COMMONWEALTH OF MASSACHUSETTS SUPREME JUDICIAL COURT No. SJC-12564 ______________________________________________________ COMMONWEALTH OF MASSACHUSETTS, Appellant, v. DENNIS JONES, Appellee. ______________________________________________________ ON INTERLOCUTORY APPEAL FROM MIDDLESEX SUPERIOR COURT ______________________________________________________ BRIEF OF AMICUS CURIAE PROFESSOR ORIN KERR IN SUPPORT OF NEITHER PARTY ______________________________________________________ Orin S. Kerr Francis and John R. Duggan Distinguished Professor Of Law UNIVERSITY OF SOUTHERN CALIFORNIA Gould School of Law 699 Exposition Boulevard Los Angeles, CA 90089 Tel: (213) 740-5617 okerr@law.usc.edu (affiliation for identification purposes only) Andrew Levchuk BBO# 545189 Lauren C. Ostberg BBO# 696883 BULKLEY, RICHARDSON AND GELINAS, LLP 1500 Main Street Suite 2700 Springfield, MA 01115 Tel: 413-272-6285 alevchuk@bulkley.com Amicus Curiae Counsel for Amicus Curiae October 11, 2018 TABLE OF CONTENTS Statement of the Issues...............................1 Interest of the Amicus................................1 Statement of the Case.................................2 Argument ............................................2 I. The Testimonial Aspect of a Gelfgatt Order Is Satisfied by Sufficient Proof that the Subject of the Order Knows the Password.....3 II. The “Reasonable Particularity” Standard Is Irrelevant to the Standard of Proof in This Case........................................6 III. The “Clear and Convincing Evidence” Standard Is an Appropriate Standard To Use for Gelfgatt Orders.............................9 IV. The Court Should Reject Appellee’s Proposed “Proof Beyond A Reasonable Doubt” Standard. ...........................................13 V. The Court Should Also Reject the Commonwealth’s Proposed Preponderance Standard for “Subsidiary Facts.”...........16 Conclusion...........................................19 Certificate of Compliance............................21 Certificate of Service...............................22 Addendum (Unpublished Cases) .................Add. 1-37 Commonwealth v. Tremblay, Mass., No. SJC-12493, slip op. (Oct. 3, 2018).....................Add. 1 United States v. Spencer, U.S. Dist. Ct. No. 17cr-00259-CRB-1 (N.D. Cal. Apr. 26, 2018)...Add. 35 i TABLE OF AUTHORITIES CASES: Commonwealth v. Gelfgatt, 468 Mass. 512 (2014) .........1, 3, 4, 6, 7, 12, 13, 16 Commonwealth v. Tremblay, Mass., No. SJC-12493, slip op. (Oct. 3, 2018). .................................17 Fisher v. United States, 425 U.S. 391 (1976).......9-11 Illinois v. Gates, 462 U.S. 213 (1983) ...........17-18 In re Winship, 397 U.S. 358 (1970) ..................14 Salinas v. Texas, 570 U.S. 178 (2013) ...............14 Tehan v. U.S. ex rel. Shott, 382 U.S. 406 (1966) ....14 United States v. Kirschner, 823 F. Supp. 2d 665 (E.D. Mich. 2010) ....................................5 United States v. Spencer, U.S. Dist. Ct. No. 17-cr00259-CRB-1 (N.D. Cal. Apr. 26, 2018)..........8, 9, 12 OTHER AUTHORITIES: Kerr, Compelled Decryption and the Privilege Against Self-Incrimination, Tex. L. Rev. (forthcoming 2019), http://bit.ly/KerrPaper...........3, 4, 7, 8, 9, 11, 12 Orin S. Kerr, A Revised Approach to the Fifth Amendment and Obtaining Passcodes, Wash. Post, Sept. 25, 2015, http://bit.ly/5Apassword ...................5 3 W.R. LaFave, J.H. Israel, N.J. King, & O.S. Kerr, Criminal Procedure (4th ed. 2015)...................6-7 ii STATEMENT OF ISSUES Amicus will address only the first issue raised by this Court’s solicitation of amicus briefs: “What is the burden of proof that the Commonwealth bears on a motion . . . in order to establish a 'foregone conclusion,' as that term is used in Commonwealth v. Gelfgatt, 468 Mass. 512, 520-526 (2014)?” INTEREST OF THE AMICUS Orin S. Kerr is the Frances R. and John J. Duggan Distinguished Professor of Law at the University of Southern California Gould School of Law. He teaches and writes in the area of computer crime law and digital evidence. His scholarship has been cited by this Court in several published decisions, including most recently Commonwealth v. Keown, 478 Mass. 232, 241 (2017); Ajemian v. Yahoo!, Inc., 478 Mass. 169, 174 (2017); Commonwealth v. Fulgiam, 477 Mass. 20, 31 (2017); Commonwealth v. Chamberlin, 473 Mass. 653, 658 (2016); Commonwealth v. Dorelas, 473 Mass. 496, 506 (2016); and Commonwealth v. Gelfgatt, 468 Mass. 512, 541 (2014). Professor Kerr’s interest is in the sound development of the law. 1 STATEMENT OF THE CASE Amicus adopts the Statement of the Case from the Brief for the Commonwealth filed on August 27, 2018. ARGUMENT The Court should hold that the Commonwealth must prove by clear and convincing evidence, based on a totality of the circumstances, that the subject of the order knows the password required to unlock the device. When the Commonwealth presents the subject of a Gelfgatt order with a password prompt to a phone, and the subject is ordered to enter the password needed to unlock the phone, the foregone conclusion doctrine is satisfied by sufficient proof that the subject knows the password. Although the precise burden of proof to use is a surprisingly open question, a clear and convincing standard is an appropriate standard for both formalist and functional reasons. The Court should reject any special standard for subsidiary facts, and it should also reject any test based on the “reasonable particularity” standard. 2 I. The Testimonial Aspect of a Gelfgatt Order Is Satisfied by Sufficient Proof that the Subject of the Order Knows the Password. It helps to begin with the underlying Fifth Amendment test. When the Commonwealth obtains an order requiring a subject to enter in a password1 to unlock a device, and it presents the device to the suspect at the password prompt, the foregone conclusion doctrine is satisfied when the Commonwealth makes an independent showing that it already knows that the subject knows the password. See Kerr, Compelled Decryption and the Privilege Against SelfIncrimination, Tex. L. Rev. (forthcoming 2019), http://bit.ly/KerrPaper (hereinafter Kerr, Compelled Decryption), at 13-21; Commonwealth v. Gelfgatt, 468 Mass. 512, 522-25 (2014). This is true because the act of entering the password at the password prompt amounts to testimony that the subject knows the password that unlocks the phone. See Kerr, Compelled Decryption at 14-16. Sufficient proof that the Commonwealth already knows 1 For convenience, this brief refers to the unique sequence of characters that decrypts a device as a “password.” The sequence may actually be a passcode, PIN, or other string of characters. 3 that the target knows the password makes that implicit testimony a “foregone conclusion” and the Fifth Amendment poses no bar to the order. See id. at 16- 21; Gelfgatt, 468 Mass. at 522-25. One caveat to this analysis is that the Commonwealth’s proposed order in this case deviates in an important way from a Gelfgatt order. The order approved in Gelfgatt required its subject to enter passwords that unlocked digital storage devices without the Commonwealth “view[ing] or record[ing] the password or key in any way.” Gelfgatt, 468 Mass. at 517, n.10. In contrast, the proposed order that the Commonwealth sought below would have required the defendant to “provide the Commonwealth in writing with the PIN code.” Proposed Order Requiring Produc. of PIN Access Code, Commonwealth’s Impounded R. App. 102.2 The distinction matters because the Fifth Amendment implications of entering a password and 2 The record appendix in this case has been impounded. In the course of preparing this brief, however, amicus asked the Commonwealth for the text of the proposed order that was denied by the trial court. In response, the Commonwealth provided amicus with a copy of the proposed order and its corresponding page numbers in the impounded appendix. The Commonwealth has not disclosed any other information about or text of any materials in the impounded record appendix. 4 revealing a password are different. The foregone conclusion doctrine does not apply when an order commands a person to write down a password and provide it to the Commonwealth. In such cases, the Commonwealth seeks direct testimony. Application of the Fifth Amendment hinges on whether the testimony is incriminating. See, e.g., United States v. Kirschner, 823 F. Supp. 2d 665, 669 (E.D. Mich. 2010) (quashing a subpoena to reveal computer passwords because the “government is not seeking documents or objects — it is seeking testimony from the Defendant, requiring him to divulge through his mental processes his password — that will be used to incriminate him”); Orin S. Kerr, A Revised Approach to the Fifth Amendment and Obtaining Passcodes, Wash. Post, Sept. 25, 2015, http://bit.ly/5Apassword (discussing the Fifth Amendment implications of disclosing passwords). Given this discrepancy, amicus assumes that any court order in this case will follow the Gelfgatt protocol of requiring its subject to enter the password without disclosing it to the Commonwealth. Applying that protocol requires a reviewing court to 5 address the burden of proof about which the Court sought briefing by amici. II. The “Reasonable Particularity” Standard Is Irrelevant to the Standard of Proof In This Case. The burden of proof to establish a foregone conclusion to enter a password raises a novel question not yet answered by appellate cases.3 understand why. It helps to Gelfgatt orders require the application of the foregone conclusion doctrine in a setting different from that with traditional production orders. The subject is “not selecting documents and producing them, but merely entering a password into encryption software.” Gelfgatt, 468 Mass. at 524 n.14. That makes a big difference. In the context of production orders, courts have often articulated the standard of proof for foregone conclusions using a “reasonable particularity” test. See, e.g., 3 W.R. LaFave, J.H. Israel, N.J. King, & O.S. Kerr, Criminal 3 This brief assumes that the relevant standard is provided by the Fifth Amendment right against selfincrimination rather than Art. 12 of the Massachusetts Declaration of Rights. See Gelfgatt, 468 Mass. at 525 (noting the greater protections of Art. 12). This brief takes no position on the proper burden of proof under the Massachusetts Declaration of Rights. 6 Procedure § 8.13(a) (4th ed. 2015). That test makes sense when the government describes the files the subject of the order must produce. The government’s description of the files with “reasonable particularity” establishes that the government knows of the files in the suspect’s possession and therefore establishes a foregone conclusion: The basic idea is that a specific description of what the government seeks necessarily reflects greater government knowledge about it. If the government’s specific description of the documents to be handed over shows that the government already knows their existence, possession, and authenticity -- the testimonial aspect of production -- then the foregone conclusion doctrine applies. If the government can pinpoint what it needs, the thinking runs, then it is not relying on the truth-telling of the person complying with the order to figure out its case. Kerr, Compelled Decryption at 9-10. That test “has no application” to a Gelfgatt order, however. Id. at 25. With a Gelfgatt order, the subject is “not selecting documents and producing them, but merely entering a password into encryption software.” Gelfgatt, 468 Mass. at 524 n.14. Because orders to decrypt do not direct their subject to produce particular files described in the order, the reasonable particularity standard is irrelevant: 7 Whatever the merits of the “reasonable particularity” standard in the specific context of subpoenaed documents, the test is notably unilluminating as to the government’s burden outside that context. The government can compel an act that has testimonial qualities but does not require the government to describe the evidence it is seeking. The act may be to do something, not to go get something. As a result, there may be no evidence for the target to retrieve that can be described with “reasonable particularity.” The nature of the burden of proof in outside of orders to compel documents remains surprisingly unclear. Kerr, Compelled Decryption at 10. As Judge Charles Breyer has explained: It is nonsensical to ask whether the government has established with “reasonable particularity” that the defendant is able to decrypt a device. While physical evidence may be described with more or less specificity with respect to both appearance and location, a defendant’s ability to decrypt is not subject to the same sliding scale. He is either able to do so, or he is not. Accordingly, the reasonable particularity standard cannot apply to a defendant’s ability to decrypt a device. United States v. Spencer, U.S. Dist. Ct., No. 17-cr00259-CRB-1, at 3 (N.D. Cal. Apr. 26, 2018). The new context of Gelfgatt orders explains why the burden of proof to apply is a new question that no appellate court has yet answered: [T]here is little or no precedent to cite. Particularity-based standards of proof 8 articulated in the context of subpoenas for documents have kept courts from confronting the degree of certainty required for a fact to be a foregone conclusion. Whatever the best answer is, a proper understanding of the foregone conclusion doctrine now requires courts to answer it. Kerr, Compelled Decryption at 26-27. III. The “Clear and Convincing Evidence” Standard Is an Appropriate Standard to Use for Gelfgatt Orders. In my view, the best choice for the standard of proof to use for decryption orders is probably the clear and convincing evidence standard used by Judge Breyer in United States v. Spencer. See United States v. Spencer, supra, No. 17-cr-00259-CRB-1 at 3. is probably the best choice for two reasons. This The first reason is formalist, and the second reason is functional. The formalist reason why the clear and convincing evidence standard appears appropriate is that it seems consistent with the Supreme Court decision that established the foregone conclusion doctrine, Fisher v. United States, 425 U.S. 391 (1976). In the course of ruling that the foregone doctrine applied, Fisher used words suggesting that the government’s burden is 9 significant but not extreme. Fisher states that “we are confident that . . . the act of producing them . . . would not itself involve testimonial selfincrimination.” Id. at 410-11 (emphasis added). Similarly, in rejecting the claim of privilege, Fisher says that the Court was “doubtful” that assertions implicit in production were enough for the privilege to apply. Id. at 11 (emphasis added). These verbal clues seem consistent with a clear and convincing evidence standard. When the Fisher Court was “confident” that the testimony was a foregone conclusion and “doubtful” that it was not, the Court concluded that the government had satisfied its burden of proof. Id. at 410-11. Although Fisher does not articulate the standard, its words appear (at least to me) consistent with something higher than a preponderance standard but lower than a proof-beyonda-reasonable-doubt standard. Functional concerns also provide some support for a clear-and-convincing-evidence standard. To see why, we need to step back and understand the purpose of the doctrine. The foregone conclusion doctrine considers whether the government would obtain any prosecutorial 10 advantage at trial from the subject’s implied assertions. See Fisher, 425 U.S. at 410-411. The doctrine prevents the mere form of testimonial assertions inherent in production from wrongly protecting nontestimonial evidence of what it produces. See Kerr, Compelled Decryption at 11-13. So viewed, a functional approach to the burden of proof might reasonably assign a burden based on what standard best approximates the elimination of prosecutorial advantage. If a conclusion is “foregone” when the government does not need the suspect’s testimonial act to prove it, perhaps a foregone conclusion doctrine applies when the government’s proof roughly matches the burden the government would have to meet for that conclusion at trial. A clear and convincing evidence standard strikes me as a fair approximation of a burden needed to eliminate that advantage. When the government can show by clear and convincing evidence that it already knows the facts implicit in the suspect’s testimonial acts, it is highly unlikely that the government will obtain any trial advantage from the testimony implicit 11 in that act. The clear and convincing standard is sufficient to be confident that the government need not and will not rely on the testimonial act at trial to prove its case. That is especially so if the court applies the estoppel rule adopted in Spencer that the government cannot rely on the foregone conclusion doctrine and then later introduce evidence of the testimonial act at trial. See Spencer, No. 17-cr-00259 at 3 (“Once Spencer decrypts the devices, however, the government may not make direct use of the evidence that he has done so.”); Kerr, Compelled Decryption at 11 (“If the government’s power to compel an act depends on not needing testimony the act implies, the government should not be allowed to later use the implied testimony it claimed not to need.”). When the government shows by clear and convincing evidence that it already knows the testimonial assertion implicit in an act, and it agrees not to use that testimonial assertion against the suspect in a criminal case, the odds that the compelled testimonial assertion will add to the government’s case are remote. Cf. Gelfgatt, 489 Mass. at 517 n.10 (listing 12 such a use limit in the order itself at Para. 7, stating that “the Commonwealth shall be precluded from introducing any evidence relating to this Order or the manner in which the digital media in this case was decrypted in its case in chief.”); Proposed Order Requiring Produc. of PIN Access Code, Commonwealth’s Impounded R. App. 103 (including a similar limit in the proposed order in this case). IV. The Court Should Reject Appellee’s Proposed “Proof Beyond A Reasonable Doubt” Standard. Appellee’s brief argues that the standard should be proof beyond a reasonable doubt. Appellee 24-27. See Br. of the According to Appellee, that standard “more aptly fits the level of certainty inherent in the phrase foregone conclusion.” Id. at 24. Amicus disagrees. Selecting the standard of proof requires broad consideration of history, precedent, and purpose, not the innate meaning of isolated words found in caselaw. That broader perspective suggests that proof beyond a reasonable doubt is a trial standard that does not align with the concerns animating the privilege against self-incrimination. Due process requires that the government must prove every element 13 of a crime beyond a reasonable doubt because an individual’s freedom is at stake. In re Winship, 397 U.S. 358, 361-64 (1970). Assigning such a high burden acts as “a prime instrument for reducing the risk of convictions resting on factual error.” Id. at 363. In contrast, “the basic purposes that lie behind the privilege against self-incrimination do not relate to protecting the innocent from conviction, but rather to preserving the integrity of a judicial system[.]” Tehan v. U.S. ex rel. Shott, 382 U.S. 406, 415 (1966). The individual must plead the Fifth and have a court assess the privilege before (often long before) any trial might occur. See generally Salinas v. Texas, 570 U.S. 178 (2013). Although the individual must show that an answer would be incriminating, the individual may not even be a criminal suspect at that time. The Commonwealth could reasonably know only a small part of the evidence that it would later learn before deciding to bring a criminal case against the individual. For these reasons, the trial standard of proof beyond a reasonable doubt standard is not wellsuited for the foregone conclusion doctrine. 14 Significantly, the prosecutorial-advantagedefeating role of the foregone conclusion doctrine does not support using the proof-beyond-a-reasonable doubt standard. The reason is that an assertion implied by an act is generally distinct from the element of a crime to which the assertion may relate. The likelihood that the suspect’s implied testimony will give the government a prosecutorial advantage is ordinarily not connected to the likelihood that the implied testimony will prove an element of the offense. An example helps show the point. Imagine that the Commonwealth has probable cause to believe that a suspect’s locked phone contains evidence of his robbing a jewelry store. The Commonwealth uses that cause to obtain both a search warrant to search the phone for robbery evidence and an accompanying Gelfgatt order requiring the suspect to enter the password and unlock the phone. The suspect pleads the Fifth, and the reviewing court must evaluate the strength of the Commonwealth’s case that it knows the suspect knows the password. 15 In such a case, the prosecutorial advantage in the implied testimony of the suspect knowing the password has no obvious connection to whether the Commonwealth can establish the elements of robbery. Whether the suspect knows the password to a phone has no direct bearing on whether the evidence found on the phone establishes the suspect’s guilt. For example, say the phone contains photographs of the suspect with the robbery proceeds. Although the photographs would be highly incriminating, their incriminating nature derives from the photographs themselves rather than the suspect’s knowledge of the password that was used to access them. V. See Gelfgatt, 468 Mass. at 524 n.14. The Court Should Also Reject the Commonwealth’s Proposed Preponderance Standard for “Subsidiary Facts.” The Commonwealth’s brief proposes a clear-and- convincing-evidence standard with the following twist: In the Commonwealth’s view, a court applying the foregone conclusion standard to review a Gelfgatt motion should apply a preponderance of the evidence standard to subsidiary facts. See Br. for the Commonwealth 22-27. Under the Commonwealth’s proposal, a trial court would apply the foregone 16 conclusion doctrine by first assessing the factual picture based on a preponderance standard, and then, after establishing the relevant facts, considering whether those facts amount to clear-and-convincing evidence that the suspect knows the password to the device. See id. at 22-27. This is a mistake. Identifying a standard for “subsidiary facts” arises when a trial court must determine what facts to consider to assess whether evidence is admissible at trial. But as explained above, Gelfgatt motions do not involve trial evidence. The only question is whether the Commonwealth can show by clear and convincing evidence that the subject of the order knows the passcode needed to unlock the device. There are no subsidiary facts to consider.4 Instead, as with an assessment of probable cause under the Fourth Amendment, the evidence should all be considered as possible evidence under the totality of 4 The issue here is distinct from the appellate standard of review for subsidiary facts analyzed in Commonwealth v. Tremblay, Mass., No. SJC-12493, slip op. (Oct. 3, 2018). Tremblay considers the appellate standard of review for subsidiary facts found by a trial court. In such a case, the question is how much deference an appellate court should give to a trial court’s determination of the facts needed to apply the legal standard. See id. at 1. In contrast, the Commonwealth’s proposal is for the trial court to use a special standard for analyzing subsidiary facts. 17 the circumstances in light of its apparent likelihood. Cf. Illinois v. Gates, 462 U.S. 213, 234 (1983) (noting, in the context of a warrant affidavit, that a “totality-of-the-circumstances analysis . . . permits a balanced assessment of the relative weights of all the various indicia of reliability” of the evidence). Adopting the Commonwealth’s preponderance standard for “subsidiary facts” would only water down the clear and convincing evidence standard. To see why, imagine a reliable informant previously told an officer that a criminal suspect regularly used one of two phones found in an abandoned car. But there’s a catch: The officer’s memory is hazy about which phone the informant identified. Based on the officer’s uncertain testimony, the trial court concludes that there is a 51% chance that the informant identified a particular phone as the one the suspect regularly used. Now imagine the Commonwealth obtains a Gelfgatt order, the suspect pleads the Fifth, and the reviewing judge must determine if the Commonwealth has proven the suspect’s knowledge of that phone’s password by clear and convincing evidence. 18 How should the court treat the officer’s uncertain testimony? Under the Commonwealth’s approach, the judge must treat as 100% certain that the informant identified that particular phone as the one regularly used by the suspect. Because the Commonwealth established the identification of that phone by a preponderance of the evidence, the judge would have to assume it as a true “subsidiary fact” and consider whether the identification showed by clear and convincing evidence that the suspect knew the phone’s password. That would make little sense. Doubt about the existence of subsidiary facts should not be ignored simply because the existence of those facts is more likely than not. CONCLUSION A court reviewing a Gelfgatt motion should consider whether the Commonwealth has established by clear and convincing evidence that the suspect knows the password based on a totality of the circumstances. 19 October 11, 2018 Respectfully submitted, ORIN S. KERR Francis and John R. Duggan Distinguished Professor of Law USC Gould School of Law 699 Exposition Boulevard, Room 462 Los Angeles, CA 90089 Tel: 213-740-5617 okerr@law.usc.edu Amicus Curiae By His Attorneys, ___________________________ Andrew Levchuk, BBO# 545189 Lauren C. Ostberg, BBO# 696883 BULKLEY, RICHARDSON AND GELINAS, LLP 1500 Main Street, Suite 2700 Springfield, MA 01115-5007 Tel: 413-272-6285 alevchuk@bulkley.com 20 CERTIFICATE OF COMPLIANCE Pursuant to Rule 16(k) of the Massachusetts Rules of Appellate Procedure, the undersigned counsel that that this brief complies with the applicable rules of court that pertain to the filing or briefs, including but not limited to Mass. R. App. P. 16(e), 16(f), 16(h), 17, and 20. _____________________ Lauren C. Ostberg 21 CERTIFICATE OF SERVICE I, Lauren C. Ostberg, hereby certify that on October __, 2018, I caused two copies of the foregoing Brief of Amicus Curiae by Professor Orin Kerr to be delivered via first-class U.S. Mail, postage prepaid, to each party separately represented as identified below: Gabriel Pell, Esq. Office of the District Attorney, Middlesex County 15 Commonwealth Avenue Woburn, MA 01801 Counsel for the Commonwealth George Ohlson, Jr., Esq. Ohlson & Ohlson, Attorneys at Law 92 Montvale Avenue, Suite 2600 Stoneham, MA 02180 James A. Reidy, Esq. P.O. Box 920365 Needham, MA 02492 Counsel for Defendant-Appellee _____________________ Lauren C. Ostberg 3006144v1 22