Explanation versus justification

Many people conflate explanation and justification. An explanation is a theory about why something happened or why we should do one thing rather than another. A justification is a story about why we are right, or probably right, to adopt one theory rather than another or one proposal for action rather than another. Explanations are good; justifications are at best a waste of time.

Explanations are good because they provide a target for criticism. If I say why I’m doing something somebody might come up with an argument against it that will change my mind. If I take an action for which I have some explanation and it all goes horribly wrong then I may be able to criticise my actions more easily if I can explain them. An explanation can also be looked upon as an encouragement to criticise an action or idea. An explicit statement of why it seems like a good idea invites people to pick apart the list of arguments I’ve given for my preference.

Justification does not encourage criticism. A justification supposedly shows that its conclusions are correct or probably correct, which means presumably that we probably shouldn’t bother criticising them. Justification also encourages people to waste time with endless fiddly details of a particular idea or course of action when often a cold hard critical look at the theory without the details will make it seem unpromising. I will give an example. I recently witnessed a philosophy discussion in which the participants were discussing knowledge by discussing the meaning of the phrase ‘to know.’ The participants were tying themselves in knots talking about justification and so on until somebody said that maybe there was something wrong with the way people usually think about knowledge and perhaps they should change their theories. The leader of the discussion then replied by saying, ‘Ah, but then we wouldn’t be discussing the meaning of the phrase “to know” in English.’ The discussion then waded into a mire of pointless precision and logic chopping. The person who spoke up against this way of proceeding was right. Just discussing the meaning of ‘to know’ in English means uncritically taking on board a load of ideas many of which may be completely wrong. This is a completely wrongheaded way to do epistemology, but the people involved in trying to justify the way people use ‘to know’ were too busy looking at the fine details to see this problem. As such they spend all their time working on a problem they could solve by spending two minutes with a dictionary instead of spending their time trying to solve interesting epistemological problems.

This entry was posted in epistemology. Bookmark the permalink.

12 Responses to Explanation versus justification

  1. Lee Kelly says:

    The confusion of explanation and justification is deeply rooted. For example, the term “reason,” as in the question “what is the reason for your belief?” is ambiguous. Sometimes the reason demanded is a cause of somekind, like “I believe because that is what I was told in school.” However, other times the reason demanded is a justification, i.e. an authority, usually an abstract principle, which justifies the belief.

    Btw, great blog!

    An explanation often provides causes for paricular facts (whether they be physical, psychological, or even, in a metaphorical sense, methamatical), while a justification often references causes while implicitly appealing to some authority to justify belief. The fact that both explanations and justifications often appear very similar, that our language fails to make the distinction clear by using different terms, seems to me more than a coincidence.

    The conflation of causes and justification, i.e. that only beliefs which have the “right” physical or psychological causes are justified, has led to some peculiar developments in both philosophy and psychology.

  2. Alan Forrester says:

    As Popper noted in “Sources of Knowledge and Ignorance” (in Conjectures and Refutations), Western rationalism has typically been justificationist and therefore authoritarian. In Retreat to Commitment Bartley separated justification and criticism. We should also separate justification and explanation.

  3. Peter D Jones says:

    “A justification supposedly shows that its conclusions are correct or probably correct, which means presumably that we probably shouldn’t bother criticising them.”

    1. Progress is based on having an ever expanding pool of ideas that shouldn’t be criticised. If we keep going back to square one, we get nowhere

    2. Ideas that are probably correct can still be criticised.

    3. Probable but not irrefutable correctness is not authoritarianism

  4. Alan Forrester says:

    1. Progress is based on having an ever expanding pool of ideas that shouldn’t be criticised. If we keep going back to square one, we get nowhere

    2. Ideas that are probably correct can still be criticised.

    3. Probable but not irrefutable correctness is not authoritarianism.

    No. Progress depends on solving problems, i.e. – generating ideas that are not wrong in some particular respect in which past theories were wrong. Note that it doesn’t depend on proving those ideas are right or probably right or on proving they solve the problem, only on whether, in fact, they do solve the problem.

    Also, leaving all your ideas open to criticism doesn’t imply going back to square one because there is no square one. There is no foundation from which we build everything up, and such an idea is inherently and deeply anti-rational because “square one” would have to remain unexplained.

    As for the notion that probable but irrefutable correctness is not authoritarian, that seems to be contradicted by your own statement that some ideas should not be criticised. There is no way you can tell whether a criticism of any particular idea will be fruitful in advance of hearing the criticism because that would require being able to anticipate what knowledge you will have in the future and if you could do that you would already have that knowledge.

  5. Rafe says:

    Peter, can you explain what you mean by “going back to square one?” with an example from some field of research where you are currently engaged?

    Have you published in the scientific or philosophical literature anywhere so we can get a handle on your approach when you are laying out a line of thought rather than just dropping into a pub conversation?

    The reason I ask is that there was an interesting exchange in the Critical Cafe a few years ago, a person came on the list and started arguing with the resident Popperians in an interesting but frustrating manner. I thought he was probably an undergraduate and it was gratifying to see that he had read much wider than the usual, but not to the point of seeing all the defects in the views he was expounding. Someone wrote (in way that was supposed to be encouraging), something like, “it is good to see a beginner taking a wider interest and if you keep reading critically for a year or two with the benefit of a better grip on Popper and Bartley, you will be able to get ahead and leave most of this stuff behind”.

    He replied that he had been reading philosophy for thirty years.

    I gather that he was not a professional but an interested laypeson, anyway he kept arguing and he kept going back to first base with the logical positivists and the logical empiricists, and he never got the idea of non-justificationism. He complained that no other philosophers talk about it, so that seemed to make it out of court so far as he is concerned.

    Justificationism is a bit like first base, we all start there and too many people don’t have to go back to it, because they never leave it:)

  6. Peter D Jones says:

    Rafe,
    Since science does progress, it doesn’t go back to square one, and so there are no examples of this. Going back to square
    one is a theoretical idea, an unwelcome consequence of the notion that not only are absolute proofs and disproof impossible, but that nothing is ever accepted or shelved for all practical purposes, so that everything is constantly researched and criticised equally forever. I am not sure if that version of fallibilism is actually found in Popper’s works.

    I do not have a problem of failing to “get”, as in understand, Poppers ideas.
    Nonetheless , I do not agree with all of them

    I have posted a number of criticisms of them and what is needed is relevant
    responses, not repetitions of the original ideas or “if only you read Popper you would Believe as we do”.

  7. Peter D Jones says:

    Alan,

    There is a square one because there is a finite history of science.

    An idea can onnly be criticised by another idea that is somehow more probable
    or better corroborated. Leaving everything equally open to criticism leads to a hopeless impasse. Criticism requires this asymmetry as much as verification of foundationalism

    It is a matter of fact that science actually does shelve ideas, such
    as phlogiston, and that it does take some theories as unproblematically
    reliable, particularly those that are needed to make and interpret observations.
    And since we live in a fallibilist world, we must sometimes re-introduce old ideas
    and abandon previously trusted theories, but we must not do it too often, or take
    things like anti-Einstein or anti-Darwin crankery too seriously. A balance must be struck between the consequences of fallibility and requirements to make progress

  8. Alan Forrester says:

    Peter,

    Logically, there is no square one. We don’t go back to square one historically because our problems and theories change over time so we don’t have the same problems and theories as Newton did. You have not explained why we should think that Popper’s epistemology implies otherwise.

    An idea can only be criticised by another idea that is somehow more probable or better corroborated. Leaving everything equally open to criticism leads to a hopeless impasse. Criticism requires this asymmetry as much as verification of foundationalism.

    Whenever you have two ideas that clash, one of them must be wrong. If you can’t think of any criticisms of one of those ideas, say, the result of an experiment, then you start trying to solve this problem by criticising your other theories. But you needn’t say that the reason you’re not criticising the experimental results is that they are somehow magically better than other theories, it just happens that you have no criticism. This might be due to the experimental result being true, or it might be due to a lack of imagination on your part. Why is this position is unsatisfactory?

    What about anti-Darwin and anti-Einstein cranks? Some scholars and scientists read their work and take the trouble to refute their ideas. Fortunately, it isn’t difficult to refute the work of cranks because the cranks don’t hold their work to a high standard. So the amount of time needed to reply to cranks is not a serious threat to scientific productivity. And phlogiston? As far as I am aware nobody has yet found a problem for which phlogiston is the solution. It is not a good theory and criticising it doesn’t take long on the occasions when they pay attention to it. Nor is there any particular reason to tell physics students about phlogiston, so the subject usually doesn’t come up. So, again, it is not a threat to science. So I don’t see why considering and dismissing easily criticisable theories should be such a big deal.

  9. Peter D Jones says:

    Again, I am not claiming we DO go back to square one. I am claiming it WOULD BE an unwelcome consequence of never discarding (even FAPP) theories. But theories are discarded. The flat earth theory, epicycles and phlogiston have been discarded.

    “one of them must be false”–but which one?
    Popperians cannot regard uncriticised theories as true by default, because then
    they would have to regard strange claims like solipsism as true, and regard multiple contradictory theories as true. The Popperian stance is that theories must survive criticism to be taken seriously. Of course they
    do and must treat uncriticised data as true by default. There is and must be an
    asymmetry between the status of data and theories, they cannot all be treated
    equally as conjecture. Data just doesn’t work like theories. We can’t say that
    all data must survive criticism, because the same would apply to whatever data we are using to criticise it. To say that some scientists cheated or made a mistake in obtaining their results, we obviously need evidence.

    It is extremely difficult to refute the work of cranks to THEIR satisfaction because they don’t see things like “your theory makes no quantitative predictions” as any kind of problem

    That physics (or chemistry) students are never told about phlogiston is further evidence that it has been discarded FAPP, something that Popperians keep
    telling me does not and should not happpen

  10. Rafe says:

    “The Popperian stance is that theories must survive criticism to be taken seriously.”

    That is not the CR or the Popperian stance.

  11. Alan Forrester says:

    You claim that Popperians would have to regard solipsism as true. There are criticisms of solipsism explained in “The Fabric of Reality” by David Deutsch, namely that solipsists are just relabelling the rest of the world as part of their mind and so their theory complicates their worldview without explaining anything and is worse than realism. So solipsism has been successfully criticised and is now categorised as false.

    We can’t say that all data must survive criticism, because the same would apply to whatever data we are using to criticise it.

    We propose conjectures about the way the world works and conjectures about what has happened in a particular place at a particular time. Conjectures about the way the world works can be criticised by comparing them to other conjectures, including other conjectures about how the world works, as in the comparison of solipsism and realism. Our conjectures about how the world works can also be compared to the results of experiments. If there is a clash, then one of the theories must be false. Experimentalists usually take a lot of trouble to criticise possible flaws in their experiments before they run the experiment. So they will know about some of the problems that could arise and if one of those problems comes up they will redo the experiment, altering it to take account of the problem. What are the experimenters using to criticise their experiment? Other theories that have stood up to criticism, or are at least independently testable. For example, they might criticise what they see through a telescope by noticing that the image looks like the image a star that is distorted and conjecture that this is due to the lens sagging. They would then run simulations of what would happen if the lens sagged and if the results seem to explain the image the experimenters may decide that they should redesign the lens or modify it. There is no guaranteed true, rock solid bedrock of experimental results. Nor does there need to be. What we need is ways to do quality control, and that need is served just as well by using independently testable theories.

    I have explained why physics and chemistry students are not usually told much about phlogiston theory: it has been successfully criticised. This does not seem contrary to Popper and you have not explained why it is.

    As for cranks, when you criticise a person’s position and they refuse to address the criticisms you have proposed and you don’t know what other criticisms you could make that would produce an interesting discussion, you should go work on another problem. This a reasonable thing to do because you are not aware of any criticisms of your current position from the crank despite attempts to find such a criticism. Again, this position is consistent with Popper.

  12. Peter D Jones says:

    Alan

    You claim that Popperians would have to regard solipsism as true.

    I have been arguing that there are two problems with regarding theories
    and data as not merely both conjectural and falsifiable but as otherwise
    on a completely equal footing. One is that if you level up so that they are both regarded as true-by-default, you are going to have to face a morass of contradictory conjectures. (A random selection of theories is generally contradictory and inconsistent: a random collection of data is generally not).
    If you go the other way, and regard both as false-by-default, you will have
    nothing to refute a conjecture with. To say of some unproven and untested conjecture A that it contradricts some other unproven and untested conjecture B
    by no means refutes it: maybe we should keep B and discard A.

    Data and theories are different things need to be treated differently.

    So solipsism has been successfully criticised and is now categorised as false.

    Really? But when I say things should be categorised as false or true , I am told
    I am an Authoritarian Justificationist. But I think that categorising things as
    true or false For All Practical Purposes is compatible with fallibilism,
    since fallibilism only say that nothing is absolutely true or false in principle

    about the way the world works can be criticised by comparing them to other conjectures,

    ..not if all conjectures are equal

    If there is a clash, then one of the theories must be false.

    If there is a situation of complete symmetry, we won’t be able to see which one is false.

    What are the experimenters using to criticise their experiment? Other theories that have stood up to criticism, or are at least independently testable.

    …or other data.

    Your claims work fine so long as the ball has somehow got rolling,
    but the ball can never get rolling if all data and all theories have
    the same initial probability assigned to them, (and taking the asymmetry
    from prevailing non-Popperian science is a bit of a cheat).

    This does not seem contrary to Popper and you have not explained why it is.

    If you want to say that all is conjecture, but not all conjectures are equal

    As for cranks, when you criticise a person’s position and they refuse to address the criticisms you have proposed and you don’t know what other criticisms you could make that would produce an interesting discussion, you should go work on another problem. This a reasonable thing to do because you are not aware of any criticisms of your current position from the crank despite attempts to find such a criticism. Again, this position is consistent with Popper.

    It is consistent with the claim that fallibilism allows you to shelve problems FAPP.
    However, if everyone agreed with that, this discussion would have finished some
    time ago.

Leave a Reply

Your email address will not be published. Required fields are marked *

please answer (required): * Time limit is exhausted. Please reload the CAPTCHA.