Field Notes

concept

Invoking the Fifth Amendment can itself be used as evidence

StatusWorkingUpdatedStands on2 sourcesSupportsnotes

In current US law, neither staying silent nor naming the Fifth Amendment privilege against self-incrimination reliably keeps a statement out of evidence. Both risks are documented, each by a different line of cases, in You Have the Right to Remain Innocent, Duane’s book, and in Don’t talk to the police, his earlier lecture. The practical advice that follows from both risks is to invoke a different right instead: the Sixth Amendment right to counsel, which courts have not yet treated the same way.

Plain silence can be used as evidence of guilt

Salinas v. Texas, 570 U.S. 178 (2013), held that a suspect’s silence during a voluntary, non-custodial, pre-Miranda police interview — after he had already been answering other questions and then stopped responding to one specific question — was not a valid invocation of the privilege without an express statement to that effect, so a prosecutor’s comment on that silence at trial did not violate the Fifth Amendment. The decision does not reach spontaneous silence with no prior police questioning, since that fact pattern was not before the Court.

The case is commonly described as 5-4, but this understates a genuinely fractured opinion, which matters for how much weight the “rule” can bear. Only a three-Justice plurality (Alito, Roberts, Kennedy) reached the holding above. Two more Justices (Thomas, joined by Scalia) concurred in the judgment on a broader, different rationale — that even an express invocation would not have helped, since they would not treat prosecutorial comment on silence as unconstitutional “compulsion” at all — and did not join the plurality’s reasoning. Four dissenting Justices (Breyer, Ginsburg, Sotomayor, Kagan) argued that requiring an unrepresented layperson to recite specific “magic words” before silence counts as invocation is unrealistic and defeats the privilege’s practical purpose. Under Marks v. United States, the operative holding that binds lower courts is the narrowest ground commanding five votes: the plurality’s express-invocation requirement. But two of those five votes rested on reasoning the plurality itself did not adopt, so Salinas is weaker and more contested authority than a clean majority opinion would be, a point developed further in Why you shouldn’t talk to the police. Duane’s book rebuts the premise that only guilty people fall silent with a statistic that 39 percent of DNA-exonerated innocent defendants chose not to testify at their own trials. LaCorte’s video, which credits Duane as its inspiration, makes the same doctrinal point by working Salinas v. Texas as a hypothetical gun-sale interrogation.

Naming the privilege can also be used as evidence

Duane’s book argues, through United States v. Long, 721 F.3d 920 (8th Cir. 2013), that explicitly invoking the Fifth Amendment by name does not reliably solve the problem either. Gillman Long told an FBI agent, during a non-custodial, pre-Miranda interview, “I do not want to incriminate myself. I would like to stop talking,” and the prosecution used that statement at trial, telling the jury it showed what an innocent person would not have said; Long was convicted and sentenced to life imprisonment. But the book’s framing that this was “upheld on appeal” overstates what happened. The Eighth Circuit did not rule on whether admitting the statement was constitutional at all. It held only that the question was unsettled, with other circuits split, so Long’s trial counsel was not deficient for failing to object; it deferred the underlying constitutional question to collateral review. A later decision on that collateral claim, Long v. United States, 875 F.3d 411 (8th Cir. 2017), again resolved the case on the lower ineffective-assistance-of-counsel standard rather than reaching the constitutional merits. No court has held that using such a statement is proper, and none has held that it is not; the doctrinal question remains open and circuit-split. The case is genuine evidence that naming the privilege carries real, uncorrected risk at trial and on appeal, but it is not the settled precedent Duane’s framing implies.

Duane’s earlier ABA article argues a related but distinct point: that the standard courtroom script — “I decline to answer because my answer may tend to incriminate me” — sounds to an ordinary juror like a confession, because the word “incriminate” shares a root with “crime,” even though the Supreme Court has held, in Ohio v. Reiner, 532 U.S. 17 (2001), that the privilege protects innocent people too. That case is itself narrower than a general rule about defendants: it held that a prosecution witness who had been granted immunity and then denied any wrongdoing still had a valid privilege to begin with, because she retained “reasonable cause to apprehend danger” from a truthful answer, not that any witness may always claim the privilege regardless of circumstance. The article’s fix stays inside the Fifth Amendment: Duane recommends different wording that names the constitutional protection without using the word “incriminate.” The book goes further and recommends abandoning the Fifth Amendment invocation altogether in favor of the Sixth.

Duane’s book resolves both risks by recommending a Sixth Amendment invocation — a request for counsel — rather than any Fifth Amendment language at all, citing United States v. Okatan and Maine v. Lovejoy for the position that courts have not, at least as of the book’s 2016 publication, allowed prosecutors to treat an unambiguous request for a lawyer as evidence of guilt. The request must be unambiguous: courts have held phrases such as “maybe I should talk to a lawyer” or “do you think I need a lawyer” too hedged to count as an invocation at all, in cases including Davis v. United States and Burket v. Angelone. Duane’s central cautionary example is Tio Sessoms, sentenced to life without parole in California after detectives talked him out of a request for a lawyer he made 40 seconds into his interrogation, before any Miranda warning, and California’s courts held that request too ambiguous to count as an invocation under Davis. The en banc Ninth Circuit reversed in Sessoms v. Grounds, 776 F.3d 615 (9th Cir. 2015), holding that reading was an unreasonable application of clearly established law, but only 6 to 5, with two separate dissents; this was a genuinely close case about how Davis’s unambiguous-request standard applies under federal habeas review’s deference to state courts, not a straightforward vindication of the request-a-lawyer strategy. Sessoms had already served roughly 16 years on the vacated conviction by the time of that ruling, and about 18 years in total custody before his eventual release on a lesser plea in 2017, somewhat longer than Duane’s “more than 13 years.” Duane’s resulting recommendation is to say exactly “I want a lawyer” — four words, with no hedging language such as “think,” “maybe,” or “might” — and to repeat it until questioning stops.

Scope and reliability

This doctrine rests on named, independently checkable federal cases, and the case-by-case verification above found Duane’s citations accurate on their facts more often than not, but several were characterized more favorably to his argument than the actual holdings support, particularly Salinas’s fractured reasoning and Long’s undecided constitutional question. Vega v. Tekoh, 597 U.S. 134 (2022), decided after this book, further narrowed the practical value of Miranda rights generally: it held that a Miranda violation is not itself a constitutional violation, so it cannot support a civil damages claim under 42 U.S.C. § 1983. Suppression of the resulting statement remains the only remedy, which removes a mechanism that deterred police from disregarding an invocation in the first place. The underlying case law should still be checked against current Supreme Court and circuit-court authority before this note’s practical advice is treated as current across all US jurisdictions. The advice is also explicitly scoped to the United States and does not describe protections against self-incrimination in other legal systems.

Built on 2 sources (2 archived here).

Working out connections…