choicesVol. 15, no. 5, June 2009 ISSN 0711-0677 www.irpp.org

Security and Democracy

IRPP

Craig Forcese

Canada’sNational Security “Complex”

Assessing theSecrecy Rules

F ounded in 1972, the Institute for Research onPublic Policy is an independent, national,nonprofit organization.

IRPP seeks to improve public policy in Canada bygenerating research, providing insight and sparkingdebate that will contribute to the public policydecision-making process and strengthen the quality ofthe public policy decisions made by Canadiangovernments, citizens, institutions and organizations.

IRPP's independence is assured by an endowment fundestablished in the early 1970s.

F ondé en 1972, l’Institut de recherche enpolitiques publiques est un organisme canadien,indépendant et sans but lucratif.

L’IRPP cherche à améliorer les politiques publiquescanadiennes en encourageant la recherche, en mettantde l’avant de nouvelles perspectives et en suscitant desdébats qui contribueront au processus décisionnel enmatière de politiques publiques et qui rehausseront laqualité des décisions que prennent les gouvernements,les citoyens, les institutions et les organismescanadiens.

L’indépendance de l’IRPP est assurée par un fonds dedotation établi au début des années 1970.

Craig Forcese is an associate professor in theFaculty of Law at the University of Ottawa, wherehe teaches public international law, national secu-rity law, administrative law, and public law andlegislation; he also runs the annual foreign policypracticum. Much of his present research and writ-ing relates to international law, national securityand democratic accountability. He is the author ofNational Security Law: Canadian Practice inInternational Perspective (2007). Prior to joiningthe University of Ottawa law faculty, he practisedwith Hughes Hubbard and Reed LLP, aWashington, DC, firm specializing in internationaltrade law. He has law degrees from the Universityof Ottawa and Yale University, a BA from McGillUniversity, and an MA in international affairsfrom the Norman Paterson School of InternationalAffairs, Carleton University. He is a member ofthe Bars of Ontario, New York and the District ofColumbia.

This publication was produced under thedirection of Mel Cappe, President, IRPP, andWesley Wark, Fellow, IRPP. The manuscript wascopy-edited by Francesca Worrall, proofreadingwas by Mary Williams, production was byChantal Létourneau, art direction was bySchumacher Design and printing was by AGLGraphiques.

Copyright belongs to IRPP. To order or requestpermission to reprint, contact:

IRPP1470 Peel Street, Suite 200Montreal, Quebec H3A 1T1Telephone: 514-985-2461Fax: 514-985-2559E-mail: irpp@irpp.orgwww.irpp.org

All IRPP Choices and IRPP Policy Matters areavailable for download at www.irpp.org

To cite this document:

Forcese, Craig. 2008. “Canada’s National Security‘Complex’: Assessing the Secrecy Rules.” IRPPChoices 15 (5).

The opinions expressed in this paper are those of the author and do not necessarily reflect the views of IRPP or its Board of Directors.

1

Contents2 List of Acronyms and Abbreviations

3 Introduction

4 Principles of Transparency

6 Principles of National Security Confidentiality

19 Reconciling Transparency and National SecurityConfidentiality

27 Conclusion and Recommendations

30 Notes

33 References

37 Résumé

38 Summary

Security and Democracy / Sécurité et démocratieResearch Directors/Directeurs de recherche

Mel Cappe and/et Wesley Wark

T his IRPP research program explores the com-plex challenges confronting Canada withregard to the post-9/11 security environment

and its impact on domestic and international poli-cies. The research addresses issues that are in manyways new to the country and to the formulation ofCanadian national security policy, above all thethreat posed by global, transnational terrorism. Theprogram examines the interrelationships betweennew security demands and democratic norms, focus-ing in particular on the building blocks of a sounddemocratic model for national security, namely,effective intelligence; capable law enforcement;appropriate, stable laws; good governance; accounta-bility; citizen engagement and public knowledge;emergency response capability; wise economicpolicy; and public-private-sector partnerships.

C e programme de recherche s’intéresse auxdéfis de sécurité d’une grande complexité quele Canada doit relever depuis le 11 septembre,

de même qu’à leur incidence sur nos politiquesnationales et internationales. Il traitera d’enjeux sou-vent inédits pour notre pays en matière de sécuriténationale, notamment le terrorisme mondial ettransnational. Le programme vise à analyser l’inter-relation entre les nouvelles exigences de sécurité etles normes démocratiques, de manière à définir leséléments de base suivants : un modèle de sécuriténationale pleinement démocratique, notamment enmatière de renseignement ; l’efficacité du maintien del’ordre ; la stabilité et la légitimité des lois ; la gou-vernance éclairée ; la responsabilisation ; l’engage-ment citoyen et l’information du public ; l’interven-tion d’urgence ; une politique économique avisée ; etles partenariats entre les secteurs public et privé.

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

2

List of Acronyms andAbbreviations

Access Act Access to Information ActATA Anti-terrorism ActCEA Canada Evidence ActCSEC Communications Security

Establishment CanadaCSIS Canadian Security Intelligence

ServiceCSIS Act Canadian Security Intelligence

Service ActSIRC Security Intelligence Review

CommitteeCSE Communications Security commissioner Establishment commissionerDFAIT Department of Foreign Affairs and

International TradeDND Department of National DefenceFOIA US Freedom of Information ActIRPA Immigration Refugee Protection ActRCMP Royal Canadian Mounted PoliceSOIA Security of Information Act

Introduction

S ecrecy,” said Cardinal Richelieu in 1641, “is thefirst essential in affairs of the State” (1641).Constraints on information disclosure may give

governments a leg-up over their international rivals, pre-serve them from their enemies and insulate them fromdomestic opponents. Protection of the nation and itsinhabitants may depend on keeping information aboutweapons systems, troop strengths, intelligence assets orphysical vulnerabilities away from enemies. As thefamous Second World War-era admonishment warned,“loose lips…sink ships.”1 For these reasons and others, “itis difficult to think of national security without alsothinking about government secrecy” (Roberts 2004).

Of course, excessive secrecy may also be a vice intoday’s modern democracies. Information is, as thefamous US Supreme Court Justice Louis Brandeis oncequipped, “the best of disinfectants” (1914). Opennessand transparency may preserve citizens from themalfeasance, incompetence, corruption and expedientbehaviour of incumbent governments.

Relative latecomers to the open government concept,Canadians have a suspicion of government secrecy.Former auditor general of Canada Denis Desautelsurged that “information is the current that chargesaccountability in government” (InformationCommissioner of Canada 2001). Government accounta-bility, in this view, requires timely and extensive accessto government information. Without the capacity tocompel disclosure of information unfavourable to gov-ernment, citizens — including elected members ofParliament — are dependent on the potentially self-serving information the government chooses to release.

Indeed, secrecy, even when motivated by an objec-tive as fundamental as national security, may some-times create more perils than it forestalls. In 2003, theStanding Senate Committee on National Security andDefence released its report The Myth of Security atCanada’s Airports. The study documented deeply inad-equate security at Canadian airports, even in the post-

3

Ca

na

da

’s Na

tion

al S

ec

urity

“Co

mp

lex

”: Asse

ssing

the

Se

cre

cy

Ru

les, b

y C

raig

Fo

rce

se

Canada’s NationalSecurity “Complex”: Assessing the Secrecy Rules

Craig Forcese

“

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

4

section I assess how well Canadian information law rec-onciles national security with transparency, flagging anumber of structural and practical problems that plaguethis reconciliation. I conclude with a number of recom-mendations as to how secrecy and transparency might bebetter reconciled in Canadian law and practice.

Principles of Transparency

T he principles of transparency at the heart ofCanadian information law and policy fall intotwo broad classes: “open government” and

“open courts.” In Canadian practice, open govern-ment involves the executive branch of government,while the open court concept relates to the judicialsystem (and to quasi-judicial bodies that, while tech-nically part of the executive, have many of the trap-pings of courts).

Open governmentJustificationThe open government doctrine envisages access toinformation as an essential attribute of democracy.As one of the founders of the United States, JamesMadison, noted: “a popular government without pop-ular information or the means of acquiring it is but aprologue to a farce or a tragedy, or perhaps both.Knowledge will forever govern ignorance; And thepeople who mean to be their own Governors, mustarm themselves with the power which knowledgegives” (letter to W.T. Barry, August 4, 1822, inPadover 1953).

Policy-makers voiced similar sentiments in discus-sions of what would become the United StatesFreedom of Information Act (FOIA),3 introduced asthe world’s first modern information access law in1966. Proponents of the Act argued that “free peopleare, of necessity, informed; uninformed people cannever be free.”4 In signing the FOIA, PresidentJohnson noted that “this legislation springs from oneof our most essential principles: A democracy worksbest when the people have all the information thatthe security of the Nation permits. No one should beable to pull curtains of secrecy around decisionswhich can be revealed without injury to the publicinterest” (“Statement by the President” 1966). In a1978 decision under the FOIA, the US Supreme Courtechoed this comment, noting that “the basic purposeof FOIA is to ensure an informed citizenry, vital to

9/11 era, and concluded that “the front door of airsecurity…[is] now being fairly well secured, with theside and back doors wide open” (2003, 9). In thecourse of preparing its report, the committee was“criticized for calling witnesses that have sharedknowledge of these breaches with the Canadian pub-lic” (2003, 11). It rejected this criticism, observing:

You can be sure that ships really will sink ifthey have a lot [sic] holes in them. And thoseholes aren’t likely to get patched unless thepublic applies pressure to get the job done.They certainly aren’t patched yet...TheCommittee recognizes the need to balancethe public’s right to know against the inter-ests of national security. But unreasonablesecrecy acts against national security. Itshields incompetence and inaction, at a timethat competence and action are both badlyneeded. (2003, 12-13)2

National security, in other words, is not aboutinsulating governments from embarrassment.

In fact, the current director of the CanadianSecurity Intelligence Service (CSIS) has argued thatexcessive secrecy may also undermine the credibilityof the government’s national security policies. In hiswords, “there is a risk that, absent adequate publicdialogue and a surfeit of secrecy, the justification foraction by governments against terrorism will beundermined or misunderstood. This in turn can put injeopardy the legitimacy of the government response”(Judd 2007). Security specialist Wesley Wark hasechoed this concern: with injudicious secrecy, “credi-bility in the field of national security is underminedand the very idea of legitimate state secrets, oftenscoffed at by a Canadian public unused to the idea, isfurther eroded” (2007).

The dilemma of any government information regimelies in balancing the strong public interest in disclosurein all areas, including national security, against legiti-mate secrecy. As the 2003 Senate committee acknowl-edged, seeking assurances that secure doors at airportsare actually locked is a proper public concern.Demanding disclosure of the combination codes tothose doors would not be (Standing Senate Committeeon National Security and Defence 2003, 12).

In this study I assess Canada’s efforts to balance trans-parency with secrecy in the national security area. In thesecond section I highlight principles of transparency thatanimate the Canadian concepts of “open government”and “open courts.” I then posit several principles ofnational security confidentiality — that is, justificationsfor secrecy predicated on national security preoccupa-tions. I also describe how Canadian information laws seekto guard national security confidentiality. In the fourth

politicians and bureaucrats remain accountable to thecitizenry.”8 While Justice La Forest was writing in dis-sent, his approach to interpreting the Access Act wasendorsed by the majority in that case and has sincebeen followed by the lower courts.9

More recently, the Supreme Court has noted that theAccess Act makes information “equally available toeach member of the public because it is thought thatthe availability of such information, as a general mat-ter, is necessary to ensure the accountability of thestate and to promote the capacity of the citizenry toparticipate in decision-making processes.”10

Open government guaranteesAs the discussion above suggests, the open governmentconcept is protected at the federal level principally bythe Access Act.11 The Access Act creates a broad princi-ple of access in its first dozen or so sections. It thendevotes a sizeable portion of its remaining sections tothe creation of exceptions and caveats to this principle.

The express purpose of the Act is “to extend thepresent laws of Canada to provide a right of access toinformation in records under the control of a govern-ment institution in accordance with the principles thatgovernment information should be available to thepublic, that necessary exceptions to the right of accessshould be limited and specific and that decisions on thedisclosure of government information should bereviewed independently of government” (section 2).

The key provision of the Act, section 4, provides thatevery Canadian citizen and permanent resident “has aright to and shall, on request, be given access to anyrecord under the control of a government institution,”subject to other sections in the Act.

Notably, the Federal Court has referred to this rightas “quasi-constitutional” in nature.12 In part, this statusreflects language in subsection 4(1) specifying that theright in section 4 applies notwithstanding any otherstatute.13

The Act creates a mechanism for policing governmentdecisions on disclosure and its use of exemptions. Anoffice of the information commissioner is created and ischarged with investigating access complaints brought byinformation requesters.14 The commissioner — an officerof Parliament — has extensive powers to conduct inves-tigations but has no power to compel the release of theinformation if he or she feels that such release is war-ranted. Instead, to compel disclosure, the informationcommissioner, or any requester dissatisfied with the out-come of the commissioner’s investigation, must bring anapplication to the Federal Court.15

the functioning of a democratic society, needed tocheck against corruption and to hold the governorsaccountable to the governed.”5

Unlike its immediate predecessor, the new Obamaadministration appears to share this perspective. Oneof the President’s first acts was to issue a directiveinstructing federal agencies to act more transparently.That instrument reads, “transparency promotesaccountability and provides information for citizensabout what their Government is doing” (PresidentialDocuments 2009).

Canadians have expressed similar views duringdiscussions of federal information access laws. PrimeMinister Trudeau noted in 1975 that “democraticprogress requires the ready availability of true andcomplete information. In this way people can objec-tively evaluate the government’s policies. To act oth-erwise is to give way to despotic secrecy.”6 Then-president of the Privy Council Walter Baker under-scored this point in 1979, urging that “if thisParliament is to function, if groups in society are tofunction, if the people of the country are to judge ina knowledgeable way what their government isdoing, then some of the tools of power must beshared with the people, and that is the purpose offreedom of information legislation.”7

Information commissioners appointed under thefederal government’s key information law, the Accessto Information Act (Access Act), express similarviews. The then information commissioner JohnGrace used colourful language to describe this per-spective in his 1998 annual report:

Any society aspiring to be free, just and civilmust depend upon and nurture a wide array ofmethods for exposing, and imposing sanctionson, ethical failures…In one way or another, allthe checks and balances designed to limitabuses of government power are dependentupon there being access by outsiders to gov-ernments’ insider information…Yes, webs ofintrigue are more easily woven in the dark;greed, misdeeds and honest mistakes are moreeasily hidden. A public service which holdstight to a culture of secrecy is a public serviceripe for abuse. (Information Commissioner ofCanada 1998, 4)

The courts have also recognized the importance offree access to information in a democracy. In his rea-sons in Dagg v. Canada, Justice La Forest urged that“the overarching purpose of access to informationlegislation…is to facilitate democracy. It does so intwo related ways. It helps to ensure first, that citizenshave the information required to participate mean-ingfully in the democratic process, and secondly, that

5

Ca

na

da

’s Na

tion

al S

ec

urity

“Co

mp

lex

”: Asse

ssing

the

Se

cre

cy

Ru

les, b

y C

raig

Fo

rce

se

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

6

ings and extended to material in the possession ofCSIS. Here, “full” disclosure does not mean that thisinformation will be passed to the person subject tothe certificate — it may not be, because of secrecyconcerns of the sort discussed below. However, at thevery least, the full record must be disclosed to thejudge and the special advocate, a special security-cleared lawyer representing the interests of thenamed person. Moreover, for this disclosure obliga-tion to be realized, “CSIS officers must retain theiroperational notes when conducting investigationsthat are not of a general nature” — that is, “wheneverCSIS conducts an investigation that targets a particu-lar individual or group.”24

Principles of National SecurityConfidentialityThe ambiguity of “national security”

T he values reflected in the open court and opengovernment concepts notwithstanding, nationalsecurity imperatives may require limitations on

both of these forms of transparency. Exactly whenand to what extent secrecy is justified depends, how-ever, on a sound understanding of “national securi-ty.” No serious observer would advocate, for example,posting Canadian troop deployments in Afghanistanfor the forthcoming week on the Internet. Other sce-narios are, however, more fraught with ambiguity. Isnational security imperilled, for example, by therelease of Canadian government reports critical ofconditions in the Afghan prisons to which theCanadian Forces transfer their battlefield detainees?There are a number of interests that might be preju-diced by the release of this prison report; not least,relations with the Afghan authorities embarrassed bya frank and damning report by their Canadian allies.

An injury to international relations should not beautomatically conflated with an injury to nationalsecurity, nor should it necessarily be given the samelevel of protection. Precluding embarrassment of a for-eign government and the preservation of Canadianlives are on very different levels in terms of the justifi-cation for secrecy they create. The problem is, howev-er, that “national security” defies clear definition and istherefore vulnerable to mutable and very elastic mean-ings. As noted in a 2002 IRPP study, “the term nation-al security is used frequently to refer to mattersranging from domestic or internal security through to

Open courtsThe open court principle is tied to fair trials. Fair trialrights in international law require a presumptivelyopen court.16 In Canadian law also, court proceedingsare presumptively open. The Supreme Court ofCanada has repeatedly underscored this point,17

pointing to the common law and relying on theCanadian Charter of Rights and Freedoms.18 Forinstance, the Supreme Court has held that “freedomof expression in section 2(b) protects both listenersand readers.”19 It therefore supports open courts:“openness permits public access to information aboutthe courts, which in turn permits the public to discussand put forward opinions and criticisms of courtpractices and proceedings.”20

Further, it is axiomatic in international21 andCanadian criminal and constitutional law that theaccused in criminal matters be given full disclosure ofthe state’s evidence against them. Subject to legitimateexceptions for privileged evidence, the Crown inCanada has a legal duty to disclose its relevant evi-dence to the defence. As the Supreme Court of Canadanoted in the leading authority on this point, R. v.Stinchcombe, “the right to make full answer anddefence is one of the pillars of criminal justice onwhich we heavily depend to ensure that the innocentare not convicted. Recent events have demonstratedthat the erosion of this right due to nondisclosure wasan important factor in the conviction and incarcera-tion of an innocent person.”22 As a consequence, the

Crown obligation to disclose all relevant andnon-privileged evidence, whether favourableor unfavourable, to the accused requires thatthe Crown exercise the utmost good faith indetermining which information must be dis-closed and in providing ongoing disclosure.Failure to comply with this initial and con-tinuing obligation to disclose relevant andnon-privileged evidence may result in a stayof proceedings or other redress against theCrown, and may constitute a serious breachof ethical standards.23

Read together, these rules create an open, transpar-ent and fair system of adjudication. This openness isnot absolute – it may be attenuated by rules on secre-cy, not least in the Canada Evidence Act (CEA). Butopen courts and full disclosure are the starting point.

Indeed, the spirit — if not the specifics — ofStinchcombe now inform disclosure of security intel-ligence in proceedings that, while not criminal, placeat risk Charter-protected interests in life, liberty andsecurity of the person. In 2008, the Supreme Courtheld in Charkaoui II that full disclosure obligationsapplied in immigration security certificate proceed-

dentiality. Not least, doctrines developed when Canadaconfronted a very different security threat — a Cold Waradversary with a sophisticated intelligence-gatheringcapacity — have been deployed in a modern securitycontext in relation to a much more inchoate (and almostcertainly much less capable) adversary: terrorist groups.

It is possible, however, to extract from the 2004 pol-icy a rough working definition of national security tiedto the three threats that policy is designed to prevent.With this understanding in mind, justifications fornational security confidentiality in Canada fall into twobroad classes. First, there is information that, ifreleased, could increase the likelihood of a threat tonational security. Second, there is information that, ifreleased, might cause injury to an intelligence relation-ship between Canada and allied states, whether or notit would also cause injury to a more specific Canadiannational security objective.

Deterring a national security threatThere is no limit to specific threats to Canada, nor tothe activities that might exacerbate these threats. I will,however, posit three obvious subclasses of national-security-related information, the revelation of whichwould likely enhance threats to the physical security ofCanadians or of our allies, or undermine our contribu-tions to international security: technology; plans andtactics; and sources and methods.

TechnologyCertain technologies obviously provide advantages tothose who possess them and constitute potential threatsto those who do not. Preserving a technological advan-tage — and keeping potentially dangerous technologiesout of the hands of adversaries — is a key and typicallyuncontroversial national security objective. At its core,technology is a manifestation of knowledge. Limitingaccess to technologies may mean restricting access tohardware, but it also means constraining the spread oftechnological know-how. Limiting the spread of tech-nology may, therefore, require secrecy.

Plans and tacticsAdversaries gain advantage by knowing their oppo-nents’ plans or being able to predict those opponents’conduct accurately. Guarding plans of action and habitsof conduct from adversaries is a core objective ofnational security confidentiality. Plans subject to suchconfidentiality could be, for example, deploymentinstructions to the Canadian Forces in Afghanistan, orthe precise travel route of a visiting dignitary at poten-

international security, but is seldom defined”(Macnamara and Fitz Gerald 2002, 7). When someeffort is made to set out its content, the definition isoften so broad as to be meaningless. For instance,instructors at the National Defence College definednational security in 1980 as “the preservation of a wayof life acceptable to the Canadian people and compati-ble with the needs and legitimate aspirations of oth-ers.” National security, these authorities asserted,“includes freedom from military attack or coercion,freedom from internal subversion, and freedom fromthe erosion of the political, economic, and social val-ues which are essential to the quality of life in Canada”(cited in Macnamara and Fitz Gerald, 8). As is immedi-ately evident, this definition wraps much of what gov-ernments exist to do in a blanket of national security,a fact acknowledged even by those comfortable withthis definition.25

The Government of Canada’s April 2004 nationalsecurity policy offers more recent and helpful guid-ance. That document describes national security asdealing “with threats that have the potential toundermine the security of the state or society” (PrivyCouncil Office 2004, 3). The three specific threats thegovernment seeks to address are: first, “protectingCanada and the safety and security of Canadians athome and abroad” (which includes “protecting thephysical security of Canadians, our values, and ourkey institutions”); second, “ensuring that Canada isnot a base for threats to our allies”; and, third, “con-tributing to international security” (5).

While this threat-based description of nationalsecurity is narrower than the National DefenceCollege version, it is still extremely general. Noteespecially the reference to the protection of“Canadian values” (whatever these may be) as anational security objective in the government’s defi-nition. Moreover, the description does little to defineexactly when threats of the sort listed in the defini-tion constitute national security concerns. As the2004 policy acknowledges, most criminal offencesthreaten personal security but do not generally chal-lenge the “security of the state or society.” The policydoes not answer the question, however, of whenexactly a threat to the physical safety of Canadiansbecomes a legitimate national security concern ratherthan a regular policing matter, an uncertainty thathas important implications when national security isinvoked to justify secrecy.

The problem of ambiguity and imprecision is anacute one in the application of national security confi-

7

Ca

na

da

’s Na

tion

al S

ec

urity

“Co

mp

lex

”: Asse

ssing

the

Se

cre

cy

Ru

les, b

y C

raig

Fo

rce

se

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

8

the World Trade Organization and in other treaty con-texts is not, for instance, a matter of national security.

That said, there are instances where national secu-rity and international relations overlap. A dominantpreoccupation for governments is guarding thesecrets shared by allies. A failure to do so mightprompt allied agencies to restrict their informationsharing. As a middle power with limited foreign intel-ligence capacities, Canada is particularly dependenton information supplied by foreign intelligence serv-ices. It is, therefore, keenly aware of the need toobserve so-called originator control rules. Originatorcontrol puts control over the use and distribution ofthe information in the hands of the state from whichit comes. It is typically associated with the so-calledthird party rule, which bars the recipient of informa-tion from sharing it with a third party without thepermission of the originating entity.

The third party rule is a regular fixture in theCanadian approach to national security confidentiali-ty. Originator control provisions are likely common-place in international information-sharingagreements. CSIS told a Federal Court in 1996 that theinformation it receives is “invariably provided in con-fidence and on the explicit or implicit understandingthat neither the information nor its source will be dis-closed without the prior consent of the entity whichprovided it.”27 This principle is “widely recognizedwithin the policing and security intelligence commu-nities,”28 an observation confirmed by RCMP,Department of National Defence (DND) andDepartment of Foreign Affairs and International Trade(DFAIT) submissions in the same case.

CSIS told the same court that “CSIS receives sensi-tive information, not just because of the third partyrule which requires CSIS to treat the information asconfidential, but also because there is confidence onthe part of information providers that the Canadiangovernment understands the need for confidentialityand has in place practices and procedures to safe-guard information.”29 Without this confidence inCanada’s ability to restrict disclosure, some allies“may discontinue the alliance or association. Othersmay continue their alliance, but with a reluctance tobe candid.”30 Similar views were expressed in this andother cases31 by the RCMP, DND and DFAIT. TheRCMP — which reports that it receives 75 times moreinformation from partner agencies than it provides –applies third-party-rule limitations even when docu-ments do not contain emphatic constraints on shar-ing information with third parties.32

tial risk of assassination. Tactical information, for itspart, may include the doctrines that guide these deci-sions on deployment and travel planning for digni-taries. Open familiarity with these habits and patternsmay allow predictions by adversaries that are almostas prejudicial as actual knowledge of specific plans.

Sources and methodsInformation on government capabilities and capaci-ties – especially in the sensitive areas of nationaldefence and security and intelligence – is also closelyguarded. An adversary who is intimately familiar, forexample, with the technological abilities and limita-tions of Canada’s signals intelligence agency,Communications Security Establishment Canada(CSEC), could employ this knowledge to circumventdetection. Likewise, an individual who is privy to themethodologies employed by CSIS in surveilling tar-gets might rely on this knowledge to avoid monitor-ing. Indeed, chief among the cardinal secrets of anyclandestine activity is the identity of informers, bethey undercover agents or members of target commu-nities or organizations.

Court jurisprudence gives some indication of theimportance the security services place on protectingsources and methods. CSIS, for example, opposes dis-closure of any information that may “identify or tendto identify”:

a) Service employees or internal proceduresand administrative methodology of theService, such as names and file numbers…b) investigative techniques and methods ofoperation utilized by the Service…c) Service interest in individuals, groups orissues, including the existence or absence ofpast or present files or investigations, theintensity of investigations, or the degree orlack of success of investigations…d) human sources of information for theService or the content of information provid-ed by a human source...e) relationships that the Service maintainswith foreign security and intelligence agen-cies and would disclose information receivedin confidence from such sources; and…f) information concerning the telecommuni-cation system utilized by the Service.26

Preserving international relationshipsCanadian law often amalgamates justifications forsecrecy predicated on national security and those asso-ciated with international affairs. However, confidential-ity dictated by international relations preoccupationsoften has little to do with “national security” as the termis used in this paper. Canada’s negotiating position at

Legal protections for national-security-relatedinformationThe objectives of national security confidentiality set outabove are largely unassailable and self-evident. Theimplementation of these objectives in Canadian law andpolicy is much more complex. Discerning whether theinterests discussed above are truly engaged, and whetherthey are engaged to a degree that justifies secrecy, is amatter of judgment. How that judgment may bedeployed is ultimately a matter set down in law.National-security-related information is protected at sev-eral levels in Canadian information law: laws limitingopen government rules otherwise applicable to the exec-utive branch; laws that constrain the open court conceptand disclosure rules typically applied by Canada’s courts;and statutes that criminalize the wrongful disclosure ofparticularly sensitive information.

Limiting open governmentBy 2006, 68 states had enacted access to informationlaws (Banisar 2006). The right to information is notabsolute; nearly all of these laws include common excep-tions relating, inter alia, to national security and interna-tional affairs and the protection of information obtainedin confidence from other states (Banisar 2006, 22).

Canada is no different in this respect. While princi-pally an information disclosure law, the Access Actincludes important limitations on open government,several of which are tied to national security.35 Theseconstraints include exemptions, that is, excuses forredacting sensitive information from released records,and a new national security exclusion – a blanket over-ride of the Access Act and its mechanisms.

The exemptions are of venerable vintage, mostlyunchanged from the time the Access Act was enacted,in the early 1980s. They are reasonably precise in theirdrafting, although they are not without inherent ambi-guity. In comparison, the exclusion was enacted inhaste in 2001 in the immediate aftermath of 9/11. It isnot so carefully considered or drafted.

National security exceptionsThe Access Act’s national-security-related exemptionsare summarized in table 1. These exemptions can becategorized in two ways: injury-based/class-basedexemptions and mandatory/discretionary exemptions.

Injury-based exemptions may be employed onlywhere the government concludes that disclosure mayproduce the harm enumerated by the Act.36 By compari-son, class-based exemptions are triggered as soon as theinformation requested is found to fall within a certain

The current security focus on counterterrorism has,if anything, augmented international intelligenceflows, and therefore the potential application of thethird party rule. The director of operations of CSISdescribed the environment for information sharingthis way in 2004: “compromising al-Qaeda operativesrequires an unprecedented level…of cooperationbetween police, law enforcement, and immigrationofficials and the like, not just domestically but inter-nationally as well.”33 As the Arar Commission reportlater noted,

Information sharing among agencies allowsa more comprehensive picture to emerge.Viewing different pieces of informationtogether may allow a more complete andaccurate assessment of the threat beinginvestigated and the steps needed to addressthat threat. Sometimes, seemingly inconse-quential bits of information may take on animportance not otherwise apparent whenviewed alongside other information. Broadinformation sharing is therefore essential toeffective prevention. (Commission of Inquiry2006, 102)

More generally, intelligence sharing permits the“acquisition of intelligence that is valuable to decisionmakers but otherwise unobtainable at an acceptablecost” (Walsh 2007, 157). As such, it is particularlyimportant for small countries like Canada to be able touse alliance relationships to tap into the intelligencecapacities of larger states. Intelligence sharing is, how-ever, also vital for larger states, including the UnitedStates. In 1985, then defense secretary CasparWeinberger observed that the “United States has neitherthe opportunity nor the resources to unilaterally collectall the intelligence information we require. We compen-sate with a variety of intelligence sharing arrangementswith other nations in the world.”34 The more recentfocus on terrorism has likely compounded this USdependency on foreign information sharing. In thewords of one commentator, writing about terrorism:

the point of impact is not necessarily thesame as the point of origin. Therefore, inorder for an individual state to make a mean-ingful assessment of its security situation itneeds to process information pertaining toindividuals and events existing in far-offplaces at any given time. No single state hasthe power to collect and exploit this type andvolume of information; intelligence sharingbetween states has become essential evenfrom the perspective of the most powerfulintelligence producers, such as the US.(Mackmurdo 2007, 207)

9

Ca

na

da

’s Na

tion

al S

ec

urity

“Co

mp

lex

”: Asse

ssing

the

Se

cre

cy

Ru

les, b

y C

raig

Fo

rce

se

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

10

Information pertaining to threats to thesecurity of Canada Section 16(1)(a) allows thegovernment to refuse release of requested records lessthan 20 years old that contain information preparedby a government investigative body in the course oflawful investigations of activities suspected ofconstituting “threats to the security of Canada”within the meaning of the Canadian SecurityIntelligence Service Act (CSIS Act).37 The latterconcept is carefully defined, although in a mannerthat creates ambiguity of its own.

The formulation and inclusion of the concept of“threats to the security of Canada” was the subject ofsustained discussion at the time Parliament enactedthe CSIS Act, in 1984. It has also drawn the attentionof the review agency empowered to scrutinize CSIS’sactivities, the Security Intelligence Review Committee(SIRC). In one of its early reports of CSIS, the SIRCquestioned several aspects of the threat definition inthe Act. These concerns are summarized in table 2.

Even with these shortcomings, however, the defi-nition of “threats to the security of Canada” is robustwhen compared to the looser invocations of “nationalsecurity” that have figured in more recent laws.

class of information, as defined by the Act. Put anoth-er way, the exemption is triggered by the type ofinformation at issue, without any consideration ofwhether disclosure of that information would actuallybe “injurious” to some listed government interest. Thisgives class-based exceptions a broad reach.

In the national security area, all the class-basedexceptions are also mandatory. With mandatoryexemptions, the government is obliged to decline dis-closure once the information falls within the class ofprotected data, subject in a few instances to a publicinterest override. Mandatory class-based exceptionsare the most robust secrecy protection offered by theAccess Act. No doubt for exactly this reason, infor-mation received in confidence from other govern-ments falls within this category.

In fact, however, the majority of exceptions inthe Act are not mandatory, but rather discretionary.Thus, the government may (but need not) choose todecline disclosure of a document captured by theexemption.

Turning to specific examples, section 16(1)(a) andsection 15 of the Access Act are the most obviousnational-security-related exceptions.

Table 1National Security Exemptions in the Access to Information Act

Class-based exemptions Injury-based exemptions

Mandatory exemptions

Discretionary exemptions

• Section 13 — Information received in con-fidence from other governments or aninternational organization, if the bodygives disclosure permission (or this bodyhas itself made public the information),the information may be disclosed

• Section 20 — Information supplied in con-fidence by third parties concerning emer-gency management plans relating to thevulnerability of critical infrastructure, sub-ject to a public interest override

• Section 24 — Information protected underother, listed statutes

• Paragraph 16(1)(a) — Informationobtained or prepared by listed investiga-tive bodies pertaining to crime prevention,law enforcement or threats to the securityof Canada, if less than 20 years old

• Paragraph 16(1)(b) — Information ontechniques or plans for specific lawfulinvestigations

• Section 15 — Information that could reasonably be expect-ed to be injurious to the conduct of international affairs orto the defence of Canada or an allied state, or the preven-tion or suppression of subversive or hostile activities

• Paragraph 16(1)(c) — Information that could reasonably beexpected to be injurious to law enforcement or to the con-duct of lawful investigations, including information on con-fidential sources

• Paragraph 16(1)(d) — Information that could reasonably beexpected to be injurious to the security of penal institutions

• Subsection 16(2) — Information that could reasonably beexpected to facilitate the commission of an offence, includ-ing technical information relating to weapons or potentialweapons or information on the vulnerability of particularbuildings or other structures or systems

• Section 17 — Information the disclosure of which couldreasonably be expected to threaten the safety of individuals

11

Tabl

e 2

Defin

ition

s of

“Th

reat

s to

the

Sec

urity

of

Cana

da”

Type

of

thre

atDe

finiti

ons

in t

he C

SIS

Act

Criti

que

by t

he S

ecur

ity In

telli

genc

e Re

view

Com

mitt

ee1

CSIS

inte

rpre

tatio

n2

1U

nles

s ot

herw

ise

note

d, t

hese

crit

ique

s ar

e dr

awn

from

Sec

urity

Inte

llige

nce

Revi

ew C

omm

ittee

(SIR

C) (1

989)

.2

Extr

acts

cite

d in

thi

s co

lum

n dr

awn

from

Can

adia

n Se

curit

y In

telli

genc

e Se

rvic

e (2

005)

.3

S.C.

200

1, c

. 41.

4Se

e di

scus

sion

in M

acLe

od (2

006)

.5

Spec

ial S

enat

e Co

mm

ittee

on

the

Anti-

terr

oris

m A

ct(2

007,

20)

.

Espi

onag

e an

d sa

bota

ge

Fore

ign-

influ

ence

dac

tiviti

es

Polit

ical

vio

lenc

e an

dte

rror

ism

Subv

ersi

on

“Esp

iona

ge o

r sa

bota

ge t

hat

isag

ains

t Ca

nada

or

is d

etrim

enta

lto

the

inte

rest

s of

Can

ada

orac

tiviti

es d

irect

ed t

owar

d or

insu

ppor

t of

suc

h es

pion

age

orsa

bota

ge.”

“For

eign

-inf

luen

ced

activ

ities

with

in o

r re

latin

g to

Can

ada

that

are

detr

imen

tal t

o th

e in

tere

sts

of C

anad

a an

d ar

e cl

ande

stin

e or

dece

ptiv

e or

invo

lve

a th

reat

to

any

pers

on.”

“Act

iviti

es w

ithin

or r

elat

ing

toCa

nada

dire

cted

tow

ards

or i

nsu

ppor

t of t

he th

reat

or u

se o

fac

ts o

f ser

ious

vio

lenc

e ag

ains

tpe

rson

s or

pro

pert

y fo

r the

pur

-po

se o

f ach

ievi

ng a

pol

itica

l, re

li-gi

ous

or id

eolo

gica

l obj

ectiv

ew

ithin

Can

ada

or a

fore

ign

stat

e.”

“Act

iviti

es d

irect

ed to

war

ds u

nder

-m

inin

g by

cov

ert u

nlaw

ful a

cts,

ordi

rect

ed to

war

ds o

r int

ende

d ul

ti-m

atel

y to

lead

to th

e de

stru

ctio

nor

ove

rthr

ow b

y vi

olen

ce o

f the

cons

titut

iona

lly e

stab

lishe

d sy

stem

of g

over

nmen

t in

Cana

da.”

“ Det

rimen

tal t

o th

e in

tere

sts o

f Can

ada”

: The

pha

se is

“w

holly

sub

ject

ive”

as

“no

crite

ria a

re p

rovi

ded

to o

ffer

any

sta

ndar

d fo

r de

term

inin

g w

hat

is ‘d

etrim

enta

l.’”It

shou

ld, t

here

fore

, be

defin

ed in

the

Act

.

“For

eign

influ

ence

d”: T

he p

hras

e “f

orei

gn in

fluen

ced”

is b

road

, cov

erin

g “f

orei

gn in

ter-

est g

roup

s, po

litic

al o

rgan

izatio

ns, i

ndiv

idua

ls, a

ssoc

iatio

ns a

nd c

orpo

ratio

ns,”

whi

leth

e co

ncep

t of “

influ

ence

d” is

am

bigu

ous

and

shou

ld b

e re

plac

ed w

ith “d

irect

ed.”

“With

in o

r rel

atin

g to

Can

ada”

: “Th

ere

are

no c

riter

ia s

et o

ut in

the

Act

to

help

dete

rmin

e ho

w m

uch

any

part

icul

ar a

ctiv

ity m

ust

‘rela

te’ t

o Ca

nada

bef

ore

CSIS

can

take

juris

dict

ion,

cre

atin

g a

requ

irem

ent

that

may

be

too

easi

ly m

et.”

“Cla

ndes

tine

or d

ecep

tive”

: “Th

e pr

ecise

mea

ning

of t

he te

rm ‘c

land

estin

e’ is

unc

er-

tain

. It m

ay c

onno

te a

n el

emen

t of u

nder

hand

edne

ss o

r mal

e fid

es, b

ut s

ome

dic-

tiona

ry d

efin

ition

s w

ould

sup

port

an

inte

rpre

tatio

n th

at m

erel

y ‘se

cret

’ act

iviti

esm

ay b

e ‘cl

ande

stin

e.’ T

he te

rm s

houl

d be

repl

aced

with

a w

ord

like

‘surr

eptit

ious

,’w

hich

mor

e cl

early

con

note

s so

me

elem

ent o

f und

erha

nded

beh

avio

ur.”

“Det

rimen

tal t

o th

e in

tere

sts o

f Can

ada”

: The

phr

ase

is “

who

lly s

ubje

ctiv

e” a

s “n

ocr

iteria

are

pro

vide

d to

off

er a

ny s

tand

ard

for

dete

rmin

ing

wha

t is

‘det

rimen

tal.’”

It sh

ould

, the

refo

re, b

e de

fined

in t

he A

ct.

“Invo

lve

a th

reat

to a

ny p

erso

n”: T

he t

erm

“th

reat

” sh

ould

be

mod

ified

by

anad

ject

ive

like

“ser

ious

.”

“Pol

itica

l, re

ligio

us o

r ide

olog

ical

obj

ectiv

e”: T

he r

efer

ence

to

“pol

itica

l, re

ligio

usor

ideo

logi

cal o

bjec

tive”

was

add

ed t

o th

e CS

IS A

ctby

the

200

1 An

ti-te

rror

ismAc

t.3Th

e ph

rase

“po

litic

al, r

elig

ious

or

ideo

logi

cal o

bjec

tive”

whe

n us

ed in

the

cont

ext

of t

he C

rimin

al C

ode’

s de

finiti

on o

f “t

erro

rist

activ

ity”

was

dec

lare

dun

cons

titut

iona

l by

a lo

wer

cou

rt in

200

6. In

the

wak

e of

tha

t de

cisi

on, j

uris

tssp

ecul

ated

tha

t th

e eq

uiva

lent

phr

ase

in t

he C

SIS

Actm

ight

ren

der

info

rmat

ion

colle

cted

by

CSIS

con

stitu

tiona

lly s

uspe

ct, i

f em

ploy

ed in

a s

ubse

quen

t cr

imin

alpr

osec

utio

n.4

To m

inim

ize

the

pros

pect

of

ethn

ic p

rofil

ing,

the

spe

cial

Sen

ate

com

mitt

ee o

n an

titer

roris

m la

w r

ecom

men

ded

its r

epea

l in

its 2

007

repo

rt.5

SIRC

rec

omm

ende

d re

peal

of

this

pro

visi

on, u

rgin

g th

at it

pre

sent

ed t

he g

reat

est

risk

in a

dem

ocra

cy a

nd t

hat

its c

ore

cont

ent

— a

void

ing

polit

ical

vio

lenc

e —

isal

read

y co

vere

d in

the

oth

er p

arag

raph

s.

Espi

onag

e: “

Activ

ities

con

duct

ed f

or t

he p

urpo

se o

f ac

quiri

ng b

y un

law

-fu

l or

unau

thor

ized

mea

ns in

form

atio

n or

ass

ets

rela

ting

to s

ensi

tive

polit

ical

, eco

nom

ic, s

cien

tific

or

mili

tary

mat

ters

, or

for

the

purp

ose

ofth

eir

unau

thor

ized

com

mun

icat

ion

to a

for

eign

sta

te o

r fo

reig

n po

litic

alor

gani

zatio

n.”

Sabo

tage

: “Ac

tiviti

es c

ondu

cted

for

the

pur

pose

of

enda

nger

ing

the

safe

ty,

secu

rity

or d

efen

ce o

f vi

tal p

ublic

or

priv

ate

prop

erty

, suc

h as

inst

alla

tions

,st

ruct

ures

, equ

ipm

ent

or s

yste

ms.”

“Act

iviti

es d

etrim

enta

l to

the

inte

rest

s of

Can

ada,

and

whi

ch a

re d

irect

ed,

cont

rolle

d, f

inan

ced

or o

ther

wis

e si

gnifi

cant

ly a

ffec

ted

by a

for

eign

sta

te o

ror

gani

zatio

n, t

heir

agen

ts o

r ot

hers

wor

king

on

thei

r be

half.

”

“Thr

eat

or a

cts

of s

erio

us v

iole

nce

may

con

stitu

te a

ttem

pts

at c

ompe

lling

the

Cana

dian

gov

ernm

ent

to r

espo

nd in

a c

erta

in w

ay. A

cts

of s

erio

us v

io-

lenc

e ca

use

grav

e bo

dily

har

m o

r de

ath

to p

erso

ns, o

r se

rious

dam

age

to o

rth

e de

stru

ctio

n of

pub

lic o

r pr

ivat

e pr

oper

ty, a

nd a

re c

ontr

ary

to C

anad

ian

law

or

wou

ld b

e if

com

mitt

ed in

Can

ada.

”

“Act

iviti

es in

tend

ed t

o un

derm

ine

or o

vert

hrow

Can

ada'

s co

nstit

utio

nally

esta

blis

hed

syst

em o

f go

vern

men

t by

vio

lenc

e. S

ubve

rsiv

e ac

tiviti

es s

eek

toin

terf

ere

with

or

ultim

atel

y de

stro

y th

e el

ecto

ral,

legi

slat

ive,

exe

cutiv

e,ad

min

istr

ativ

e or

judi

cial

pro

cess

es o

r in

stitu

tions

of

Cana

da.”

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

12

are special rules limiting access to certain sensitiveinformation during these proceedings. For the mostpart, the decision on whether to disclose this sensitiveinformation is in the hands of the Federal Court.However, the Act empowers the attorney general toissue a certificate “in connection with a proceedingfor the purpose of protecting information obtained inconfidence from, or in relation to, a foreign entity asdefined in subsection 2(1) of the Security ofInformation Act or for the purpose of protectingnational defence or national security.”45

The minister’s certificate decision may be challengedin only a perfunctory manner, before a single judge ofthe Federal Court of Appeal. The role of this judge issimply to determine that the information covered bythe certificate “relates to” the permissible grounds forissuing a certificate, in which case the judge must con-firm the certificate.46 The concept of “relates to” isambiguous. Indeed, in 2007, a special Senate committeeexamining antiterrorism law recommended that the Actbe amended to “specify the way in which informationmust ‘relate to’ information obtained in confidencefrom a foreign entity, or to national defence or nationalsecurity, in order for that aspect of the certificate to beconfirmed by a judge” (Special Senate Committee onthe Anti-terrorism Act 2007, 65), and that the judge beempowered to consider “whether the public interest indisclosure outweighs in importance the public interestin non-disclosure” (67).

Implications for the Access Act Although thesepowers have not yet been used, amendments introducedto the Access Act by the ATA give attorney general’scertificates clear primacy over the right to access. Theydo so by creating a new exclusion. Section 69.1 nowspecifies that the Access Act “does not apply” toinformation covered by a CEA certificate issued beforean access complaint is filed with the informationcommissioner and, if issued after a complaint, quashesall proceedings in relation to that complaint.47 The effectof a certificate is, therefore, to preclude release of theinformation and deny the information commissioner arole in assessing whether nondisclosure is warranted.

Limiting open courtsAs discussed above, the Canadian judicial systemguarantees open and transparent adjudication of dis-putes in the form of open courts – with ready accessto the information at issue in them – and full disclo-sure to litigants, especially the accused in criminalmatters. National security does not negate either the

Conduct of international affairs, defence orthe prevention of subversion Under section 15 –an exception whose equivalent in the Privacy Actthe Supreme Court of Canada has labelled a “nationalsecurity”38 exemption – the government may refuseto disclose any record requested under the Act “thatcontains information the disclosure of which couldreasonably be expected to be injurious to the conductof international affairs, the defence of Canada or anystate allied or associated with Canada or thedetection, prevention or suppression of subversive orhostile activities.”39

While “international affairs” is not defined, theexpression “defence of Canada or any state allied orassociated with Canada” is limited to efforts by Canadaand foreign states “toward the detection, prevention orsuppression of activities of any foreign state directedtoward actual or potential attack or other acts ofaggression against Canada or any state allied or associ-ated with Canada.”40 Meanwhile, the expression “sub-versive or hostile activities” is also carefully delimited.41

National security exclusionRead together, the Access Act exemptions provide gov-ernment with substantial power to shield nationalsecurity secrets from the effects of the Act. It is notablethat the security and intelligence community itselfapparently had few quibbles with the scope of theAccess Act exemptions. In an admittedly now datedand pre-9/11 study prepared for the government’sAccess to Information Review Task Force, Wesley Warkreported that “both the Canadian Security andIntelligence Service and the Communications SecurityEstablishment, the two main collectors of sensitiveintelligence in the community, regard the Access Actas offering sufficient protection.” Given the breadth ofthese exemptions, Wark labels access to contemporaryintelligence records under the Act “a fiction,” and con-cludes that “the current Access exemptions providepowerful and sufficient tools” for protecting intelli-gence information (2001, section 2).

Yet, notwithstanding the breadth of long-standingCanadian exemptions from Canada’s access statutes,the government moved to enhance its power to keepinformation secret in its 2001 Anti-terrorism Act(ATA).42 Specifically, since 2001, the CEA43 now hasa central place in government secrecy law.

Canada Evidence Act amendments The CEA’sprimary purpose is to set out several evidentiarystandards for “proceedings.”44 Among its provisions

13

Ca

na

da

’s Na

tion

al S

ec

urity

“Co

mp

lex

”: Asse

ssing

the

Se

cre

cy

Ru

les, b

y C

raig

Fo

rce

seThe Canada Evidence Act The Canada Evidence

Act (CEA) is the most important statute governingsecrecy in court proceedings. Amendments made by the2001 ATA to section 38 of the CEA contain specialrules limiting access to “potentially injuriousinformation” and “sensitive information” in“proceedings,” including criminal trials.50

The concepts of potentially injurious informationand sensitive information are broadly defined: poten-tially injurious information means “information of atype that, if it were disclosed to the public, could injureinternational relations or national defence or nationalsecurity,” whereas sensitive information means “infor-mation relating to international relations or nationaldefence or national security that is in the possession ofthe Government of Canada, whether originating frominside or outside Canada, and is of a type that theGovernment of Canada is taking measures to safe-guard.”51 As this language suggests, sensitive informa-tion need not be of a sort that if released could causeinjury; it must merely “relate to” international rela-tions, national defence or security.52

Participants in a proceeding must notify the attor-ney general when they intend (or believe another par-ticipant or person intends) to disclose these classes ofinformation. The attorney general may then authorizedisclosure or, alternatively, may deny this authoriza-tion, in which case the matter is taken up by theFederal Court.

In adjudicating the section 38 dispute, the FederalCourt proceeds in three steps: first, it assesses whetherthe evidence is relevant to the proceeding in question;second, it determines whether disclosure would be inju-rious to international relations, national defence ornational security; and third, it determines whether thepublic interest in disclosure outweighs the public inter-est in nondisclosure.53 At the end of this process, thejudge may authorize disclosure (or not).

However, as already noted, the CEA allows the gov-ernment to short-circuit a court disclosure order. Theattorney general may issue a certificate “in connectionwith a proceeding for the purpose of protecting informa-tion obtained in confidence from, or in relation to, a for-eign entity as defined in subsection 2(1) of the Securityof Information Act or for the purpose of protectingnational defence or national security.”54 Issuance of thecertificate has the effect of barring any subsequent dis-closure of the information in a proceeding for 15 years(and for a further period if the certificate is renewed atthe end of that 15 years). In other words, the certificatemay reverse an order from the Federal Court authorizing

open court principle or rules on disclosure. It may,however, limit the reach of these doctrines.

Notably, the potential conflict between nationalsecurity confidentiality and the open court principlehas become more acute in the last several years, asthe dominant focus of security intelligence has shift-ed from the actions of sovereign states to an era ofcounterterrorism. The intelligence activities requiredto combat the interests of potential state adversariesare very different from those associated with coun-tering a diffuse, more autonomous and more individ-ualized terrorist movement. Counterintelligenceagainst states occurs within the broader envelope ofinternational relations; not least, foreign state agentsmay be accredited diplomats, and those exposed asspies are expelled rather than prosecuted.

Terrorists are, first and foremost, criminals, and unlessa state is prepared to take extralegal action against them,the response to terrorists among the populace is a crimi-nal law one, as bolstered by administrative law mecha-nisms like immigration proceedings. Legal responsesrequire appropriate regard for due process, whichensnares intelligence agencies in a judicial process towhich they are unaccustomed, producing what CSISdirector Jim Judd has called the “judicialization” of intel-ligence (2008). The net result is acute tensions betweenthe classic secrecy doctrines of the intelligence world andthe overt transparency of the judicial process.

Closing courts for national security reasonsIn Re Vancouver Sun, the Supreme Court observedthat even in national security matters “the open courtprinciple is a fundamental characteristic of judicialproceedings, and that it should not be presumptivelydisplaced in favour of an in camera process.”48 Thequestion of closed proceedings must be left in thehands of judges, not mandated in advance by statute.Further, where judges exercise that discretion, theymust balance the national security impulse againstthe public interest in openness, and not simply giveprimacy to one principle over the other. A constitu-tional doctrine, this approach has been “read into”several existing statutory rules for closing courts onnational security grounds.49

Limiting disclosure of evidence for national securityreasonsRestrictions on full disclosure of information by thegovernment to other parties in court proceedings arealso permissible in Canada. These rules on secret evi-dence are summarized in table 3.

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

14

To compel trials to proceed even where the only evi-dence available to establish the accused’s innocenceis withheld from him or her would be an unquestion-able violation of section 7 of the Charter of Rightsand Freedoms.56

The Act sidesteps this possible clash between legalrights and the state’s secrecy preoccupation by providingcriminal trial judges with an escape from the dilemma:

disclosure, subject to a limited appeal before a singlejudge of the Federal Court of Appeal.55

The system established by the Act would be vul-nerable to constitutional attack in those circum-stances where a person’s innocence in a criminal trialcan only be proven by evidence that the FederalCourt refused to disclose, or that is subject to anattorney general’s certificate barring that disclosure.

Table 3Secret Evidence Rules in Canadian Law

Context Basis Application Scope Limitation

1 R. v. Leipert, [1997] 1 S.C.R. 281 at para. 21. 2 Ibid. at para. 18.3 R. v. McClure, 2001 S.C.C. 14 at para. 45. 4 Canada Evidence Act, s. 38. For a discussion of the scope of s. 38 in relation to

the earlier doctrine of “public interest immunity,” see Stewart (2003).5 Immigration Refugee Protection Act, s. 83(1)(e). 6 Criminal Code, s. 83.05.7 Criminal Code, para. 83.05(6)(a). 8 Charities Registration (Security Information) Act, S.C. 2001, c. 41, s. 113, s.6.9 Regulations Implementing the United Nations Resolutions on the Suppressionof Terrorism, SOR/2001-360, s.2.2.

Informerprivilege

Section 38

Securitycertificates

Terrorist grouplisting

Terrorist-financingcertificate

Terrorism-financingregulations

Common law

CanadaEvidence Act

Immigrationand RefugeeProtection Act

Criminal Code

CharitiesRegistration(SecurityInformation)Act

United NationsAct

Typically criminalproceedings

All “proceedings,”including criminaltrials4

Immigration proceedingsin which the govern-ment issues a securitycertificate (as well as apossible use in regularremoval proceedings)

Listing of groups as“terrorist groups” underthe Criminal Code,6

appealable before ajudge.

Issuance of a certificateattesting that a charity(or applicant for charita-ble status) is involved interrorist financing, adju-dicated before a judge

Listing of a personbelieved to be affiliatedor associated with ter-rorist activity, appeal-able before a judge

The informer privilege rule“prevents not only disclosureof the name of the informant,but of any information whichmight implicitly reveal his orher identity.”2

Allows government to refusedisclosure of “potentially injuri-ous” and “sensitive” information.

A judge may refuse to discloseanything that, in the judge’sopinion, would be injurioustnational security or endangerthe safety of any person ifdisclosed.

A judge may, at the request ofthe attorney general, hear allor part of the government’sevidence or information in theabsence of the applicant andany counsel representing theapplicant, “if the judge is ofthe opinion that the disclosureof the information wouldinjure national security orendanger the safety of anyperson.”7

See entry for “Terrorist grouplisting” above.8

See entry for “Terrorist grouplisting” above.

Informer privilege is subject to the “innocence at stake”exception; that is, secrecy will give way where “disclo-sure of the informer’s identity is necessary to demon-strate the innocence of the accused.”1 Where informerprivilege yields to the innocence at stake doctrine, “theState then generally provides for the protection of theinformer through various safety programs.”3

A federal court judge decides whether the public inter-est in disclosure outweighs the public interest innondisclosure

The interests of persons subject to security certificatesare now defended by "special advocates" entitled tosee all the evidence, but who may only have limitedcommunication with the person whose interest theyare defending.5

The judge must provide a summary of the confidentialinformation.

The judge must provide a summary of the confidentialinformation.9

Same as terrorist group listing process above.

15

Ca

na

da

’s Na

tion

al S

ec

urity

“Co

mp

lex

”: Asse

ssing

the

Se

cre

cy

Ru

les, b

y C

raig

Fo

rce

seis often so oblique as to be useless in understanding thegovernment’s case. This problem becomes acute – andindeed a constitutional matter – when the risk to theinterested person is taken into account: detentionsunder security certificates have been lengthy,61 and sev-eral of the most recent cases have involved individualssaid to present terrorist threats who may well be tor-tured (or worse) if returned to their countries of origin.

In its 2007 Charkaoui decision,62 the Supreme Court con-cluded that the secrecy in the security certificate processviolated the Charter, underscoring that “before the state candetain people for significant periods of time, it must accordthem a fair judicial process.”63 In particular, the “magistratemust make a decision based on the facts and the law”64 and“the affected person be informed of the case against him orher and be permitted to respond to that case.”65 The Courtproposed a special counsel model as a possible solution tothe conundrum of national security confidentiality and fair-ness in security certificate proceedings.

Criminalizing wrongful disclosureThe constraints on open government and open courtsdescribed to this point limit what information may bedisclosed on national security grounds under rulesfavouring full disclosure. A third plank in Canada’snational security information law relates to punishingthose who wrongfully release (and potentially receive)sensitive information. To grapple with this prospect,Canadian law includes a number of information disclo-sure offences. First, it criminalizes unauthorized disclo-sure of sensitive information. Second, it proscribesphysical spying – that is, the physical infiltration of sen-sitive locations. The SOIA66 is the key statute creatingboth of these types of offence. In this paper I focus on itsfirst objective — criminalizing unauthorized disclosure.

Background to the Security of Information ActOriginally enacted in 1939 as the Official Secrets Act,67

the SOIA was substantially amended and renamed inDecember 2001, as part of the ATA. The 1939 Act was avariant on the United Kingdom’s 1889 Official SecretsAct, and it had two main focuses. First, in section 3, itmade espionage or spying an offence, and second, insection 4, it criminalized wrongful dissemination ofinformation, sometimes called “leakage” (CanadianSecurity Intelligence Service 2004).

From the 1960s or earlier, this wartime statute wasroundly condemned for its breadth and ambiguity. The1969 report of the Royal Commission on Security (theMackenzie Commission) called the 1939 law “an unwieldystatute, couched in very broad and ambiguous language”

The person presiding at a criminal proceed-ing may make any order that he or she con-siders appropriate in the circumstances toprotect the right of the accused to a fair trial,[other than ignoring a Federal Court determi-nation on disclosure or a Attorney General’scertificate].

Among the permissible orders are those:(a) dismissing specified counts of the indict-ment or information, or permitting theindictment or information to proceed only inrespect of a lesser or included offence…(b) effecting a stay of the proceedings and…(c) finding against any party on any issuerelating to information the disclosure ofwhich is prohibited.57

Immigration proceedings The CEA process isquite different from that applied in the controversialimmigration security certificate process under theImmigration Refugee Protection Act (IRPA).58 Thatprocess allows the government to issue a certificateagainst a foreign national or permanent resident,seeking that person’s removal (and detention pendingremoval) on national security grounds, among otherthings. The reasonableness of this certificate (and theduration of any detention pending removal) isadjudicated before a Federal Court judge. Thegovernment is entitled, however, to presentinformation to that judge in closed sessions andwithout the person subject to the certificate (and hisor her lawyer) being present. Moreover, the personreceives only a summary of this secret information,sometimes of the most general sort.59

The criterion for what is released is based exclu-sively on the national security (or public safety) pre-occupation: in providing the summary,

the judge shall ensure that the permanentresident or foreign national is provided witha summary of information and other evi-dence that enables them to be reasonablyinformed of the case made by the Minister inthe proceeding but that does not includeanything that, in the judge’s opinion, wouldbe injurious to national security or endangerthe safety of any person if disclosed. [empha-sis added]60

Unlike the CEA, under the IRPA there is no balanc-ing of the secrecy interest against the fair trial impera-tive. Nor is there any fair trial protection (also presentin the CEA) authorizing a judge to throw out the gov-ernment case if disclosure to the interested person isinsufficient to meet basic fairness expectations.

The security certificate process obviously con-strains the interested person’s ability to meet the caseagainst him or her; the information that is provided

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

16

impose upon a citizen. A reading of ss.3 and4 of the Official Secrets Act amply demon-strates its failure to do so; the provisions areambiguous and unwieldy…A completeredrafting of the Canadian Official SecretsAct seems appropriate and necessary.70

Unauthorized disclosureIn fact, the ATA subsequently repealed and replacedthe much-criticized 1939 espionage provision. Thecurrent SOIA now includes a list of offences underthe heading “Special Operational Information andPersons Permanently Bound to Secrecy.” Mostnotably, persons employed at a number of securityand intelligence government agencies are deemed (orare named by the government as) permanently boundto secrecy.71 As table 4 suggests, these persons arecriminally liable for the communication of “specialoperational information” subject to a public interestdefence in limited circumstances. “Special operationalinformation” is defined as and basically meansmilitary- and intelligence-related information thatthe government seeks to “safeguard.”72 The 2001changes also introduced other provisions, governingother, more specific forms of unauthorized disclosure.These SOIA provisions supplement sections of theCSIS Act that criminalize unauthorized disclosure ofinformation from which the identity of any CSISinformant or operative may be discerned. Theseoffences are summarized in table 4.73

The new SOIA provisions might reasonably becritiqued for the breadth and uncertainty of the“special operational information” concept. At least,however, there is a definition — something notablefor its absence in the section 4 antileakage provi-sion. Most critically, the public interest override isan important feature, allowing secrecy to give wayin exigent circumstances. Still, that overridedeserves renovation: except where necessary toavoid grievous bodily harm, the public interestdefence exists only when two prerequisites are met.First, prior to disclosure, the whistle-blower musthave provided all relevant information to his or herdeputy head or the deputy attorney general ofCanada and have received no response within a rea-sonable time. Subsequently, the whistle-blower musthave also provided the information to the SIRC or,where the alleged offence concerns the CSEC, to theCSE commissioner, and must have not received aresponse within a reasonable time.

The Act leaves open the question of what wouldconstitute a “reasonable time.” Likewise, it does not

(1969, para. 204). In 1986, the Law Reform Commissioncondemned the statute as “one of the poorest examplesof legislative drafting in the statute books” (Law ReformCommission of Canada 1986, 30). It called the OfficialSecrets Act and other laws that criminalized “crimesagainst the state” as “out of date, complex, repetitive,vague, inconsistent, lacking in principle and over-inclu-sive,” as well as potentially unconstitutional under theCharter of Rights and Freedoms (38-9).

In particular, the commission took issue with sec-tion 3 of the Act, relating to spying, which could beinterpreted as imposing an onus of proving inno-cence on the accused. This reverse onus, the commis-sion speculated, was inconsistent with subsection11(d) of the Charter, which guarantees the presump-tion of innocence until proven guilty (39). The gov-ernment itself criticized the statute in 1998, when thethen solicitor general called the Act “badly outdatedand overbroad.”68

Perhaps for these reasons, the Official Secrets Acthas rarely been invoked. CSIS reports that since 1939there have been two dozen prosecutions under theAct, but only six in the past 40 years (2004). In oneof these cases, Stephen Ratkai pleaded guilty in 1989to charges under the espionage provisions of thestatute for spying for the USSR. In sentencing Ratkaito two concurrent terms of nine years, the court com-mented that the object of the Official Secrets Act is“to protect the safety and interests of the state. Everycountry has an obligation to protect its citizens andits territory and countries must depend and rely uponits citizens to ensure [their] safety and security. Whatis disturbing and despicable about offences of thisnature is that a citizen betrays his country which hehas a duty to protect and defend.”69

However, in R. v. Toronto Sun – probably the lead-ing case on the Official Secrets Act – the court wasmoved much less by the Act’s objectives than by itsawkward structure. At issue in this pre-Charter casewas whether a newspaper and its editors had violatedthe Act by printing excerpts of a top secret documentabout Soviet intelligence activities in Canada. Thecourt concluded that they had not, as the allegedlysecret information had already been invoked in thepublic domain. However, the court was also critical ofthe Act itself. In the court’s words,

since the Official Secrets Act is a restrictingstatute, and seeks to curb basic freedoms,such as freedom of speech and the press, itshould be given strict interpretation…Thestatute must, in clear and unambiguous lan-guage, articulate the restriction it intends to

17

Ca

na

da

’s Na

tion

al S

ec

urity

“Co

mp

lex

”: Asse

ssing

the

Se

cre

cy

Ru

les, b

y C

raig

Fo

rce

se

information. This omission provoked substantial con-troversy soon afterward.

Criticisms of the leakage provision As withthe now repealed section 3, dealing with espionage,the precise scope of section 4 of the 1939 OfficialSecrets Act is difficult to discern from the drafting ofthe section itself. In Keable v. Canada (Attorney-General),74 the Supreme Court held that “Section 4 ofthe Official Secrets Act makes it clear that it is theduty of every person who has in his possessioninformation entrusted in confidence by a governmentofficial and subject to the Act, to refrain fromcommunicating it to any unauthorized person.”

address whether the public interest defence wouldapply were the responses received from thesereview bodies inadequate. Indeed, it does not spellout exactly how the review bodies are to respond tothe disclosures, a point made by the Special SenateCommittee on the Anti-terrorism Act in 2007.

Criminalized leakageThe ATA was notable as much for what it did not doas for what it changed in SOIA. The 2001 amend-ments left intact section 4, criminalizing leakage; thatis, the simple unauthorized disclosure (or, in somecases, receipt) of secret (or sometimes merely official)

Table 4Information Offences under the Security of Information Act (SOIA) and the CSIS Act

Type of information Person to whom prohibition applies Prohibition in the offence

1 The economic espionage offence in section 19 is constrained in subsection 19(3) by certain defences protecting independent development of trade secrets or reverse engineering.

“Special operationalinformation”

Information the govern-ment is “taking measuresto safeguard”

Any information obtainedor accessed in the courseof the performance ofduties and functionsunder this CSIS Act orparticipation in theadministration orenforcement of the Act

Trade secrets

Persons permanently bound by secrecy

Every person

Every person with a security clearance given bythe Government of Canada

Every person

Every person

Every person

Intentionally and without authority communicates or confirms infor-mation that, if it were true, would be special operational information(SOIA, s. 13)

Intentionally and without authority communicates or confirms specialoperational information (SOIA, s. 14)

Intentionally and without lawful authority communicates specialoperational information to a foreign entity or to a terrorist group ifthe person believes (or is reckless as to whether) the information isspecial operational information (SOIA, s. 17)

Intentionally and without lawful authority communicates, or agrees tocommunicate, to a foreign entity or terrorist group any informationthat is of a type that the Government of Canada is taking measures tosafeguard (SOIA, s. 18)

Without lawful authority communicates to a foreign entity or terroristgroup information while believing (or reckless as to whether) thatinformation is safeguarded and in order to increase the capacity ofthat foreign entity or terrorist group to do harm to Canadian interests(SOIA, s. 16)

Without lawful authority communicates to a foreign entity or to aterrorist group information while believing (or reckless as to whether)that information is safeguarded, and harm to Canadian interestsresults (SOIA, s. 16)

Discloses this information and from it the identity of “(a) any otherperson who is or was a confidential source of information or assis-tance to the Service,” or “(b) any person who is or was an employeeengaged in covert operational activities of the Service” can be inferred(CSIS Act, s. 18)

At the direction of, for the benefit of or in association with a foreigneconomic entity, fraudulently and without colour of right, communi-cates a trade secret to another person or organization or obtains,retains, alters or destroys a trade secret “to the detriment of” Canada’seconomic interests, international relations or national defence ornational security1

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

18

Tabl

e 5

The

Leak

age

Prov

isio

n in

Sec

tion

4 of

the

Sec

urity

of

Info

rmat

ion

Act

(SO

IA)

Type

of

info

rmat

ion

Pers

on t

o w

hom

pro

hibi

tion

appl

ies

Proh

ibiti

on in

the

off

ence

Any

secr

et o

ffic

ial c

ode

wor

d, p

assw

ord,

sket

ch, p

lan,

mod

el, a

rtic

le, n

ote,

doc

umen

t or

info

rmat

ion

that

rel

ates

to

or is

use

d in

a p

ro-

hibi

ted

plac

e

Any

secr

et o

ffic

ial c

ode

wor

d, p

assw

ord,

sket

ch, p

lan,

mod

el, a

rtic

le, n

ote,

doc

umen

t or

info

rmat

ion

Any

sket

ch, p

lan,

mod

el, a

rtic

le, n

ote,

doc

u-m

ent

or in

form

atio

n th

at r

elat

es t

o m

uniti

ons

of w

ar

Any

offic

ial d

ocum

ent

Ever

y pe

rson

in p

osse

ssio

n

Ever

y pe

rson

rec

eivi

ng t

his

info

rmat

ion

Ever

y pe

rson

in p

osse

ssio

n

•Pe

rson

for

who

se u

se t

he in

form

atio

nw

as is

sued

•Ev

ery

pers

on

•Fa

ils “

to t

ake

reas

onab

le c

are

of,”

or e

ndan

gers

the

“sa

fety

of”

the

info

rmat

ion

(SOI

A. p

ara.

4[1

][d]1

•Co

mm

unic

ates

the

info

rmat

ion

“to

any

pers

on, o

ther

tha

n a

pers

on t

o w

hom

he

[or

she]

is a

utho

rized

to

com

mun

icat

ew

ith, o

r a

pers

on t

o w

hom

it is

in t

he in

tere

st o

f th

e St

ate

his

[or

her]

duty

to

com

mun

icat

e it”

(SOI

A, p

ara.

47[

1][a

])•

Use

s th

e in

form

atio

n “f

or t

he b

enef

it of

any

for

eign

pow

er o

r in

any

oth

er m

anne

r pr

ejud

icia

l to

the

safe

ty o

r in

tere

sts

of t

he S

tate

” (S

OIA,

par

a. 4

[1][b

])•

Reta

ins

the

info

rmat

ion

in t

he a

bsen

ce o

f a

“rig

ht t

o re

tain

it o

r w

hen

it is

con

trar

y to

[the

rec

eivi

ng p

erso

n’s]

dut

y to

reta

in it

or

[he

or s

he] f

ails

to

com

ply

with

all

dire

ctio

ns is

sued

by

law

ful a

utho

rity

with

reg

ard

to t

he r

etur

n or

dis

posa

lth

ereo

f” (S

OIA,

par

a. 4

[1][c

])

Know

s, or

has

rea

sona

ble

grou

nd t

o be

lieve

, at

the

time

he o

r sh

e re

ceiv

es it

, tha

t th

e in

form

atio

n is

com

mun

icat

ed t

o hi

mor

her

in c

ontr

aven

tion

of t

his

act,

unle

ss h

e or

she

pro

ves

that

the

com

mun

icat

ion

of t

he in

form

atio

n w

as c

ontr

ary

to h

is o

rhe

r de

sire

(SOI

A, s

ubs.

4[3]

)

Com

mun

icat

es t

he in

form

atio

n to

“any

for

eign

pow

er”

or “

in a

ny o

ther

man

ner

prej

udic

ial t

o th

e sa

fety

or

inte

rest

s of

the

Stat

e” (S

OIA,

sub

s. 4[

2])

•Al

low

s an

y ot

her

pers

on t

o ha

ve p

osse

ssio

n of

the

doc

umen

t (S

OIA,

par

a. 4

[4][b

])•

Has

pos

sess

ion

“with

out

law

ful a

utho

rity

or e

xcus

e” o

f “a

ny o

ffic

ial d

ocum

ent…

issu

ed f

or t

he u

se o

f a

pers

on o

ther

than

” hi

m o

r he

rsel

f (S

OIA,

par

a. 4

[4][b

])•

Upo

n “o

btai

ning

pos

sess

ion

of a

ny o

ffic

ial d

ocum

ent

by f

indi

ng o

r ot

herw

ise,

” fa

ils t

o “r

esto

re it

to

the

pers

on o

rau

thor

ity b

y w

hom

or

for

who

se u

se it

was

issu

ed, o

r to

a p

olic

e co

nsta

ble”

(SOI

A, p

ara.

4[4

][b])

•Re

tain

s “f

or a

ny p

urpo

se p

reju

dici

al t

o th

e sa

fety

or

inte

rest

s of

the

Sta

te”

an “

offic

ial d

ocum

ent,

whe

ther

or

not

com

-pl

eted

or

issu

ed f

or u

se, w

hen

he h

as n

o rig

ht t

o re

tain

it, o

r w

hen

it is

con

trar

y to

his

dut

y to

ret

ain

it” o

r in

con

trav

en-

tion

of in

stru

ctio

ns f

rom

the

gov

ernm

ent

to r

etur

n or

dis

pose

of

it (S

OIA,

par

a. 4

[4][a

])

19

Ca

na

da

’s Na

tion

al S

ec

urity

“Co

mp

lex

”: Asse

ssing

the

Se

cre

cy

Ru

les, b

y C

raig

Fo

rce

seAnti-terrorism Act in 2007 (94ff). That body urgedmuch clearer definitions of the information protectedby the section and a public interest override where thepublic interest in disclosure exceeds the public interestin nondisclosure.

As it stands, the 2006 Ontario Superior Court ofJustice decision remains the final word on section 4. Itis critical to note, however, that this decision comesfrom a single lower court in one province. It wouldundoubtedly be considered by other such lower courts,were they presented with a question about the constitu-tionality of section 4. However, in Canada’s commonlaw system, the 2006 decision has no binding prece-dential effect. It could be rejected by another lowercourt, or by an appeal court. Put another way, section 4lurks on the statute books and remains potentially offorce and effect in any court other than the one thatdecided the Juliet O’Neill case.

Reconciling Transparency andNational Security Confidentiality

I now turn to assessing how well Canadian law andpractice reconcile transparency and national secu-rity confidentiality in both the open government

and open court contexts. Given the subject matter ofthis paper – government secrecy – this is no easy task.It is difficult to arrive at definitive conclusions aboutwhether government has properly balanced secrecy anddisclosure, since an external researcher does not haveaccess to the full range of information guarded onnational security grounds by the government.Conclusions depend, therefore, on anecdote and par-tially tested hypothesis. With that caveat, in this sec-tion I nevertheless contend that Canada’s record inmaximizing open government and courts while pro-tecting national security confidentiality is less thanfully satisfactory. The Canadian approach suffers fromboth a failure of design and serious difficulties inimplementation. Reform efforts, meanwhile, are bur-dened with their own shortcomings.

Difficulties of designOpen governmentThe already sufficient Access ActPut simply, the national security limitations on opengovernment are poorly planned and integrated. The mostthoughtful law is the Access Act. The national securityexceptions in this statute are fairly precisely defined,

However, the section is much broader in scope thanthis interpretation suggests.

Indeed, communication of information is criminal-ized in a fashion likely to render most public servicewhistle blowing a crime. As the Law ReformCommission noted in 1986, the Official Secrets Act“always treats the loquacious public servant and thesecret agent alike: both may be charged under thesame section (section 4), the punishment is the same,and, more importantly, the terrible stigma of prosecu-tion under the [Act] is identical for both, because thepublic and the news media are unable to discernwhether it is a case of calculated espionage or care-less retention of documents” (1986, 37).

So broadly crafted is section 4 of the SOIA that itwas difficult to imagine that the government would,for example, fail to secure convictions for the almostdaily “leaks” of written government information thatfill newspaper pages. More than that, it seems likelythat it could be used to secure the conviction of thejournalists and newspapers reporting these leaks. Theprecise parameters of section 4 are set out in table 5.

Constitutionality of the leakage provision Thehistorical absence of prosecutions brought undersection 4 of the SOIA likely reflected a soberappreciation of the political consequences flowingfrom aggressive uses of secrecy law, coupled with aprincipled realization by prosecutors and others thatthe section was unusable. However, a law of thisbreadth could be used to either threaten a prosecutionor obtain warrants, both of which are tactics that raisecivil liberties issues. Most notoriously, in January2004, the RCMP raided Ottawa Citizen reporter JulietO’Neill’s home and office looking for leakedinformation pertaining to Maher Arar, the Canadiandeported by US officials to Jordan and thenincarcerated and tortured in Syria. The warrant allegeda violation by O’Neill of section 4 of the SOIA.

That warrant, and the resulting search of O’Neill’shome and office, sparked a constitutional challengeto section 4. Specifically, O’Neill and the OttawaCitizen contended that section 4 violated section 2(b)of the Charter by infringing on the freedom of thepress to gather and disseminate information of publicinterest and concern, and it contravened section 7 ofthe Charter on the basis of vagueness and over-breadth. Those claims were accepted by the OntarioSuperior Court of Justice in 2006,75 a decision thegovernment chose not to appeal. Section 4 was alsocritiqued by the Special Senate Committee on the

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

20

security or the other justifications listed in the Act(whatever that might mean).

The added government secrecy muscle that theCEA grafts onto the Access Act is particularly trou-bling, given the failure to carefully define the nation-al security grounds justifying the issuance of acertificate. In other words, the careful attention todetail and balancing found in the Access Act isentirely circumvented by a CEA provision that pro-vides minimal guidance on when governments areempowered to issue certificates.

The then information commissioner considered the2001 amendments to be an unnecessary overreaction:“the Access to Information Act posed no risk of possi-ble disclosure of sensitive intelligenceinformation…no such information had ever been dis-closed under the Act in the 18 years of its lifeand…the Access to Information Act régime offered asmuch or more secrecy to intelligence information asdo the laws of our allies.”76 This conclusion is sup-ported by Wesley Wark’s assessment concerning thesufficiency of national security protection under theregular Access Act exemptions, noted above.

The unconstitutional antileakage provisionLurking in the background, meanwhile, is the SOIA,and its unconstitutional section 4. Under section 4 ofthat act, unauthorized disclosure of even nonsecretbut “official” government documents brings with itthe possibility of criminal prosecutions. Where thedocument is “secret” within the (undefined) meaningof the Act, the prospect of being found criminallyculpable increases and could include a situationwhere the document is shared internally within thegovernment itself. Further, since the Act extends to“persons” and not just civil servants, and because itcriminalizes receipt as much as disclosure, it makesleaked government information a “hot potato” thatmost risk-averse people would rather not receive. Thenet effect cannot be other than to chill the sharing ofinformation, even when there may be a clear publicinterest in disclosure and dissemination.

The risks posed by the inarticulate section 4 arereduced by the O’Neill decision, but not eliminated.O’Neill is merely persuasive in other courts, it is nottechnically binding. Section 4 remains on the statutebooks, and it should not be expected that every offi-cial inclined to seek (or issue) a warrant under thisprovision, or threaten a prosecution, will be deterredby (or necessarily aware of) the constitutional critiqueof a single Ontario lower court, especially as time

creating intelligible standards and not simply invoking“national security” or “secret” or some other murkyconcept. In many instances, the Act includes an injuryrequirement, precluding disclosure only where there issome deleterious impact on a usually defined nationalsecurity interest associated with the release of informa-tion. Moreover, almost all of the national securityexemptions are discretionary, which implies that inmaking decisions on disclosure the government wouldhave to balance the interests involved rather thanimpose a strict nondisclosure requirement.

The excessive Canada Evidence ActUnfortunately, the relative clarity of the Access Act isnot matched by the 2001 changes to the CanadaEvidence Act (CEA). First, as the discussion abovemakes evident, the CEA creates three different classesof information: potentially injurious information,defined as “information of a type that, if it were dis-closed to the public, could injure international rela-tions or national defence or national security”;sensitive information, defined as “information relat-ing to international relations or national defence ornational security” that the government is safeguard-ing; and, in the context of attorney general’s certifi-cates, “information obtained in confidence from, orin relation to, a foreign entity as defined in subsec-tion 2(1) of the Security of Information Act or for thepurpose of protecting national defence or nationalsecurity.” The first concept – potentially injuriousinformation – is generally consistent with the injury-based exemption in the Access Act. However, theother two concepts are significantly broader and veryuncertain as to their precise scope. In no case are theterms “international relations,” “national defence” or“national security” defined, leaving these concepts tobe decided by the eye of the beholder.

Second, while the Canada Evidence Act provisionsanticipate adjudication of government national secu-rity claims in court, and expressly enable the FederalCourt to balance public interests in arriving at out-comes, this process may be short-circuited by theissuance of an attorney general’s certificate. This cer-tificate not only quashes any decision by a FederalCourt judge to order the release of information underthe CEA, but it also may be employed to quash pro-ceedings under the Access Act. Subsequently, thegovernment’s judgment in issuing a certificate is sub-ject to scrutiny only via a limited appeal to a singlejudge at the Federal Court of Appeal, who considersonly if the government’s reasons “relate to” national

21

Ca

na

da

’s Na

tion

al S

ec

urity

“Co

mp

lex

”: Asse

ssing

the

Se

cre

cy

Ru

les, b

y C

raig

Fo

rce

seservant is liable if the information is of the sort thatdisclosure is “likely to cause such damage.”81 Ignoranceof the security and intelligence nature of the informa-tion is a defence, as is the reasonable absence of beliefthat disclosure would be damaging.

Parallel provisions regulating disclosure of informa-tion relating to “defence” and to “international rela-tions” are contained in sections 2 and 3 of the UK Act.The concepts of “defence” and “international relations”are both defined. Further, in both sections the disclo-sure is only an offence if it causes damage, a conceptspelled out in detail in each instance. A lack of knowl-edge (or reasonable belief as to such knowledge) of thesubject matter of the information is again a defence.

The UK Act also creates other offences for secondaryleaking of secrets by recipients of wrongfully leakeddocuments. Thus, a person who receives a documentrelating to defence or international relations commitsan offence under section 5 if he or she subsequentlydiscloses it, knowing or having reasonable cause tobelieve that the information is protected by section 2 or3. However, this subsequent disclosure must itself bedamaging and the person must know, or have reason-able cause to believe, the disclosure to be damaging.

Thus, unlike the draconian section 4 of the SOIA,the UK Act carefully defines the sorts of information atissue in the criminalization of disclosure. Also unlikethe Canadian law, it includes a requirement that disclo-sure of even this sensitive information be “damaging”(within the meaning of the Act) before there is criminalculpability.

Open courtsCanada’s open government laws are not alone in beinginternally incoherent. A similar lack of consistencyplagues Canada’s rules on disclosure in court proceed-ings. As noted, section 38 of the CEA’s definitions ofinformation that can be guarded on national securitygrounds are ambiguous. However, that Act at leastestablishes a balancing regime, in which national secu-rity interests can be weighed against, inter alia, thepublic interest in fair trials. That balancing is potential-ly truncated by the nuclear bomb of an attorney gener-al’s certificate (to date, never used), but a trial court incriminal matters retains the discretion to toss the gov-ernment’s case if the government refuses to discloseenough information to preserve a fair trial.

In comparison, the information disclosure regimeestablished by the security certificate system underthe IRPA has no such nuance. Information is withheldif necessary on national security grounds (without

passes. The recent DFAIT report on the Bernier-Couillard matter appears to be a striking case inpoint. As is now well known, then foreign affairsminister Maxime Bernier left classified briefing docu-ments at the home of his partner at the time — JulieCouillard. Apparently, because Couillard delayedreturning the documents, despite opportunities to doso, the DFAIT report declares that “Ms. Couillard mayhave put herself in potential jeopardy of having con-travened a provision of the Security of InformationAct” (Foreign Affairs and International Trade Canada2008). The only conceivable provision of the SOIA atissue in the Bernier-Couillard affair is section 4.77 Anofficial government report has, therefore, intimatedcriminal liability — and likely precipitated reputation-al and other consequences — apparently based on astatutory provision declared unconstitutional by anOntario trial court.

In sum, Canada is left with an antileakage provi-sion whose doubtful reliability in a prosecutionmakes it an uncertain, counterproductive but stillfrightening tool. In an era when Canada is anxiousabout its international credentials in the security andintelligence community, it is worth noting that theSOIA compares unfavourably to its closest equiva-lent, the UK Official Secrets Act of 1989.78 Certainly,in some respects, this UK law is less measured thanits Canadian counterpart. Thus, the UK Act doesinclude provisions that cover the security servicesand that are broadly equivalent to the Canadianstatute’s “persons permanently bound by secrecy”sections. Here, the UK Act is more unforgiving,imposing a blanket prohibition on unauthorized dis-closure of “any information, document or other arti-cle relating to security or intelligence” in the person’spossession by virtue of his or her employment in thesecurity services.79 There is no requirement that thedamage stem from the disclosure. Further, unlike theCanadian law, in the UK law there is no public inter-est exception. The sole defence anticipated by the Actis if the person did not know of the security or intel-ligence nature of the information.80

Yet, insofar as its other leakage provisions areconcerned, the UK Act is much more moderate (andintelligible) than section 4 of the SOIA. Thus, the UKAct makes it an offence, in section 1, for a civil ser-vant to disclose information relating to security orintelligence, but only if this disclosure is damaging.This damage is measured by any actual damage itcauses to “the work of, or of any part of, the securityand intelligence services.” Alternatively, that civil

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

22

compounded by differing “cultures” of secrecy versustransparency that evolve in different governmentdepartments in different eras. Discussing this problemin the American context as it existed in the Cold War,then senator Daniel Patrick Moynihan noted that “Topreserve an open society it was deemed necessary totake measures that in significant ways closed itdown” (United States, Commission on Protecting andReducing Government Secrecy 1997). In a “culture ofsecrecy,” the propensity is to overclassify, and poweris sometimes derived from withholding secrets andrevealing them only to a select few.

Establishing the effects of this postulated subjectiv-ity (and culture) is difficult. Anecdotal evidence tends,however, to point to its existence and importance. Forexample, critics have repeatedly complained of theDepartment of National Defence’s ready reliance onnational security exceptions under the Access Act. Insome instances, exemptions have been applied toinformation that is already publicly available.According to media reports from 2006, “In an exami-nation of 23 access requests made to the DefenceDepartment over the last 18 months, 87 pieces ofinformation, now censored, had been previouslyreleased to the public or are still on government andDefence Department websites” (Pugliese 2006). Thisobservation adds empirical substance to a chorus ofopinion accusing the current government of excessivesecrecy, especially in relation to Afghanistan.84 It alsoshows how mores and views on the need to censorparticular bits of information vary over time.

The “mosaic effect” and overclaimingAs a related issue, there is serious reason to believethat national security confidentiality produces “over-claiming” — that is, a propensity to assert secrecywhere it is not truly warranted. Again, proving a sys-tematic propensity to overclaim is next to impossible.Overclaiming has, however, been an acute issue in anumber of recent cases. Most famously, in his factualreport in the Arar inquiry, Justice O’Connor com-mented on the government’s propensity to overclaimas follows:

the public hearing part of the Inquiry couldhave been more comprehensive…if theGovernment had not, for over a year, assert-ed NSC [national security confidentiality]claims over a good deal of information thateventually was made public, either as aresult of the Government’s decision to reredact certain documents beginning in June2005, or through this report…This “over-claiming” occurred despite the government’sassurance at the outset of the inquiry that its

any effort to define this concept). There is no bal-ancing, and there is no possibility that a court candismiss the proceeding because the level of disclo-sure has been inadequate.

For the latter reason, the security certificate systemis out of line not just with the approach adopted bythe CEA, but also with that evolving in the UnitedKingdom in the wake of the House of Lords’ recentdecision of Secretary of State v. MB.82 In that deci-sion, the House of Lords examined the compatibilityof the UK system of “control orders” – limitations onliberty imposed on the basis that a person is suspect-ed of posing a terrorist threat – with the EuropeanConvention on Human Rights. Control orders pro-ceedings include the use of secret evidence. In rea-soning that will almost certainly affect UKnational-security-related immigration proceedings aswell, the law lords suggested that a residual discre-tion rests with the judge to determine whether thelevel of disclosure to the interested person suffices tomeet fair trial standards. Where the proceeding fallsshort of a fair hearing, the matter might come to anend, unless the government is prepared to make fullerdisclosure.83

Difficulties in executionThese shortcomings in design are augmented by acuteproblems in execution. Even with a thorough re-drafting of Canada’s secrecy laws, with ample atten-tion to definition and precision, problems ofexecution would almost certainly continue to plagueefforts to reconcile transparency with security.Indeed, problems exist in the Canadian system, eventhough one of the most critical structural shortcom-ings identified above — the attorney general’s certifi-cate under the CEA — has not been used to date.

The eye of the beholderThe single largest challenge in execution is probablythe problem of subjectivity. Information in the pos-session of government is, in the first (and often final)instance, assessed for disclosure by government offi-cials themselves. Even acting in the utmost goodfaith, individual officials will vary in their assess-ments of whether a given document contains, forinstance, information that could reasonably beexpected to be injurious to the defence of Canada oran allied state. This variability almost certainlyincreases as definitions of information to be withheldbecome more ambiguous; that is, simply invoking“national security.” Individual subjectivity may be

23

Ca

na

da

’s Na

tion

al S

ec

urity

“Co

mp

lex

”: Asse

ssing

the

Se

cre

cy

Ru

les, b

y C

raig

Fo

rce

se(among other things) that the government was aware ofthe American policy of “rendition” and the likelihoodthat Arar was being removed to Syria to be tortured.Commenting on the Arar experience, Kent Roach hasobserved:

The subsequent release of the majority of thedisputed passages in the Arar Commission’sreport affirmed for many that the governmenthad overclaimed NSC and used it to shieldinformation that was embarrassing to the gov-ernment and its allies but that did not protectthe lives of informers or ongoing operations.Such findings affect the government’s credibil-ity. There is a danger that the wolf of nationalsecurity confidentiality will be called toooften. (2008)

Indeed, the Arar case is apparently not exceptional.In several more recent CEA cases, the Federal Courthas disagreed with the government’s secrecy claimsand ordered information released. In Khawaja, forexample, Justice Mosley offered the following obser-vation about the government’s propensity to exciseevery reference to CSIS from material at issue in thatcriminal proceeding:

CSIS has acknowledged an investigative inter-est in the respondent up to his arrest on March29, 2004, assistance and involvement with sur-veillance of his movements and their presenceon the date of the RCMP search of his home.But those holding the black pens seem to haveassumed that each reference to CSIS must beredacted from the documents even where thereis no apparent risk of disclosure of sensitiveinformation such as operational methods orinvestigative techniques or the identity of theiremployees. I acknowledge that there can beinstances in which an informed reader willconnect the dots to obtain such informationbut that does not apply in every case.88

It may be that overclaiming is almost an inevitableoutcome of national security confidentiality — nogovernment official wishes to be the one to discloseinformation that ultimately proves prejudicial. Thenatural propensity may be, therefore, to err on theside of secrecy. However, it seems likely that thisunderstandable human instinct is compounded byofficial doctrines that encourage overclaiming.Justice Mosley’s reference to connected dots is nopassing reference. Chief among government secrecydoctrines is the so-called “mosaic effect,” a conceptthat has been invoked frequently by the Canadiangovernment.89

Put simply, the mosaic effect posits that the releaseof even innocuous information can jeopardize nationalsecurity if that information can be pieced together withother data by a knowledgeable reader. The result is amosaic of little pieces of benign information thatcumulatively discloses matters of true national security

initial NSC claims would reflect its “consid-ered” position and would be directed at max-imizing public disclosure. The Government’sinitial NSC claims were not supposed to bean opening bargaining position…It is perhapsunderstandable that, initially, officials choseto err on the side of caution in making NSCclaims. However, in time, the implications ofthat overclaiming for the Inquiry becameclear. I raise this issue to highlight the factthat overclaiming exacerbates the trans-parency and procedural fairness problemsthat inevitably accompany any proceedingthat can not be fully open because of NSCconcerns. It also promotes public suspicionand cynicism about legitimate claims by theGovernment of national security confiden-tiality. It is very important that, at the outsetof proceedings of this kind, every possibleeffort be made to avoid overclaiming.(Commission of Inquiry 2006, 301-2)

The Arar inquiry is not, of course, the only oneprobing government national security secrets. Othersinclude the 2006 Commission of Inquiry into theInvestigation of the Bombing of Air India Flight 182(Air India inquiry),85 and the 2006 Internal Inquiryinto the Actions of Canadian Officials in Relation toAbdullah Almalki, Ahmad Abou-Elmaati andMuayyed Nureddin (Iacobucci Commission).86

Anecdotally, I have the impression that the variousjudicial inquiries on national security matters haveexposed overclaiming, pressed the government logicfor secrecy in many instances and perhaps produceda slightly more critical internal assessment by gov-ernment of secrecy justifications. It is impossible tomeasure how meaningful this adjustment has been, orwhether it will have lasting impact. Nor is it clearthat a key countervail to this slight opening – thedeep secrecy generated by Canada’s military opera-tions in Afghanistan – will sweep away whatevergains were made through the era of inquiries.

At any rate, it is true that whatever impact pres-sure mounted on the government through the inquiryprocess had on government secrecy doctrine, govern-ment overclaiming persisted even in relation to thefinal report in the Arar matter. In a CEA proceeding,the Federal Court ultimately ordered disclosure of ahandful of additional lines that the governmentwished censored from the report. In the course of hisreasoning, Justice Noël noted that the “Court will notprohibit disclosure where the Government’s sole orprimordial purpose for seeking the prohibition is toshield itself from criticism or embarrassment.”87 Thecourt did not connect this statement to the informa-tion ultimately released. That information was, how-ever, embarrassing to the government, demonstrating

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

24

Some information is so innocuous that it strainsplausibility to maintain that it must be kept secret. InO’Neill v. Canada, for example, the RCMP resisteddisclosure on national security grounds of the loca-tion of an RCMP building, “even though it has anexterior sign indicating that it is an RCMP build-ing.”93 Moreover, the location of the building hadalready been disclosed in the Arar inquiry. Not sur-prisingly, the Ontario Superior Court of Justiceordered the release of the information in the O’Neillproceedings.

The Federal Court also appears attuned to the diffi-culties in applying the mosaic effect theory and hasnow expressed frank skepticism of the doctrine:“Witnesses from the intelligence community may takethe mosaic effect theory as an article of faith, relyingupon it as a complete answer to the release of infor-mation they consider sensitive or potentially harmful.As stated by Justice Noël in Arar…‘Simply allegingthe effect is not enough. There must be some basis orreality for such a claim based on the particulars of agiven file.’”94 Thus, “by itself the mosaic effect willusually not provide sufficient reason to prevent thedisclosure of what would otherwise appear to be aninnocuous piece of information. Something furthermust be asserted as to why that particular piece ofinformation should not be disclosed.”95

The “third party” trump cardAs noted above, secrets shared by allied governmentsare given robust protection in Canadian informationlaw. However, blanket prohibitions on disclosurebecause of simple foreign provenance give secrecylaw an extended reach in an intelligence-importingcountry like Canada. It is not likely that every pieceof foreign information must be held secret because itwould jeopardize the security interests of the provid-ing state. Guarding information simply because offoreign origin inevitably “launders” innocuous infor-mation and restricts the universe of disclosableinformation.

Recently, a number of court decisions have raisedquestions about the reach of the third party rule. InKhadr, Justice Mosley commented as follows: “it ismy view that too much of the routine communica-tions between foreign and Canadian agencies is pro-tected by the Attorney General in application of thethird party principle. In this case there were examplesthat simply did not stand up to scrutiny.”96

Accordingly, in this and other cases, the courtshave tempered the circumstances in which the third

significance. As urged by one CSIS official in a publicaffidavit in the Federal Court:

assessing the damage caused by disclosure ofinformation cannot be done in the abstractor in isolation. It must be assumed thatinformation will reach persons with a knowl-edge of Service targets and the activitiessubject to this investigation. In the hands ofan informed reader, seemingly unrelatedpieces of information, which may not inthemselves be particularly sensitive, can beused to develop a more comprehensive pic-ture when compared with informationalready known by the recipient or availablefrom another source.90

The mosaic effect has been accepted by Canadiancourts, and it has guided decisions on disclosure. Asnoted by the Federal Court in one of the first cases toapply the doctrine,

in security matters, there is a requirement tonot only protect the identity of humansources of information but to recognize thatthe following types of information mightrequire to be protected:…information per-taining to the identity of targets of the sur-veillance whether they be individuals orgroups, the technical means and sources ofsurveillance, the methods of operation of theservice, the identity of certain members ofthe service itself, the telecommunicationsand cypher systems and, at times, the veryfact that a surveillance is being or is notbeing carried out. This means for instancethat evidence, which of itself might not be ofany particular use in actually identifying thethreat, might nevertheless require to be pro-tected if the mere divulging of the fact thatCSIS is in possession of it would alert thetargeted organization to the fact that it is infact subject to electronic surveillance or to awiretap or to a leak from some humansource within the organization.91

There is an obvious logic in this reasoning, espe-cially if one is confronted with a well-resourced andsophisticated adversary with the means to pull togeth-er inchoate pieces of information into a prejudicialwhole. However, the use of the mosaic effect outsidethe context of security investigations involving confi-dential informers and techniques could greatly erodeopen government and open courts. Since the doctrineapplies to innocuous information, the future use ofwhich can never be predicted, it could be deployed tostave off disclosure of virtually any piece of informa-tion. Further, the mosaic effect is a creature of theCold War, when the chief adversaries were EasternBloc intelligence services. Its uncritical application tocases involving, for example, much more diffuse andindefinite terrorist threats may not be warranted.92

There must therefore be an outer limit where eveninformation related to an investigation is disclosable.

25

Ca

na

da

’s Na

tion

al S

ec

urity

“Co

mp

lex

”: Asse

ssing

the

Se

cre

cy

Ru

les, b

y C

raig

Fo

rce

seof injury should the information be disclosed.(House of Commons Standing Committee onAccess to Information, Privacy and Ethics2008)

Unfortunately, the Access Act complaint process ismired in significant delays – in 2007-08, the informationcommissioner’s office closed complaints in eight months,on average (Information Commissioner of Canada 2008),a period almost double that at the beginning of thedecade (Information Commissioner of Canada 2000).These delays are in addition to the now regular delays bythe government in processing the initial request. In 2007-08, the responses to over 40 percent of access requeststook longer than the statutory default of 30 days – and asizeable minority (13 percent) took longer than sixmonths (Treasury Board of Canada Secretariat 2008).

In sum, the vast bulk of government balancing oftransparency versus national security confidentiality iseither unexamined by arm’s length observers, or exam-ined only after prolonged delay.

Reform minimalismThe challenge of addressing problems in reconciling trans-parency with national security confidentiality is mademore difficult by the Canadian approach to reform. Thereis little evidence that the government or legislators havemuch interest in revising, refining and improvingCanada’s information laws. We are burdened, it wouldseem, with inconsistent and (in the case of section 4 of theSOIA), incoherent and ineffectual laws. Moreover, wherechanges are made, they tend to be minimalist; that is, thegovernment selects (and Parliament endorses) approachesthat maximize secrecy at the expense of transparency. Acase in point is the new “special advocate” system intro-duced into the security certificate system in the wake ofthe Supreme Court’s 2007 Charkaoui decision.

The Supreme Court released the Charkaoui decisionon February 23, 2007, finding the immigration securitycertificate system (as it then existed) unconstitutional.Specifically, the Court held that the system’s secrecyrules violated section 7 of the Charter. This secrecy wasnot saved by section 1; the government had shown noreason why it had failed to adopt some sort of model inwhich an independent “special counsel” represented theinterests of the named person in the secret portions ofthe proceedings.

Critically, in the course of its decision, the Court can-vassed a number of different special counsel options,but without expressing a preference between thesealternatives. Indeed, it expressly left it to Parliament todecide “what more should be done.”102 These alterna-

party rule will be accepted. At the very least, the gov-ernment should be obliged to seek permission to dis-close from the foreign source in an effort to separatetruly sensitive information from more benign data.This is an obligation that probably already exists inthe Access Act and Privacy Act contexts97 and hasalso been invoked by the Federal Court in relation tothe CEA.98 In the latter context, the court examineswhether “good faith efforts were made and continueto be made to obtain such consent.”99

Moreover, the government cannot claim “third-party” confidentiality over information that it hasobtained from a foreign partner, but it has receivedalso by other means. Refusal to disclose this informa-tion must instead be grounded in another justifica-tion.100 Nor can third party confidentiality be used “toprotect the mere existence of a relationship betweenCanada and a foreign state or agency, absent theexchange of information in a given case.”101

Overwhelmed checks and balancesAs this discussion suggests, many of the shortcomingsin Canadian law and policy could be overcome byrobust checks and balances that query subjectivity andoverclaiming in national security confidentiality. TheFederal Court has been particularly active in the pastthree years in developing more demanding standardsfor government secrecy in the national security area.

The Federal Court’s influence in this area is,however, limited to judicial proceedings, such asthose under the CEA. While in principle disputesover transparency under the Access Act could endup in front of the court, they do so infrequently,and usually only when backed by requesters withunusually deep pockets. Most denied requests orredacted disclosures go unchallenged or, at best,are funnelled into the information commissioner’scomplaints process. The commissioner’s scrutinyof government decisions can be meaningful. Asdescribed by the Commons access to informationcommittee,

where the Commissioner investigates a dis-cretionary exemption, such as section 15(1)of the ATIA [Access Act], his office willrequire the head of the government institu-tion to explain the factors, pro and con dis-closure, that were weighed in exercising thediscretion. The Commissioner will then assesswhether these were the correct factors to betaken into consideration and whether theywere given the appropriate weight. Wherethe exemption is injury-based, theCommissioner will also look to ensure thatthere is a reasonable basis for an expectation

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

26

SIRC) are permitted to communicate throughout theprocess with the interested party and his or her counsel.They are obliged, of course, not to reveal secret informa-tion but may continue to pose questions of a sort thatdoes not betray such a secret in order to better under-stand the facts. No SIRC in-house or outside counsel hasever reportedly received any complaints from the gov-ernment that this contact with the named person hasresulted in an involuntary disclosure injurious to nation-al security (see Forcese and Waldman 2007). As GordonCameron, an experienced SIRC legal agent, has noted:“To my knowledge, these interviews have been conduct-ed without any inadvertent disclosure of information andI am not aware that CSIS has ever complained about orattempted to change this practice.”104

At the same time, this continued contact has been ele-mental in preserving a fair process. In Cameron’s words:

In my work in SIRC hearings, there havebeen a number of occasions on which theability to question the complainant afterreviewing the confidential information hasallowed SIRC in-house counsel and me toadvance the interests of complainants onboth of the afore-mentioned bases; that is, todisprove or cast incriminating allegationsinto doubt, and to rebut or qualify allega-tions of deceitfulness.105

In comparison, the UK-style “special advocate”model imposes extremely strict restrictions on contin-ued contact between the special advocate and thenamed person once the former has reviewed the secretinformation. In effect, there is no communication atall. The Supreme Court itself was alive to criticisms ofthe UK system, including those concerning restrictionson contact between the named person and the specialadvocate.106 Nevertheless, for reasons that have neverbeen satisfactorily explained to this author, the gov-ernment opted for the secrecy-maximizing and fair-trial-minimizing United Kingdom approach.107

As Kent Roach observes, in so doing the governmentselected, and Parliament ratified, “the only alternativethat the [Supreme] Court recognized [in Charkaoui] hadbeen subject to criticism and the one alternative thatarguably achieves the worst job of all the alternatives inensuring fair treatment of the affected person” (Roach,forthcoming). Parliamentarians themselves were clearlyaware of this problem. When the Bill was debated in theSenate, even a Conservative Senator expressed concern:“Sometimes we hold our noses when it comes time toadopt bills, and we have done so in the past with otherlegislation, knowing that in the near future we will cor-rect the errors we have agreed to let through. That is thesort of legislation we have before us now.”108

tives were (in the order in which they appeared): theSIRC process; the Air India trial process; the ArarCommission process; and the United Kingdom systemof special advocates. It also discussed the approachtaken by the CEA in the disclosure of information saidto raise national security issues.

The Court suspended its declaration of constitutionalinvalidity for one year, until February 23, 2008, inorder to allow a reaction from Parliament. Not surpris-ingly, in the immediate aftermath of the Charkaouidecision, the policy focus was on “special counsel.” Forinstance, a month after Charkaoui, a special Senatecommittee recommended that a special counsel processbe extended to all proceedings where “information iswithheld from a party in the interest of national securi-ty and he or she is therefore not in a position to makefull answer and defence,” including under the IRPA, theCriminal Code terrorist group listing process, theCharities Registration (Security Information) Act103 andthe CEA (Special Senate Committee on the Anti-terror-ism Act 2007, 42). Moreover, the committee urged thatthe special counsel be empowered to communicate withthe affected parties after receiving confidential infor-mation, subject to guidelines designed to bar the releaseof secret information. The counterpart Commons com-mittee also recommended a comprehensive “panel ofspecial counsel” for national security cases (House ofCommons Standing Committee on Public Safety andNational Security 2007, 81), but without weighing in onthe precise design of this system.

The government, for its part, remained largelysilent. Despite the notoriety of the security certificatesystem and the controversy sparked by it, no publicconsultations were held and no formal notice wasgiven of the government’s approach to the specialcounsel issue until Bill C-3 was tabled in the Houseof Commons and received first reading, on October22, 2007 – fully eight months after the Charakouidecision and four months prior to the expiry of theone-year suspension of the declaration of invalidity.

As noted above, the Supreme Court of Canada can-vassed a number of special counsel options in Charkaouiwithout mandating a particular model. The governmentwas therefore presented with a choice in the crafting ofBill C-3, which included building on Canada’s indige-nous experience with SIRC in immigration and othermatters. This SIRC approach was employed in securitycertificates against permanent residents up until 2002,and it is still applied in relation to SIRC’s investigation ofcomplaints against CSIS. In the SIRC model, SIRC coun-sel (whether in-house or outside legal agents retained by

27

Ca

na

da

’s Na

tion

al S

ec

urity

“Co

mp

lex

”: Asse

ssing

the

Se

cre

cy

Ru

les, b

y C

raig

Fo

rce

seor antiterrorism law per se, and courses in antiterrorismor national security law are taught in only a handful oflaw schools. Instead, academics tend to have more spe-cialized expertise in criminal, privacy or immigrationlaw, for example.

All of these actors are asked to grapple, in a democrat-ic policy-making environment, with a threat that is ofteninchoate in nature. As Lynch notes, “it is extremely diffi-cult for parliamentarians to know the true extent of thethreat to the nation’s security when trying to insist eitherupon a more consultative, careful process or the inclusionof specific powers and safeguards” (2006). For all of thesereasons, national security law is an area in which treesmay sometimes attract scrutiny more readily than doesthe forest; that is, attention is drawn to a handful of par-ticularly controversial areas, leaving the vast bulk ofnational security measures unexamined in any detail.National security confidentiality is unquestionably one ofthese orphans – the Security of Information Act, forexample, has attracted relatively little attention from par-liamentarians, even with the sweeping renovations madeto it in 2001 as part of the massive Anti-terrorism Act,and even after the O’Neill decision.109

Nevertheless, despite these difficulties, I envisagefour fixes that would go a long way in rebalancing thesecrecy law regime and undergirding the culture ofopenness required in a functioning democracy.

Correcting the Security of Information ActFirst, Parliament should formally repeal section 4 of theSecurity of Information Act (SOIA), replacing it with a muchmore measured provision. Despite the condemnation of theOntario Superior Court of Justice, section 4 sits lurking onthe statute book as a big stick to be deployed by inattentiveofficials. The recent provisions relating to persons bound bysecrecy penalize leakage from the intelligence services.With the repeal of section 4, a more measured provisioncovering other civil servants and persons receiving protect-ed information would have to be drafted. In this respect, thenew law could follow the precedent of the UK OfficialSecrets Act by defining extremely carefully and narrowlythe sorts of secrets covered by criminal provisions, and byintroducing a prerequisite that damage, as defined by theAct, stem from disclosure. The amended SOIA could thenapply a public interest override of the sort currently avail-able to persons bound by secrecy.

Coherent redrafting and integration ofinformation lawSecond, the government should standardize its definitionof “national security” (and similar terms, such as

Conclusion and Recommendations

I n sum, Canada’s information and secrecy laws andpractices leave much to be desired. Past commen-tary from the security and intelligence community

itself suggests that the breadth of secrecy laws is morethan is necessary to protect legitimate national securitysecrets. At the same time, the incoherence of theselaws, the uncertainty this incoherence produces andthe overclaiming it allows create conditions likely tocurb benign information exchanges, at great expenseto the credibility of the security services. Finally, thelimits they impose on information access – and thedraconian penalties they level in some instances – aredeeply inconsistent with the democratic society theyare supposed to protect. They are broad enough to letgovernment sidestep embarrassment and mask incom-petence, all in the name of national security.

In these concluding sections I propose several“quick fixes” for this problem. It is worth acknowl-edging, however, that fixes are easy to imagine anddifficult to implement, in a significant measurebecause of the difficult policy environment in whichnational security issues are addressed. Not least,national security law responds to uncertain dangers,while at the same time straddling conventional sub-ject-matter divisions in the policy and parliamentarycommunities. The Commons, for example, has a (rela-tively new) national security committee, but it alsohas separate committees dedicated to justice, immi-gration, and privacy and access to information.

At the bureaucratic level, the development of dif-ferent statutes grappling with different national-security-related issues is centred in differentdepartments. Meanwhile, operating independently ofthese departments are the auditor general and the pri-vacy and information commissioners, specializedofficers of Parliament, whose mandates extend toseveral areas of law affected by antiterrorism statutes.There is, however, no officer of Parliament whosewrit reaches across all of national security or evenantiterrorism law. Within the security and intelligencecommunity itself, arm’s length review responsibilitiesare “stovepiped” between the Security IntelligenceReview Committee (for CSIS) and the commissionerof Communications Security Establishment Canada(for CSE). No equivalent body existed at the time ofthis writing for the RCMP.

Meanwhile, within the academic community, onlya handful of scholars specialize in national security

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

28

“national defence” and “international relations”) acrossthe statute book. Currently, exemptions from disclosureon national security grounds are conveyed by a con-fusing array of terms, such as “international affairs,national defence and national security,” “nationalsecurity,” “security of Canada,” “Canadian interests,”“information that is of a type that the Government ofCanada is taking measures to safeguard.”

It makes sense, as a first step, to harmonize what thegovernment means by national security. To a certainextent, the government has already done this with theterm “security of Canada” in the CSIS Act. Employingthis definition in lieu of other, undefined references to“national security” or its similes in the Canadian statutebook would have two salutary effects. First, it wouldprovide a necessary metre stick against which to meas-ure the legitimacy of national security justifications inthe many statutes that lack a definition of the term.Second, it would standardize and centralize the under-standing of national security throughout Canadian fed-eral law. Debate could then focus on the adequacy ofthis standardized and centralized definition and not bedistracted by questions of whether national securitymight be approached differently in the other, sometimesobscure circumstances in which statutes invoke it.

It is true, however, that the concept of the “securityof Canada” defining the mandate of CSIS may notalways overlap with the classes of information that thegovernment seeks legitimately to protect on nationalsecurity grounds. Special definitions of national securi-ty secrets will also have to be articulated. For instance,the SOIA has attempted to provide a definition ofnational security secrets with its concept of “specialoperational information.” The Access Act, meanwhile,has comprehensive definitions for its national securityexemptions. Lining up these two sources of definitionsof national security secrets and then using this recon-ciliation to contribute greater certainty to the invoca-tion of international relations, defence and nationalsecurity in the CEA and other statutes, such as theIRPA, will require some modest redrafting.

In this respect, section 15 of the Access Act is thelogical nexus point for a common understanding ofnational security secrets. First, its use of the expression“international affairs” should be defined. A startingpoint might be the UK Official Secrets Act of 1989,which defines international relations as “any matterrelating to a State other than the United Kingdom or toan international organisation which is capable ofaffecting the relations of the United Kingdom withanother State or with an international organisation.”110

Second, the reference to “prevention or suppression ofsubversive or hostile activities” in subsection 15(1)should be replaced with “national security.” The sec-tion should then define “national security” accordingto the proposed, standard definition in the CSIS Act.Finally, section 15 should also capture “special opera-tional information,” as defined by the SOIA.

The CEA should be amended to eliminate its cre-ation of three new and slightly different classes ofnational security information. Instead, the Act shouldonly apply to “potentially injurious information,”defined to mean information specified in section 15 ofthe newly amended Access Act. A similar changeshould be made to the IRPA’s security certificate rules.

Notably, fixing section 15 as the litmus test wouldincorporate that section’s injury test into the CEAand IRPA. Such a change would probably add little toregular CEA section 38 determinations. This sectionalready incorporates an injury test and allows aFederal Court judge to contemplate the balance ofpublic interests in reaching a disclosure decision. Theproposed amendment would, however, providegreater latitude for a Federal Court of Appeal judge tocontemplate the balance of interests in assessing themerits of an attorney general’s certificate issuedunder the Act. Likewise, judges in the IRPA contextcould also complete a balancing.

In essence, the amended CEA would allow the min-ister to certify Access Act section 15 information asexempt from the Access Act, disallow its use in pro-ceedings, and bring it under the different appeal andreview regime established by the CEA. These are stillextremely potent powers. In the absence of compellingevidence that more is needed, these changes seem morethan sufficient to secure legitimate government secrecy.

These amendments would not put national securityat risk. Disclosure would still be carefully circumscribedby the now even more generous Access Act exemptions— exemptions that the Canadian security services sawas sufficient even before the new post-9/11 restrictions.On the other hand, these changes would simplify andstandardize government secrecy law, eliminate muchuncertainty as to its scope and leave good governancein Canada less dependent on benign executive branchinterpretations of today’s perplexing secrecy laws.

Institutional reformsThird, the government should address the institution-al failings of the access regime. Better resourcing ofthe information commissioner’s office to permit theexpeditious review of complaints is a pressing need.

the US Freedom of Information Act and to apply a pre-sumption of disclosure (Presidential Documents 2009).

Sustained parliamentary attentionLast, Parliament itself should demonstrate a healthyskepticism of government secrecy claims on nationalsecurity grounds. The Standing Senate Committee onNational Security and Defence did so in its 2003 reporton airport security. Legislators were, however, willingto pass amendments to the security certificate processthat overreach in the defence of secrecy at the expenseof fairness in the judicial process.

Given the later experience, I may be accused of beingirremediably naive to expect much of parliamentariansin the areas at issue in this paper. On the other hand,contempt of Parliament and its ability to show meas-ured contemplation on national security questions willproduce nothing but a vicious circle in which a distract-ed and disinterested Parliament, treated as inept in thisarea, is given short shrift and is stripped of any oppor-tunity to gain experience, insight and authority. In otherworks, I have expressed support both for a standingreviewer of national security law and policy and a sepa-rate committee of parliamentarians, with legislatedpowers and access to secret information (and subject,concomitantly to nondisclosure obligations) (Forcese2008). I renew that call as a conclusion to this paper.

Last wordsIndividually or collectively, there is no reason tobelieve that any of the changes I have proposed wouldproduce the feared “sinking ships” said to be producedby “loose lips.” Legitimate secrets would be protected.At the same time, these reforms might remove govern-ment secrecy laws as an obstacle to legitimate publicscrutiny, and in so doing actually enhance security. Toparaphrase the Standing Senate Committee on NationalSecurity and Defence, public pressure could do exactlywhat secrecy might fail to do: motivate governments torepair the holes that do sink ships.

That office should be particularly attentive to over-claiming in the national security area, taking its leadfrom the more critical eye being cast on governmentsecrecy claims by the Federal Court.

Enhancing the information commissioner’s capacityis not, however, enough. Part of the burden on thatoffice would be reduced if the government’s applica-tion of the Access Act was more credible in the firstinstance. In this regard, better resourcing and trainingof the front-line access personnel who manage infor-mation requests received by their departments is desir-able. The anecdotal evidence cited above suggests thatconsistent and standardized application of nationalsecurity exemptions remains a challenge. Sincererecognition that timely responses to informationrequests are part of a public official’s obligations – andthe resources to back that recognition up – might alsoresolve some of the Access Act’s challenges.

Attentiveness to overclaiming should also figureamong the responsibilities of the various specializednational security review bodies — such as the SIRC,the CSE commissioner and other security services(should the Arar Commission’s policy recommenda-tions ever be implemented). Annual reviews ofagency performance by these bodies should includeclose attention to agency information disclosurephilosophies and practices.

Lastly, the security intelligence services themselves(and central government agencies such as the PrivyCouncil) should carefully consider whether, on balance,information can reasonably be withheld. As then USrepresentative Lee Hamilton observed in 1997, “If wehave too much secrecy, we cannot focus enough onprotecting the truly important secrets. Secrecy can bestbe preserved when the credibility of the system isassured” (Hamilton 1997). Overclaiming does no serv-ice to those instances where bona fide national securi-ty preoccupations require secrecy. The risk, whenoverclaiming is revealed, is that claims to secrecythereafter are viewed as cries of wolf by the public, themedia and (potentially) the courts. To its credit, there issome evidence that the security and intelligence com-munity understands this, and I and a number of aca-demics have recently been party to unusually frankdiscussions with at least one agency that have servedto enlighten without imperilling national security.

At core, changing government attitudes are ulti-mately at the heart of a renewed culture of openness.Indeed, such a sea change seems now to be underwayin the United States, with the Obama administration’sfirm pronouncement on the need to adhere fully to

29

Ca

na

da

’s Na

tion

al S

ec

urity

“Co

mp

lex

”: Asse

ssing

the

Se

cre

cy

Ru

les, b

y C

raig

Fo

rce

se

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

30

17 For example, see Attorney General of Nova Scotia v.MacIntyre, [1982] 1 S.C.R. 175, at 187; CanadianBroadcasting Corp. v. New Brunswick (AttorneyGeneral), [1996] 3 S.C.R. 480 at paras. 21 & 22;Edmonton Journal v. Alberta (Attorney General),[1989] 2 S.C.R. 1326.

18 Being Schedule B to the Canada Act 1982 (U.K.) 1982,c. 11.

19 Ruby v. Canada, 2002 S.C.C. 75 at para. 52.20 Ibid. at para. 53, citing Canadian Broadcasting Corp. v.

New Brunswick (Attorney General), [1996] 3 S.C.R.480, at para. 23.

21 ICCPR, Art. 14(3)(b) specifies that an accused in crimi-nal cases is “[t]o have adequate time and facilities forthe preparation of his defence.” The UN Human RightsCommittee interprets “facilities” as including “access todocuments and other evidence which the accusedrequires to prepare his case.” UN Human RightsCommittee, General Comment 13, U.N. Doc.HRI/GEN/1/Rev.6 at 135 (2003) at para. 9.

22 R. v. Stinchcombe, [1991] 3 S.C.R. 326 at para. 17.23 R. v. Chaplin, [1995] 1 S.C.R. 727 at para. 21.24 Charkaoui v. Canada (Citizenship and Immigration),

2008 S.C.C. 38 at para. 43.25 See, for example, Douglas Bland, appearing at the

Standing Senate Committee on National Security andDefence, Evidence (October 29, 2001) (indicating that“if a broad definition of ‘national security’ is taken,there is a danger that there will be no obvious limits topolicy,” but then citing with a measure of approval theNational Defence College definition).

26 CSIS affiant’s testimony, reported in Khawaja v.Canada, 2007 F.C. 463 at para. 132.

27 Affidavit of Margaret Ann Purdy (then director generalof the Counter-terrorism Branch) (31 Oct. 1994), citedin Ruby v. Canada, [1996] 3 F.C. 134 at para. 25(F.C.T.D.).

28 Ibid. (court summary of affidavit).29 Ibid. at para. 26.30 Ibid. at para. 27.31 See, for example, Ribic v. Canada, 2003 F.C.T. 10 at

para. 10 (F.C.); Khawaja v. Canada, 2007 F.C. 490 (F.C.)at para. 122 et seq.

32 Ibid. at para. 127.33 Jack Hooper, Assistant Director of Operations, CSIS,

Testimony, Arar Commission, June 22, 2005, at 485.34 Declaration of the Secretary of Defense, United States

of America v. Jonathan Jap Pollard, United StatesDistrict Court for the District of Columbia, CriminalNo. 86 at 22, cited in Richelson (1999, 291).

35 These limitations are echoed in the Privacy Act’s pro-visions on disclosure to applicants of applicants’ ownpersonal information in the possession of government.See Privacy Act, R.S., 1985, c. P-21, ss. 12 et seq.Because the Privacy Act’s disclosure functions are lim-ited to an individual’s own personal information, thefocus in this paper is on the Access Act.

36 See, for example, Rubin v. Canada (Minister ofTransport), [1998] 2 F.C. 430 at para. 36, 154 D.L.R.

Notes1 The phrase was an allusion to the dangers to Atlantic con-

voys of German foreknowledge of sailing times and routes. 2 Similar comments have been made by academic

observers. “Freedom of expression and access to infor-mation, by enabling public scrutiny of governmentaction, serve as safeguards against government abuseand thereby from a crucial component of genuinenational security” (Coliver 1999, 11-12); ”The problemwith the ‘national security state’ is not so much that itviolates [fundamental] rights, although it sometimesdoes just that, but that it can lead to the repetition ofirrational decisions” (Chevigny 1991, 138).

3 Freedom of Information Act of 4 July 1966, Pub. L. No. 89-487, 80 Stat. 250 (5 U.S.C. § 552).

4 Freedom of Information: Hearings on S. 1666 and S.1663 before the Subcomm. on Admin. Practice andProcedure of the Senate Comm. on the Judiciary, 88thCong. 3 (1964) (statement of Sen. Edward Long), citedin Wichmann (1998, 1217).

5 NLRB v. Robbins Tire and Rubber Company, 437 U.S.214 at 242 (1978), 57 L Ed 2d 159 at 178.

6 Pierre Elliott Trudeau, quoted by G. Baldwin, MP, inMinutes of Proceedings and Evidence of the StandingJoint Committee on Regulations and other StatutoryInstruments, 30th Parl., 1st Sess. (1974-75) 22:7, citedin Rankin (1979, 1).

7 House of Commons Debates (29 Nov. 1979) at 1858,cited in House of Commons Standing Committee onJustice and the Solicitor General on the Review of theAccess to Information Act and the Privacy Act (1987, 4).

8 [1997] 2 S.C.R. 403 at para. 61, 148 D.L.R. (4th) 385.9 For example, see Canada (Attorney General) v. Canada

(Information Commissioner), [2004] F.C. 431 at para.22 (F.C.T.D.); Yeager v. Canada (Correctional Service),[2003] 3 F.C. 107 at para. 39, 2003 F.C.A. 30; Rubin v.Canada (Minister of Transport), [1998] 2 F.C. 430 atpara. 36, 154 D.L.R. (4th) 414 (F.C.A.).

10 Canada (Information Commissioner) v. Canada(Commissioner of the Royal Canadian Mounted Police),[2003] 1 S.C.R. 66 at para. 32.

11 R.S.C., 1985, c. A-1.12 AstraZeneca Canada Inc. v. Canada (Health), 2005 F.C.

1451 at para. 49 (F.C.), aff’d, 2006 F.C.A. 241.13 Canada Post Corporation v. Canada (Minister of Public

Works), [1995] 2 F.C. 110 at 129, F.C.J. No. 241 (F.C.A.)(“subsection 4(1) contains a ‘notwithstanding clause’which gives the Act an overriding status with respectto any other Act of Parliament”).

14 Access Act, s. 30. 15 Access Act, ss. 41 and 42. 16 For example, see International Covenant on Civil and

Political Rights (ICCPR), G.A. res. 2200A (XXI), 21 U.N.GAOR Supp. (No. 16) at 52, U.N. Doc. A/6316 (1966),999 U.N.T.S. 171, entered into force Mar. 23, 1976, Art.14 (applicable to criminal proceedings and any suit atlaw determining rights or obligations, and requiring a“public hearing,” except in limited circumstances suchas when required by national security).

55 CEA, s. 38.131. The government urges that the AttorneyGeneral certificate process is necessary to bar release ofinformation obtained in confidence under internationalintelligence-sharing arrangements (Government ofCanada 2007, 16).

56 For an assessment of s. 38 of the CEA’s overall compli-ance with s. 7 of the Charter, see Grant (2003).

57 CEA, s. 38.14.58 Immigration and Refugee Protection Act (IRPA), S.C.

2001, c. 27.59 IRPA, ss. 76 et seq.60 IRPA, s. 83(1)(e).61 The men subject to security certificates at the time of

this writing spent (and, in one case, continue to spend)lengthy periods incarcerated: by the beginning of 2007,the average period of detention for the men still impris-oned at that time was almost six years. This is a periodof detention longer than the median sentence for con-victed attempted murderers in Canada. In fact, it fell justshort of the seven-year median sentence for convictedmurders (see Statistics Canada, “Sentenced Cases andOutcomes in Adult Criminal Court, by Province andYukon, 2003” (http://www40.statcan.ca/l01/cst01/legal21a.htm, accessed May 4, 2009). Almrei —the individual still in detention — has spent much of hisperiod in jail in solitary confinement and, until relativelyrecently, in provincial detention facilities ill-equipped forlong-term detentions. His period of detention is nowlonger than the median sentence in Canada for a con-victed murderer.

62 Charkaoui v. Canada, 2007 S.C.C. 9.63 Ibid. at para. 28. 64 Ibid. at para. 48.65 Ibid. at para. 53.66 R.S.C., 1985, c. O-5.67 R.S.C., 1970, c. O-3. This Act, in turn, is an “adoption of

the English statutes as enacted in Great Britain (1911[U.K.] c. 28, and 1920 [U.K.], c. 75).” R. v. Toronto SunPublishing Limited, (1979) 24 O.R. (2d) 621 at 623 (Ont.Prov. Ct.) [Toronto Sun].

68 House of Commons Debates, 096 (30 April 1998) at 1010(Hon. Andy Scott).

69 R v. Ratkai, [1989] N.J. No. 334 (Nfld. S.C. [T.D.]) (Q.L.).70 Toronto Sun, (1979) 24 O.R. (2d) 621 at 632. Notably, the

question of ambiguity in the Act became the key pointon which the Juliet O’Neill matter, described below,turned. By the time of the latter case, however, excessiveambiguity was unconstitutional, given the way in whichsection 7 of the Charter of Rights and Freedoms hadbeen interpreted.

71 SOIA, s. 8 and accompanying schedule. Further, undersection 10, other persons may be designated “a personpermanently bound to secrecy” if certain senior govern-ment officials believe that “by reason of the person’soffice, position, duties, contract or arrangement…the per-son had, has or will have authorized access to specialoperational information; and…it is in the interest ofnational security to designate the person.”

72 SOIA, s. 8.

(4th) 414 at para. 30 (F.C.A), citing Canada PackersInc. v. Canada (Minister of Agriculture), [1989] 1F.C. 47 at 60 (F.C.A.) (“Subsection 2(1) provides aclear statement that the Act should be interpretedin the light of the principle that government infor-mation should be available to the public and thatexceptions to the public’s right of access should be‘limited and specific.’ With such a mandate, Ibelieve one must interpret the exceptions to accessin paragraphs [20(1)] (c) and (d) to require a reason-able expectation of probable harm”) (emphasisadded).

37 R.S.C. 1985, c. C-23. See also Privacy Act, s. 22.38 Ruby, 2002 S.C.C. 75 at para. 5.39 Access Act, s. 15. See also Privacy Act, s. 21.40 Access Act, subs. 15(2).41 Ibid. The expression means: “espionage against Canada

or any state allied or associated with Canada…sabo-tage…activities directed toward the commission of ter-rorist acts, including hijacking, in or against Canada orforeign states…activities directed toward accomplish-ing government change within Canada or foreignstates by the use of or the encouragement of the use offorce, violence or any criminal means…activitiesdirected toward gathering information used for intelli-gence purposes that relates to Canada or any stateallied or associated with Canada, and…activities direct-ed toward threatening the safety of Canadians,employees of the Government of Canada or propertyof the Government of Canada outside Canada.”

42 S.C. 2001, c. 41.43 R.S.C., 1985, c. C-5 (CEA).44 CEA, s. 38. A “proceeding…means a proceeding before

a court, person or body with jurisdiction to compel theproduction of information.”

45 CEA, s. 38.13.46 CEA, s. 38.131.47 Access Act, s. 69.1.48 2004 S.C.C. 43 at para. 4.49 See, for example, s. 486 of the Criminal Code, R.S.,

1985, c. C-46, the central provision in the CriminalCode curtailing the open court principle for criminalproceedings on national security grounds, as interpret-ed by the Supreme Court in Vancouver Sun, at para.71. See also Toronto Star Newspapers v. Canada, 2007F.C. 128 (F.C.), reading down those provisions obligingthe Federal Court to hold a Canada Evidence Act appli-cation and records confidential and to hold the hear-ing in private to only those circumstances where thegovernment requests an ex parte hearing.

50 CEA, s. 38. 51 CEA, s. 38. 52 For example, see Rosenthal (2003, 191) for a critique

of this language.53 Khawaja, 2007 F.C. 490 at paras. 62 et seq.; Canada

(Attorney General) v. Commission of Inquiry into theAction of Canadian Officials in Relation to Maher Arar,2007 F.C. 766 (Arar Commission) at paras. 37 et seq.

54 CEA, s. 38.13.

31

Ca

na

da

’s Na

tion

al S

ec

urity

“Co

mp

lex

”: Asse

ssing

the

Se

cre

cy

Ru

les, b

y C

raig

Fo

rce

se

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

32

requests for information about various aspects of theAfghanistan mission, such as operations, relatedevents and activities, treatment of detainees, and poli-cies” (Information Commissioner of Canada 2008). Theinformation commissioner expressed at least partialsatisfaction with the government’s performance on theAfghan file. For its part, the House of CommonsStanding Committee on Access to Information, Privacyand Ethics flagged a number of concerns in the han-dling of at least one Afghan-related informationrequest. See House of Commons Standing Committeeon Access to Information, Privacy and Ethics (2008).

85 Order-in-Council, P.C. 2006-0293 (2006-05-01).86 Order-in-Council, P.C. 2006-1526 (2006-12-11).87 Canada (Attorney General) v. Commission of Inquiry

into the Action of Canadian Officials in Relation toMaher Arar, 2007 F.C. 766 (F.C.) (Arar Commission) atpara. 58.

88 R. v. Khawaja, 2007 F.C. 463 at para. 150 varied 2007F.C.A. 388.

89 See, for example, ibid.; Canada v. Kempo, 2004 F.C.1678 (F.C.); Re Zundel, 2005 F.C. 295 (F.C.); Cemerlic v.Canada (Solicitor General), 2003 FCT 133 (F.C.); Canada(Minister of Citizenship and Immigration) v. Singh,[1998] F.C.J. No. 978 (F.C.T.D.); Ternette v. Canada,[1992] 2 F.C. 75 (F.C.T.D.); Henrie v. Canada (SecurityIntelligence Review Committee), [1989] 2 F.C. 229, aff’d[1992] F.C.J. No. 100 (F.C.A.); Re Jaballah, [2003] 4 F.C.345 (F.C.); Ruby v. Canada (Solicitor General), [1998] 2F.C. 351.

90 Reproduced in Kempo, 2004 F.C. 1678 at para. 62.91 Henrie, [1989] 2 F.C. 229 at para. 29.92 In this last regard, see Pozen (2005) for a sustained cri-

tique of how the mosaic effect has been employed inthe United States.

93 O’Neill, [2004] O.J. No. 4649 at para. 69.94 Khadr v. Canada, 2008 F.C. 549 at para. 77.95 Khawaja, 2007 F.C. 463 at para. 136.96 Khadr, 2008 F.C. 549 at para. 98.97 See Cemerlic, 2003 F.C.T. 133 at paras. 18 et seq. (dis-

cussing s. 19 of the Privacy Act).98 Khawaja, 2007 F.C. 463 at para. 146 (“it is not open to

the Attorney General to merely claim that informationcannot be disclosed pursuant to the third party rule, ifa request for disclosure in some form has not in factbeen made to the original foreign source”).

99 Ibid. at para. 152.100 Ibid. at para. 147; Ottawa Citizen Group Inc. v. Canada

(Attorney General), 2006 F.C. 1552 (F.C.) at para. 66.101 Khawaja, 2007 F.C. 463 at para. 148.102 Charkaoui, 2007 S.C.C. 9 at para. 87.103 R.S.C., 1985, c. O-5.104 Affidavit of Gordon Cameron, in Almrei, Court File

DES-3-08, dated July 20, 2008 at para. 22.105 Ibid. at para. 36.106 Charkaoui, 2007 S.C.C. 9 at para. 83.107 At the Commons committee hearing on Bill C-3, the

government asserted a preference for the UK modelbecause “the Supreme Court mandate was to make sure

73 CSIS Act, s. 18.74 [1979] 1 S.C.R. 218 at 250–51.75 O’Neill v. Canada (2006), 82 O.R. (3d) 241 (On. Sup. Ct

of Jus.) at para. 62 and 71 (holding, inter alia, that sec-tion 4 failed “to define in any way the scope of what itprotects and then, using the most extreme form ofgovernment control, criminalizes the conduct of thosewho communicate and receive government informa-tion that falls within its unlimited scope including theconduct of government officials and members of thepublic and of the press” and that “the lack of delin-eation of a zone of risk by these sections gives noguidance to law enforcement officials to be able todetermine whether a crime has been committed underthem, with the result that there are no controls on theexercise of their discretion and there is the danger ofarbitrary and ad hoc law enforcement”).

76 Information Commissioner of Canada (2002, 20). Foran academic critique of the amendments, seeMcMahon (2002).

77 The same report acknowledges that the information inquestion was not “special operational information asdefined by the Security of Information Act.” In thesecircumstances, the only provisions of SOIA that couldapply are sections 4 and 16. However, section 16 —which relates to information the government is “safe-guarding” — is only triggered if the person communi-cates the information to a foreign entity or to aterrorist group in either an effort to harm Canadianinterests (a defined concept that involves real risks) orif such harm actually occurs. There is absolutely nobasis to conclude — and nothing in the DFAIT report tosuggest — that Couillard was communicating theBernier documents to a foreign entity or terroristgroup. Even if she had, the DFAIT report asserts that“Our review of the documents in question concludedthat their disclosure would not have caused significantinjury to the national interest.” Neither of the require-ments for section 16 were, in other words, in play inthe Bernier-Couillard matter. By default, the only pro-vision of SOIA that could apply to Couillard’s casewould be the (unconstitutional) section 4.

78 1989 c. 6 [UK Official Secrets Act]. For a full discus-sion of the UK Act, see Wadham and Modi (2003).

79 UK Official Secrets Act, s. 1(1).80 Ibid., s. 1(5).81 Ibid., s. 1(4).82 [2007] U.K.H.L 46.83 For a summary of the effect of MB, see Carlile (2008).84 See, for example, Toughill (2008); Pugliese (2008a, b,

c); Naumetz (2008); Woods (2008); “Harper’s Secrecy”(2008); “Too Much Secrecy” (2008); “Afghanistan:Selling the Mission’s Merits” (2008); “Harper’s UnwiseAfghan Blackout” (2008); Mayeda (2007); “SecretDocuments, Secret Challenges” (2007); “A Few MoreBricks” (2007). In his 2007-08 report, the informationcommissioner reported receiving “more than 100 com-plaints in 2007-2008 from the media, members ofParliament, academics and the public related to access

Coliver, Sandra. 1999. “Commentary on the JohannesburgPrinciples on National Security, Freedom of Expressionand Access to Information.” In Secrecy and Liberty:National Security, Freedom of Expression and Access toInformation, edited by Sandra Coliver, Paul Hoffman,Joan Fitzpatrick, and Stephen Brown. The Hague:Martinus Nijhoff Publishers.

Commission of Inquiry into the Actions of Canadian Officialsin Relation to Maher Arar (Arar Commission). 2006.Report of the Events Relating to Maher Arar: Analysisand Recommendations. Ottawa: Public Works andGovernment Services Canada.

Forcese, Craig. 2008. “Fixing the Deficiencies of ParliamentReview of Anti-terrorism Law: Lessons from the UnitedKingdom and Australia.” IRPP Choices 14 (4). AccessedMay 1, 2009. http://www.irpp.org/fasttrak/index.htm

Forcese, Craig, and Lorne Waldman. 2007. “Seeking Justice inan Unfair Process: Lessons from Canada, the UnitedKingdom, and New Zealand on the Use of ‘SpecialAdvocates’ in National Security Proceedings.” Studycommissioned by the Canadian Centre for Intelligenceand Security Studies. Accessed May 1, 2009.http://aix1.uottawa.ca/~cforcese/other/sastudy.pdf

Foreign Affairs and International Trade Canada. 2008. “FinalReport on the Administrative Review into the SecurityIncident Reported by Maxime Bernier — ClassifiedDocuments Left at a Private Residence.” Accessed May 1,2009. http://www.international.gc.ca/about-a_propos/report-rapport.aspx

Government of Canada. 2007. “Government Response to theSeventh Report of the Standing Committee on PublicSafety and National Security.” 39th Parl., 1st Sess.Accessed May 1, 2009. http://www2.parl.gc.ca/ commit-teebusiness/ReportsResponses.aspx?Cmte=SECU&Language=E&Mode=1&Parl=39&Ses=1

Grant, Kathy. 2003. “The Unjust Impact of Canada’s Anti-ter-rorism Act on an Accused’s Right to Full Answer andDefence.” Windsor Review of Legal and Social Issues 16:137-68. Accessed May 1, 2009. http://faculty.law.ubc.ca/Pue/pdf/Kathy%20Grant%20anti%20terror%20evidence%20act.pdf

Hamilton, Lee. 1997. “Secrecy.” Congressional Record, March19, E514.

“Harper’s Secrecy on Detainees.” 2008. Times Colonist,January 26, A14.

“Harper’s Unwise Afghan Blackout.” 2008. Toronto Star,January 25, AA4.

House of Commons Standing Committee on Access toInformation, Privacy and Ethics (Paul Szabo, Chair).2008. “Access to Information Request for theDepartment of Foreign Affairs and International TradeInternal Report Entitled ‘Afghanistan 2006: GoodGovernance, Democratic Development and HumanRights.’” 39th Parl., 2nd Sess., report 4. Accessed May 1,2009. http://www2.parl.gc.ca/committeebusiness/ReportsResponses.aspx?Cmte=ETHI&Language=E&Mode=1&Parl=39&Ses=2

House of Commons Standing Committee on Justice andSolicitor General on the Review of the Access to

that the special advocate represents the interests of theindividual. The only living example of this was the UKspecial advocate system, so that was essentially how wegot to the UK model as the starting point” (DanielTherrien [acting Assistant Deputy Attorney General,Citizenship, Immigration and Public Safety Portfolio,Department of Justice], Evidence, 39th Parl., 2nd Sess.,Tuesday, November 27, 2007). On a technical level, it istrue that in the SIRC process, SIRC counsel representsthe interests of SIRC, and those interests tend also thento dovetail with the complaint to the extent that bothhave an interest in the truth. The government’s explana-tion for its reliance on the UK model does not, however,answer the question of why they did not graft onto thatmodel the more rights-protecting aspects of the SIRCapproach, not least in terms of continued access to thenamed person and robust full-disclosure powers.

108 Hon. Pierre Claude Nolin, Debates of the Senate(Hansard), 39th Parl., 2nd Sess., volume 144, issue 32,Tuesday, February 12, 2008. In direct response to thescant time available to it when Bill C-3 was promul-gated, the Special Senate Committee on Anti-terrorismis now engaged in an extensive, supplemental study ofthe security certificate process.

109 To be fair, the Special Senate Committee on the Anti-terrorism Act did discuss both the SOIA and the CEAin some detail in its 2007 report. Nothing, however,has happened since.

110 UK Official Secrets Act, s. 3(5).

References“A Few More Bricks in Wall of Secrecy.” 2007. Toronto Star,

May 1, A14. “Afghanistan: Selling the Mission’s Merits.” 2008. Windsor

Star, January 25, A8.Banisar, David. 2006. Freedom of Information around the

World 2006. London: Privacy International. AccessedApril 30, 2008. http://www.freedominfo.org/documents/global_survey2006.pdf

Brandeis, Louis D. 1914. Other People’s Money and How theBankers Use It. New York: Frederick A. Stokes Company.

Canadian Security Intelligence Service (CSIS). 2004.Backgrounder 12 – Security of Information Act.Accessed May 1, 2009. http://www.csis.gc.ca/nwsrm/bckgrndrs/bckgrndr12-eng.asp

------------. 2005. Backgrounder 1 – CSIS Mandate. AccessedJune 1, 2009. http://www.csis-scrs.gc.ca/nwsrm/bckgrndrs/bckgrndr01-eng.asp.

Carlile, Alex (Lord Carlile of Berriew). 2008. Third Report ofthe Independent Reviewer Pursuant to Section 14(3) ofthe Prevention of Terrorism Act 2005. Accessed April31, 2009. http://security.homeoffice.gov.uk/news-publications/publication-search/general/report-control-orders-2008?view=Binary

Chevigny, Paul H. 1991. “Information, the Executive andthe Politics of Information.” In Speech and NationalSecurity, edited by Shimon Shetreet. Boston: MartinusNijhoff Publishers.

33

Ca

na

da

’s Na

tion

al S

ec

urity

“Co

mp

lex

”: Asse

ssing

the

Se

cre

cy

Ru

les, b

y C

raig

Fo

rce

se

IRP

P C

ho

ice

s, V

ol.

15

, n

o.

5,

Jun

e 2

00

9

34

McMahon, Patricia. 2002. “Amending the Access toInformation Act: Does National Security Require theProposed Amendments of Bill C-36?” University ofToronto Faculty of Law Review 60 (1): 89-101.

Naumetz, Tim. 2008. “Ombudsman Admits Aiding DNDSecrecy; Afghan Detainees.” National Post, May 28, A9.

Padover, S., ed. 1953. The Complete Madison: His BasicWritings. New York: Harper, cited in T. MurrayRankin,1979. Freedom of Information in Canada: Willthe Doors Stay Shut? Ottawa: Canadian BarAssociation.

Pozen, David E. 2005. “The Mosaic Theory, NationalSecurity, and the Freedom of Information Act.” YaleLaw Journal 115 (3): 628-79.

Presidential Documents. 2009. “Transparency and OpenGovernment.” Federal Register 74 (15): 4685-6.Accessed April 26, 2009. http://frwebgate4.access.gpo.gov/cgi-bin/PDFgate.cgi?WAISdocID=857024457699+ 13+2+0&WAISaction=retrieve

Privy Council Office. 2004. Securing an Open Society:Canada’s National Security Policy. Accessed May 1,2009. http://www.pco-bcp.gc.ca/docs/information/Publications/natsec-secnat/natsec-secnat-eng.pdf

Pugliese, David. 2006. “Ottawa Declares Publicly AvailableData a State Secret.” Edmonton Journal, October 4, A5.

_____. 2008a. “Once Public, Now Secret.” Ottawa Citizen,June 16, A1.

_____. 2008b. “DND Releases Secret Files on AfghanMission Ammo; Military Careful about Sharing InfoBecause Taliban Read Citizen, Top General Says.”Ottawa Citizen, February 7, A1.

_____. 2008c. “Reasons for Secrecy Are Secret, MilitarySays; Canadian Forces Refuses to Give Details onAmmo Used in Afghanistan — or Say Why.” OttawaCitizen, February 5, A4.

Rankin, T. Murray 1979. Freedom of Information inCanada: Will the Doors Stay Shut? Ottawa: CanadianBar Association.

Richelieu, Duc De. 1641. “Maxims.” Testament Politique,cited in The Columbia World of Quotations, New York:Columbia University Press: 1996. Accessed May 1,2009. http://www.bartleby.com/66/34/46534.html

Richelson, Jeffrey. 1999. The U.S. Intelligence Community.4th ed. Boulder, CO: Westview Press.

Roach, Kent. 2008. “The Role and Capacities of Courts andLegislatures in Reviewing Canada’s Anti-terrorism Law.”Windsor Review of Legal and Social Issues 24:5-56.

------------. Forthcoming. “Charkaoui and Bill C-3: SomeImplications for Anti-terrorism Policy and Dialoguebetween Courts and Legislatures.” Supreme Court LawReview.

Roberts, Alasdair. 2004. “National Security and OpenGovernment.” Georgetown Public Policy Review 9 (2):69-85.

Rosenthal, Peter. 2003. “Disclosure to the Defence afterSeptember 11: Sections 37 and 38 of the Canada EvidenceAct.” Criminal Law Quarterly 48 (2): 186-204.

Royal Commission on Security. 1969. Report of the RoyalCommission on Security. Ottawa: Queen’s Printer.

Information Act and the Privacy Act. 1987. Open andShut: Enhancing the Right to Know and the Right toPrivacy. Ottawa: Queen’s Printer.

House of Commons Standing Committee on Public Safetyand National Security, Subcommittee on the Review ofthe Anti-terrorism Act. 2007. “Rights, Limits, Security:A Comprehensive Review of the Anti-terrorism Actand Related Issues,” 39th Parl.,1st Sess., report 7.Accessed May 1, 2009. http://www2.parl.gc.ca/committeebusiness/ReportsResponses.aspx?Cmte=SECU&Language=E&Mode=1&Parl=39&Ses=1

Information Commissioner of Canada. 1998. Annual Report1997–1998. Ottawa: Minister of Public Works andGovernment Services. Accessed May 1, 2009.http://www.infocom.gc.ca/reports/default-e.asp

------------. 2000. Annual Report 1999-2000. Ottawa:Minister of Public Works and Government Services.Accessed May 1, 2009. http://www.infocom.gc.ca/reports/default-e.asp

------------. 2001. Annual Report 2000-2001. Ottawa:Minister of Public Works and Government Services.Accessed May 1, 2009. http://www.infocom.gc.ca/reports/default-e.asp

------------. 2002. Annual Report 2001–2002. Ottawa:Minister of Public Works and Government Services.Accessed May 1, 2009. http://www.infocom.gc.ca/reports/default-e.asp

------------. 2008. Annual Report 2007-2008. Ottawa:Minister of Public Works and Government Services.Accessed May 1, 2009. http://www.infocom.gc.ca/reports/default-e.asp

Judd, Jim. 2007. “How Should a Democracy Respond toDomestic Terrorist Threats.” Talking points for 2007 RaoulWallenberg International Human Rights Symposium,January 19. Canadian Security Intelligence Service.Accessed May 1, 2009. http://www.csis.gc.ca/nwsrm/spchs/spch19012007-eng.asp

------------. 2008. “Remarks at the Global Futures ForumConference in Vancouver, May 15,” Canadian SecurityIntelligence Service. Accessed May 1, 2009.http://www.csis-scrs.gc.ca/nwsrm/spchs/spch15042008-eng.asp

Law Reform Commission of Canada. 1986. Crimes againstthe State. Ottawa: Law Reform Commission of Canada.

Lynch, Andrew. 2006. “Legislating with Urgency.”Melbourne University Law Review 30 (3): 747.Accessed May 22, 2009. http://search.austlii.edu.au/au/journals/MULR/2006/24.html

Mackmurdo, Christopher. 2007. “Intelligence andInternational Security after 9/11.” In UnderstandingGlobal Terror, edited by Christopher Ankersen andMichael O’Leary. Oxford, UK: Blackwell Publishing.

MacLeod, Ian. 2006. “Ruling Threatens Law That Lets CSISProbe Terrorism.” Ottawa Citizen, November 27.

Macnamara, W.D., and Ann Fitz-Gerald. 2002. “A NationalSecurity Framework for Canada.” IRPP Policy Matters3 (10).

Mayeda, Andrew. 2007. “MPs Push for Contract Details;Harper Government under Fire for SecrecySurrounding Hired Contractors in Afghanistan.”Vancouver Sun, November 26, A6.

“Secret Documents, Secret Challenges.” 2007. Globe andMail, August 8, A16.

Security Intelligence Review Committee. 1989. AnnualReport 1988-89. Accessed May 1, 2009.http://www.sirc-csars.gc.ca/anrran/index-eng.html

Special Senate Committee on the Anti-terrorism Act. 2007.Fundamental Justice in Extraordinary Times: MainReport of the Special Senate Committee on the Anti-Terrorism Act. Accessed May 1, 2009.http://www.parl.gc.ca/39/1/parlbus/commbus/senate/com-e/anti-e/rep-e/rep02feb07-e.htm

Standing Senate Committee on National Security andDefence. 2003. The Myth of Security at Canada’sAirports. Accessed May 4, 2009.http://www.parl.gc.ca/37/2/parlbus/commbus/senate/com-e/defe-e/rep-e/rep05jan03-e.htm

“Statement by the President upon Signing Bill RevisingPublic Information Provisions of the AdministrativeProcedure Act.” 1966. Weekly Compilation ofPresidential Documents 895 (July 4). Accessed May 4,2009. http://www.gpoaccess.gov/wcomp/index.html

Stewart, Hamish. 2003. “Public Interest Immunity after BillC-36.” Criminal Law Quarterly 47 (3): 249.

“Too Much Secrecy.” 2008. Winnipeg Free Press, January26, A18.

Toughill, Kelly. 2008. “Tentacles of Secrecy Grip Tightly.”Toronto Star, August 9, AA6.

Treasury Board of Canada Secretariat. 2008. “Access toInformation Requests — April 1, 2006 to March 31,2007.” InfoSource Bulletin no. 30. Accessed May 4,2009. http://www.infosource.gc.ca/bulletin/2008/bulletin03-eng.asp

United States. Commission on Protecting and ReducingGovernment Secrecy.1997. Report of the Commissionon Protecting and Reducing Government Secrecy,Senate Document 105-2. Washington, DC: GovernmentPrinting Office.

Wadham, John, and Kavita Modi. 2003. “National Securityand Open Government in the United Kingdom.”National Security and Open Government: Striking theRight Balance. Syracuse, NY: Campbell Public AffairsInstitute. Accessed May 4, 2009. http://www.maxwell.syr.edu/campbell/events/pastevents.htm

Walsh, James Igoe. 2007. “Defection and Hierarchy inInternational Intelligence Sharing.” Journal of PublicPolicy 27 (2): 151-81.

Wark, Wesley. 2001. “The Access to Information Act and theSecurity and Intelligence Community in Canada.”Access to Information Review Task Force Report 20.Accessed May 4, 2009. http://www.atirtf-geai.gc.ca/paper-intelligence1-e.html

------------. 2007. “It’s Official: The Government Abused ItsSecrecy Authority.” Globe and Mail, August 10, A17.

Wichmann, Charles J., III. 1998. “Ridding FOIA of Those‘Unanticipated Consequences’: Repaving a NecessaryRoad to Freedom.” Duke Law Journal 47:1213-56.

Woods, Allan. 2008. “Watchdog Calls Detainee Inquiry; SaysTories Stalling Attempts to View Documents on Transferof Afghan Prisoners.” Toronto Star, March 13, A3.

35

Ca

na

da

’s Na

tion

al S

ec

urity

“Co

mp

lex

”: Asse

ssing

the

Se

cre

cy

Ru

les, b

y C

raig

Fo

rce

se

37

L’auteur note que les limites imposées par les loisencadrant l’accès à l’information — et les pénalités dra-coniennes prévues dans certains cas — sont nettementincompatibles avec la démocratie qu’elles sont censéesprotéger. Leur ambiguïté est telle qu’elles pourraient êtreutilisées par le gouvernement pour éviter les situationsembarrassantes et masquer les cas d’incompétence, aunom de la sécurité nationale. Si cela se produisait, ces loisamoindriraient la confiance de la population, ce qui ris-querait d’entraîner des correctifs qui rendraient plus diffi-cile la protection de secrets légitimes. L’auteur formule enterminant une série de recommandations visant à concilierplus efficacement les exigences de secret et de trans-parence dans la pratique et le droit canadiens.

E n matière d’information, le principal dilemme querencontrent les gouvernements consiste à concilierl’intérêt public pour la divulgation de l’information

et la légitime confidentialité de certains renseignements.Dans cette étude, Craig Forcese évalue les efforts du

Canada en vue de trouver un juste équilibre entre trans-parence et secret dans le domaine de la sécurité nationale.Il décrit d’abord les principes sur lesquels reposent lesnotions de « gouvernement transparent » et d’« audiencespubliques », puis les principes de confidentialité enmatière de sécurité nationale, c’est-à-dire les motifs quisont invoqués pour justifier la nécessité de préserver lesecret. L’auteur examine ensuite comment les lois enmatière d’information au Canada protègent la confiden-tialité des questions de sécurité nationale et dans quellemesure celles-ci réussissent à concilier les exigences detransparence et de sécurité nationale, en signalant cer-tains problèmes pratiques et structurels qui compromet-tent cette conciliation. Ces problèmes font dire à l’auteurque les lois en matière d’information et leur mise en pra-tique laissent grandement à désirer au Canada.

Certaines observations des milieux mêmes du ren-seignement et de la sécurité amènent à penser que l’éten-due de la confidentialité prévue dans les lois dépasse lesbesoins de protection légitime. L’incohérence de ces lois —de même que l’incertitude et les réclamations excessivesqu’elles entraînent — crée des conditions susceptibles defreiner les échanges d’information utiles et de miner parconséquent la crédibilité des services de sécurité.

RésuméCanada’s National Security “Complex”:Assessing the Secrecy Rules

Craig Forcese

38

In addition, he says, the limits these laws impose oninformation access, and the draconian penalties theylevel in some instances, are deeply inconsistent with thevery democracy they are supposed to protect. They aresufficiently ambiguous to let government sidestep embar-rassment and mask incompetence, all in the name ofnational security. If they were used to this end, theywould jeopardize confidence in government and poten-tially stimulate changes that make protecting legitimatesecrets more difficult. The study concludes with a numberof recommendations as to how secrecy and transparencymight be reconciled better in Canadian law and practice.

Among other things, he urges Parliament to formallyrepeal section 4 of the Security of Information Act andreplace it with a much more measured provision. He alsocalls on government to standardize its definition of“national security” across the statute book, to avoid theconfusing array of terms that currently exist. Forcese alsoexpresses support for a standing reviewer of nationalsecurity law and policy, and a separate committee of par-liamentarians with legislated powers and access to secretinformation.

T he dilemma of any government information regimelies in balancing the strong public interest in dis-closure in all areas, including national security,

against legitimate secrecy. In this study, Craig Forceseassesses Canada’s efforts to balance transparency withsecrecy in the national security area. He first highlightsthe principles of transparency that animate the Canadianconcepts of open government and open courts. He thenposits several principles of national security confidential-ity – that is, justifications for secrecy predicated onnational security preoccupations. He also describes howCanadian information laws seek to guard national securi-ty confidentiality. In the final section of the study, heassesses how well Canadian information law reconcilesnational security with transparency, identifying a numberof structural and practical problems that plague this rec-onciliation.

Forcese concludes that Canada’s information laws andpractices leave much to be desired. Past commentaryfrom the security and intelligence community suggeststhat secrecy laws are broader than is necessary to protectlegitimate national security secrets. At the same time, theincoherence of these laws, the uncertainty this incoher-ence produces and the overclaiming it allows, create con-ditions that are likely to curb benign informationexchanges. This would be at great expense to the credi-bility of the security services.

Summary Canada’s National Security “Complex”: Assessing the Secrecy Rules

Craig Forcese