Philosophy of Science

We know that science is a thing that can be studied, so bring in the philosophers.


The philosophy of science deals with the practice, methods, workings, implications, and values of science. Philosophers of science are concerned with what distinguishes science from nonscience, and about topics such as reasoning, induction, realism, explanation, reductionism, empiricism, validation, and the "social impact" of science.

Read more at:

What is Science?

You'll get lots of definitions if you look up this word in a dictionary (, Merriam-Webster) but Wikipedia's got a good one:

Science is a systematic enterprise that builds and organizes knowledge in the form of testable explanations and predictions about the universe.

You probably already heard about how "modern day" science can trace back to Mikołaj Kopernik, and how others such as Johannes Kepler and Galileo Galilei and Isaac Newton contributed their ideas and methods and we now have this system that values:

Philosophers of science look at these aspects, and more.

History of Science

In addition to the scientific method and the body of knowledge we now have in scientific fields, it's helpful to have a sense of the history of discoveries in science, people's immediate reaction to them, how they became accepted, how they were applied, and sometimes, how they were extended or even supplanted.

Often knowledge evolves through better measurement. Sometimes there are true scientific “revolutions,” for example:

Why does physics keep popping up here?

Philosophers of science pay great attention to the historical aspects of scientific discovery and application.


Science differs from pseudo-science, junk science, voodoo science, bad science, pathological science, nonscience and just plain nonsense, to paraphrase Michael Shermer. But how?

Is there a clear line of demarcation between science and nonscience, or is science perhaps just a set of activities that loosely share some main features?

There have been claims in the past about what constitutes a demarcation:

But all of these have problems.

Exercise: Research the problems with all of these proposals. Wikipedia's article on the Demarcation problem will be a great place to start. See if you can come up with your own proposals for demarcation.

Some have even said the whole demarcation problem is without merit.

Exercise: Who said this? What are their reasons?

Some philosophers of science are interested in this; many are not.


Philosophers of Science are interested in the nature of scientific reasoning. What kinds of reasoning do we know about?

Abduction explicitly appeals to explanatory or causal effects; induction appeals only to quantitative effects.

Read about these at Wikipedia and the SEP.


Induction and abduction (a.k.a inference to the best explanation) are not precise, and may lead to false conclusions. But we have to use them, right?

Exercise: Karl Popper once tried to show science needed only deductive reasoning, since falsifying a theory by finding a counterexample deductively works. But what is this argument missing?

Hume's Problem of induction: Induction can only be justified... inductively! Wow, good stuff. Possible solutions:

But probability doesn't solve the problem, regardless of the interpretation: frequency, subjective, or logical. The frequency interpretation ultimately relies on induction! Ouch! Subjective probability doesn't even begin to address the problem (it's all about hunches). So that leaves the logical interpretation which....

Exercise: Why do many people think the logical interpretation of probability cannot solve Hume's Problem?
Exercise: If Hume's Problem is unsolvable, what does that imply about the nature of rationality and the degree to which we can be confident of scientific reasoning?

Finally, speaking of probability:


What might a philosopher of science wonder about abduction? A couple things, at least:

Ockham's Razor

Why should we generally make only the simplest assumptions (or assume the simplest causes)? Because we can keep on adding further assumed "causes" until — voila — the theory is explained.

Ockham's Razor is the principle that one should make no more assumptions than necessary.

It is a heuristic not a law. And it is often misued! The RationalWiki article advises: "It's probably best to avoid ever citing Occam's razor unless you're damn well sure you know what you're doing with it."

Exercise: William of Ockham did not invent this principle. So why did it get named for him?

Coherence and Foundationalism

If a scientist makes a claim C then there should be a justification C1. But that's a claim, too, so we need to justify it with C2. Another claim? When does it end?

We can regress down to a basic claim, something we take as inherently true, not in need of any justification. This is foundationalism.

Or we regress down to a claim which is part of a coherent set of claims. This is coherentism.

Exercise: Research the problem of the criterion, the regress argument, foundationalism, and coherentism. How are foundationalism and coherentism opposed?


What makes an explanation "scientific"?

Hempel's Covering Law

Hempel offered the following as a structure for scientific explanations:

  General Law
  Specific Facts
∴ Phenomenon to be explained

Sometimes it seems to work, as in this example from Okasha:

  Laws of trigonometry
  The sun is 37° over the horizon
  The flagpole is 15 feet high
∴ The flagpole shadow is 20 feet long

But there are at least two obvious problems with the covering law, as in these two examples from Okasha:


Can we fix the covering law by requiring relevance, and ... causality? It's plausible but not everyone would accept this.

To see how large of a topic causality is in philosophy, try this Google search.

What can science not explain?

If you are going to claim something can not be explained, you should have a good reason for saying so, not just "well we don't know yet." But there are a few serious directions people have taken:

  1. Some things are just really mysterious and at least to some defy explanation. For example, there's the Hard problem of consciousness.
    Exercise: Contrast the views of (1) Dennett and (2) McGill on this issue.
  2. Chains of explanations keep on going, so maybe some things are just too fundamental?
    • Why do electrons behave the way they do?
    • Where did the electrons come from? And if we ever explain that, then how to explain that clause?
    • Why is there gravity? How did it get separated from the other forces?
    Exercise: Got more of these?
  3. There's the problem of multiple realizations. It's hard to say biology is precisely reducible to physics because the abstraction of the "cell" isn't exactly describable in physics. What exactly is and isn't a cell in terms of molecular arrangements? What is and is not pain? An economy?
    Exercise: Read the multiple realizability article at the SEP. How have reductionists challenged this doctrine?
    Exercise: Read Dawkins's essay The Tyranny of the Discontinuous Mind, in which he discusses the continuum of intermediates. How is this related to multiple realizability?


The ancient philosophical debate between realism (nature exists independent of conscious observers) and idealism (no it doesn't) isn't part of the philosophy of science, but the debate between scientific realism and instrumentalism is.

Scientific Realism and Anti-Realism

Traditionally realism refers to the idea that the universe has a true existence outside of the perception of conscious beings. It's opposite is idealism. But when it comes to science we have a related distinction:

Scientific RealismAnti-Realism
Science aims to provide a true description of the universe. Science aims to describe only the observable part of the universe.
Quarks, leptons, bosons, and hadrons really exist. Unobservable things like quarks, leptons, bosons, and hadrons are only convenient fictions that help describe things and make predictions .
Science can help us discover reality. The unobservable is beyond our knowledge; we are limited by our senses.
Why is the atomic theory so successful (e.g., lasers work as predicted based on the energy state transitions of electrons), were it not actually real? We don't want to believe in miracles. Not so fast! Phlogiston turned out to not be real. The "ether" turned out not to be real. So did lots of other empirically successful things. Maybe there is another non-miraculous explanation for lasers. We just don't know.
What's wrong with tools to enhance one's powers of observation? Are binoculars, telescopes, microscopes, electron microscopes, particle detectors not appropriate? Why would you object to these? Okay so the dividing line between observable and unobservable is fuzzy, but we still have clear-cut cases.
Underdetermination is not a problem: this argument rules out the reality of unobserved entities as well as unobservable. It's just a complaint about induction. Underdetermination is a serious philosphical problem. The claims science makes about the unobservable are justified by only by observable behavior. Therefore, there will always be alternative explanations.
Even though multiple explanations may exist regarding the unobservable, there's often a best explanation (e.g., the simplest), and sometimes it's hard even to find one. Best and simple don't mean true.

Scientific Progress

Philosophers of science are interested in how sciences changes, or progresses. When does it just slowly evolve? What constitues a scientific revolution? Are there patterns in the way ideas change?

From Logical Positivism to Kuhn

The Logical Positivists (Wikipedia, SEP)

Exercise: What are the objectives to logical positivism? Why was it considered dead? What parts of it are still relevant?

Thomas Kuhn's 1963 book The Structure of Scientific Revolutions was hugely influential. He said

Okay, cool so far. Not very controversial. Even though the Darwinian and Einsteinian revolutions fit Kuhn's model, but the molecular biology one doesn't, the ideas are plausible. But Kuhn made some claims that did generate controversy....

Adoption of New Paradigms and Theory Dependence

Kuhn also claimed:

Exercise: Can a relativist make the claim "It is objectively true that truth is paradigm-relative"?


Do Kuhn's claims (if valid) show science to be a rather non-rational endeavor? He didn't exactly mean that...

Social Aspects of Scientific Progress

After Kuhn, we now consider

But some have misinterpreted Kuhn, or taken his ideas in radical directions, sometimes by:

Problems in Specific Disciplines

Absolute Space

Newton's concept of space was absolutist: space always existed and has a fixed 3-D coordinate system; at some point perhaps matter was placed into this absolute space. Leibniz's was relational: space is defined by the relations between what it contains, and did not exist before the creation of all things.

This question is still being debated.

Exercise: Which "narrative" appeals to you? What does relativity contribute to the debate? Can they both be right?

Philosophy of Biology

See Wikipedia for this.

Biological Classification

How should we choose to classify organisms? Two opposing schools of thought:

See the difference? Let Randall Munroe explain it:

So cladists are really into the phylogenetic relationships. See the Wikipedia article on the phylogenetic tree for more information, and some lovely diagrams.

Speaking of diagram, please take the time to study this beautiful tree of life diagram, because it shows more than just today's species; it gives you a good sense of the history of life on Earth (including those pesky mass extinctions!)

Exercise: What are some of the best guesses for which percentage of species that have lived on this planet have gone extinct? Cite your sources.

Why does this debate matter?

Exercise: Ponder this related philosophical question: Why do these classification schemes in biology always seem to be hierarchical?

Modularity of the Mind

Is the mind a general-purpose problem solver, or is it made up of specialized modules that are (to some degree) independent of each other? Or a bit of both?

Fodor noted some properties of cognitive modules, including:

Brain injuries, language learning, and some experimental data strongly support modularity, but is the whole brain modular? How does a juror go about determining a defendant's guilt or innocence? What module is that? What module recognizes a second cousin?

But how is any of this of concern to philosophy? Some ideas:

Exercise: (For computer scientists) How is the Unix philosophy similar to the modularity of mind hypothesis?

Ideology and Morality

Because the philosophy of science is concerned with social aspects of science, we sometimes see science and ideology and (perhaps ethics getting mixed up together). Samir Okasha writes "advocates of sociobiology have tended to be politically right-wing, while its critics have tended to come from the political left."

Wait, what? Left and right? What do those terms mean? Jonathan Haidt has a pretty good answer:

Interesting, eh? Haidt is introducing 5 dimensions of morality: harm avoidance, fairness, in-group loyalty, authority, and purity. (For more, see Haidt's essay in Edge.) Now the next question is, can science explain morality? Sam Harris says it can:


Science is effective for many reasons: it is a "collective pursuit," it tends to root out mistakes over time, etc. Are there any criticisms? The philosophy of science looks at criticisms, too.

Case Study

Chris Mooney has written a book called The Republican War on Science, about the "conservative agenda to put politics ahead of scientific truth." Fine. But the left can attack "scientific truth" too. How?

Start by watching this video, by Steven Pinker. He cites real numbers!

Pinker even has a book on this topic too. Note the claims made about rates of violence (or at least the chance of dying a violent death at the hands of another person) being higher in certain societies than others.

Exercise: Do some research to see how these claims are backed up.

Now, check out this article. Note the interesting tagline: "Respected author's book condemned by Survival International as 'completely wrong, both factually and morally'."

Exercise: Is the "leftwing" group accurate or off-the-mark to claim that Diamond's work is factually incorrect? Why do they make such a claim? Did Pinker address in his video why someone might make such an attack?