Underdetermination

Evidence may support multiple theories

Underdetermination of theory by data (UTD) suggests that evidence available at any given time may not be sufficient to determine which scientific theory to believe in. This occurs when the available evidence is insufficient to uniquely identify a single theory, allowing for multiple theories to be consistent with the same evidence.

Example

If all that was known was that exactly $10 were spent on apples and oranges, with apples costing $1 and oranges $2, one could eliminate some possibilities (e.g., 6 oranges could not have been purchased), but not know the specific combination of apples and oranges purchased. This illustrates how evidence can support multiple theories (e.g., different combinations of apples and oranges).

Understanding underdetermination is crucial for recognizing the limitations of empirical evidence in theory choice and the potential for multiple valid interpretations of the same data.

Related concepts

Falsifiability

Popper introduced falsifiability as a criterion for scientific theories

Instrumentalism

Instrumentalism views scientific theories as useful tools, not as descriptions of unobservable reality

Karl Popper

Lakatos's research programmes improve over Popper's by protecting hard cores with auxiliary hypotheses

Replication crisis

Replication crisis undermines scientific credibility

Michel Foucault

Foucault's theories link power, knowledge, and social control

Logical positivism

Logical positivism's verification principle claims only empirically verifiable statements are meaningful

One email a day: 5 concepts + the 5 stories that matter →

Swipe through 100 ML concepts daily

Open TickerNews