of methodology and the term “reproducibility” as an act of language development. Shouldn’t these steps also count as science?
I would say no. It’s an extremely useful and practical idea, and closely related to science. If something cannot be reproduced, then it seems to me that it cannot be falsified (and so is not scientific, in the Karl Popper view).
I would say the “idea of reproducibility” itself, and discussions about it, is not science but philosophy. Part of “the philosophy of science” (Wikipedia:Philosophy_of_science).
Ad cutting too low/too high – human activities can be observed on a number of different levels of abstraction. Ergonomics, for example, is concerned even with single moves of our bodies. The moves can them be grouped into whole “meta” moves, like “come here” or “take this”. At higher levels you’re concerned with intentions, plans, world view, etc., up to esoteric questions about life, universe and all the rest. Now, not all ideas exist in all these levels, and sometimes the meanings can change at different levels (moves that are “bad” from the ergonomics point of view, because they are tiring or even painful, can be “good” from a dance critic point of view). We are trying to isolate the human activity known as “science”, so we are looking at what people do when they are “doing science”. But if we look too low or too high, it’s hard to make sense of these activities, or isolate the ones characteristic for science. For example, myself, I think that scientific method is an “activity” that is often involved in “doing science”, but is not absolutely necessary, nor is only suitable for science. In other words, it’s not this that makes people scientists.
Radomir, if it’s not that, then what is it that makes people scientists? What is it then that is necessary or suitable for science?
Ah, interesting. I’ve been thinking about “cutting too low/too high”, but calling it “looking at things at the wrong level of detail”. I’m tempted to make a page about it – “DetailLevel?” or “AbstractionLevel”, perhaps.
In this particular case, however, I think it’s a bit more sideways. At “my” college, classes in machine vision (image analysis, etc.) are categorized in the Electrical Engineering department – although, in my opinion, what I learn in those classes is more related to biology and even more closely related to computer programming than it is to, say, the Unified National Smart Grid.
I suspect this is because of a historical accident. The electrical engineers building cameras and computers were the first to succeed in the simplest kinds of machine vision, merely because they were the first ones with the necessary hardware.
Likewise, “scientists” were the first people to actually have their hands on a huge variety of other tools, and so any activity involving those tools is often categorized (or mis-categorized) as “doing science”.
I think it’s a certain kind of curiosity and inquisitiveness, plus personal beliefs that put Truth (or whatever close to it they believe exists) at a very high place. The difference is nicely captured in this little strip: http://xkcd.com/242/ Note that in my opinion the Academia is full of non-scientists. They do the same things, but the reason for which they do them (fame, dominating position, fear, competition, habit) is non-scientific.