Observation and Theory‐ladenness
In the philosophy of science, observations are said to be “theory‐laden” when they are affected by the
theoretical presuppositions held by the investigator. The thesis of theory‐ladenness is most strongly
associated with the late 1950s and early 1960s work of N.R. Hanson, T.S. Kuhn, and P. Feyerabend, and was
probably first put forth (at least implicitly) by P. Duhem about 50 years earlier. Although often run together,
at least two forms of theory‐ladenness should be kept separate: (i) the meaning of observational terms is
partially determined by theoretical presuppositions; (ii) the theories held by the investigator, at a very basic
cognitive level, impinge on the perceptions of the investigator. The former may be referred to as semantic
and the latter as perceptual theory‐ladenness. The thesis of theory‐ladenness, if true, has troublesome
consequences for theory‐testing. If there are no theory‐neutral observations, then this raises doubts about
whether empirical tests can truly decide between competing theories. So, if theories partially determine
the meaning of observation terms, two investigators holding incompatible theories will mean different
things when they use the same observational vocabulary, and, if theories partially determine ‘what we see’,
two investigators holding incompatible theories will see the objects relevant for discriminating between
their theories differently.
A thesis that also goes under the heading of theory‐ladenness may be (more appropriately) referred to theory‐dependence of instruments, on which much discussion has focused on: the investigator’s
confidence in the truthfulness of the results obtained with certain instruments depends on her having
sound theories of how these instruments work. Such theories are also referred to as ‘background’ theories.
The theory‐dependence of instruments is particularly problematic when the background theories are the
very theories that the investigator seeks to test, for in those scenarios the testing procedure is rendered
circular.