On the one hand, scientific realism posits that scientific theories aim to describe reality, and that progress in science is marked by our theories providing a more and more accurate picture of what the world is like. The problem-solving approach, on the other hand, views scientific progress as the increasing ability of theories to effectively solve more problems.

These views are commonly supposed to be incompatible because whether a hypothesis solves a problem to our satisfaction does not seem to depend on whether it is, in fact, true.

Are there any viable attempts at reconciliation?


  • See Theories of Scientific Progress. But how to measure the amount of "effective solutions" to problem? With some sort of approximation to what? Commented Dec 12, 2023 at 12:50
  • 1
    @Turtur Could you please elaborate a bit why you consider both views as being incompatible? Why are they not complementary?
    – Jo Wehler
    Commented Dec 12, 2023 at 15:14
  • 1
    @Turtur however I'm not seeing the friction the other way -- if in fact scientists are making more accurate descriptions of the world that would seem to necessarily mean better ability to solve problems.
    – Dave
    Commented Dec 13, 2023 at 14:39
  • 1
    @Turtur I've never seen science described or practiced in a way where a problem solved "to our satisfaction" isnt accurate. The whole idea of problem solving is to get answers closer to "true" (as much as we can even determine what is true). No one is satisfied with a scientific problem solving process that gives incorrect solutions. Its trivial to come up with ways to "solve" problems if you dont care about truth.
    – JMac
    Commented Dec 15, 2023 at 14:46
  • 1
    because whether a hypothesis solves a problem to our satisfaction would not seem to depend on whether it is, in fact, true. -- to the contrary, it seems intuitive to me that if two people have very different ideas about how to solve a problem, the person who's ideas more closely match reality will more often (but not always) be a better candidate for solving the problem as well. The practical considerations don't always perfectly align with ontological truth, but it would seem to me that they align well enough.
    – TKoL
    Commented Dec 18, 2023 at 11:02

4 Answers 4


Following the suggestion of @MauroALLEGRANZA I read the explanation from Realism and Instrumentalism as follows:

  1. The distinction is between realism and instrumentalism. Realism can be glossed as the attempt to discover how the world “really” is, i.e. to discover its ontology. While instrumentalism is already satisfied with a theory which successfully predicts the result of observations. Instrumentalists do not make any ontological claims, while realists specially claim that their findings have an ontological meaning.

    These two positions are not contradictory. In both cases, first the scientists develop and test scientific theories in order to predict - or to explain in retrospect - certain observations from the domain under investigation.

    The difference between realism and intrumentalism becomes obvious when one considers the interpretation of a successful theory. IMO the realist claims too much while the instrumentalist stops too early.

  2. Example: What is an electron?

    To the best of our knowledge an electron is an elementary particle which satisfies the Dirac theory - more precisely the laws of quantum electrodynamics.

    We cannot say to a greater extent what an electron is, because we do not have direct access to electrons. Electrons do not belong to the mesocosmos, we do not know them from our daily contact and handling.

    There are no other concepts to describe electrons than the scientific concepts of the Dirac theory. But even concerning the fundamental physical property of an electron, its spin: I have never seen the spin as part of a philosophical ontology - at best it is described in a metaphorical way. I do not expect that such a philosophical ontology is possible and I do not see if it would be of any use.

    On the other side, mature science is not satisfied if there are rival theories which successfully predict the same observations, but are incompatible with each other. In this case scientists strive to prove the equivalence of both on a theoretical level. That’s what happened with Heisenberg’s matrix mechanics and Schroedinger's wave mechanics. The equivalence of both was proved by Dirac.


I am afraid that your question describes a false dichotomy. You are correct that science attempts to discover reality or truth. After all, it grew from philosophy. How could a false hypothesis, that is one that has been falsified, ever be useful? Truth and usefulness are compatible. Reconciliation is unnecessary. I would recommend Karl Popper. He explained that a hypothesis should be falsifiable, not verifiable. Science is useful, but only when it is true.

  • I didn't downvote, but a falsified hypothesis is useful, as Popper spells out in Conjectures and Refutations, because it is the only way to be deductively certain about hypotheses at all. On his reading, neither confirmation nor verification as proposed by the LPs holds any water given questions of relevancy and defeasibility.
    – J D
    Commented Jan 22 at 15:27

The core problem that you are asking about, is that we have no ability to establish logical "truth", only model utility. Karl Popper, who presumed that improving model utility meant that our models were increasingly closer to "truth", attempted to provide a proof of this. He used the term Verisimilitude to characterize how close our models are to truth, and developed a proof that increasing modeling success was also associated with increased verisimilitude. Popper's failure to demonstrate the validity of verisimilitude has been followed by a series of failures by other authors, using three different strategies. See https://plato.stanford.edu/entries/truthlikeness/

The SEP author of the Truthlikeness entry remains optimistic in the face of these repeated failures. An alternative conclusion is that logical "truth" is an impossible standard, and utility is all we can ever achieve: therefore our truth standard should shift from the logical absolute "truth" to the pragmatic truth standard of utility == truth.


These two aspects you are talking about are the two side of the same coin. Since everyone is focused on either one side and not to the coin the reconciliation cannot be made. Theories have nowdays become abstract mathematical representational aspects of measurements. Reality has been decomposed into static concepts that acquire a meaning inside power-structured academic frameworks agreed upon by consensus.

  • In this context, the "problem-solving approach" seems close to what is going on; the problem though is that the problems that are to be solved are not solved, but are expressed in scientifically canonical procedural symbolisms.

  • On the other hand "scientific realism" has - from an illusion - become a disorted view of reality that is leading us in a continuous (in)accurate picture of what the world is like; we are heading away and away from the true meaning of things.

Don't get obscured by scientific achievements, look inside you, and around you in our world to see the despair in which this polarization has led us.

However, in recent years I see some attempts to bring these two views together. These attepts can be divided into two groups. The first is attempts to give a meaning to already established scientific knowledge by focusing on a problem-solving approach, outside of mainstream conceptual frameworks, by thinking out-of-the-box. The second is attempts to provide an ontological substance to existing theories, in order to give them a meaning closer to reality.

I have come accross many papers and in different scientific fields, from physics and biology to psychology and philosophy in general that are compatible with this aim. The problem though is that these propositions seem not to be taken into account, because science is now focused only in measurable accounts and because everything that is not conforming with mainstream views is discarded. I cannot avoid stating that science has become in a large extent a totalitarian dependant entity.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .