digital privacy at risk under modern self-incrimination jurisprudence

Late last year I read some interesting reporting from BBC about privacy in the digital age. The case at issue involved the state of Florida compelling an alleged voyeur to reveal the lock code of his iPhone which had been seized with probable cause. The defendant refused, citing his Fifth Amendment privilege not to incriminate himself, but a Florida appeals court has ruled that the privilege could not have been invoked to block surrender of the code.

bbc stock footage never looked so corporate and scary.

That’s an interesting result by itself; though, as a state-level appeal, that decision only has controlling precedent (for now) in Florida. But the BBC reporting really caught my attention for revealing some of the rationale of the appeals court’s decision, authored by Judge Anthony Black, in the story’s second half:

Judge Black referred to a famous Supreme Court case, Doe v US 1988, in which Justice John Paul Stevens wrote that a defendant could be made to surrender a key to a strongbox containing incriminating documents but they could not “be compelled to reveal the combination to his wall safe”.

“We question whether identifying the key which will open the strongbox – such that the key is surrendered – is, in fact, distinct from telling an officer the combination,” wrote Judge Black. “More importantly, we question the continuing viability of any distinction as technology advances.”

It happens that Justice Stevens was a favourite of mine before his retirement from the bench in 2010, and in fact this blog reviewed his 2014 book detailing his proposed constitutional amendments. I understood him to be a left-leaning civil libertarian, so it was surprising to hear an opinion of his cited to diminish the reach of one of the pillars of the Bill of Rights. I found the Doe opinion to be a quite interesting exercise in linguistic gymnastics, particularly, as Judge Black noted, amid the rapid advance of technology and the numerous and contentious questions about privacy that it has created. But beyond that, the case history led me down a 60-year-long rabbit hole of jurisprudence that I believe serves as a cautionary tale of how the Court can get it very wrong and the dramatic effects it can have on basic rights. With great care taken to exclude superfluous detail, I summarize that epic journey for you here.

Doe v. United States (1988)

The first thing I noted, to my relief, was that Stevens’s opinion was actually a dissent. A few additional facts of this remarkable 1988 case are also worth noting. Doe, the defendant in a federal grand jury probe over oil cargo fraud, had produced some bank records but took the Fifth when questioned about the “existence or location of additional records.” The US branches of the banks (headquartered in the Cayman Islands and Bermuda) were served subpoenas but refused to comply with them citing their governments’ bank secrecy laws which required the customer’s consent [1]. The Government then filed a motion to get Doe to sign disclosure consent forms for 12 accounts, explicitly citing their account numbers, that it suspected Doe maintained. The District Court denied the motion since so doing would constitute a testimonial act: implicit admission (the Government could argue) that Doe controlled the accounts, which could lead to an indictment against Doe not yet obtained.

It turns out that under then-fairly-recent precedent, Fisher v. United States (1976), the “testimonial” nature of a compelled and incriminating statement is the legal standard [2] for whether Fifth Amendment privilege may be invoked to refuse to make it. The Government subsequently reformulated the consent form to “apply to any and all accounts over which Doe had a right of withdrawal, without acknowledging the existence of any such account.” This Orwellian formulation is all the more worth partially reproducing here (emphasis mine) since this was going on in 1984:

I, [Doe], …do hereby direct any bank or trust company at which I may have a bank account of any kind or at which a corporation has a bank account of any kind upon which I am authorized to draw… to disclose all information and deliver copies of all documents of every nature in your possession or control which relate to said bank account to Grand Jury 84-2… and this shall be irrevocable authority for so doing… and shall be construed as consent with respect [to the relevant Cayman Island and Bermuda laws] as the same shall apply to any of the bank accounts for which I may be a relevant principal.

Thence some entertaining back and forth [3] before the actual holding.

Majority Opinion

The Court eventually ruled that signing the consent from was not testimonial, the test referenced from Fisher, and made several variously interesting and surprising arguments.

…the directive itself does not point the Government toward hidden accounts or otherwise provide information that will assist the prosecution in uncovering evidence. The Government must locate that evidence ‘by the independent labor of its officers’ [citation omitted]. As in Fisher, the Government is not relying upon the ‘truth-telling’ of Doe’s directive to show the existence of, or his control over, foreign bank account records.

Further, Doe’s “execution of the form [would not] admit the authenticity of any records produced by the bank… authentication evidence would have to be provided by bank officials.”

Addressing Doe’s extra-Fifth defense arguments, the Court found the right to remain silent [4] applies to “custodial interrogation” but not absolutely, as evidenced by several recent cases [5]. Doe’s “heavy reliance” on Kastigar v. United States (1972), which upheld a federal law permitting compelled testimony contingent upon subsequent immunity from its use, “as well as evidence derived directly and indirectly therefrom,” was “misguided” since that prohibition “assumes that the suspect’s initial compelled communication is testimonial.” Take note of the repeated use of that word ‘testimonial.’ But finally, and perhaps most striking, executing the “consent” form somehow doesn’t actually “admit or assert Doe’s consent.” Explains the majority:

The form does not state that Doe “consents” to the release of bank records. Instead, it states that the directive “shall be construed as consent” with respect to Cayman Islands and Bermuda bank secrecy laws. Because the directive explicitly indicates that it was signed pursuant to a court order, Doe’s compelled execution of the form sheds no light on his actual intent or state of mind.

Stevens’s Dissent

Stevens’s dissent, exemplifying an approach which Wikipedia says “has usually been characterized as idiosyncratic,” is a lot more straightforward. A footnote addresses that matter of consent thus:

That the directive asserts that it was executed “pursuant to” court order does not save petitioner from this compelled admission. Only the most sophisticated bank officer could be expected to understand the phrase ‘pursuant to that certain order,’ ibid., to mean “executed involuntarily under pain of contempt.”

But in the dissent proper, Stevens quickly gets to the line which sent me down this path of discovery:

[A defendant] may in some cases be forced to surrender a key to a strongbox containing incriminating documents, but I do not believe he can be compelled to reveal the combination to his wall safe — by word or deed.

Here the majority explicitly disagrees, believing the compulsion in this case to be closer to surrendering a key [6]. Stevens continues, explicitly addressing the “voice exemplar” style of argument that the “forced execution of this document differs from the forced production of physical evidence.” [7] Citing both Miranda v. Arizona (1966) (see also [4]) and Schmerber v. California (1966) (see also [5], [7]), Stevens delivers what I find to be his most powerful argument:

The deviation from this [right to silence] principle can only lead to mischievous abuse of the dignity the Fifth Amendment commands the Government afford its citizens… In allowing the Government to compel petitioner to execute the directive, the Court permits the Government to compel petitioner to speak against his will in answer to the question “Do you consent to the release of these documents.” Beyond this affront, however, the Government is being permitted also to demand that the answer be “yes.”

Fisher v. United States (1976)

Much of the majority opinion in Doe (1988) rests on Fisher, a 1976 case ruling on two cases about whether taxpayers could retain their Fifth amendment privilege about documents they didn’t author and didn’t then posses. In both cases, the defendants had accountants who prepared their tax returns. The Government investigated them for civil and criminal lawbreaking, and for the sake of their defense the defendants had their accountants forward to their lawyers the documents the accountants retained in the course of that preparation. The defendants’ lawyers later refused to produce those documents to the Government’s lawyers, citing their clients’ Fifth amendment privilege.


In a unanimous 8-0 decision [8] that can easily be read as “activist”, the Court ruled that the client-attorney privilege actually did enable the refusal on the part of the lawyers [9], but, citing the scholarly work of John Wigmore, only if it would have allowed the refusal on the part of their clients if the documents had remained in their possession. Case law around document production rested squarely with a then-ninety-year-old case, Boyd v. United States (1886), which plainly held that “seizure or compulsory production of a man’s private papers to be used in evidence against him is equivalent to compelling him to be a witness against himself.” Broader but germane guidance for Fifth Amendment jurisprudence came from Counselman v. Hitchcock (1892) which declared that the privilege “is as broad as the mischief against which it seeks to guard.”

But, citing (among others) a flurry of cases from the late 1960s, the majority declared that “several of Boyd‘s express or implicit declarations have not stood the test of time.” Notwithstanding Boyd‘s pronouncement of the sanctity of private papers “often [appearing] as dictum in later opinions of this Court”, the Fisher Court essentially if not explicitly overturned Boyd, declaring:

…the prohibition against forcing the production of private papers has long been a rule searching for a rationale consistent with the proscriptions of the Fifth Amendment against compelling a person to give “testimony” that incriminates him.

I find this to be a shocking passage since the quoted word “testimony” is in fact nowhere to be found in the text of the Fifth Amendment! That doctrine instead derives from the much-referenced Schmerber case, to be discussed shortly.

Perhaps the most interesting result of Fisher is the “act of production” doctrine, which dictates that the Fifth Amendment “applies only when the accused is compelled to make a testimonial communication that is incriminating” and that merely producing documents is only a testimonial communication if it provides information not already known about those documents’ existence, custody, or authenticity. In other words, documents prepared in the past cannot themselves be testimonial; only metadata about those documents can be. Central to this doctrine is the “foregone conclusion” doctrine, named for a single, seemingly minor passage (also referenced in part by Doe) which again cites Wigmore in exploring the possible testimonial aspects of production:

Surely the Government is in no way relying on the “truthtelling” of the taxpayer to prove the existence of or his access to the documents. [Wigmore]. The existence and location of the papers are a foregone conclusion, and the taxpayer adds little or nothing to the sum total of the Government’s information by conceding that he, in fact, has the papers. Under these circumstances, by enforcement of the summons, “no constitutional rights are touched. The question is not of testimony, but of surrender.” [citation omitted].

Brennan’s Concurrence

While the judgement of the Court was unanimous, it’s well worth noting that Justice Brennan filed a lengthy concurring opinion. Waiving doctrine from Boyd due to the documents’ “wholly business, rather than personal, nature”, and briefly noting that the accountants also had access to the papers anyway, Brennan agrees that the summonses were unprotected by Fifth amendment privilege. But he doesn’t join the majority opinion, saying “it is but another step in the denigration of privacy principles settled nearly 100 years ago” by Boyd. Before a detailed examination of Boyd and other case law (including Schmerber, for which Brennan had written the majority opinion), Brennan declares (citing two 20th century cases) “History and principle, not the mechanical application of its wording, have been the life of the Amendment.”

Brennan’s specific rejection of the majority opinion in Fisher is deserving of its own lengthy analysis. Of particular relevance to the arguments in Doe based upon it (and by extension, Judge Black’s analysis in the iPhone case) is this passage (emphasis mine):

I perceive no principle which does not permit compelling one to disclose the contents of one’s mind but does permit compelling the disclosure of the contents of that scrap of paper by compelling its production. Under a contrary view, the constitutional protection would turn on fortuity, and persons would, at their peril, record their thoughts and the events of their lives. The ability to think private thoughts, facilitated as it is by pen and paper, and the ability to preserve intimate memories would be curtailed through fear that those thoughts or the events of those memories would become the subjects of criminal sanctions however invalidly imposed. Indeed, it was the very reality of those fears that helped provide the historical impetus for the privilege.

And about the foregone conclusion doctrine limiting the testimonial aspect of production, and hence the scope of the Fifth amendment privilege:

I know of no Fifth Amendment principle which makes the testimonial nature of evidence and, therefore, one’s protection against incriminating himself, turn on the strength of the Government’s case against him… “the privilege ‘is as broad as the mischief against which it seeks to guard.’…” [Schmerber, citing Counselman]

Schmerber v. California (1966)

As so much of Fisher and its progeny rested on Schmerber, studying that case was the next step in understanding how the plain language of the Fifth Amendment had so unraveled. In Schmerber, the petitioner had been driving drunk, struck a tree, and was thence taken to hospital. While conscious [10] and against both his will and advice of counsel, the petitioner was blood tested and judged to be intoxicated on that basis, and that evidence was entered into the trial which resulted in his conviction. Significantly, the case was decided by a 5-4 Supreme Court with a majority opinion (and a separate concurring opinion) met by four different dissenting opinions.

Majority Opinion

Many arguments were unsuccessfully raised to the admission of the blood sample, but the argument against self-incrimination alone is relevant here. Addressing Miranda, the narrow majority of the Schmerber Court significantly asserted that “the privilege has never been given the full scope which the values it helps to protect suggest” by citing in part Holt v. United States (1910), a case in which a defendant was made to wear a blouse by the prosecution:

the prohibition of compelling a man in a criminal court to be witness against himself is a prohibition of the use of physical or moral compulsion to extort communications from him, not an exclusion of his body as evidence when it may be material. The objection in principle would forbid a jury to look at a prisoner and compare his features with a photograph in proof.

Reconciling that observation about a class of physical evidence sources with the clear prohibition on the production of private papers in Boyd, the majority formalized the language that would later form the basis of Fisher:

The distinction which has emerged, often expressed in different ways, is that the privilege is a bar against compelling “communications” or “testimony,” but that compulsion which makes a suspect or accused the source of “real or physical evidence” does not violate it.

Even while ruling against the defendant, the majority seemed to be aware of the precipice they approached with this language, and they included several caveats:

To compel a person to submit to testing in which an effort will be made to determine his guilt or innocence on the basis of physiological responses, whether willed or not, is to evoke the spirit and history of the Fifth Amendment… we [therefore] reach this judgment only on the facts of the present record. The integrity of an individual’s person is a cherished value of our society.

Black’s Dissent

Justice Black, joined by Justice Douglas — both liberal stalwarts appointed by FDR and among the longest-serving justices in history — was unsatisfied by those caveats and the basis they defended, and said so in plain language that is worth reproducing at length  especially since it addressed Wigmore (emphasis mine):

I think it unfortunate that the Court rests so heavily for its very restrictive reading of the Fifth Amendment’s privilege against self-incrimination on the words “testimonial” and “communicative.” These words are not models of clarity and precision, as the Court’s rather labored explication shows. Nor can the Court, so far as I know, find precedent in the former opinions of this Court for using these particular words to limit the scope of the Fifth Amendment’s protection. There is a scholarly precedent, however, in the late Professor Wigmore’s learned treatise on evidence. He used “testimonial” which… means “communicative” [citation omitted], as a key word in his vigorous and extensive campaign designed to keep the privilege against self-incrimination “within limits the strictest possible.” [Wigmore]. Though my admiration for Professor Wigmore’s scholarship is great, I regret to see the word he used to narrow the Fifth Amendment’s protection play such a major part in any of this Court’s opinions.

I am happy that the Court itself refuses to follow Professor Wigmore’s implication that the Fifth Amendment goes no further than to bar the use of forced self-incriminating statements coming from a “person’s own lips.” It concedes, as it must so long as Boyd v. United States stands, that the Fifth Amendment bars a State from compelling a person to produce papers he has that might tend to incriminate him. It is a strange hierarchy of values that allows the State to extract a human being’s blood to convict him of a crime because of the blood’s content, but proscribes compelled production of his lifeless papers…

A basic error in the Court’s holding and opinion is its failure to give the Fifth Amendment’s protection against compulsory self-incrimination the broad and liberal construction that Counselman and other opinions of this Court have declared it ought to have.

The liberal construction given the Bill of Rights’ guarantee in Boyd v. United States, supra, which Professor Wigmore criticized severely, makes that one among the greatest constitutional decisions of this Court.

I find it profoundly significant that two of the most prolific justices to ever serve regretted that Wigmore’s word ‘testimonial’ could (emphasis mine) “play such a major part in any of this Court’s opinions” and opined that Boyd‘s liberal construction of the Bill of Rights from eighty years before was among the “greatest constitutional decisions” the Supreme Court ever made; yet only ten years later that very Court functionally overturned its own great constitutional decision to replace it with a test substantially based on Wigmore’s regrettable word. That’s exactly what Fisher did, with contortions based upon it appearing in Doe just over ten years later still.

United States v. Hubbell (2000)

Somewhere along my trek through these cases I actually read Judge Black’s opinion in the motivating state district court case about iPhone voyeurism, Florida v. Stahl (2016). It cites all the cases I’ve reviewed here, several that those cases themselves reference, and many more besides. One such significant case that piqued my interest — not least since it sprang from Ken Starr’s independent counsel investigation of the Whitewater scandal during the Clinton administration — was the near-unanimous decision in United States v. Hubbell (2000).

Webster Hubbell — an attorney Bill Clinton had made Arkansas Supreme Court chief justice in the 1980s — had plead guilty to charges of mail fraud and tax evasion from when he worked at a private practice. As part of his plea deal he agreed to spill about Whitewater, but later Starr doubted he had fully done so and subpoenaed 11 broad categories of documents from Hubbell. Hubbell took the Fifth on that request, but Starr granted him immunity to produce the documents anyway, which he then did to the tune of some 13,120 pages. Those pages eventually led to a grand jury indictment on new and unrelated mail and wire fraud charges. [11]

At issue in Hubbell were two questions: how vague may a document production request be while still affording the availability of the self-incrimination privilege, and must grants of immunity in exchange for such production prevent their subsequent use to prepare criminal charges unrelated to the reason for their subpoena against the producer? The former question evokes the disclosure forms from Doe (now more clearly understood under Fisher‘s act of production doctrine); while the latter implicates immunity principles codified under federal law that were refined under Kastigar (also referenced from Doe) and Counselman (also referenced from Schmerber).

Majority Opinion

This time Stevens was speaking for the Court rather than against it when finding in favor of the defendant, Webster Hubbell. Revisiting many of the decisions already discussed here, the Hubbell Court affirmed the original district court’s finding with a nod to Stevens’s own strongbox argument from Doe:

What the District Court characterized as a “fishing expedition” did produce a fish, but not the one that the Independent Counsel expected to hook… It was only through respondent’s truthful reply to the subpoena [Wigmore] that the Government received the incriminating documents of which it made “substantial use … in the investigation that led to the indictment.” … The assembly of those documents was like telling an inquisitor the combination to a wall safe, not like being forced to surrender the key to a strongbox.

Addressing the Government’s argument that the act of production in this case wasn’t sufficiently testimonial to trigger the privilege, Stevens somewhat snarkily asserts that the Government “misreads Fisher” (emphasis mine):

Whatever the scope of this “foregone conclusion” rationale, the facts of this case plainly fall outside of it. While in Fisher the Government already knew that the documents were in the attorneys’ possession and could independently confirm their existence and authenticity through the accountants who created them, here the Government has not shown that it had any prior knowledge of either the existence or the whereabouts of the 13,120 pages of documents ultimately produced by respondent. The Government cannot cure this deficiency through the overbroad argument that a businessman such as respondent will always possess general business and tax records that fall within the broad categories described in this subpoena.

So absent the Government demonstrating an ability to obtain those documents through a “wholly independent” source — which it admitted it plainly couldn’t do, not least since among the categories subpoenaed were “any and all” receipts, bank and wire transfer records, timesheets, appointment books, and even diaries [12] — the privilege was available and the immunity grant was enforceable.

Rehnquist’s Dissent from Williams

Chief Justice Rehnquist alone dissented in Hubbell, stating with just a single sentence that he did so for the same reasons that Judge Williams of the Appeals Court dissented from that decision to remand Hubbell over act-of-production questions. That seems a bit of a cop-out since it doesn’t address any of the majority arguments, but it does highlight some interesting arguments from the lower court.

Williams is off to the races with the first sentence of his dissent in that court:

The commentator who predicted that Fisher and Doe would “inevitably lead” to “metaphysical speculation” was apparently all too prescient. [13]

Williams goes on to make several colourful but at times both compelling and labored criticisms of Fisher’s testimonial and foregone conclusion tests. The entire dissent is well worth the entertaining read — not least since it touches on the horrifying case of Baltimore Deptartment of Social Services. v. Bouknight (1990), wherein a woman (unsuccessfully) asserted the Fifth Amendment privilege to refuse to turn over a child she was believed to have been abusing since that “act of production” would tend to incriminate her. One of the more cogent and informative passages about the practical effect of these subjective tests — which evokes Brennan’s concurrence in Fisher and references several other precedents we’ve already seen — is worth reproduction in full (emphasis mine):

For a district judge, the challenge of the majority’s view is to determine the “quantum” of relative prosecutorial ignorance that triggers a self-incrimination violation. Prosecutors know that businessmen keep business records (just as they know that living humans have blood and literate persons have handwriting); this is plainly too little information for the majority. But evidently the prosecutor need not have advance knowledge of the details that he is interested in. Somewhere in that range is an imaginary line which, unlike the equator, can never be fixed or defined with clarity. Henceforth, therefore, the operational meaning of the “act of production” doctrine in our circuit will largely turn on district courts’ discretion in this metaphysical classification of prosecutors’ knowledge.

And Williams takes note of a seemingly arbitrary distinction throughout these precedents which led to a newly explicit test introduced by the appeals court’s decision, namely that the fruits of the body are unprotected but the fruits of the mind are not:

Though recognizing that no other court has applied [the appeals court’s] mind/body distinction explicitly, the majority claims that the existing lower-court cases can be lined up to fit. If so, this seems to me only because the factual detail of the cases is so skimpy and the majority’s test so elastic.

Thomas’s Concurrence

Perhaps the most interesting result of Hubbell was the concurring opinion of Justice Thomas, joined also by Justice Scalia. Very recently I’ve been quite dismissive of Thomas; and though I have admitted grudging respect for certain of his arguments, Scalia faced strong criticism until and following his sudden death last year. The stubborn and self-satisfied originalism of both justices has made them chief enemies of advocates of a more fluid reading of a “living” Constitution. But with a stunningly straightforward and compelling opinion, Thomas makes the wholesale return to the plain language of Boyd a hard argument to oppose.

Stating at the outset his approval of the majority opinion due to its proper application of the “act-of-production doctrine”, Thomas introduces his separate writing “to note that this doctrine may be inconsistent with the original meaning of the Fifth Amendment’s Self-Incrimination Clause” and that “in a future case” he would be willing to reconsider its “scope and meaning.”

A lengthy review of the relevant eighteenth and nineteenth century dictionaries and case law makes it very clear that the “key word at issue”, namely witness, in the Fifth Amendment’s guarantee that no one shall be made “to be a witness against himself” was understood to be a “person who gives or furnishes evidence.” Moreover, “the phrase ‘furnish evidence’ is repeatedly used to refer to the compelled production of books, records, and archives in response to a government request.” Thomas thence asserts:

Such a meaning of “witness” is consistent with, and may help explain, the history and framing of the Fifth Amendment. The 18th-century common-law privilege against self-incrimination protected against the compelled production of incriminating physical evidence such as papers and documents.

But it isn’t just dictionaries and common law. According to Thomas, four states with similarly expansive constructions of self-incrimination protections in their own constitutions lobbied the 1787 constitutional convention which led to the United States to provide similar protections federally. Writes Thomas:

In response to such calls [for self-incrimination protection], James Madison penned the Fifth Amendment. In so doing, Madison substituted the phrase “to be a witness” for the proposed language “to give evidence” and “to furnish evidence.” But it seems likely that Madison’s phrasing was synonymous with that of the proposals. The definitions of the word “witness” and the background history of the privilege against self-incrimination, both discussed above, support this view. And this may explain why Madison’s unique phrasing — phrasing that none of the proposals had suggested — apparently attracted no attention, much less opposition, in Congress, the state legislatures that ratified the Bill of Rights, or anywhere else.

If all this weren’t enough, Thomas cites a pair of famous federal cases. In the 1807 treason case United States v. Burr, Burr requested under the Sixth Amendment’s compulsory process clause for “obtaining witnesses in his favor” a letter from President Thomas Jefferson “that was said to incriminate Burr.” Jefferson objected on the grounds that a document was not a witness, but Thomas writes that Chief Justice Marshall rejected that argument, “holding that the right to compulsory process includes the right to secure papers — in addition to testimony — material to the defense.” Thomas adds that the Court subsequently (and recently) reaffirmed that reading by citing United States v. Nixon (1974), the case that led directly to President Nixon’s resignation ahead of imminent impeachment. Concludes Thomas:

if the term “witnesses” in the Compulsory Process Clause has an encompassing meaning, this provides reason to believe that the term “witness” in the Self-Incrimination Clause has the same broad meaning. Yet this Court’s recent Fifth Amendment act-of-production cases implicitly rest upon an assumption that this term has different meanings in adjoining provisions of the Bill of Rights.

I take such a reading to support questioning the basis for cases like Holt (compulsion to don a blouse) and Schmerber (compulsion to provide a blood sample), but Thomas is more circumspect and focuses just on Fisher in his final assertion:

None of the parties in this case has asked us to depart from Fisher, but in light of the historical evidence that the Self-Incrimination Clause may have a broader reach than Fisher holds, I remain open to a reconsideration of that decision and its progeny in a proper case.

Stahl and Contemporary Implications

Bringing this all back to the present day, we’re left with the digital privacy concerns evoked by compelled production of phone passcodes in Stahl. The facts of that case are far more interesting — and the State court’s decision much easier to follow — now that we understand a lot more about the controlling case law. Stahl may have explicitly invoked his privilege or Florida might have preemptively asserted it didn’t apply (the record doesn’t show) but the trial court found that the privilege controlled because “production of the passcode was testimonial and that the State had not sufficiently established that the foregone conclusion doctrine applied.” Despite other failings in the arguments had at the trial court, Judge Black found that “the only issue before [his] court was whether it could compel Stahl to provide the passcode.”

Interestingly, and recalling the Schmerber-line’s physical exemplars, Black notes that

The State made no mention of whether it had attempted to compel Stahl to unlock the phone using his fingerprint. At least one court has held that compelling a witness to use his fingerprint to unlock or access his cellphone is not testimonial… When pressed by the court, the State conceded that “in the most technical sense” the court would be forcing Stahl to “use the contents of his mind” in compelling him to provide the passcode.

This leads to what I find to be the most perverse implication yet of Fisher‘s departure from Boyd. In a footnote, citing Hubbell paraphrasing Fisher, Black recalls the

settled proposition that a person may be required to produce specific documents though they contain incriminating assertions of fact or belief because the creation of those documents was not “compelled.”

From there he questions whether producing the password is compelled within the meaning of the privilege [emphasis in the original] “even though it may be testimonial and incriminate him because the creation of the password was not compelled.” In other words, and this is huge: voluntarily encrypting one’s phone with a password might not bar the government from demanding that password be divulged, even though that action is exactly what one might do to have a reasonable expectation of privacy about the phone’s contents, because the password itself wasn’t created under Government compulsion.

Of course there’s more. Black again cites Hubbell and Doe to argue that protected compelled communication must itself have testimonial significance, but that in this case the communication is sought for its content alone: “by providing the passcode, Stahl would not be acknowledging that the phone contains evidence.” And addressing Fisher‘s foregone conclusion doctrine, Black strikes down all three hurdles to the Government compelling production of evidence: namely its existence, custody, and authenticity. Since the phone is clearly locked, a password must exist; since the phone has been linked to Stahl through carrier records and his own testimony, he must possess the password; and if the password unlocks the phone, it must be authentic.

The Future of the Fifth

This tangled web of jurisprudence leaves us in an interesting position. I think it’s helpful to remember that while the 60 years since many of the cases which led to Fisher represent a long time — in fact, most of a lifetime — that’s still only about a quarter of the lifetime of the Bill of Rights. All things considered, a flurry of recent decisions, frequently concerning technologies which were quite new at the time, are now forming the basis for more decisions around even newer technologies.

In this chronological view, colour-coded lines indicate references to earlier cases. The graph would be a little more complex if I included references found only in footnotes.

In light of the facts, the controlling precedents do not seem to be very helpful. To the contrary: the contorted language of Fisher, itself built on then-recent and divided opinions, has undermined the plain language of the Framers which was liberally construed for much longer before it. My head spins just scratching the surface of recent Fifth Amendment jurisprudence, and I’ve devoted considerable time to understanding it. How is a common citizen to make any practical sense of the prescription of law that this mess has mandated? Not without the help of a lawyer when push comes to shove, certainly. The implications for privacy in the digital age are both clear and concerning.

Sadly, this state of affairs was feared more than 100 years ago by the Counselman Court. Writing in 1892, it observed:

illegitimate and unconstitutional practices get their first footing in that way, namely, by silent approaches and slight deviations from legal modes of procedure. This can only be obviated by adhering to the rule that constitutional provisions for the security of person and property should be liberally construed. A close and literal construction deprives them of half their efficacy, and leads to gradual depreciation of the right, as if it consisted more in sound than in substance. It is the duty of courts to be watchful for the constitutional rights of the citizen, and against any stealthy encroachments thereon.

In wild times such as these, after much time for those stealthy encroachments to find purchase, I think a return to the clear protections of Boyd is warranted. In other words, I surprisingly find myself casting my lot in with Justice Thomas.


[1] How was such an arrangement initially possible for a customer conducting business over which the US ostensibly has jurisdiction? A footnote: “The [US] Government has not yet sought contempt charges against the banks.”

[2] The significant holding for Doe (1988) came from Fisher and United States v. Doe (1984), which together held that “the privilege protects a person only against being incriminated by his own compelled testimonial communications” and that “in order to be ‘testimonial,’ an accused’s oral or written communication, or act, must itself, explicitly or implicitly, relate a factual assertion or disclose information.”

[3] The district court still rejected the consent form since it “might lead to”, as well as “admit signatory authority over” and “implicitly authenticate”, records which the grand jury would otherwise not known about. An appeals court reversed and remanded, so the district court ordered Doe sign it. Doe refused and was held in civil contempt and ordered confined until he signed, but confinement was stayed pending an appeal. The appeals court affirmed the contempt ruling and the appeal went to the Supreme Court. After all that, a footnote in this (the Supreme Court’s near-unanimous majority) opinion unironically notes that “a compelled consent, such as the one at issue in this case, is not sufficient to authorize the release of confidential financial records protected by Cayman law.” But since the “effectiveness of the directive under foreign law has no bearing on the constitutional issue in this case,” the show must go on!

[4] Later also cited by Stevens in his dissent, that protection, elaborated in Miranda v. Arizona (1966), states that “the [Fifth Amendment] privilege is fulfilled only when the person is guaranteed the right ‘to remain silent unless he chooses to speak in the unfettered exercise of his own will.'”

[5] From the majority opinion: “…the prosecution is allowed to obtain and use… evidence which, although compelled, is, generally speaking, not ‘testimonial.’ [Schmerber v. California (1966)]… If the societal interests in privacy, fairness, and restraint of governmental power are not unconstitutionally offended by compelling the accused to have his body serve as evidence that leads to the development of highly incriminating testimony, as Schmerber and its progeny make clear, it is difficult to understand how compelling a suspect to make a nonfactual statement that facilitates the production of evidence by someone else offends the privilege.” Stand by for much more on Schmerber.

[6] Given that, as earlier noted, the Government of the Cayman Islands maintained (still pending appeal in 1988) that compelled consent did not qualify for purposes of its secrecy laws, the “consent” form seems to me closer to obtaining a wall safe combination under duress.

[7] Stevens elaborates on this, also citing Schmerber:

The forced production of physical evidence, which we have condoned, see Gilbert v. California (1967) (handwriting exemplar); United States v. Wade, (1967) (voice exemplar); Schmerber v. California, (1966) (blood test); Holt v. United States, (1910) (lineup), involves no intrusion upon the contents of the mind of the accused. See Schmerber (forced blood test permissible because it does not involve “even a shadow of testimonial compulsion upon or enforced communication by the accused”). The forced execution of a document that purports to convey the signer’s authority, however, does invade the dignity of the human mind; it purports to communicate a deliberate command. The intrusion on the dignity of the individual is not diminished by the fact that the document does not reflect the true state of the signer’s mind… the accused is being compelled “to be a witness against himself”; indeed, here he is being compelled to bear false witness against himself.

[8] In an interesting wrinkle, Justice Stevens “took no part in the consideration of decision of the cases” because, though they were decided in April 1976, they had been argued in November 1975, about six weeks before Stevens joined the bench.

[9] The majority elaborated at length that it was the special nature of the client-attorney privilege alone which was the saving grace for the defense in this case. The Court had very recently held in a similar case, Couch v. United States (1973), that the Fifth didn’t excuse a defendant’s accountant from responding to an IRS summons for documents pursuant to tax returns he’d prepared  for her.

[10] The fact that the defendant was conscious at the time the blood sample was taken is significant. An earlier case, Breithaupt v. Abram (1957), was found to be controlling precedent in this case despite the fact that the defendant was unconscious at the time that sample was collected. Writing for the majority in Schmerber, Justice Brennan noted that since Breithaupt was unconscious he therefore “had no opportunity to object to the procedure” and that “under such circumstances, the withdrawal did not offend ‘that sense of justice’ of which we spoke in Rochin v. California,” a somewhat similar case from 1952. While the Rochin Court forbade brutal police practices which “shocked the conscience,” the Breithaupt Court countenanced non-consensual fluid extraction in the name of highway safety. Justice Douglas, channeling more contemporary views on whether consent for insertion can be granted by an unconscious person, noted that he found it “repulsive… for the police to insert needles into an unconscious person” in his dissent in Breithaupt.

[11] The district court found the charges to be at odds with the immunity deal and dismissed the indictment, but the appeals court reversed and remanded because it believed the district court had wrongly focused on the content of the documents rather than the act of their production. It instructed the district court to find out whether the Government already knew with “reasonable particularity” about the existence, possession by Hubbell, and authenticity of the documents, in which case the privilege could never have been invoked. It found that the Government did not, so the case then went to the Supreme Court to determine the scope of the immunity grant.

[12] Imagine the view a Freedom of Information Act court might take of a similar request of the Government!

[13] That commentator was none other than Samuel Alito, who would join the Supreme Court just seven years after Williams mentioned him. Alito was writing in 1986, so it is significant to point out that the “Doe” to which he refers is not Doe (1988) analyzed here, but its related and influential predecessor Doe (1984) only mentioned here in footnotes. That case held that “the contents of the subpoenaed records in question are not privileged under the Fifth Amendment” because such records were not prepared under compulsion, but that “the act of producing the subpoenaed documents cannot be compelled without a statutory grant of use immunity” of the kind involved in Hubbell.

Leave a Reply