# Theory-ladenness

> Observations depend on prior ideas about what to notice, how to measure it, and what the result means.

- HTML version: https://robbiepalmer.me/ideas/theory-ladenness
- Source: https://plato.stanford.edu/entries/science-theory-observation/

Theory-ladenness is the claim that observation depends on prior theory. A
measurement does not arrive with its own interpretation. Someone has chosen
what to observe, built or selected an instrument, decided which conditions
matter, and learned how to read the result.

This does not make observation useless or arbitrary. Background theory is what
lets an observation carry information at all. A number on a thermometer becomes
evidence about temperature only because we rely on claims about the instrument,
its calibration, its contact with the object, and the conditions under which it
works.

## Where theory enters

Theory can shape observation at several points:

* **Attention.** Expectations affect which features an observer notices and
  which records they keep.
* **Production.** An instrument turns a physical event into a mark, image, or
  number by relying on assumptions about its mechanism.
* **Classification.** A result counts as an instance of a category only within
  some scheme for recognising that category.
* **Use as evidence.** Connecting a result to a claim requires further
  assumptions about relevance, reliability, and possible causes.

Suppose two gardeners inspect a yellowing plant. One records a nutrient
deficiency and the other records heat stress. They have seen the same leaves,
but each selects and organises details through a different causal account. A
soil test and a weather record may help them decide between those accounts.
Those further observations also depend on instruments, categories, and
assumptions. Inquiry improves the chain; it does not step outside every chain.

## Theory-laden is not theory-determined

An observer's expectations do not dictate every result. Instruments surprise
their designers, repeated observations expose mistakes, and people who favour
different explanations can agree that a reading occurred. Reality can resist
the account used to investigate it.

The practical question is therefore not whether evidence is perfectly neutral.
It is whether the assumptions that produced and interpreted the evidence are
good enough for the claim being made. Independent instruments, calibration,
blind assessment, shared records, and criticism from people with different
expectations can reveal dependencies that one observer misses.

This is an [epistemological](/ideas/epistemology) point. Confidence should
depend on how an observation was made and what assumptions connect it to a
conclusion, not on a fiction that data speaks without interpretation.

## Datasets preserve prior judgements

A table or archive can hide theory-ladenness because it presents settled rows
instead of the decisions that created them. Before an entry exists, someone has
defined the unit of observation, drawn category boundaries, chosen what to
exclude, handled ambiguous cases, and decided how much error is acceptable.

Consider records kept by two coastal settlements. In a village of fragile homes
on an exposed beach, strong wind may flood rooms, stop fishing, and close the
only road. Residents call it a storm because it interrupts daily life. The same
wind reaching a town of modern homes farther inland may be logged as bad
weather. If a later authority replaces those lived classifications with a wind
speed threshold, the old and new counts track different things. Comparing
periods requires the classification rules as well as the figures.

[Postpositivism](/ideas/postpositivism) takes this dependence seriously while
retaining empirical inquiry. The response is to expose assumptions, test them,
and make revision possible. Theory-ladenness is a reason for better records and
stronger criticism, not a reason to give up on evidence.

## Sources

* Boyd, Nora Mills, and James Bogen. ["Theory and Observation in
  Science"](https://plato.stanford.edu/archives/fall2024/entries/science-theory-observation/).
  *The Stanford Encyclopedia of Philosophy*, 2024.
* Hanson, Norwood Russell. *Patterns of Discovery*. Cambridge University
  Press, 1958.
* Longino, Helen E. *Science as Social Knowledge*. Princeton University Press,
  1990\.

## Related ideas

- [Epistemology](https://robbiepalmer.me/ideas/epistemology.md): How we know what we claim to know, what justifies a belief, and when new evidence should change it.
- [Falsifiability](https://robbiepalmer.me/ideas/falsifiability.md): A claim is falsifiable when some possible observation could count against it, though real tests examine a bundle of assumptions.
- [Incommensurability](https://robbiepalmer.me/ideas/incommensurability.md): Rival frameworks may lack a complete common measure because their concepts, problems, or standards differ.
- [Paradigm shift](https://robbiepalmer.me/ideas/paradigm-shift.md): A scientific community can replace the shared examples, standards, and concepts that organise its normal work.
- [Postpositivism](https://robbiepalmer.me/ideas/postpositivism.md): Empirical inquiry can pursue objectivity while accepting fallible evidence, theory-laden observation, and the influence of values.

## Where it appears

- Blog post: [The Philosophy of Data Science](https://robbiepalmer.me/blog/2022-03-02-the-philosophy-of-data-science.md)

---

Markdown index of this site: https://robbiepalmer.me/llms.txt