The statement that defines derivative classification is the one describing the act of incorporating, paraphrasing, restating, or generating classified information from an existing classified source into a new or updated document, and marking that new material consistently with the source's classification markings. In practice, this means the definer must name both the act (deriving from a source) and the obligation (carrying forward the source's classification level, category, and handling caveats). A statement that omits either half is incomplete.
That definition matters because derivative classification is the most common path by which classified material enters the federal system. Most classified documents are not originally classified by an original classification authority. They are derived from sources that already carry markings. Understanding the definition therefore means understanding a chain: source marking, derivative incorporation, review, and declassification.
The mechanism: how a derivative decision actually happens
Original classification creates classified information from scratch. Derivative classification does not create new secrets. It propagates existing ones. The person performing derivative classification is not deciding whether the information is sensitive enough to classify. They are deciding that the information already is classified somewhere, and that it is now appearing in a new place.
The mechanism runs in three steps. First, the classifier identifies the source document or portion and reads its markings. Second, the classifier determines that the information being used is the same information, not a new judgment about the same topic. Third, the classifier applies the source's level, category, and dissemination controls to the new portion.
The subtle part is step two. Paraphrasing does not launder classification. If you restate a classified fact in your own words, the restatement is still derivative. If you combine several unclassified observations into a conclusion that reveals a classified fact, that synthesis is derivative too. The test is whether the new text communicates information already classified in the source, not whether the words match.
A second subtlety involves the "classified by" and "derived from" lines. A properly marked derivative document cites the source and its date, rather than the original classifier's name. This preserves the audit trail. If the source is later declassified or upgraded, reviewers can trace every document that inherited from it.
Quantified context: how large the derivative chain is
Derivative classification dominates the classified system by volume. Every guidance document that tells employees "mark derivatives consistently with the source" presupposes that most markings are inherited, not original. Training programs treat derivative classification as a distinct competency because the failure mode is mechanical: people copy markings incorrectly or drop them.
The practical stakes are measurable in review time. A single improperly marked derivative forces a reviewer to stop, trace the source, and correct the portion. In organizations that process thousands of documents, a small error rate produces a large correction backlog. That is why marking discipline is audited separately from classification judgment.
The timing pattern matters more than any single count. Derivatives are marked at creation, reviewed before release, and reconsidered whenever the source changes. A derivative that inherits a marking today can become over-classified tomorrow if the source is declassified. The chain must therefore remain visible throughout the document's life.
Plain language: derivative classification is inheritance, and inheritance failures compound.
Edge cases: where the definition gets tested
Several scenarios push the definition to its limits, and each clarifies what derivative classification actually requires.
Compilation is the classic edge case. A document that assembles many unclassified items can become classified if the assembly reveals a classified relationship. The individual facts are not classified; the pattern is. The derivative classifier must recognize that the compiled whole communicates something the parts do not.
Upgrades and downgrades are another edge case. If a source is upgraded, existing derivatives do not automatically change level overnight. They are corrected through a review process. The definition covers the moment of derivation, but the obligation to keep markings consistent continues afterward.
Foreign government information and controlled unclassified information add category complexity. The level may be the same, but the category and dissemination controls differ. A derivative statement that captures only "classified" but ignores handling caveats is functionally incomplete.
The professional caveat worth flagging: copying markings from a document whose own markings are wrong propagates the error. Derivative classification assumes the source is correctly marked. When it is not, the derivative classifier inherits a defect.
Red-flag warnings to watch for:
A derivative document with no "derived from" citation is untraceable and should be treated as improperly marked. Markings that are stricter than the source suggest over-classification, not caution. Markings that are looser than the source are the most serious failure because they leak. A paraphrase with no marking is not automatically unclassified; check the underlying fact. Compilations assembled from open sources can still be classified if the pattern reveals protected information. Marking copied from a draft source may not survive the source's final review.
Misconceptions: what derivative classification is not
The most common misconception is that derivative classification requires authority. It does not. Original classification requires a designated authority; derivative classification is an obligation that applies to anyone who incorporates classified source material into a new document. The authority question belongs to the source, not the derivative.
A second misconception treats paraphrasing as declassification. Rewording a classified fact produces classified text. The information, not the phrasing, carries the classification. Employees who rewrite to "soften" the language are still deriving.
A third misconception confuses derivative classification with declassification review. They are different processes with different decision-makers. Declassification review asks whether protection is still warranted. Derivative classification asks where existing protection must travel.
A fourth misconception assumes that unclassified sources cannot feed a classified derivative. They can. The combination of unclassified facts can itself be classified, which is why compilation review exists as a separate discipline.
The counterintuitive detail: a derivative can be more restrictive than its source without being wrong, if a separate authority governs the new context. But it cannot be less restrictive. The floor set by the source is the binding constraint.
Comparison: original versus derivative classification
| Dimension | Original classification | Derivative classification |
|---|---|---|
| Who decides | Designated original classification authority | Anyone incorporating classified source material |
| What is created | New classified information | New placement of existing classified information |
| Basis for decision | Judgment that disclosure would cause identifiable damage | Consistency with an existing source marking |
| Marking citation | Classifier's name and authority | Source document and date |
| Failure mode | Over- or under-classifying a genuinely new fact | Dropping, loosening, or misattributing inherited markings |
| Review trigger | Periodic review by the original authority | Correction when the source changes |
The table's practical lesson: original classification is a judgment problem, derivative classification is a consistency problem. Training and audits should treat them differently.
Timeline: the life of a derivative marking
| Stage | What happens | Governing question |
|---|---|---|
| Source marked | Original authority classifies and marks the source | Is protection warranted? |
| Derivation | Classifier incorporates source material into a new document | Does the new text communicate classified information? |
| Portion marking | Each portion receives level, category, and caveats | Are markings consistent with the source? |
| Citation | "Derived from" line records the source and date | Can a reviewer trace the chain? |
| Review | Reviewer checks markings against the source | Is anything looser than the source? |
| Source change | Source is upgraded, downgraded, or declassified | Do derivatives need correction? |
| Correction | Derivatives are remarked to match the new source | Is the chain still accurate? |
| Declassification | Derivatives are reviewed alongside the source | Is protection still needed? |
The timeline explains why derivative markings are never truly finished. They are maintained, not merely applied.
Practical actions: getting the definition right in real work
The definition becomes useful only when it changes behavior. These steps translate it into concrete practice.
First, before marking anything, locate the source. If you cannot name the source document, its date, and its markings, you are not ready to derive.
Second, classify at the portion level. Whole-document markings hide the specific facts that carry protection and make later correction harder.
Third, write the "derived from" line at the moment of creation. Retrofitting citations after review is error-prone and breaks the trace.
Fourth, compare your marking to the source, not to your comfort level. If your marking is looser than the source, stop and correct. If it is stricter, document why.
Fifth, treat compilations as a separate review pass. Ask what the assembled document reveals that no single source does.
Sixth, when a source changes, revisit derivatives. A declassified source with still-marked derivatives is an inconsistency that reviewers will find.
Thresholds to reconcile: different agencies publish different marking guidance, and the definitions converge on the same core but differ in citation format and caveat handling. When instructions conflict, follow the most restrictive applicable guidance and document the choice. Consistency with the source is the floor; the governing instruction sets the ceiling.
Limitations: what the definition cannot settle
The definition tells you what derivative classification is and what it obligates. It does not tell you whether a specific fact is classified; that requires reading the source. It does not resolve conflicts between agencies' marking conventions; that requires the governing instruction. And it does not substitute for the review process that catches inherited errors.
A definition is a starting point for correct marking, not a guarantee of it. The chain is only as reliable as each link's citation and each reviewer's comparison against the source. Organizations that treat derivative classification as clerical work tend to discover, during audits, that clerical work is where leaks begin.
For any specific document, marking decision, or release question, consult your organization's classification guidance and a qualified security professional. This article is general educational information about how derivative classification is defined, not a marking determination for any particular document.