COMP6216What Is Science?

Outline of these two lectures

Practical concerns: third year projects, Msc theses, getting a PhD

Origins of science in philosophy; the problem of induction

Popper and falsificationism; Kuhn and scientific revolutions; Feyerabend's scepticism on method

Science as model making

Quine's pragmatic view

A warning from Richard Feynman

Feynman famously said that:

"Philosophy of science is about as useful to scientists as ornithology is to birds."

And yet here I am trying to talk to you about philosophy of science.

Why do I disagree with Feynman?

Why talk about the nature of science?This module is about how to build useful computational models.

Before learning about how to do science via computer simulation, we need to start with a clear picture of what science is.

And a practical reason -- in about 3.5 years you will all face external examiners of your doctoral thesis.

External examiners

Examiners like to ask things like:

"What were your research questions?"

"Convince me that your methods were the right way to answer these questions."

"What specific hypotheses did you have?"

"What do we know now that we didn't know when you started?"

"Sum up your original contribution to scientific knowledge."

Your PhD/Msc/Third Year project in contextAnswering these questions is easier if you have a clear view of how science works and how your own PhD fits into the bigger picture.

Let's consider a graphical view of what a PhD is:http://matt.might.net/articles/phd-school-in-pictures/

Imagine a circle that contains all of human

knowledge.

By the time you finish primary school, you know a little.

By the time you finish secondary school you know a bit more.

With a bachelor's degree, you gain a specialty.

A master's degree deepens that specialty.

Reading research papers takes you to the edge of human

knowledge.

Once you're at the boundary, you focus.

You push at the boundary for a few years.

Until one day the boundary gives way.

And that dent you've made is called a Ph.D.

Of course, the world looks different to you now.

So don't forget the bigger picture.

Do we agree with this picture?

Certainly helps put one person's PhD in perspective.

Note that your contribution is put in the context of a much larger body of research -- "one more brick in the wall", if you like.

But nothing much is said about the structure of what's already there.

A bigger reason

Getting you to pass your viva is not our only concern here.

As doctoral students you're all part of a great human project.

You deserve some space to reflect on how that project works.

How do we know things?

Figure it out in your head, e.g., 21 / 7 = 3.

Instinct. Born knowing it, e.g., that snakes are dangerous.

Revelation? Religious or meditative experience?

From authority: somebody told us, we were taught that it was true, e.g., Ulan Bator is the capital of Mongolia.

Tradition: something has always been held to be true.

Experience: see for yourself.

Putting questions to nature

The development of science was a move away from knowledge gained from authority, revelation, instinct, tradition, etc.

Instead the idea was to look carefully at the natural world in order to find out how it really worked.

This is often dated to the 1500s but whether it was revolutionary or evolutionary is open to debate.

The scientific revolution?

Certainly 1543 saw the publication of two remarkable books.

"On the revolutions of the heavenly spheres" by Nicolaus Copernicus.

"On the fabric of the human body"by Andreas Vesalius.

Origins in philosophy

Some of you're studying for the degree of PhD: to become a "doctor of philosophy". But why philosophy?

What we now know as science developed as an offshoot of philosophy.

In the middle ages there were specific areas of practical professional knowledge, e.g., stonemasons or book-keepers.

All of the big abstract questions about how the world worked were the domain of philosophers.

Science as a new method in philosophyConsider the methods we would use today regarding a question of ethics, e.g., animal rights.

Reasoned argument, invented examples, personal intuition, etc.

These methods were once considered the right way to learn about the world also.

Then "natural philosophy" emerged: the idea that we should look carefully at the world itself to find out how it works.

As scientists, you are all "natural philosophers".

Do I want to be doing philosophy?

Short answer: you can't avoid it.

There are no scientists without philosophical positions, just scientists with implicit or unexamined philosophical positions.

Wilfrid Sellars: philosophy is the study of "how things, in the broadest possible sense, hang together, in the broadest possible sense."

No bad thing to be involved in!

From natural philosophy to science

Older meaning of the word "science": a body of knowledge that could give reliable or predictable results, e.g., cooking or navigation.

This sense of the word led John Locke to express scepticism that natural philosophy could ever become a science.

The irony was not intended.

Precise methods?We now associate science with precision in measurement and, sometimes, prediction. So Locke's scepticism seems strange.

But if we look at scientific processes such as:

Coming up with a research questionDesigning an experimentFormulating a hypothesisAdvancing or modifying a theoryComing up with a whole new scientific paradigm

... they can seem like black arts.

Locke's scepticism becomes a lot more understandable.

Some questions to come back to

Science seeks to explain natural events -- but what counts as a good explanation?

Are good theories synonymous with successful predictions?

Science: a special kind of knowledge practiced by a distinct group and pursued through a unique method? Or just an extension of common sense?

Consider library science, political science, and computer science. Is the word "science" here being used in the modern or the older sense?

Inducing natural law

The earliest picture of how science worked was that it was all about induction.

You would go out into the field, contemplate anything and everything, notice some kind of regularity, and a discovery would emerge. (This view lasted right up until the Vienna Circle in the 1920s.)

From lots of observations, you would induce a natural law, e.g., the sun always comes up in the east.

Inductive as opposed to deductive logic.

Deduction versus induction

Deduction goes like this:

All men are mortal Alan is a man Therefore Alan is mortal

Induction goes like this:

Alan is mortal Bob is mortal Chris is mortal ... Therefore all men are mortal

The problem of induction

David Hume was the first to point out that justifying knowledge gained via induction is tricky.

That the sun came up every day last week does not guarantee it will come up tomorrow.

Bertrand Russell's example of a turkey making (false) inductive conclusions about the regularity of its being fed.

The problem of induction

Many people object to Hume's attack on inductive reasoning by pointing out that it must work because science has done very well in the past using this method.

The fatal flaw in this argument is that it is using an inductive argument to justify the use of induction.

The conclusion is not that scientific progress is impossible. It's that we must be doing something other than sitting back and inductively discovering laws of nature.

Falsificationism: Karl Popper

Popper understood the problem of induction.

Could see that the observe-and-induce model wasn't a good account of how science worked.

Noticed an important asymmetry in scientific law discovery:

Lots of positive observations prove nothing, but one negative observation can disprove a theory.

Falsificationism: Karl Popper

Popper therefore claimed that science was all about trying to falsify hypotheses.

This idea has three major consequences:

1.Experiments should be designed so that they can test and potentially reject hypotheses.

2.Scientific knowledge is the list of hypotheses that have not yet been falsified: we're not 100% confident of anything.

3.To count as a scientific proposition, something has to be potentially falsifiable.

Problems for falsificationism

Is this what scientists really do? Try as hard as they can to falsify their most cherished hypotheses and then throw them away immediately? In practice scientists hold on to ideas for a long time despite empirical grounds for rejection.

Popper is clear on how hypotheses can be rejected but doesn't say much on where hypotheses come from in the first place. Why test this and not that idea?

Are we content with an account of knowledge that allows any proposition not yet falsified a potentially equal status? Are we confident that the Popperian process will lead us closer to truth?

Good points to take from Popper

Falsifiability is important: if an idea cannot be tested through experiment (e.g., psychoanalysis, religion) it is outside the scientific realm. This helps us to define what science is.

Absolute certainty is something we probably have to give up on. The argument against induction shows that we cannot have mathematical precision. The history of science shows that even the most elegant theories get modified or replaced.

Confidence in our hypotheses and theories is something that builds up over time as they stand up to repeated testing.

Scientific revolutions: Thomas KuhnKuhn began with the observation that real scientists don't do what Popper says they should do.

They don't drop their favourite theories when new empirical evidence suggests those theories are wrong.

Most science is done within a particular "paradigm" and scientists are resistant to changing paradigms.

Scientific revolutions: Thomas KuhnKuhn described three stages of science:

1.Pre-science: sorting out an initial perspective on a new problem.

2.Normal science: doing experimental work that usually supports the dominant perspective. Inconsistent results are often treated as mistakes.

3.Revolutionary science: after enough results come up that are difficult to explain, some creative individual initiates a "paradigm shift" that results in a whole new way of seeing the problem.

Scientific revolutions: Thomas KuhnExamples of paradigm shifts: Ptolemaic to Copernican astronomy, the move from Newtonian to Einsteinian physics, the "cognitive revolution" in psychology.

Kuhn said that opposing paradigms are incommensurable, i.e., they're fundamentally different ways of viewing the world -- they can't both be entertained at the same time.

Problems for the Kuhnian view

People doing normal science sound a bit like they're getting nowhere, i.e., they're just supporting the orthodox view.

If opposing paradigms are incommensurable, how are we supposed to rationally choose between them?

Doesn't this make scientific paradigms sound a bit like changing fashions in clothes or music? The new paradigm is "cool", the old one is boring? How can this square with the idea of scientific progress, of closer approximations to the truth?

Some points to take from Kuhn

Science does have movements, fashions, and paradigms. Your work will be better if you're clear on which ones you belong to.

A minority of us will be involved in starting a brand-new paradigm: an exciting prospect! But how can we tell the difference between controversial but fruitful ideas and ideas that are just crazy?