Do not confuse this with proof: verification connects a statement to the result of a test, whereas proof establishes the necessity of a conclusion from accepted premises.
Initial Definition
Verification is the testing of a statement by means of a predefined procedure and determining whether the result obtained meets an accepted criterion.
It has a conditional form:
If statement A is tested using procedure V under conditions C, the test result either meets or fails to meet criterion K.
Positive verification does not mean the unconditional truth of the statement. Rather, it means that:
the statement has passed this particular testing procedure under these particular conditions and according to these particular criteria.
What Makes Verification Possible
Verification does not begin with direct contact between a statement and reality. The following must already have been defined:
- the statement being tested;
- the object of testing;
- the testing procedure;
- the conditions of execution;
- instruments of observation or measurement;
- the criterion for a positive and negative result;
- permissible error;
- the method of interpreting the result;
- the criterion for the sufficiency of the test.
Remove these conditions, and no pure or neutral verification remains. The very possibility of determining what was tested, how it was tested, and with what result disappears.
What Verification Establishes
Verification establishes correspondence between three elements:
- the statement being tested;
- the result of the procedure performed;
- the criterion according to which the result is recognized as confirming or refuting the statement.
It does not connect a statement to reality without intermediaries. Between them stand the procedure, instrument, conditions, and interpretation.
Therefore, the precise formula is not:
The statement was verified by reality.
But:
The result obtained through procedure V under conditions C satisfies criterion K adopted for testing statement A.
Verification and Proof
Proof and verification are different operations.
Proof establishes the necessity of a conclusion from accepted premises:
If A and the rules of inference R are accepted, then B necessarily follows from A.
Verification establishes that a statement has passed a selected test:
If A is tested by procedure V, the resulting outcome meets the accepted criterion.
| Proof | Verification |
|---|---|
| Establishes the necessity of an inferential transition | Establishes that a test has been passed |
| Connects premises to a conclusion | Connects a statement to the result of a procedure |
| Relies on axioms and rules of inference | Relies on a procedure and a testing criterion |
| The conclusion is necessary within the accepted system | The result is valid within the limits of the test performed |
| May not require empirical observation | Often requires observation, measurement, or comparison |
| Answers: “Does the conclusion follow?” | Answers: “Is the statement confirmed by this test?” |
A verification result may become a premise in a proof. Proof, in turn, may justify the design of a procedure or the transition from a measurement result to a conclusion. But neither operation replaces the other.
Proof connects a premise to a conclusion. Verification connects a statement to the result of a test.
Verification Is Not Neutral
Verification always already takes the side of a chosen testing construction. Before a result is obtained, it has been determined:
- what counts as an object;
- which property is recognized as essential;
- what will be measured or observed;
- which conditions are considered controlled;
- which instrument is permissible;
- which result will be recognized as significant;
- which margin of error is accepted;
- how the result will be translated into a conclusion about the statement.
This does not make verification false or useless. It means verification is not a view from nowhere.
Verification establishes correspondence not between a statement and unmediated reality, but between a statement and the result of a particular procedure for distinguishing reality.
Verification and Repeatability
A single successful test shows only that the statement withstood one test under particular conditions.
Repetition makes it possible to establish reproducibility of the result, but this, too, depends on what counts as a repetition of the same conditions, the same procedure, and the same result.
Thus, repeatability strengthens confirmation of the statement but does not turn a conditional procedure into unconditional access to reality.
Can Verification Be Final?
A positive result from a finite number of tests does not mean that a statement will withstand all possible tests under all possible conditions.
This is especially true of general statements. Testing individual cases may confirm those cases, but it does not in itself exhaust the set of cases that have not yet been tested.
Thus, the expression:
the statement has been definitively verified,
often conceals the boundaries of the procedures performed.
It is more precise to say:
the statement has been confirmed by a specified procedure under specified conditions with specified accuracy.
Finality can exist only relative to a previously delimited set of tests and criteria. It does not automatically extend to every possible way of examining the statement.
Verification, Confirmation, Testing, and Falsification
These concepts should for now be distinguished provisionally:
- testing: performing a procedure without specifying the result in advance;
- verification: establishing whether the result meets the criterion for the statement being tested;
- confirmation: a result that strengthens the grounds for accepting the statement;
- falsification: a result incompatible with the statement under accepted testing rules.
Verification may end in confirmation or non-confirmation. Therefore, it should not be identified in advance with a positive test result.
The definitive boundaries between these four concepts require separate investigation.
Central Formulas
Verification does not prove a statement with logical necessity. It establishes that the statement withstood a particular test.
The result of verification belongs not only to the statement but also to the procedure, conditions, instrument, and criterion of interpretation.
The more precisely the conditions of testing are specified, the less verification passes off a local result as unconditional truth.
Analyses That Are Still Missing
The following questions have not yet received independent examination.
1. The Bearer of Verified Status
It has not been established what the property of “being verified” belongs to: the statement, an individual result, the relationship between them, or the entire testing construction.
2. The Procedure’s Access to Reality
It is necessary to examine how the result of a procedure permits conclusions about an object beyond the procedure itself, and where the risk arises of substituting the measurement model for the reality being measured.
3. Observation and Interpretation
It has not been determined whether a result exists before interpretation or becomes a result only after a language, scale, and criterion of distinction are applied.
4. A Negative Result
It is necessary to distinguish refutation of a statement, instrument malfunction, procedural error, violation of conditions, and incorrect interpretation of the data obtained.
5. Degrees of Confirmation
It has not been examined whether verification is a binary operation or allows degrees, probabilities, and changes in confidence following a series of tests.
6. Independence of Repeated Testing
It is necessary to determine when a repeated test is genuinely independent and when it merely reproduces the same initial assumptions, instruments, and systematic errors.
7. Boundaries of Applicability
It is necessary to establish how testing conditions limit the transfer of a result to other objects, times, scales, and environments.
8. Self-Verification of the Procedure
A procedure cannot be completely tested using only itself without risking circularity. The role of calibration, control procedures, and the metalevel, along with the grounds for trusting them, must be examined.
9. Different Kinds of Verification
Scientific, technical, documentary, software, legal, and everyday verification have not yet been compared. It is unclear whether they share a common structure or merely share a name.
10. Relationship to Falsification
It remains to be determined separately whether confirmation and refutation are symmetrical, whether every statement is in principle testable, and what the untestability of a statement means.
