FAS | Government Secrecy | Library || Index | Search | Join FAS

Security Classification of Information: Table of Contents

Chapter 3.
CLASSIFICATION OF INFORMATION—AN OVERVIEW

INTRODUCTION

This chapter provides an overview of classification of information. The difference between the three categories of classified information—National Security Information, Restricted Data, and Formerly Restricted Data—with respect to "original" classification is discussed in the next section of this chapter. A following section briefly describes the three major actions that are required for the classification of information: (1) determining whether it should be classified, (2) determining its classification level, and (3) determining its duration of classification.

Determining whether information should be classified is the most difficult of the three major actions. Five steps should be completed to determine whether information should be classified. The final section to this chapter outlines those five steps and also describes in more detail the easiest two of those steps. Three subsequent chapters and an appendix are required for a complete discussion of the other three steps of the classification process. The other two major actions that are required when classifying information, determining classification level and determining classification duration, are described in Chapters 7 and 8, respectively.

NSI, RD, FRD, AND ORIGINAL CLASSIFICATION DECISIONS

Only three categories of classified information are currently used by the U.S. Government—Restricted Data (RD), Formerly Restricted Data (FRD), and National Security Information (NSI).^* The authority for classifying information as RD or FRD comes from the Atomic Energy Act of 1954.¹ The current authority for classifying information as NSI comes from Executive Order (EO) 12356.² RD and FRD are sometimes referred to as "classified atomic energy information" since those two categories deal solely with atomic energy information. Classified atomic energy information can be placed only in the RD or FRD categories and cannot be placed in the NSI category.^{+ 3} All other (non-atomic energy) information that is classified by the government must be classified as NSI. The major classifiable areas of NSI are briefly discussed in Appendix B.

^* "Restricted" was a category of classified information until Dec. 15, 1953. It is still used as a classification category by some foreign governments and international organizations. "Official Use Only" was a category of classified information from July 18, 1949, through Oct. 22, 1951.

⁺ There is one exception to this statement. Some atomic energy intelligence information concerning foreign countries can be transclassified to NSI [42 U.S.C. §2162(e)].

Only NSI can be originally classified. RD and FRD are born classified. That is, if information meets the description of RD or FRD as defined by the Atomic Energy Act of 1954⁴ and if that information has not been declassified (or transclassified to NSI⁵), then that information is RD or FRD. No classification action is required—only a determination by the Department of Energy (DOE) Director of the Office of Classification (OC) that the information fits within the Atomic Energy Act's definition of RD or FRD.^{*, 6} In contrast, an original classifier—a government employee specifically given original classification authority—must make an affirmative decision to classify information as NSI.⁺

^* There appears to be one anomaly to the "born classified" concept. That anomaly is for those methods of isotope separation that can be used to produce special nuclear materials (e.g., to enrich uranium in the U-235 isotope, an isotope useful in nuclear weapons). According to current DOE procedures, research and development on methods of isotope separation other than gaseous diffusion or gas centrifuge can be carried out on an unclassified basis until that research and development shows a "reasonable potential for the separation of practical quantities of special nuclear material." At that point, classification restrictions must be applied [Fed. Reg., 37, 15393 (Aug. 1, 1972); Fed. Reg., 32, 20869 (Dec. 28, 1967)]. Thus, this area of atomic energy information is not "born classified" but is classified only when it reaches "adolescence."

⁺ Government-contractor classification personnel do not have original classification authority. DOE-contractor classification officers had original classification authority for many years. This authority was withdrawn in 1991, partly as a result of a General Accounting Office study [Nuclear Security: DOE Original Classification Authority Has Been Improperly Delegated, GAO/RCED-91-183, U.S. General Accounting Office, Washington, D.C., July 1991].

Since RD and FRD are born classified, one might ask why a classifier who is mostly concerned with atomic energy information needs to understand classification principles. The response is that all classified information must be periodically reviewed for possible declassification,⁷ and declassification of information, whether NSI, RD, or FRD, involves most of the same principles as classification of information. The DOE OC relies on advice from DOE and DOE-contractor classification officers and their staffs to help decide whether to recommend to the DOE Director of Security Affairs (to whom the Secretary of Energy has delegated the authority to make declassification decisions) that certain atomic energy information should be declassified. Therefore, persons who are primarily concerned with atomic energy information will sometimes provide declassification advice to the OC and should be aware of classification and declassification principles. For "newly developed" atomic energy information, those declassification recommendations might be requested shortly after the information is discovered, to help the classifier decide whether to immediately declassify this information and thereby eliminate, as soon as possible, the costs of unnecessary classification of information.

One might also ask why classification management personnel employed by contractors who do classified work for other government agencies need to be concerned with original classification decisions since only government employees have original classification authority. The answer is the same as that given above for DOE-contractor classification personnel. Other government agencies, such as the Department of Defense (DoD), when preparing classification guidance frequently request advice from their contractor classification personnel on this guidance. ⁺⁺

⁺⁺ See, for example, D. F. Rankin, "Panel—Classification in the Department of Defense Today," J. Natl. Class. Mgmt. Soc., 3(2), 80–83 (1967), p. 82; G. A. Zacharias, "Remarks," J. Natl. Class. Mgmt. Soc., 7, 14–16 (1971) p. 15; D. L. Edwards, "The Air Force Classification Process," J. Natl. Class. Mgmt. Soc., 22, 109–117 (1986), pp. 114–115; but see M. C. Anderson, "Department of Defense Overview—Information Security Program," J. Natl. Class. Mgmt. Soc., 21, 20–24 (1985), p. 23, who indicated that contractor participation in preparing classification guidance for DoD was limited because the responsibility for that guidance was the government's.

MAJOR DECISIONS IN CLASSIFYING INFORMATION

The original classification of NSI requires three major decisions: (1) Should the information be classified? (2) What level of classification does it require? (3) What should be the duration of its classification?^* Only the second of those decisions is strictly applicable to the classification of RD and FRD. Atomic energy information is classified by the Atomic Energy Act ("born classified"), and a classification determination by an original classifier is not required. The comparable first step for information that may be classified atomic energy information is to determine whether it falls within the Atomic Energy Act definition of RD (or FRD)⁸ and has not been previously declassified (or transclassified to NSI⁹). The second step, determining the classification level of the classified information, is applicable to RD and FRD as well as to NSI. The third step, determining the classification duration, is not applicable to classified atomic energy information. For both RD and FRD, the classification duration is indefinite.

^* A derivative classifier makes three similar decisions with respect to documents or materials. (1) Should the document or material be classified because of the information that it contains or reveals? (2) What level of classification does the document or material require? (3) What should be the duration of classification of the document or material? Those decisions are all based on existing, relatively detailed (ideally), written classification guidance.

Classification principles useful in making the first decision are discussed in the next section of this chapter and in Chapters 4, 5, and 6. Principles for determining classification level are discussed in Chapter 7, and principles for determining classification duration are discussed in Chapter 8.

DETERMINING WHETHER INFORMATION SHOULD BE CLASSIFIED

Determining whether information not covered by existing classification guidance should be classified as NSI generally involves five major steps:

1. Precisely defining the information to be classified (optional but recommended).

2. Determining whether the information falls within one of the areas permitted to be classified by EO 12356.^{+ 10}

⁺ For atomic energy information, a comparable step is determining whether the information falls within the definition of RD or FRD and has not been declassified or transclassified to NSI.

3. Determining whether the information is under the control of the government.^{++ 11}

⁺⁺ For atomic energy information, the government does not have to control it at the time the information is discovered, developed, or created, in order to classify that information. The Atomic Energy Act gives the government the authority to classify and control that information, no matter who discovers, develops, or creates that information. Of course, the person has to be subject to U.S. law before that control can be exerted.

4. Determining whether disclosure of the information reasonably could be expected to cause damage to the national security.^{**, ++, 12}

^** Actually, the first step in a classification determination may be an order-of-magnitude estimate as to whether the information is important to national security. Only if the information seems to be important to national security would an original classifier expend the effort to go through these steps in the order given. See also Chapter 6 and the proposed de minimis classification principle.

⁺⁺ Such a determination is not required for RD or FRD. The Atomic Energy Act presumes that the unauthorized disclosure of atomic energy information that has not been declassified will cause undue risk to the common defense and security.

5. Precisely specifying why the information is classified (optional but recommended).

As noted above, Steps 1 and 5 are optional. Step 1 is necessary to ensure that the scope of a classification determination is well defined and that its boundaries are well established. Step 5 is necessary to ensure that the rationale of the classification decision can be readily understood by the derivative classifiers who will subsequently apply this classification guidance to a wide variety of fact situations. Knowledge of that rationale will assist those derivative classifiers in reaching correct classification decisions. Knowledge of that rationale will also help ensure that different derivative classifiers reach consistent derivative classification decisions. Therefore, Steps 1 and 5 are strongly recommended.^{*, +, 13, 14}

^* Steps 1 and 5 are also applicable to the classification (identification) of RD and FRD.

⁺ Making original classification decisions is analogous to establishing legal precedents. Therefore, most of the reasons for including Steps 1 and 5 in the classification decision process, as given above and in the following subsections of this chapter, are very similar to the reasons why appellate courts of our legal system provide written decisions for most cases brought to them for resolution.

Since Steps 1 and 5 are optional and since they require less discussion that the other three steps, for convenience they will be discussed in the following subsections of this chapter. The determination required by Step 2 (whether the information falls within a classifiable NSI area as defined by EO 12356) is relatively straightforward and will not be further discussed. For convenience, the ten classifiable areas of NSI and a brief description of each are given in Appendix B. Steps 3 and 4 are the subjects of Chapters 4, 5, and 6.

Precisely Defining the Information to be Classified

The information requiring protection should be described in clear, precise language. This classification guidance should leave no doubt about exactly what information is classified and what information is unclassified.¹⁵ "All written guidance must communicate the same ideas to all users." ^{++, 16} Ambiguities in classification guidance will subsequently cause difficulties for derivative classifiers when they try to reach derivative classification decisions based on that guidance. The value of this step is supported by trade secret law, wherein trade secrets are said to be more likely to be given the protection of trade secret law if they are well defined (see Appendix A).

⁺⁺ As stated by Mr. Marsh in the same article from which this quotation was taken, "Any rules or guidance that can be misunderstood will be misunderstood." This might be considered one of Murphy's Laws of Classification.

The importance of specifically indicating the information that is unclassified as well as specifying the information that is classified was recognized during the early days of the Atomic Energy Commission (AEC). Because all atomic energy information is "born classified," and because few declassification actions were implemented in the first years of the AEC, it was difficult for scientists working in atomic energy fields to be certain about what information in their fields was not classified. Therefore, it was recommended that the AEC "publish explicit and detailed catalogues of types of data not included in the restricted category," thereby relieving those scientists "of the intolerable fear that publication of every research finding is a violation of the [Atomic Energy Act of 1946]." ¹⁷

A goal sometimes urged by managers of a classified program or project is that the classification guidance for the project be prepared as unclassified so that it can be easily and widely distributed. However, unclassified classification guidance is usually too generalized.¹⁸ For example, a guide may state that the operating speed of an aircraft is classified but does not indicate the value for that speed. Even worse is when a guide states that significant advances in technology are classified and does not indicate what constitutes a significant advance in that technology.¹⁹ Generalization in classification guidance usually leads to many problems. Generalized guidance is ambiguous, requires additional interpretation, and causes mistakes in applying that guidance (either underclassification or overclassification). Precisely defining the information to be classified usually requires, at least with respect to scientific and technical areas, that the classification guidance be classified.

Precisely Specifying Why the Information is Classified

There should be a definite, identifiable reason or rationale for classifying information or materials.^{20, 21} If a reason is definite, then it should be expressible. If a reason cannot be expressed or can only be given in vague terms, then the information or material probably should not be classified. "Precise classification guidance is prerequisite to effective and efficient information security. . . ." ²² In 1970, DoD listed "lack of specific rationale as to what information should be classified" as one of three major factors in the overclassification of information.²³

In the late 1970s, DOE began a program to prepare better classification guides for its nuclear weapon activities. That effort was started because numerous declassification errors occurred in the early and middle 1970s during a large-scale declassification review of several million classified DOE documents.^* Also, other information indicated that the then-existing classification guides for nuclear weapon activities were not adequate.²⁴ The new guides for the DOE nuclear weapon programs attempt to specifically explain why the classified information was classified.⁺

^* About 2,440,000 documents were reviewed between July 1971 and June 1975 under this Comprehensive Classification Review Program. About 1,244,000 of those documents were declassified [J. A. Griffin, "ERDA—A New Structure," J. Natl. Class. Mgmt. Soc., 11, 14–29 (1975), p. 27]. Most of those document declassification decisions were subsequently declared to be of no effect, and an additional classification review was required for those documents before they could be considered as declassified.

⁺ Consistent with the previously mentioned first step in determining whether information is classified, the new DOE guidelines provide better definitions of the classified information and also indicate related information that is not classified.

If specific reasons for classification actions are required, then the classifier will be required to be well informed on what he or she is doing and to carefully think through the classification decision;^{21, 25} this will lead to a better classification decision.⁺⁺ When a reason that can be examined by others is provided, erroneous reasons (bad classification decisions) are more easily identified. Indicating why specific information is classified enables others to better understand the rationale behind the classification decision. Derivative classifiers are better able to make their decisions if they know the rationale for the original classification decisions. ²⁶ Additionally, people are generally more willing to follow guidance if they know the reasons for the guidance.²⁷ "Classification is demotivating and of questionable value when people do not understand its purpose."²⁸ Finally, as mentioned earlier, if the classification rationale is known by derivative classifiers, there will be more consistency in subsequent derivative classification decisions that are made by different derivative classifiers.

⁺⁺ One of the major reasons for DoD's 1964 decision to emphasize paragraph marking, both within DoD and DoD's civilian contractors, was a belief that such marking would require disciplined analysis and evaluation and would lead to better classification decisions [W. Skallerup, "Panel—The Executive Views Classification Management," J. Natl. Class. Mgmt. Soc., 1(2, 3, 4), 68–72 (1965) p. 71].

Providing specific reasons for classification will also assist a subsequent declassification or downgrading process.^{21, 29} Declassification or downgrading decisions are easier to make if the reason for the original classification decision is known. Those reasons should be recorded so they will not be lost when the person(s) who made the original classification decision leaves the organization.²⁸

REFERENCES

1. Atomic Energy Act of 1954, 68 Stat. 919; 42 U.S.C. §§2011–2296.

2. Executive Order 12356, Fed. Reg., 47, 14874 (Apr. 6, 1982). Hereafter cited as EO 12356.

3. 42 U.S.C. §2166.

4. 42 U.S.C. §2014(y).

5. 42 U.S.C. §2162(e).

6. U.S. Department of Energy, DOE Order 5650.2B, "Identification of Classified Information," Chap. II, Part A, §4(a)(5), Dec. 31, 1991.

7. Atomic Energy Act of 1954, §142(a), 142(b); 32 CFR Part 2001.21(b); U.S. Department of Energy, DOE Order 5650.2B, "Identification of Classified Information," Chap. II, Part A, §4.a(6)(d), Dec. 31, 1991.

8. 42 U.S.C. §§2014(y), 2162(d).

9. 42 U.S.C. §2162(e).

10. EO 12356, §1.3(a).

11. EO 12356, §6.1(b).

12. EO 12356, §1.3(b).

13. G. MacLain, "The Road Ahead," J. Natl. Class. Mgmt. Soc., 1(1), 6–9 (1965), p. 7.

14. W. Skallerup, "Panel—The Executive Views Classification Management," J. Natl. Class. Mgmt. Soc., 1(2, 3, 4), 68–72 (1965), p. 71. 15. J.C. Cotton, "Panel—Preparation of Classification Guidance," J. Natl. Class. Mgmt. Soc., 4(2), 107-110 (1968), p. 107.

16. J. Marsh, "Workshop on Training of Classification Managers," J. Natl. Class. Mgmt. Soc., 8, 43–47 (1972), p. 47.

17. J. R. Newman and B. S. Miller, The Control of Atomic Energy, McGraw-Hill Book Company, Inc., New York, 1948, p. 220.

18. G. M. Stephan, "Military Operational Secrecy and Classification," J. Natl. Class. Mgmt. Soc., 24, 196–199 (1988), p. 199.

19. J. W. Leonard, "Classification Management Panel," J. Natl. Class. Mgmt. Soc., 23, 21–38 (1987), p. 35.

20. F. W. May, "Classification Futures," J. Natl. Class. Mgmt. Soc., 6, 94–100 (1970), p. 95. Hereafter cited as "May."

21. S. J. Lukasik, "Remarks, Workshop A—Lifetime Cycles for Security Classification," J. Natl. Class. Mgmt. Soc., 7, 56–58 (1971), p. 58. Hereafter cited as "Lukasik."

22. U.S. Department of Defense, Department of Defense Handbook for Writing Security Classification Guidance, DoD 5200.1-H, March 1986, p. 1-1.

23. D. O. Cooke, Department of Defense, in U.S. Government Information Policies and Practices— The Pentagon Papers (Part 2), Hearings Before a Subcommittee of the Committee on Government Operations, House of Representatives, 92nd Congress, 1st Session, June 28 and 29, 1971, U.S. Govt. Printing Office, 1971, p. 660.

23. R. R. Fredlund, "Developments in Classification in DOE," J. Natl. Class. Mgmt. Soc., 17, 42–44 (1981).

24. G. MacLain, "Comments at a National Classification Management Society Meeting," J. Natl. Class. Mgmt. Soc., 1(2, 3, 4), 43–44 (1965).

25. F. J. Thomas, "Remarks, Workshop A—Lifetime Cycles for Security Classification," J. Natl. Class. Mgmt. Soc., 7, 58–62 (1971), p. 61.

26. May, p. 96.

27. J. S. Foster, in Government Secrecy, Hearings Before the Subcommittee on Intergovernmental Relations of the Committee on Government Operations, U.S. Senate, 93rd Congress, 2nd Session, May 22, 23, 29, 30, and 31 and June 10, 1974, U.S. Govt. Printing Office, 1974, p. 268.

28. R. C. Arnold, "Panel—Classification in the Department of Defense Today," J. Natl. Class. Mgmt. Soc., 3(2), 83–87 (1967), p. 85.

FAS | Government Secrecy | Library || Index | Search | Join FAS