In bartley, critical rationalism, empiricism, fideism, quine, van fraassen on 12/08/2011 at 11:18 am
Empiricism in the philosophy of science, even in its modern variants, takes observation reports of phenomenal states as a reliable route to knowledge. The problem for empiricists is as follows: Why adopt the theory (T1) that observation reports are a reliable way to sort out true from false theories and not adopt the theory (T2) that theories are a reliable way to sort out true from false observation reports? If the empiricist rejects T1 and adopts T2, the empiricist rejects empiricism and becomes a Kantian; however, this does not address the problem, for why should the Kantian adopt T2 and reject T1? The epistemological pendulum swings back.

If a basic statement and theory are incoherent, then observation reports cannot inform us that theories are wrong and theories cannot inform us that observation reports are wrong. Either the theories or observation reports are wrong — or both. Neither T1 nor T2 should be adopted a priori, for they effectively annihilate one another: all we can see is an incoherence between T1 and T2.

Furthermore, even after we recognize an incoherence between an observation report and a theory, this ‘recognizing’ is relative to a given system of logic, background assumptions, language, and cognitive framework: we are even fallible in recognizing incoherence. Whatever method or route we follow that leads to preferring one over the other will either not rest on an Archimedean point, or will not be known to rest on an Archimedean point. All is theory-laden and subject to error. We must admit that it is possible to change the logic we employ, statements we adopt, theories we accept, methods we follow, language we use, or the cognitive frameworks we inhabit, for none of them are privileged.


  1. It’s important to have a good explanation of why the observation contradicts the theory. For example, Einstein’s theories not only made better predictions than Newton’s, they could also be used to explain why Newton’s theories made some wrong predictions. Before that explanation, the wrong predictions were just anomalies, but afterward they became falsifiers. In other words, the shift from T2 to T1 thinking normally requires a good explanation, in light of a rival theory, why the old theory made wrong predictions. Until then, all the wrong predictions can only be explained in the context of the old theory, even if nobody can think of a good explanation for the time being.

    • Lee,

      You’re right. Preferring one theory over another rests not in an a priori preference for theories over the set of observation statements (or vice versa), but the reconciliation of a new or modified theory with the set of observation statements by way of expanding the logical content of the theory.

      There was no way (without significant ad hoc adjustments to Newton’s theory) to reconcile Newton’s theory with our unproblematic observation statements. Therefore, even though observation reports cannot inform us that theories are wrong and theories cannot inform us that observation reports are wrong, we must decide, and this decision rests on whether or not our observation statements can withstand criticism.

      But how can we criticize our observation statements outside appealing to other observation statements, unproblematic theories, or background assumptions? Quine gave the best explication of the problem, but his solution was inadequate. Protecting the ‘hard core’ depicts theology, pseudo-science, or other hermeneutic practices, not science. Popper’s methodological solution is, I think, superior.

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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