# Hierarchy of Theories There is an oft held assumption that theories can be classified into a hierarchy, in order of decreasing intrinsic reliability:  Many people take the existence of this hierarchy for granted, despite the fact that these judgements of comparative reliability depend entirely on philosophical arguments, arguments that classify themselves as quite unreliable! The idea of this hierarchy is a cousin of the [reductionist mistake](Reductionism.md). The same assumption occurs in [inductivism](Induction.md), which supposes that we can be absolutely certain of the conclusions of mathematical arguments because they are deductive, reasonably sure of scientific arguments because they are ‘inductive’, and forever undecided about philosophical arguments, which it sees as little more than matters of taste. But none of that is true. Explanations are not justified by the means by which they were derived; [they Are Justified By Their Superior Ability to Solve Problems They Address](Explanations%20Are%20Justified%20By%20Their%20Superior%20Ability%20to%20Solve%20Problems%20They%20Address.md). ## Self-Reference Is the hierarchy of theories an example of [Bad Self-Reference](Self-Referential.md#Bad%20Self-Reference)? Here we have a system (the hierarchy) that talks about itself (philosophy, part of the system, talks about the system - it determines the rank ordering of the theories).  The problem is that the system, as it is constructed, is [Incoherent](Incoherent.md) and [self-contradictory](Contradiction.md). The system uses philosophy to justify the reliability of math and science, yet it simultaneously claims that philosophy is the *least* reliable method of reasoning. It also exhibits a [Circular](Circular%20Reasoning.md) justification: the system attempts to justify itself using its own standards, creating a *closed loop* of justification. You may rightly ask: is the self-reference really the problem here? For instance, say we move philosophy to the *top* of the hierarchy. In that case the contradiction would be gone, but self-reference would remain. Would we still have a problem? Yes, in fact we would. The problem is that our system is trying to [Justify](Justification.md) itself through its own means! This is a [Circular justification](Circular%20Reasoning.md). Any system should be *open to evaluation* based on criteria that are *external* to the system itself. Otherwise, our system may be *closed* and avoid criticism and thus [Error Correction](Error%20Correction.md). --- Date: 20241025 Links to: [Fabric of Reality](Fabric%20of%20Reality.md) pg 84 Tags: References: * []()