The debate between Evidential Decision Theory (EDT) and Causal Decision Theory (CDT) is one of the most significant schisms in modern formal philosophy. While both theories aim to guide agents toward rational choices, they diverge sharply when correlation and causation come apart—a conflict perfectly illustrated by Newcomb's Paradox.
Key Differences at a Glance
To understand how these two frameworks compare, we can examine their core tenets side-by-side:
| Feature | Evidential Decision Theory (EDT) | Causal Decision Theory (CDT) |
|---|---|---|
| Core Question | "What does my choice predict about the world?" | "What will my choice physically cause in the world?" |
| Mathematical Basis | Standard conditional probability: P(O | A) | Causal/Counterfactual probability: P(A âž” O) |
| Newcomb's Paradox Choice | One-box: Take only Box B to secure $1,000,000. | Two-box: Take both boxes to secure an extra $1,000. |
| Primary Strength | Pragmatic success; its adherents end up richer in Newcomb's setup. | Causal realism; avoids acting on non-causal correlations. |
| Primary Weakness | Vulnerable to the "Smoking Lesion" or "Medical Newcomb's" error. | Vulnerable to the "Why Aren't You Rich?" objection. |
The Core Philosophical Divide
The split between EDT and CDT represents two fundamentally different views of what it means to act rationally:
1. The Evidential View: News Value
EDT is concerned with "news value." A rational agent should want to receive good news. If choosing to one-box provides excellent news (that you are highly likely to be a millionaire), then that is the rational choice. EDT treats the agent's own decision as a diagnostic piece of data. If your actions are a reliable symptom of a great state of affairs, EDT tells you to perform those actions, regardless of how that state of affairs was brought about.
2. The Causal View: Efficacy
CDT is concerned with "efficacy." A rational agent should only care about what they can physically affect. Because the past is fixed, a rational agent must treat the state of the boxes as a constant, not a variable. CDT argues that EDT is guilty of "magical thinking"—acting as if your current choice can retroactively change what the Predictor did yesterday. To a causal theorist, it is irrational to choose a worse option (one-boxing) just because it correlates with a better state that you cannot causally influence.
The Unresolved Dilemma
Neither theory has achieved complete dominance in philosophy because both have intuitive failure modes. EDT seems to fail in the Smoking Lesion case, where it advises people to avoid benign behaviors (like drinking tea or smoking, in a non-causal genetic model) just to avoid getting "bad news" about their genes.
On the other hand, CDT seems to fail in Newcomb's Paradox by leaving its followers poorer than their evidential peers. This has led some modern philosophers to seek hybrid frameworks or "decision-theoretic pluralism" to bridge the gap between evidence and cause.
If you found this valuable, consider supporting our work.
Join PhiloCrux community.
Unlock high-density masterclasses and investigations into ideas surviving outside the algorithmic consensus. Support independent thought and get full access to our digital library.
Join Now