Coherentism is a theory that challenges the presuppositions of Foundationalism and of the Regress Problem.


      The Regress Problem presupposes that justification has a linear, inferential structure.  What does this mean:


      Justification is understood on the model of a proof in mathematics.  This presupposes that justification flows in one direction:  From premises to a conclusion.  We will call this assumption the assumption that the flow of justification is uni-directional.  (Dancy calls this asymmetric.)


      Another feature of linear justification is that it is monotonic.  It begins with accepted premises and ends with accepting the conclusion.


      The Coherentists’ radical claim:  Justification (and reasoning) are not linear.  There is no purely uni-directional, purely monotonic kind of justification—that is, there are no linear inferences.

(Dancy’s version:  Justification is symmetric.)


      How could this be?  We seem to make linear inferences all the time.  Are the Coherentists denying that we ever accept a belief on the basis of other beliefs that we accept?  No.  They are denying that the reasoning involved should be understood as uni-directional and monotonic, even when it seems to be.









      The Coherentist claims that all reasoning is coherence or equilibrium reasoning.  Whenever you are presented with a derivation of a conclusion from premises you believe, you are not rationally compelled to accept the conclusion.  It may make more sense to give up one (or more) of the premises.  Consider the following “proof”:


Let x = 1.

Square both sides:  x2 = 1

Subtract 1 from both sides:  (x2 – 1) = 0

Divide both sides by (x-1):  (x+1) = 0

Substitute for x:  1+1 = 2 = 0!


Consider also the example of Frege and Russell.





      The coherence theorist replaces linear inferential justification with holistic relations of mutual support.  (Compare Chisholm’s the role of confirmation and concurrence in Chisholm’s theory.  These are examples of coherence reasoning in his theory.)


      Recall the Regress Problem.  The regress gets started by the presupposition that justification relations among beliefs are linear—that is, uni-directional and monotonic.  To someone who thinks that justification relations among beliefs are linear, it seems that the coherentist is advocating circles of justification.  The coherentists are NOT advocating circular reasoning.  They are advocating a new holistic way of understanding reasoning.  (What I called “equilibrium reasoning” in the handout on reasoning could just as well be called “coherence reasoning”, with one qualification that I discuss in class.)

Dancy's Coherence Theory of Justification


      1.  Justification is symmetric

      Why is justification asymmetric for the Foundationalist?

      Why is the notion of a linear inference asymmetrical? 


      2.  A set of beliefs is coherent to the extent that the members are mutually explanatory and consistent.


      3.  The full account:  If a's belief set is more coherent with the belief that p as a member than without it or with any alternative, a is (or would be) justified in believing that p.

      Why is this account of justification holistic?





Dancy's Coherence Theory of Truth:


      A proposition is true ó It is a member of a coherent set. 

      The plurality objection. Fumerton argues that this makes truth relative (217).


      Why not combine the Coherence Theory of Justification with the Correspondence Theory of Truth?














Fumerton’s Objections to a

Coherence Theory of Justification



      1.  Fumerton argues that on a coherence theory of justification is arbitrary.  Why?


      2. Fumerton also argues that we must distinguish between the coherence of a set of beliefs and the subject’s apprehension of that coherence.  What is this problem?


      3.  Fumerton argues that justification does not require logical consistency:  the Lottery Paradox.


      But the biggest problem of all is this: 






III.  The Input Problem for Coherence Theories of Empirical Justification


      According to the coherentist, empirical justification is solely a product of relations of mutual support among beliefs.  So for the coherentist, empirical justification does not require any connection to the way the world actually is.  But there are many different coherent sets of propositions that are incompatible with each other.  Coherentism provides no rational basis for preferring any of them over any of the others.  But it is hard to see how empirical justification could be this disconnected from the real world.  For empirical beliefs to be justified, it seems necessary that there be some appropriate process by which the world can affect those beliefs.  This would be an additional constraint on empirical justification, in addition to any coherence considerations.




III.  Dancy's Reply:  Antecedent and subsequent security

      Pure coherentism vs. weak coherentism

(Is weak coherentism a form of foundationalism?)

      Genetic asymmetry vs. justificatory asymmetry

      Justificatory symmetry:  Equal antecedent security for all beliefs

      What is Dancy's final response to the Input Problem?

      The coherentist can be an empiricist.  Is this correct?


      What would Dancy say about Fumerton's example of back pain?






IV.  Another Problem:  The Problem of Too Much Coherence:  Berkeley’s Theory of Perception or the Paranoid Schizophrenic


      Perhaps the coherence theory is not a complete theory of justification.  Perhaps it is only a theory of reasoning.  The revised claim would be that all reasoning fits an equilibrium model.


      If this is true, then, reasoning is not linear.  Strictly speaking, there are no linear inferences.


      Linear inferences are uni-directional. Reasoning is multi-directional.  Linear inferences are monotonic.  Reasoning is often non-monotonic.


      This suggests that a complete account of justification, would combine a coherence theory of reasoning with a solution to the Input Problem.