In this note we will explain explanations—one of the most useful, if not the most useful, epistemic praxis.

Explanation is the practice of (extracting) identifying regularities in the connection between observable effects (phenomena) and the underlying ontology (states, structures, and dynamics), making these connections explicit, clear, and intelligible, thereby enabling answers to why things are the way they are and not otherwise. Another way to see it is - removing the magic from the underlying ontology → by compressing it - reducing it to the core - to what’s matter.

An explanation is a structured mapping that connects a phenomenon or entity (explanandum) to an intelligible conceptual, formal, or mechanistic structure (explanans), which renders the phenomenon understandable, predictable, or actionable. It serves the function answering why questions - and grounding the answers that we give about a given explanandum.

An explanation is a set of statements that represents these connections.

Note: By this account, explanations are not about particular (singular) cases; they concern the relationship between effects and classes of cases. Explanations require testing, and singular cases generally cannot be tested because their conditions often cannot be recreated.

When analysis focuses on singular cases, it often drifts into deterministic narratives or speculative accounts. The only reliable way to avoid this is to analyze classes of cases, where regularities can be identified and tested.

This is also the only way to strip away the apparent “magic” of the ontology: singular cases can often be grounded in many different underlying ontological accounts, making them insufficient for establishing explanatory validity.

Note: The term explanation is often applied to singular cases (e.g., a fire → smoke). However, in such instances what occurs is not the construction of a new explanation but the instantiation or application of a general explanatory relation to a particular case. The explanatory work is performed at the level of the general class of events, while the singular case merely recasts that general explanation under specific conditions.

An explanation is a rule or principle derived from observed regularities, which helps to assemble coherent accounts of individual observed events. The stronger the regularity linking a class of phenomena to its underlying substrate, the more concise and unified these accounts can be.

What remains in this note is the task of explaining how to do this and how to evaluate our explanations.

Note: The term ‘Explanation’ is often used loosely to refer both to the connection between a phenomenon and its underlying ontic structure, and to the observation that emerges from this connection. It is also used to denote the application of this understanding to particular cases. However, there is no true explanation of singular cases—only applications of the general principle.

Guiding Questions:

• What kinds of statements constitute an explanation?
• What makes an explanation good?
• What do we mean by “Why?”
• Are all forms of explanation ultimately contrastive—answering why things are this way rather than some relevant alternative?
• How to evaluate explanations - of the past? We only know one realization - not the different future that were not realized?
• Which are the limits of evolutionary explanation in the light of stochasticlity?
• What is the role of counterfactuals in explanations?

Index

Term