You do not have Javascript enabled. Some elements of this website may not work correctly.

In standard decision theory, different ways that the world could be are called states. States are made up of all of the facts that are true in that state, including all empirical facts and all facts about how other agents will act.

The actions available to a person are the various different things that they can do in the world. For example, suppose that a person can buy a painting for $100 and then sell it, or they can keep their $100. The agent has two options: to buy the painting or not.

Standard decision theory assumes that people have accurate beliefs about how likely each possible state of the world is. For example, suppose the person above has a credence of 0.6 that they are in a state of the world in which the painting in question will increase in value by $20, and a credence of 0.4 that they are in a state of the world in which it will decrease in value by $50.

The two actions that this person can take will have different consequences, depending on what state they are actually in. The different states that could result from people’s actions are called outcomes, and people assign values to all outcomes.

For example, the acts, states, outcomes, and outcome values of the decision described above can be represented as follows:

states, acts, outcomesstate 1 (60% probability): painting appreciatesstate 2 (40% probability): painting depreciates
act 1: buy and then sell paintingoutcome[a1,s1]: $20outcome[a1,s2]: - $50
act 2: don’t buy paintingoutcome[a2,s1]: $0outcome[a2,s2]: $0

The available actions can then be ranked based on the value of each of the outcomes that they produce, weighted by how likely they think it is that they are in the state of the world that will produce that outcome. For example, the expected value of buying the painting and selling it is - $8 (0.6 x $20 + 0.4 x -$50), while the expected value of not buying it is $0. So it is better for them not to buy this painting, according to expected value theory.

Idealized ethical decision-making has the same basic acts, states, and outcomes structure as standard decision theory, but it is concerned with the ethical value of outcomes, such as the number of lives improved, rather than personal gains and losses.

Further reading

Joyce, James M. 1999. Causal decision theory. Cambridge: Cambridge University Press.

Steele, Katie & H. Orri Stefánsson. 2015. Decision theory. In Edward Zalta (ed.), Stanford encyclopedia of philosophy.