What is Science? (2)

Posted in Change in Science, Philosophy, Philosophy of Science, Science by e1saman on August 24, 2010

Francis Bacon’s scientific method consists of a procedure to compile scientific results in a set of rules or principals which will try to describe all the identical phenomena to the ones under experiment. The step between experimental results and theories involves what we call reason; in order to decide on future events we use reason to study the past events and create a rule that will guide us to the future.  So for example, since the sun is rising every day, for the past thousand of years, therefore it is going to rise tomorrow. It’s easy to understand that this is a “tricky” point in the scientific procedure, it involves a little “guessing”.

We either have to assume that nature is really using a set of rules and laws , or even -more extreme- that laws and rules are more important parts of nature than the phenomena and experiences  themselves. And if we want to take the second view  to a more extreme we can also assume (or believe) that we can trust our own thought and reach these secret ways of nature without experience.

David Hume

“Humans have knowledge only of things they directly experience”

David Hume (7 May 1711 – 25 August 1776), Scottish philosopher considered as one of the most important figures of Western philosophy. Hume attacked the views that ideas can exist outside of human experience and he tried to show that the root of all ideas is experience. His main “weapon” was the so-called “problem of induction” which can be (rather poorly) described  with the above example of the sun. He suggested that something else other than reason suggests us to suppose that nature is uniform (i.e. since the sun is rising every day it is going to rise tomorrow, that the pattern will continue) and this  something else is natural instinct. Since knowledge is based on experience, knowledge is subjective. In that view science can only be if there is collective experience that has structure, stability and repeatability. These regularities are the true subject of Science according to Hume.

“As far as the laws of mathematics refer to reality, they are not certain, and as far as they are certain, they do not refer to reality.” Albert Einstein


12 Responses

Subscribe to comments with RSS.

  1. beyondanomie said, on August 24, 2010 at 4:33 pm

    Put another way, Science is an agreed, widely applicable, frame of reference to interpret subjective experiences. The problem (to many rationalists, at least) with that definition is that there is little to actively distinguish it, from magic (viz. Arthur C. Clarke’s famous dictum), or religion, or political belief systems. They’re all more or less inherently consistent, though often anathema to others using a different frame of reference.

    I would suggest that to worry about this is less important than analysing what the various systems have to offer us. In other words what is the benefit to each system of belief, rather than how true it is.

    Science, as a system, has allowed for some wonderful improvements to our ability to influence the world. That’s enough to justify continuing its development.

  2. e1saman said, on August 24, 2010 at 7:57 pm

    I agree with the definition, and I like provocative ideas too; one of them is that science does not defer actually from magic that was elaborated by Paul Feyerabend .
    I know that realists don’t like that but they are so many in the areas of “positive” science that we need to breathe by reminding them that science is not what they believe only :-).
    But don’t take me wrong my purpose is not in any case the denunciation of science, on the contrary is to search on the boundaries of science in order to make it broader and more appealing. Things that we don’t know we will know! Thanks for the comment.

  3. beyondanomie said, on August 24, 2010 at 11:23 pm

    Absolutely; it would be daft to denounce science.

    Equally, I find it daft to put it on a pedestal on its own and assume it’s more rational than others; it’s simply a very useful way of processing & interpreting data, amongst others.

    Thanks for an interesting post!

  4. aretae said, on August 28, 2010 at 5:06 am

    I think you can do better than Hume. Or at least, I’ve been trying to for 20 years. I say Hume didn’t understand Bayesian updating, and that is sufficient to fix Hume.

    If one starts from sense perception, beliefs with p<1, and Bayesian updating, then our current world falls out quickly. The existence of the external world becomes a hypothesis with a near unitary probability. So do the laws of logic and math, and the existence of self and other minds.

    Hume's primary objection is that we have no reason to believe that the future will be like the past…perhaps instead we could assume that the future will be unlike the past. He misses, though, that his also is a Bayesian hypothesis that experiment can show to be silly. When compared to the alternate hypothesis: the future is like the past, his hypothesis approaches p=0 very quickly.

    Also, if one asks reforms the question into one of prediction…I think one does better overall. Truth is far over-rated. How good are your predictions.

    • beyondanomie said, on August 30, 2010 at 3:49 pm

      “Also, if one asks reforms the question into one of prediction…I think one does better overall. Truth is far over-rated. How good are your predictions.”

      This comment is spot-on, but this doesn’t conflict with Hume – or with Science – at all.

      What it conflicts with is the assumption that there is ever certainty. To use the language of statistics, p can never equal 1 or 0, no matter how close it gets (not even using a Bayesian model). There can never be 100% confidence in an external world. Science only applies within confidence intervals. Absolutist faith is the only entity that operates on a binary yes/no basis.

      Science is not free of faith however, it simply adopts a consensus approach to the limit faith is allowed to influence a judgement (eg a 95% CI in a more classical hypothesis test, which could be thought of as a special subset of a more general Bayesian model). That’s what I meant in my previous reply, by it being “an agreed frame of reference”.

      • aretae said, on August 30, 2010 at 4:04 pm


        Fair enough, mostly. Hume argues that there is no certainty in deduction…but he goes further and suggests that we have no reason to believe that the future will resemble the past. I think Bayes kills this.

        Further, I think that he argues that deductive reasoning does give certainty…which again, I object to. The line is 0<p<1 for all propositions…which you nail in your comment.

        I've been arguing over chez moi that science divorced from tested predictivity isn't science.

        As to CI of 95%…I think that's an agreement about publishing, not about science.

  5. e1saman said, on August 28, 2010 at 6:49 pm

    I am sorry but I am not familiar with Bayesian updating I will try to give it a try though! But I don’t understand what possibilities have to do with philosophy, I am sure I don’t get something right!

  6. aretae said, on August 28, 2010 at 7:20 pm

    It makes for a really interesting approach to knowledge to start from mostly Descartes (I observe that I exist, and that I have observations) and apply Bayes. Read Eliezer at LessWrong for the primers on Bayes.

    Super short version:

    All beliefs have a (subjective) probabilty. I believe (predict) that the sun is coming up tomorrow with p=.99999

    If it comes up again tomorrow, the subjective probability goes up to .999991 (not much of difference).

    If the sun appears NOT to come up tomorrow, my subjective probabilty for the next day drops to something like .80 — an unforseen event like that would screw with my worldview.

    If ALL beliefs are approached that way, life is easy.

  7. aretae said, on August 29, 2010 at 6:23 pm

    Eli doesn’t like penrose for 2 reasons:

    1. Eli is trying to build an AI, and considers building friendly AI his super-important, civilization-determining lifework. Penrose says this is impossible. Kind of a conflict.
    2. There’s really only 2 choices in theory of Intelligence. Either it’s material and evolved, or it’s not. And the anti-material position is (a) Penrose’s, and (b) essentially anti-scientific. Eli is extra-smart, and offended by other smart people (Penrose) who are effectively mystical.

    • e1saman said, on August 29, 2010 at 10:51 pm

      Hmmm, I like the “mystical” part. Actually I have to admit that this is the original purpose for starting this blog; I want to study the mystical part of great Scientists. (for example the “alchemical” background of Sir Isaac Newton, or the “mystical” side of Carl Young). The problem is that I do not know the correct point to start from.

      Don’t take me wrong I am not anti-science or religious, but i like the tricky questions, and special the origins of discovery.

  8. beyondanomie said, on August 30, 2010 at 5:12 pm

    Replying to aretae, from their reply to me above (I can’t reply in the threaded comment, as software doesn’t allow nested replies beyond 3 layers deep. 😦 )

    “he goes further and suggests that we have no reason to believe that the future will resemble the past. I think Bayes kills this.”

    OK, I may be mis-remembering Hume here, so apologies if so. But I thought Hume would argue that you cannot KNOW that the future would resemble the past, not that you cannot BELIEVE it will. Probabilistic theories allow for a (very!) strong belief it will, but never a _knowledge_ that it will.

    To Hume, “relations of ideas” (things removed of belief, and existing only in truth) were basically pure mathematical constructs, because they were independent of subjectivity. This is of course highly arguable on many different fronts, but I don’t think science would automatically fall under Hume’s “relations of ideas” sphere rather than the more subjective “matters of fact” one.

    (when I used “science” I mean science as it is commonly perceiving by the world at large; which itself a construct very influenced by publication agreement/bias, as you rightly hint).

    I think we’re actually arguing from a fairly similar core position to be honest, but from a slightly different facet of the topic. Interesting stuff, for sure!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: