Raw Thought

by Aaron Swartz

The Hard Sciences

If we say that science is the goal of trying to figure things out about the world, then we see the sciences broadly classified into two categories: “hard” and “soft”. In the former are subjects like physics, biology, and perhaps the honorary inclusion of mathematics. The “soft” sciences, by contrast, include fields like history, psychology, sociology, and economics.

As you might gather from the terms involved, partisans of the hard sciences often look down upon the softer sciences, considering them barely worthy of the term science at all. Indeed, the soft sciences rarely formulate general laws or clear predictions, as the harder sciences sometimes do. But why is that?

The reason is, because the “soft” sciences are, in fact, harder. Humans are far more complicated than atoms, trying to figure out how they work is a great deal more difficult than coming up with the rules of mechanics. As a result, the social sciences are less well developed, which means there’s less to study, which means the fields are easier to learn.

Nonetheless, since the field is so much harder, the people who make progress in it should get more respect. Physicists can isolate atoms and run an experiment; historians have to try to find clever ways to make a “natural experiment”.

Obviously, the progress has to be actual, rather than simply perceived, which is indeed a common confusion in the social sciences, but real observers of science should reconsider who they esteem.

You should follow me on twitter here.

July 11, 2006


It all balances out really. One way in which hard sciences are more difficult than soft sciences is that, since they’re more well-developed, it becomes progressively harder and harder to discover new and interesting things about them. For example, it’s pretty much a dead end to try and discover new chemical elements, since elements with high atomic numbers have half-lives too short to be of use.

Of course, whether it’s a “hard” or a “soft” science, the practitioner shouldn’t abandon the scientific method and resort to telling just-so stories. A historian may not be able to formulate any general Laws of History, but it’s no excuse for him to make up stuff about the founding of the Ottoman Empire, for example.

posted by bi on July 11, 2006 #

“the people who make progress in it should get more respect.”

I think the keywords there are “make progress”.

Too many people don’t.

And I believe part of the disrespect comes from the attitude of many in the social sciences that since it’s about humans, having mathematical and experimental rigor is somehow offensive, or inversely, that it validates nonsense.

posted by Seth Finkelstein on July 11, 2006 #

Another thing to note is that there is no “objective” experiment in the soft sciences in any reasonable sense of the word. Indoctrination, culture, affects the result of any experiments. Any “laws” at the level the soft sciences deal with things, only hold as long as people believe in them. So, perhaps “soft” should be understood in the sense that there are less “hard facts” (without getting too philosophical about the concept of “fact” here) than in the hard sciences, or any “hard fact” would be too complex to be meaningful.

posted by tuomov on July 11, 2006 #

the attitude of many in the social sciences that since it’s about humans, having mathematical and experimental rigor is somehow offensive, or inversely, that it validates nonsense.

Who has that attitude? Everyone I’ve seen has the exact opposite attitude (with the possible exception of postmodern anthropologists).

posted by Aaron Swartz on July 11, 2006 #

Good thinking, i agree with it for the most part; but i think you goofed on this one … “As a result, the social sciences are less well developed, which means there’s less to study, which means the fields are easier to learn.” … keeping with your train of though here me thinks that though should have come out as “As a result, the social sciences are less well developed, which means there’s more wrong information to sort through, which means the fields far harder to learn.”

posted by Seth Russell on July 11, 2006 #

I always thought the “soft” sciences were labeled that way because the claims made by their practitioners are presented as backed by research that isn’t actually science. (Usually it isn’t really reproducible, and this is because, like you say, humans are really complicated.)

posted by Matt Von Bargen on July 11, 2006 #

Could for example psychology be called a science at all if it did not make heavy use of statistics?

posted by Peter Sheldrick on July 11, 2006 #

Aaron Swartz: well, I know of some names in the disciplines of UI design and linguistics, but for now the names have been withdrawn to protect the guilty.

In a way, the fact that the foundations of these fields aren’t well-known, does make it a bit easier for people to get away with pseudo-scientific junk.

History is a sort of special case, I think: the goal of a historian isn’t to discover general laws, but to ascertain which events happened when. And in this task the typical historian is quite scientific indeed.

posted by bi on July 11, 2006 #

Sheldrick: Of course! If I find that making a whole in a certain part of your head causes you to stop forming words, that’s scientific information even if I didn’t use serious statistics. Even in epidemiology, tests of statistical significance didn’t become mandatory until recently.

I don’t think it’s the job of a scientist to discover general laws, especially a social scientist. I think that’s another artifact of physics envy.

posted by Aaron Swartz on July 11, 2006 #

Well, one does need to find general laws, in order to e.g. make health recommendations to the general public regarding smoking. Knowing that this one guy smoked and died of lung cancer isn’t very helpful.

posted by bi on July 11, 2006 #

I guess I’m using the term differently. I don’t consider “smoking leads to lung cancer” a general law, but a specific finding. “Mankind’s inventions harm mankind” was more along the lines of genericity I was thinking about.

posted by Aaron Swartz on July 11, 2006 #

If there is some universal truth about the difficulty rankings of various subjects, I don’t suggest that you try to prove it scientifically. (What would happen if you scientifically prove that history is harder than math?) Instead, how about some Godel?

posted by Felipe on July 11, 2006 #

“Nonetheless, since the field is so much harder, the people who make progress in it should get more respect.”

Can you clarify what you mean by the words “harder” and “progress” in this sentence? I didn’t get a clear idea of what either meant in my four years of social science study.

posted by Scott Reynen on July 12, 2006 #

X field is harder than Y: To come up with an interesting result in X you have to be more clever than you do to come up with an interesting result in Y.

Progress: Coming up with an interesting result.

I’m not sure that really answers your question, though. Perhaps you’re wondering what interesting results are? Here’s an interesting result from psychology: people are less motivated and do worse on tasks when they’re given rewards for doing a good job on them.

posted by Aaron Swartz on July 12, 2006 #

“I’m not sure that really answers your question, though.”

It does, thanks. That’s pretty much what I got out of my four years (thought it was never explicitly stated), and I guess I liked that because I liked interesting more than predictable at the time, but I don’t think that’s the same criteria that’s used to measure progress in the hard sciences.

The big advances in the hard sciences, seems to me, are incredibly boring because they’re so obvious in hindsight - they just formally describe the familiar physical world in ways that make it more predictable. So I’m not sure it makes any sense to compare the hard and soft sciences like this.

posted by Scott Reynen on July 12, 2006 #

The big advances in the hard sciences, seems to me, are incredibly boring because they’re so obvious in hindsight - they just formally describe the familiar physical world in ways that make it more predictable.

Really? As Howard Gardner notes (The Unschooled Mind) a lot of physics is not what you’d expect. Most everyone has a model of the world, for example, in which things “use up” force and eventually run out. The idea of Newtonian mechanics — that a body in motion will stay in motion without any additional force — is pretty shocking; even more so for the fact that outer space proves it to be so.

posted by Aaron Swartz on July 12, 2006 #

I take some exception to the claim that to come up with interesting results in the hard sciences is easy. The basics of the hard sciences are accessible, sure, but they have either been explored beyond hope for a lot of new and interesting discoveries (physics) or are still understood so little that it’s hard to get answers to many simple questions (biology).

Even where the basics are now well understood, the picture is not as simple as meets the eye. Look at what physics was like in 1700AD: most theories were based on intuition, many things were simply mysterious, and the idea of rigor was in diapers. Sure looks a lot like the soft sciences do now, no? And you know, except from the fact that we now know more about what proper rigor entails (though it would seem hubristic to claim that we’ve learned all about it), that’s still what the “hard” sciences look like at their forefronts. Read up on dark matter, dark energy, accelerated expansion, and the like. Noone has the faintest idea about any of these things, despite how long we’ve known about them. (No less than three decades in the case of dark matter; who knows if it’s even there. Maybe it’s just like the ether, and we need a new Einstein to bring about another radical realignment of our existing understanding.) Or take string theory: it’s been in discussion for some half a century, and it remains mostly without any avenue of experimental confirmation. Besides all that, there’s the mundane fact that to make any interesting discoveries in particle theory or astronomy requires millions in equipment.

So it’s not at all easy to make material progress in the hard sciences.

Now consider medicine, which, like many of the soft sciences, has been practiced in some form for the entire history of mankind. Like the soft sciences, it consisted almost entirely of conjecture, superstitution and plain rubbish until the then emerging understanding of rigor began to inform it. Like the soft sciences, it looked difficult to achieve rigor in, until people started losing their superstitions and knowing more.

Some sciences are soft now because we live in their Middle Ages. They are full of conjecture and superstititions. Their Newtons and Einsteins, if there ever will be any, are far in the distant future.

So I would say it’s understandable why some look down on the soft sciences; but I agree with you that to do so is mistaken, even as I disagree that there is any substantial basis on which to rank any of the sciences against each other. All of them are important.

posted by Aristotle Pagaltzis on July 12, 2006 #

I am fascinated to be reading non-scientific arguments about science by seemingly very intelligent people. Are you all making true truth statements about science that cannot proven using science? Thus, you have to use arguments from a system “more powerful” than science to talk about science?

In very simple math terms, Aaron argues that psychology > physics. But what are the basic axioms? Let’s simplify. How do you go about proving English > Java formally?

Is a “hard” scientist more likely to ask these stupid questions?

posted by MIT Guy on July 13, 2006 #

There’s essentially a chain of envy in these disciplines. The scientists and engineers envy the mathematicians’ freedom from boring experiments/raw data and their intellectual rigor (the guy who commented before me is obviously an example, with his use of mathematical symbols and call for axioms). The sociologically-oriented disciplines in turn envy the certainty that physical experiments can bring and their mathematical rigor. However, as you point out, these disciplines are fundamentally different, as the level and types of complexity that they deal with varies. What’s funny is that they often just ape those that they envy, without understanding either how to properly wield the intellectual tools they appropriate or the larger structure of the field they operate in. Take for example economics, and how they try to use a lot of math that isn’t really necessary. They do this because they have seen the physicists do the same and it obscures the intellectual murkiness of their arguments.

posted by Ajay on July 13, 2006 #

“If people do not believe that mathematics is simple, it is only because they do not realize how complicated life is.” — John von Neumann

Occasionally you’ll find sensible scientists and mathematicians, who explain science & math’s place in the world without puffery. Aleksandrov/Kolmogorov’s book on math is one such book, which I found to have a particularly interesting intro.

I don’t have any opinion though on what field is more “worthy of respect.” (I think they’re all potentially useful, modulo any pressures on getting the Right Answers, as Feynman explained in “Cargo Cult Science.”)

posted by Tayssir John Gabbour on July 14, 2006 #

The soft sciences are harder than the hard sciences only in the sense that trying to nail jelly to a tree is harder than trying to nail a piece of wood to a tree. That is, frequently, the soft sciences are engaged in endeavors that don’t make sense. Meanwhile, the hard sciences are almost exclusively engaged in useful prediction. Examples:

  • Physics made all the predictions we now use to make cars, computers, etc, etc.

  • The field of “ethnic studies” has pondered the nature of various marginalized groups for the purpose of being politically correct. No useful results have come from them. How do I know? Well, unlike physics with cars, I can assuredly point to no aspects of everyday life that have been in any way changed or improved by the study of ethnic studies.

There are exceptions to this rule, such as the 1980’s approach to AI (an example of a semi-hard science (CS) nailing jelly to a tree), but they’re fairly rare, and tend to end up inescapably abandoned.

posted by on July 20, 2006 #

MIT Guy: What makes you think that these statements aren’t provable using science? You have a very constrained notion of science if you think it only includes statements that can be proven formally from axioms. In fact, as Ajay points out, that sounds more like math… (For once, Ajay and I agree!)

posted by Aaron Swartz on July 31, 2006 #

Oh, I think we agree more than you think. Nonwithstanding certain disagreements, I wouldn’t be here if I didn’t think you had some things right and were capable of figuring more out.

posted by Ajay on August 1, 2006 #

lmao @humans being more complicated than atoms,quantum theory doesn’t apply on the human scale,there are only a mere 6 billion people on earth.The social sciences are an art,not a science,they’re case/society specific,there are no hard and fast rules…fuzzy/messy/imprecise, it’s like a field based on jello, vs a field based on a rock,Social Science and most of economics is common sense whereas most things in science boggle the mind,but they just happen to be true

posted by zanzabarjones on July 23, 2011 #

You can also send comments by email.

Email (only used for direct replies)
Comments may be edited for length and content.

Powered by theinfo.org.