United States v. Christie

717 F.3d 1156, 2013 WL 2477252, 2013 U.S. App. LEXIS 11704

• Court: Court of Appeals for the Tenth Circuit
• Decided: June 11, 2013
• Docket: 11-2106, 11-2221
• Status: Published

Opinions

GORSUCH, Circuit Judge.

For Rebecca Christie, life must have seemed more virtual than real. She usually awoke around noon, settled in before her computer, and logged on to World of Warcraft for gaming sessions lasting well past midnight. There she assumed a new identity in a fantastical world filled with dragons and demons where players staged heroic adventures with and against other players. All the while back in the real world Ms. Christie ignored the needs of her three-year-old daughter. The neglect didn’t prove fatal so long as Ms. Christie’s husband was around to provide some care. But nine days after her husband left for an out-of-state deployment, the child was dead from dehydration.

Appealing second-degree murder and child abuse convictions, Ms. Christie raises significant questions about computer searches under the Fourth Amendment and the exclusion of witnesses from trial under the Sixth Amendment. The government’s cross-appeal raises important questions, too, touching on the Assimilative Crimes Act and the Fifth Amendment’s double jeopardy guarantee. In the end, however, we think the district court handled all these questions well and carefully and we see no grounds on which we might reverse its judgment in this tragic case.

I

Ms. Christie’s child began life a healthy baby girl. But by twenty-one months, something appeared badly wrong. She plummeted to the bottom fifth percentile in weight for her age and began suffering from chronic diarrhea. A pediatrician prescribed PediaSure, a nutritional drink that helps children gain weight. That seemed to do the trick: the diarrhea soon stopped and BW (the district court and parties refer to the child by her initials) began gaining weight. By all appearances, she had turned a corner.

But even the best cure won’t work if it isn’t administered. And it seems Ms. Christie and her then-husband Derek Wulf, himself a zealous gamer, weren’t up to the job. For her part, Ms. Christie would put BW to bed each night around 10 p.m., shut BW’s door, and failed to retrieve the child until noon or later the following day. Because BW couldn’t open the door herself, Ms. Christie effectively locked the child away without food or water for fourteen or more hours a day. Even when BW was free to seek food and water it appears little was available to her. Ms. Christie let slip to investigators that the child was always hungry and would sometimes try to eat the food she left out for the family cats. Ms. Christie’s stepdaughter (who visited on occasion) testified that Ms. Christie often wouldn’t feed her or BW until noon and that BW’s obvious hunger drove her to share her own food with BW.

Eventually, Mr. Wulf faced a deployment on the other side of the country. With the little care he provided BW now gone with him and the child’s fate entirely in Ms. Christie’s hands, the child was in trouble. Already badly malnourished, she succumbed to dehydration in nine days. An autopsy revealed that no inborn disorder was to blame. BW simply died from being ignored.

Medical experts testified that BW’s desperate condition in the days before her death would have been blindingly clear. BW would have sought out water as a survival instinct. When that failed, she would have become lethargic and, on the day before her death, too weak to move. Her diapers wouldn’t have needed changing. She would have had sticky saliva and then no¹ saliva at all. She would have developed cracked lips, sunken eyes, and a sunken abdomen.

First responders confirmed that this is exactly what they saw when they found the child. They testified that 'BW’s lips were cracked and blue,-her eyes glassy, and her eyelids so dry they couldn’t close. They said bones protruded from her body and her gums had turned black. ■

Because BW died on an Air Force base, federal authorities bore the responsibility to investigate and the power to prosecute. They proceeded against Ms. Christie and Mr. Wulf separately. In our proceeding, a federal jury found Ms. Christie guilty of second-degree murder, two assimilated state law homicide charges, as well as an assimilated child abuse charge. After trial, the district court dismissed the two assimilated homicide charges and entered a twenty-five year sentence on the remaining second-degree federal murder and the assimilated child abuse charge.

It is this judgment both sides now appeal.

II

Much of the evidence presented at trial against Ms. Christie came from the computer she so prized. From their forensic analysis, FBI investigators learned that Ms. Christie’s online activities usually kept her busy from noon to 3 a.m. with little pause. They learned that she was in a chat room only an hour before finding BW near death, and that she was back online soon afterwards. They learned from Ms. Christie’s messages to other gamers that she was annoyed by her responsibilities as a mother and “want[ed] out of this house fast.” When Mr. Wulf was slated for deployment, she announced to online friends that she would soon be free to “effing party.”

Ms. Christie contends this evidence and more from her computer was uncovered in violation of her Fourth Amendment rights and the- district court should have suppressed it from her trial. Because the court didn’t, because it admitted the proof against her, Ms. Christie says a new trial is required. To be precise, Ms. Christie doesn’t question whether the government’s seizure of the computer satisfied the Fourth Amendment. The government took possession of the computer in May 2006 with Mr. Wulf s consent. Everyone accepts that he was at least a co-owner of the computer—it was a gift from his father—and everyone accepts he had at least apparent authority to relinquish its control. Instead, Ms. Christie attacks the propriety of the two searches the government undertook once it had control of the computer. To justify its searches the government does not seek to rely on Mr. Wulfs consent but points to a pair of warrants it sought and received, one for each search. It is these warrants Ms. Christie challenges, arguing they were issued in defiance of the Fourth Amendment.

A

The first warrant came in October 2006, some five months after authorities seized the computer. Ms. Christie argues this investigative delay—between seizure and search—was constitutionally impermissible, should have precluded any warrant from issuing, and itself requires the suppression of everything the government found.

We do not doubt that an unreasonable delay in obtaining a search warrant can sometimes violate the Fourth Amendment. This much surely flows from the Fourth Amendment’s guarantee against “unreasonable searches and seizures.” U.S. Const, amend. IV. What, after all, is “reasonable” about police seizing an individual’s property on the ground that it potentially contains relevant evidence and then simply neglecting for months or years to search that property to determine whether it really does hold relevant evidence needed for trial or is totally irrelevant to the investigation and should be returned to its rightful owner? See, e.g., United States v. Burgard, 675 F.3d 1029, 1033 (7th Cir.2012); United States v. Laist, 702 F.3d 608, 613-14 (11th Cir.2012); United States v. Martin, 157 F.3d 46, 54 (2d Cir.1998); United States v. Respress, 9 F.3d 483, 488 (6th Cir.1993).

In assessing the reasonableness of a delay in seeking a warrant, however, we must take account of “the totality of the circumstances” in each case as it comes to us, United States v. Sokolow, 490 U.S. 1, 8, 109 S.Ct. 1581, 104 L.Ed.2d 1 (1989), wary of the temptation to impose “rigid rules, bright-line tests, and mechanistic inquiries,” Florida v. Harris, — U.S. -, 133 S.Ct. 1050, 1055, 185 L.Ed.2d 61 (2013). What reasonably justifies a brief delay, for example, may not reasonably justify a longer one. Our task in each case is to “balance the nature and quality of the intrusion on the individual’s Fourth Amendment interests against the importance of the governmental interests alleged to justify the intrusion.” United States v. Place, 462 U.S. 696, 703, 103 S.Ct. 2637, 77 L.Ed.2d 110 (1983).

Looking first to Ms. Christie’s side of the ledger, it’s hard to see a significant invasion of her Fourth Amendment interests flowing from the government’s delay. That’s not to say she never had an interest in the computer. Far from it. She was its primary user and she stored a great deal of personal data on the computer. We don’t question that is enough to establish some possessory and privacy interest in the computer the Fourth Amendment would recognize and protect. See United States v. Jacobsen, 466 U.S. 109, 113, 104 S.Ct. 1652, 80 L.Ed.2d 85 (1984); United States v. Angevine, 281 F.3d 1130, 1133-34 (10th Cir.2002).

The difficulty Ms. Christie faces comes from her husband’s actions and her own inactions. The fact is Mr. Wulf was at least a co-owner of the computer, he consented to its seizure,. and Ms. Christie herself raised no objection to the seizure either at the time or in the following weeks and months. No call to the authorities, no letter, no motion to the court. Put differently, one individual with common and at least apparent authority over the computer freely gave it to authorities and the other individual never objected. In these circumstances, the government was entitled to assume under long-standing Supreme Court teachings that any Fourth Amendment interest in the computer’s continued possession had been voluntarily relinquished. See, e.g., United, States v. Matlock, 415 U.S. 164, 171, 94 S.Ct. 988, 39 L.Ed.2d 242 (1974); United States v. Andrus, 483 F.3d 711, 716-20 (10th Cir.2007).¹

These features of our case also distinguish it from the primary case on which Ms. Christie relies: United States v. Mitchell, 565 F.3d 1347 (11th Cir.2009). There, officers seized a computer mthout the consent of the (only) owner and, in doing so, at least presumptively impaired his possessory interest in it. Id. at 1349-50. But when a computer is seized with a co-owner’s consent and no one else bearing a potential interest in the computer complains about the resulting delay, the degree of prejudice appears in an altogether different—and paler—light. See United . States v. Johns, 469 U.S. 478, 487, 105 S.Ct. 881, 83 L.Ed.2d 890 (1985) (defendants who “never sought return of the property” failed to identify how “the delay in the search ... adversely affected legitimate interests protected by the Fourth Amendment”); United States v. Stabile, 633 F.3d 219, 235-36 (3d Cir.2011) (distinguishing Mitchell on the same grounds as we do); United States v. Burgess, 576 F.3d 1078,1097 (10th Cir.2009).

Shifting to the other side of the ledger, the government makes out at least a color-able case for holding onto the.computer so long before undertaking a search. The agent who seized the computer and later searched it says that in between he was called upon to help with out-of-town undercover operations in other cases. No one before us disputes the - existence of these operations. No one questions that they amounted to a higher law enforcement priority than the computer search in this case. Neither has Ms. Christie (who bears the burden of proof on her motion to suppress) developed evidence suggesting her case or the others could have been transferred to another available agent. In these circumstances, we hesitate to say the government lacked any colorable grounds for its delay.

Having set out the parties’ competing (maybe incommensurable) interests, we must now try to “balance” or “weigh” them. As we see it, the government’s side of the ledger reveals a colorable interest in prioritizing law enforcement efforts while Ms. Christie can point to little harm to her

interests in light of her husband’s express consent and her lack of objection. Given this, we hold the delay in our case falls inside the bounds of constitutionally reasonable conduct, if not by a very great margin. In reaching the decision we do, we hardly mean to suggest a government agent’s need to attend to other assignments will always justify delay of this length. If, for example, the evidence showed the agent could have readily transferred his current case or his newly assigned responsibilities to someone else, our judgment -might be different. If, too, Mr. Wulf had not voluntarily relinquished the computer or someone had sought its return; our judgment might be different. To decide this case we need only hold (and only hold) that when an investigator obtains property by consent and retains it without objection, and when his search is delayed by virtue of having to attend to other and higher law enforcement priorities and no evidence suggests a reassignment was reasonably possible, a five-month delay isn’t constitutionally unreasonable. See Stabile, 633 F.3d at 236 (holding three-month search delay reasonable when officer was reassigned to another matter); Mitchell, 565 F.3d at 1353 (holding a three-week delay unreasonable, but noting the case would be different if the government had shown the need to “diver[t] ... law enforcement personnel to another case”).

B

Next Ms. Christie attacks the validity of the second warrant, this one issued in May 2009, to conduct a more thorough search of her computer. She argues this warrant violated the Fourth Amendment’s promise that “no Warrants shall issue” without “particularly describing the place to be searched, and the persons or things to be seized.” The concern originally animating this so-called “particularity requirement” was a wish to prevent the sort of warrants English kings once favored, ones that proscribed few limits on the scope of the search to be conducted and included no explanation why they were issued. Even today the particularity requirement remains a vital guard against “wide-ranging exploratory searches,” a promise that governmental searches will be “carefully tailored to [their] justifications.” Maryland v. Garrison, 480 U.S. 79, 84, 107 S.Ct. 1013, 94 L.Ed.2d 72 (1987). Probable cause to believe a suspect possesses a stolen car may justify issuing a warrant authorizing the search of her garage but not necessarily her attic, and probable cause to believe a gang member is hiding weapons may merit a warrant to conduct a search of his home but not necessarily his grandmother’s. A warrant isn’t ever supposed to be a license for just “a general ... rummaging.” United States v. Brooks, 427 F.3d 1246, 1251 (10th Cir.2005).

No doubt the particularity requirement and its underlying purposes are fully engaged when investigators seek to search a personal computer. Personal computers can and often do hold “much information touching on many different areas of a person’s life.” United States v. Walser, 275 F.3d 981, 986 (10th Cir.2001). They can contain (or at least permit access to) our diaries, calendars, files, and correspondence—the very essence of the “papers and effects” the Fourth Amendment was designed to protect. U.S. Const, amend. IV. In today’s world, if any place or thing is especially vulnerable to a worrisome exploratory rummaging by the government, it may be our personal computers.

This court’s efforts to apply the Fourth Amendment’s particularity requirement to computer searches are still relatively new. But the parties seem to agree that our cases already draw at least one recognizable line. On the one hand, we have held invalid warrants purporting to authorize computer searches where we could discern no limiting principle: where, for example, the warrant permitted a search of “ ‘any and all’ information, data, devices, programs, and other materials,” United States v. Otero, 563 F.3d 1127, 1132-33 (10th Cir.2009) (alteration omitted), or “all computer and non-computer equipment and written materials in [a defendant’s] house,” Mink v. Knox, 613 F.3d 995, 1011 (10th Cir.2010). On the other hand, we have said warrants may pass the particularity test if they limit their scope either “to evidence of specific federal crimes or [to] specific types of material.” United States v. Riccardi, 405 F.3d 852, 862 (10th Cir.2005); see also United States v. Campos, 221 F.3d 1143, 1147 (10th Cir.2000); United States v. Burke, 633 F.3d 984, 992 (10th Cir.2011); Brooks, 427 F.3d at 1252.

Having more or less agreed on the current state of our jurisprudence, the parties proceed to debate on which side of the line our warrant falls. Is it more like Otero and Mink, or more like Campos, Brooks, and Burke? The October 2009 warrant authorized a search of Ms. Christie’s computer for

[a]ll records and information relating to the murder, neglect, and abuse of [BW] from June 19, 2002 (date of birth) to May 4, 2006, (date computer seized), including:

1. All photographs of [BW].

2. All correspondence and/or documents relating to [BW].

3. All records and information, including any diaries or calendars, showing the day-to-day activities of Rebecca Christie and/or [BW].

4. All addresses and/or contact information of friends, families, or acquaintances who may have had regular contact with Rebecca Christie and/or [BW].

According to Ms. Christie, paragraph 3 effectively permitted law enforcement to search any and all records and information on her computer for any and all purposes. And this, she says, authorized an unprincipled and unlimited search of the sort we held invalid in Otero and Mink.

The government responds that all of its search efforts, including those authorized in paragraph 3, were restricted by the warrant’s opening language. In support of its interpretation, the government points to Brooks where this court read a similar warrant as “naturally” restricting all subsequently enumerated searches to the scope set forth at the warrant’s outset. See Brooks, 427 F.3d at 1252. Advocating a parallel construction here, the government says paragraph 3 didn’t authorize it to rifle through Ms. Christie’s files looking for any sort of incriminating evidence. Instead, it had to direct all of its search efforts, including those specified in paragraph 3, to information related “to the murder, neglect, and abuse” of BW. And that limiting direction,, the government submits, is particularity enough under our case law.

We find it hard to fault the government’s reasoning. Though the ' warrant before us is surely open to interpretation, Brooks did approve a similar one on the basis that its opening language limited the scope of all later enumerated searches to seeking evidence of' particular federal crimes. At the very least, in light of Brooks we cannot deny that an objectively reasonable officer acting in good faith could have read the warrant before us in this same manner—as restricting the scope of any search to information “related to the murder, neglect, and abuse of’ BW and thus analogous to the warrants we upheld in Brooks, Campos, and Burke. And to know just that much is to know suppression cannot follow. The Fourth Amendment’s exclusionary rule will “not bar the use of evidence obtained by police officers acting in good faith and with reasonable reliance on a facially valid search warrant.” United States v. Tisdale, 248 F.3d 964, 972 (10th Cir.2001) (citing United States v. Leon, 468 U.S. 897, 919-20, 104 S.Ct. 3405, 82 L.Ed.2d 677 (1984)). In light of our past approval of a similar warrant, it seems more than a little difficult for us to say now that “a reasonably well trained officer would have known that the search” in this 'case “was illegal despite the magistrate’s authorization.” Id.

Ms. Christie rejoins that if our current case law endorses the warrant in this case then our case law needs to be reexamined. In .an age where computers permit access to most every “paper and effect” a person owns, she fears that merely restricting the government to a search topic or objective does little to prevent it from examining along the way virtually every bit and byte of our lives. Risking with it the possibility the government will claim to find “in plain view” evidence of crimes totally unrelated to the reasons spurring their search in the first place. The text of the Fourth Amendment says the government must identify with particularity “the place to be searched” and requiring it to describe that place tersely as “a computer” is to allow the government to traipse willy-nilly through an entire virtual world. To prevent that, Ms. Christie suggests a warrant must go further: it must specify limitations not just what the government may search for but how the government should go about its search.

In reply, the government argues that it’s often difficult to know what search protocols might be reasonably required at the time of a warrant application, before the computer has been examined. Computer files can be misnamed by accident, disguised by intention, or hidden altogether, leaving investigators at a loss to know ex

ante what sort of search will prove sufficient to ferret out the evidence they legitimately seek. See Burgess, 576 F.3d at 1093; Brooks, 427 F.3d at 1252. Neither has the particularity requirement, the government points out, ever been understood to demand of a warrant “technical precision,” United States v. Simpson, 152 F.3d 1241, 1248 (10th Cir.1998), or “elaborate detail,” United States v. Triplett, 684 F.3d 500, 504 (5th Cir.2012), but only “practical” limitations, Simpson, 152 F.3d at 1248, affording “reasonable] specificity],” United States v. Adjani, 452 F.3d 1140, 1149 (9th Cir.2006). Though this court’s efforts to apply the Fourth Amendment’s particularity requirement to computer searches are still relatively new and our existing treatment is far from comprehensive, it also seems difficult to square Ms. Christie’s demand for a how with our existing cases (Brooks, Burke, Campos, and Riccardi), all of which suggest a what may be particular enough. Indeed, we’ve gone so far as to suggest that it is “unrealistic to expect a warrant to prospectively restrict the scope of a search by directory, filename or extension or to attempt to structure search methods—that process must remain dynamic.” Burgess, 576 F.3d at 1093.

This isn’t to say the Fourth Amendment has nothing to say on how a computer search should proceed. Even putting aside for the moment the question what limitations the Fourth Amendment’s particularity requirement should or should not impose on the government ex ante, the Amendment’s protection against “unreasonable” searches surely allows courts to assess the propriety of the government’s search methods (the how) ex post in light of the specific circumstances of each case. See, e.g., United States v. Ramirez, 523 U.S. 65, 71, 118 S.Ct. 992, 140 L.Ed.2d 191 (1998) (“The general touchstone of reasonableness ... governs the method of execution of the warrant.”); United States v. Angelos, 438 F.3d 738, 746 (10th Cir.2006). So even if courts do not specify particular search protocols up front in the warrant application process, they retain the flexibility to assess the reasonableness of the search protocols the government actually employed in its search after the fact, when the case comes to court, and in light of the totality of the circumstances. Unlike an ex ante warrant application process in which the government usually appears alone before generalist judges who are not steeped in the art of computer forensics, this ex post review comes with the benefit, too, of the adversarial process where evidence and experts from both sides can be entertained and examined. See Burgess, 576 F.3d at 1094; United States v. Carey, 172 F.3d 1268, 1275-76 (10th Cir.1999); Orin S. Kerr, Searches and Seizures in a Digital World, 119 Harv. L.Rev. 531, 574-75 (2005).

Whenever courts should review search protocols and whatever the scope of our authority to do so may be, one thing is certain. To undertake any meaningful assessment of the government’s search techniques in this case (the how), we would need to understand what protocols the government used, what alternatives might have reasonably existed, and why the latter rather than the former might have been more appropriate. Unfortunately, however, that we do not have in this case. Though Ms. Christie bore the burden of proof in her suppression proceeding, she offered little evidence or argument suggesting how protocols the government followed in this case were unreasonable or insufficiently particular, especially when compared with possible alternatives. Without more help along these lines, we simply cannot assess rationally her challenge to the government’s search procedures in this case and must leave the' development of the law in this arena to future cases.

C

In her final, Fourth Amendment attack, Ms. Christie returns to the theme of faulting the government for its delay in seeking a warrant. This time the focus of her complaint is the second, May 2009 warrant. But the government’s delay in seeking this warrant poses much less of a concern than its delay in seeking the first warrant. After the government executed its first search warrant and found incriminating evidence, it was presumptively entitled to retain the computer through trial. “[T]he general rule is that” lawfully seized property bearing evidence relevant to trial “should be returned to its rightful owner once the criminal proceedings have terminated,” not before. United States v. Rodriguez-Aguirre, 264 F.3d 1195, 1212 (10th Cir.2001); see also United States v. Thomson, 354 F.3d 1197, 1201 (10th Cir.2003). Ms. Christie offers no reason to think her case an exception to this general rule. And without a right to the computer’s return until after trial, it’s hard to see how she had any interest that could be impaired by a second (and itself otherwise lawful) search during the period of the government’s lawful possession, let alone by the timing of that second search. Neither does Ms. Christie identify any authority or reason suggesting otherwise.²

III

When it comes to the conduct of the trial itself, Ms. Christie objects to the district court’s decision to exclude Mr. Wulf from the courtroom during the brief testimony of his ten-year-old daughter. Mr. Wulfs exclusion, Ms. Christie insists, violated her Sixth Amendment right to a public trial. See U.S. Const, amend. VI (“in all criminal prosecutions, the accused shall enjoy the right to ... a public trial”). According to Ms. Christie, before closing the courtroom in this way the district court had—but failed—to satisfy the detailed four-part test announced in Waller v. Georgia: “[1] the party seeking to close the hearing must advance an overriding interest that is likely to be prejudiced, [2] the closure must be no broader than necessary to protect that interest, [3] the trial court must consider reasonable alternatives to closing the proceeding, and [4] it must make findings adequate to support the closure.” 467 U.S. 39, 48, 104 S.Ct. 2210, 81 L.Ed.2d 31 (1984).

Waller's stringent test, however, applies only to “the total closure of a trial.” United States v. Galloway, 937 F.2d 542, 546 (10th Cir.1991). Here the courtroom was closed only to a single person (not the entire public), and only while a single witness (Mr. Wulfs daughter) testified. Of course it’s still no small thing to exclude anyone from any portion of a trial. But where the potential invasion of Sixth Amendment interests is so much less likely, the test a district court must satisfy is correspondingly less onerous. We have explained that to support a partial closure of a trial, the district court need only identify a “ ‘substantial’ interest, rather than a ‘compelling’ one” and document it with “sufficient findings to allow the reviewing court” to assess the decision. Id.

Our cases have yet to inscribe a definitive and exhaustive list of what qualifies as a “substantial interest” and today we don’t claim the insight necessary to craft a definition sufficient to cover all future cases. But in light of existing precedent we can say this much: “safeguarding the physical and psychological well-being of a minor” qualifies not just a substantial interest but a compelling one. See Globe Newspaper Co. v. Superior Court for Norfolk Cnty., 457 U.S. 596, 607, 102 S.Ct. 2613, 73 L.Ed.2d 248 (1982). Factors that may play into this calculus include the witness’s age and relationship to the person excluded from the courtroom, the nature of the alleged offense, and the harm the witness may reasonably expect to face without the court’s protection. See Galloway, 937 F.2d at 546.

In our case, there is no suggestion that Mr. Wulfs presence would have posed a physical danger to his daughter: there was no need to close the courtroom to protect the girl in that sense. But there were concerns about protecting the girl’s psychological well-being. The district court observed that Mr. Wulfs daughter was only ten years old. It heard her testify that “she would be afraid to talk” if her father was present, that she feared having to explain he too was sometimes negligent in caring for BW. And the court found that forcing the girl to testify in her father’s presence would place at risk her psychological well-being by undermining her long-term relationship with her father. No evidence before us suggests the district court’s assessment was wrong in any respect.

Of course we must still consider this interest in “balance” against those the Sixth Amendment seeks to protect. A public trial is a potent disinfectant, a republican assurance against the corruption that breeds in collusive environments. Though the benefits public scrutiny brings to a trial may be diffuse and “frequently intangible,” Waller, 467 U.S. at 49 n. 9, 104 S.Ct. 2210, the history of secret tribunals teaches that the right to a public trial serves many important interests, including at least these: it does much to ensure the trial is conducted fairly, to remind the lawyers and judge of their responsibilities to the accused and the importance of their functions, to encourage possible witnesses to come forward, and to discourage perjury. See Nieto v. Sullivan, 879 F.2d 743, 753 (10th Cir.1989) (citing Waller, 467 U.S. at 46, 104 S.Ct. 2210); In re Oliver, 333 U.S. 257, 266-70, 68 S.Ct. 499, 92 L.Ed. 682 (1948) (discussing the evolution of the public trial guarantee); Akhil Reed Amar, Sixth Amendment First Principles, 84 Geo. L.J. 641, 677-81 (1996); Max Radin, The Right to a Public Trial, 6 Temp. L.Q. 381 (1932).

But in this case, none of these underlying Sixth Amendment interests was meaningfully impaired. The girl’s testimony lasted only minutes in a trial that lasted days. Save Mr. Wulf, all members of the public were free to attend even the girl’s testimony. Ms. Christie doesn’t claim the prosecutors or judge took their responsibilities any less seriously or that a witness was less likely to come forward. Neither does Ms. Christie argue this limited closure prompted the girl (let alone any other witness) to testify falsely, and she does not suggest the closure deprived her of an opportunity for effective cross examination.

We do not mean to suggest in any way that the temporary exclusion of one individual from a trial will always be constitutionally acceptable. The Sixth Amendment is more demanding than that. But in the circumstances here we simply discern no meaningful Sixth Amendment interest to weigh against the need to protect a child’s psychological well-being, an interest that qualifies under the precedent of the Supreme Court as substantial enough to justify even the temporary closure of a courtroom let alone the exclusion of a single witness.³

IV

That takes us to the government’s cross-appeal. At the end of Ms. Christie’s trial, the jury convicted her of one federal charge (second-degree murder) and three crimes assimilated from New Mexico state law (intentional child abuse resulting in death, negligent child abuse resulting in death, and negligent child abuse not resulting in death). After receiving the verdict, the district court proceeded to dismiss the two assimilated homicide charges, leaving Ms. Christie responsible for second-degree murder and negligent child abuse not resulting in death. In explaining its dismissal of the assimilated homicide charges, the court pointed to the fact New Mexico does not permit more than one homicide conviction per death and, in this case, 'Ms. Christie already stood convicted for second-degree murder.

This logic, the government now argues, was in error and the assimilated homicide charges should be reinstated. Ms. Christie claims error too but in the opposite direction. She argues the district court was right in dismissing the assimilated homicide charges at the end of trial but should never have permitted them to go to the jury in the first place.

For our part, we affirm the district court’s chosen middle path between these extremes. It’s clear to us that the district court had to dismiss the assimilated charges as a matter of law after trial for two separate but related reasons: the first grounded in the statutory commands of the Assimilative Crimes Act (ACA), 18 U.S.C. § 13, and the second based on our received double jeopardy doctrine. Whether the district court might have been obliged to dismiss the charges before trial as Ms. Christie argues is a question we don’t have to tangle with. We don’t because even assuming some conceivable error arose from the district court’s decision to dismiss the assimilate homicide counts later rather than sooner, any such error was surely harmless.

Take the ACA first. With few exceptions, federal courts have abjured the power to fashion a federal common law of crime, holding that the Constitution generally assigns the job of specifying federal crimes and punishments to the Legislative Branch. See, e.g., Liparota v. United States, 471 U.S. 419, 424, 105 S.Ct. 2084, 85 L.Ed.2d 434 (1985); United States v. Hudson & Goodwin, 11 U.S. (7 Cranch) 32, 34, 3 L.Ed. 259 (1812). In the early days of the republic when federal criminal statutes were few (different days than our own), this meant federal enclaves—places like the military base in this case where the federal government alone enjoys sovereign power—were pretty literally lawless.

Responding to this problem in the 1820s, Daniel Webster introduced the ACA. Instead of trying to write an exhaustive criminal code for federal enclaves, Webster sought more expeditiously and cleverly to borrow from preexisting state law. A testament to its efficacy and economy of design, the ACA remains today little changed from its original form, providing that anyone “guilty of any act or omission which, although not made punishable by any enactment of Congress, would be punishable if committed ... within the jurisdiction of the State ... in which [the federal enclave] is situated, ... shall be guilty of a like offense and subject to a like punishment.” 18 U.S.C. § 13(a); see also Lewis v. United States, 523 U.S. 155, 160-61, 118 S.Ct. 1135, 140 L.Ed.2d 271 (1998); id. at 179, 118 S.Ct. 1135 (Scalia, J., concurring in the judgment).

Sometimes questions about the propriety’ of an assimilated charge will arise and can be resolved early in a criminal case. If, for example, it’s clear that a federal statute applies to the defendant’s conduct and that the assimilation of a state law applying to that same conduct would “interfere with the achievement of a federal policy” or “effectively rewrite an offense definition that Congress carefully considered,” or enter a field Congress has expressed an “intent to occupy,” then the need for dismissing an assimilated crime may be evident even before trial. See id. at 164,118 S.Ct. 1135.

Sometimes, though, it will be difficult to resolve ACA questions until after trial. The statute after all, addresses not only which charges may be brought but also which punishments may be meted out, instructing that for crimes properly assimilated into federal proceedings courts may impose only “like punishment[s]” to those state courts can impose. See 18 U.S.C. § 13(a). In this way the statute plays an important role not just at trial but at sentencing, requiring federal courts “to provide a method of punishing a crime committed on government reservations in the way and to the extent that it would have been punishable if committed within the surrounding jurisdiction.” United States v. Sain, 795 F.2d 888, 890 (10th Cir.1986) (emphasis added). And, of course, not infrequently sentencing issues are difficult to anticipate in advance of trial.

In this case we are certain at least this latter sentencing-related feature of the ACA required the district court to dismiss the assimilated state law claims after trial. New Mexico, it turns out, prohibits the entry of convictions for child abuse-resulting-in-death alongside a conviction for any other form of homicide. See State v. Santillanes, 130 N.M. 464, 27 P.3d 456, 468 & n. 3 (2001); State v. Mann, 129 N.M. 600, 11 P.3d 564, 570 (2000). It is, the parties before us both acknowledge, the state’s avowed penal policy that one death should result in only one homicide conviction: if the jury convicts on multiple homicide charges, all but the most serious must be vacated after trial. Here, Ms. Christie stood convicted of second-degree murder as well as the two child-abuse-resulting-in-death charges. Had she been tried in state court, there’s no question a state court would have dismissed the latter two charges after trial. The only way the district court could obey the ACA’s textual “like punishment” directive was to do the same. This circuit and others have unanimously interpreted the ACA’s “like punishment” mandate to require federal sentencing courts to abide any maximum and minimum prison terms proscribed by state law for an assimilated crime. See United States v. Garcia, 893 F.2d 250, 254 (10th Cir.1989); see also United States v. Martinez, 274 F.3d 897, 906 & n. 12 (5th Cir.2001) (collecting cases). When the maximum punishment a state court can impose for a conviction is “nothing,” it follows that the same must hold true in a federal ACA proceeding.

Neither can there be any question that a conviction on a defendant’s record amounts to a form of “punishment” subject to the ACA’s “like punishment” command. Even when it results only in a concurrent prison sentence or no prison sentence at all, a conviction surely bears “adverse collateral consequences that may not be ignored”: its presence on a defendant’s record may delay her “eligibility for parole or result in an increased sentence under a recidivist statute for a future offense,” it may be “used to' impeach the defendant’s credibility” and it “certainly carries the societal stigma accompanying any criminal conviction.” Ball v. United States, 470 U.S. 856, 865, 105 S.Ct. 1668, 84 L.Ed.2d 740 (1985). All of this is plainly “punishment” indeed. Id.

To be sure, the ACA requires like punishment, not precisely the same punishment. It is not necessary, for example, for federal courts to “duplicate every last nuance of the sentence that would be imposed in state court.” Garcia, 893 F.2d at 254. And federal courts must depart from state guidance when Congress has expressed a specific and contrary penal policy. Id. So, for example, we do not require district courts to follow state parole policies given Congress’s express abolition of parole in the federal system. See United States v. Pinto, 755 F.2d 150, 154 (10th Cir.1985). Neither may federal courts ignore congressionally mandated sentencing guidelines in favor of state sentencing guidelines. See Garcia, 893 F.2d at 254. And federal courts cannot impose the death penalty relying on state law when doing so would upset Congress’s judgment about when capital punishment is and is not warranted. See Lewis, 523 U.S. at 170, 118 S.Ct. 1135. These examples of mandatory departures from the “like punishment” principle are all particular illustrations of the broader principle that when it comes to resolving competing congressional directions the specific (here, regarding parole, sentencing policy, and the death penalty) controls over the general (here, ACA’s generic “like punishment” principle). See RadLAX Gateway Hotel, LLC v. Amalgamated Bank, — U.S. -, 132 S.Ct. 2065, 2071, 182 L.Ed.2d 967 (2012); Elwell v. Oklahoma ex rel. Bd. of Regents, 693 F.3d 1303, 1310 (10th Cir.2012).

Whether or not a conviction stands or falls, however, cannot be sensibly described as a nuance and we see no conflicting and more specific congressional direction in play in our case. Congress has not mandated, or even suggested, that a federal court must enter separate convictions for murder and the assimilated crimes. Instead, the only congressional direction we have is the ACA’s general command to impose only “like punishment” and, in the absence of any more specific direction, that command must be respected.

Apart from the ACA but not wholly unrelated to it, there remains the question of double jeopardy. The Supreme Court has said that the Double Jeopardy Clause of the Fifth Amendment protects not only against successive trials on the same charges but also against “multiple punishments for the same offense.” Whalen v. United States, 445 U.S. 684, 688, 100 S.Ct. 1432, 63 L.Ed.2d 715 (1980) (quotation marks omitted). Sometimes called a guarantee against “multiplicitous punishments,” this feature of due process doctrine “prevents] the sentencing court from prescribing greater punishment than the legislature intended.” Missouri v. Hunter, 459 U.S. 359, 366, 103 S.Ct. 673, 74 L.Ed.2d 535 (1983). Put plainly, federal courts cannot impose cumulative punishments unless Congress has authorized them. It may be a nice question whether this rather simple and intuitive rule flows more naturally from the double jeopardy clause or from the Constitution’s separation of powers, see Whalen, 445 U.S. at 689, 100 S.Ct. 1432, but whatever its source, the result in this case is the same: the assimilated charges against Ms. Christie must go.

In arguing as much, however, Ms. Christie makes a mistake. Because the New Mexico legislature precludes cumulative punishment for homicide, she reasons that the federal district court had no choice but to dismiss the state homicide charges in light of her federal second-degree murder conviction. But this line of thinking misconceives the assimilated homicide charges as state law charges and looks directly to the intent of the New Mexico legislature. This is a misconception because an assimilated charge is a federal charge. Offenses properly assimilated into federal law are no less part of federal law than, say, Congress’s statutorily defined second-degree murder charge. See Sain, 795 F.2d at 891. Indeed, if assimilated charges were truly state charges, a federal court would lack the authority to enforce them in the first place. Knowing this much, we see it is Congress ’s directions about cumulative punishment—not New Mexico’s—that must control our double jeopardy analysis.

How can we discern whether Congress meant to authorize cumulative punishment in this case? Especially when the federal murder statute gives us no express direction on the issue? The Supreme Court has instructed us to apply the “same elements test” as a proxy for congressional intent in the usual case. See Blockburger v. United States, 284 U.S. 299, 304, 52 S.Ct. 180, 76 L.Ed. 306 (1932); see also Hunter, 459 U.S. at 367-68, 103 S.Ct. 673. This test requires us to “inquire[ ] whether each offense” at issue “contains an element not contained in the other.” United States v. Dixon, 509 U.S. 688, 696, 113 S.Ct. 2849, 125 L.Ed.2d 556 (1993). If the answer is yes, we are told to say Congress authorized punishments for both offenses; if the answer is no, we are instructed to say Congress authorized punishment for only one offense. Id. But whatever virtue the Blockburger “same elements test” may have as a proxy for congressional intent, it surely depends on an assumption that Congress wrote all the statutes in question, choosing whether to frame one offense with elements that do or do not overlap with elements of another. Our case, however, poses a quandary, given that Congress did not define the assimilated homicide offenses and could not have crafted their elements in light of the text of its own murder statute. Given this complication, what to do?

The answer lies in the statute Congress did write: the ACA. When a federal statute contains a plainly expressed direction on the question of multiple punishments, it—rather than Blockburger ’s proxy—controls. See Garrett v. United States, 471 U.S. 773, 779, 105 S.Ct. 2407, 85 L.Ed.2d 764 (1985) (“[T]he Block-burger presumption must of course yield to a plainly expressed contrary view on the part of Congress.”). After all, why employ a proxy for congressional intent when we have direct evidence of Congress’s wishes? See, e.g., United States v. Lovett, 964 F.2d 1029, 1042 (10th Cir.1992); United States v. Hampton, 786 F.2d 977, 979-80 (10th Cir.1986). And in this case, the ACA tells us specifically and directly how much punishment Congress wants us to impose.

All this takes us now full circle to our earlier statutory interpretation discussion. The ACA authorizes federal courts to impose “like punishment” and no more. This textual command expresses Congress’s judgment that federal punishment for assimilated crimes should be no more severe than what a state court can impose. And as we’ve already seen, “like punishment” in this case means no punishment at all because in similar circumstances New Mexico courts would have to dismiss both of the child-abuse-resulting-in-death convictions given the presence of another homicide conviction. So, as a matter of federal law and congressional direction— whether one conceives of the issue merely as one of statutory interpretation or more grandly as one of double jeopardy (or separation of power for that matter)—Ms. Christie cannot stand convicted of both second-degree murder and the dismissed offenses.

One final issue we’ve alluded to remains to be confronted. Ms. Christie applauds

the district court’s dismissal of the state homicide charges after the jury returned its verdict. But in her view the district court erred by failing to dismiss the charges before trial. The “like punishment” (ACA) and “multiplicitous punishment” (double jeopardy) problems we’ve discussed arose only at the end of the case when it came clear the jury found Ms. Christie guilty of second-degree murder under federal law and guilty as well of the two assimilated state homicide offenses. At that point, the district court had to act, it had to dismiss the assimilated homicide offenses, to honor the ACA’s “like punishment” command. But, as we’ve seen, the ACA speaks not only to the punishments federal courts may impose but also to the charges government may bring. And Ms. Christie contends an independent and equally apparent ACA charging problem arose before trial, a problem that the district court should have seen and failed to confront. She says the district court ignored the ACA’s textual command to assimilate “like” state offenses only when a defendant “is guilty of any act or omission ... not made punishable by any enactment of Congress.” In her view, the federal murder and manslaughter statutes so completely occupy the field of homicide that they obviously leave no room for assimilating state law child abuse crimes. Accordingly, the district court should have dismissed those claims before the case was placed before the jury. For its part, the concurrence agrees with Ms. Christie at least that far.

We confess we are less sure. Lewis certainly teaches that the federal murder statute, 18 U.S.C. § 1111, occupies the field of killing with “malice aforethought” and, in doing so, precludes the assimilation of state charges covering that same ground. But it’s far from clear to us whether the two state charges at issue here trench on that territory. “Malice aforethought” is a common law term that denotes an intent to kill or to cause serious injury, or reckless disregard for human life. Wood, 207 F.3d at 1228. That mental state clearly does not extend to mere negligence, yet one of the state homicide charges at issue here aims specifically at negligent child abuse resulting in death. Even the other state child abuse offense at issue requires only an intent to abuse, see N.M. Stat. Ann. § 30-6-l(D), and it’s not entirely clear whether that intent, as the state defines it, must amount to a reckless disregard for human life or an intent to kill or cause serious injury.

Alternatively, we might examine whether the state offenses conflict with the federal definition of felony murder. But that course turns out to raise thorny questions of its own. For example, we have suggested that a killing in the course of committing any felony satisfies the malice requirement for second-degree murder. United States v. Pearson, 159 F.3d 480, 486 (10th Cir.1998). The assimilated charges are, of course, in precisely that form: a killing in the course of the assimilated felony of child abuse. But the common-law definition of felony murder endorsed in Pearson might be read more narrowly, as requiring the predicate felony to be one that was recognized at common law or one whose required mens rea would, if transferred to the homicide, meet the “malice aforethought” standard. See Wayne R. LaFave, Substantive Criminal Law § 14.5(b); cf. United States v. Chischilly, 30 F.3d 1144, 1159-60 (9th Cir.1994) (finding no scope for federal felony murder outside the enumerated list of crimes amounting to first-degree murder). And it is entirely unclear whether the state homicide crimes could meet this narrower standard, or whether we’ve ever actually upheld a felony murder conviction that hasn’t met this narrower standard.

Neither are we sure the assimilation of the two state offenses at issue before us necessarily offend the federal manslaughter statute. The two offenses do not purport to address situations in which a homicide occurs in the course of “a sudden quarrel or heat of passion,” as the federal voluntary manslaughter statute does. 18 U.S.C. § 1112(a). The two offenses, as well, are clearly unlawful and felonious acts under New Mexico law, and this may mean that one who commits them is necessarily not guilty of federal involuntary manslaughter. The federal involuntary manslaughter statute, after all, covers only killings “[i]n the commission of an unlawful act not amounting to a felony, or in the commission in an unlawful manner, or without due caution and circumspection, of a lawful act which might produce death.” Id. (emphasis added). Additionally, the ACA expressly contemplates the assimilation of at least some state homicide statutes that seem to overlap with federal manslaughter, see 18 U.S.C. § 13(b)(2)(A)® (increasing the punishment for assimilated DUI offenses resulting in death), so it’s not even clear Congress has any intent to occupy the field of unintentional homicides.

Some day this court may have to resolve these tricky questions. We have no doubt that when it does, the court and litigants alike will be aided by the concurrence’s thoughtful discussion. But for now we don’t need to go so far. We don’t because even assuming (without deciding) that the district court committed an error by failing to dismiss the state homicide charges before trial as Ms. Christie and the concurrence contend, any such error was surely harmless. And on this more modest and entirely dispositive point, we and the concurrence find ourselves in full agreement.

To be sure, Ms. Christie says she was prejudiced in the eyes of the jury. She acknowledges, of course, that the two state homicide charges were eventually dismissed after trial. But she says her defense was compromised by their presence until the end of trial. And at least theoretically she has a point. We do not doubt, for example, that the government could unduly prejudice a defendant in the eyes of the jury by presenting a proliferation of charges at trial that clearly have no lawful place there. It is not impossible to fathom a situation in which the sheer quantity of legally unauthorized charges might induce a jury to think the defendant must be guilty of something. See United States v. Johnson, 130 F.3d 1420, 1426 (10th Cir.1997); United States v. Ketchum, 320 F.2d 3, 8 (2d Cir.1963) (Friendly, J.); Pointer v. United States, 151 U.S. 396, 403, 14 S.Ct. 410, 38 L.Ed. 208 (1894).

That, however, is not our case. The district court took care to instruct the jury to consider each count independently without regard to others. The number and character of the contested charges were relatively modest: just two challenged charges were placed before the jury, and both were lesser species of homicide than the federal murder charge everyone agrees was properly at issue. The evidence against Ms. Christie on each count of conviction was, as well, frankly overwhelming. Given all of these reassuring facts and the absence of any contrary indication that the jury was swayed by the number of charges rather than the evidence presented, we believe Rule 52(a) of the Federal Rules of Criminal Procedure permits us to “disregard[ ]” any conceivable error here on the .ground that it did not affect Ms. Christie’s “substantial rights.” We simply do not see how the presence of these two charges during this trial with this evidence and these jury instructions could have made any difference, In concluding as much, we reach the same destination and by much the same reasoning we did in Johnson, where we held that any conceivable error arising from the district court’s failure to force the government to choose between charges before trial was harmless given the “overwhelming” evidence of guilt presented on the charges that properly remained by the time of final judgment. 130 F.3d at 1426.

So it is we decline to reverse the district court on this score, just as we find no other reversible error anywhere else in its careful treatment of this sad case. The judgment is affirmed.

1 . Ms. Christie says she did object to the seizure. But the district court found otherwise and that factual finding can only be overturned if clearly erroneous. See United States v. Ludwig, 641 F.3d 1243, 1247 (10th Cir.2011). Seeking to meet that demanding standard, Ms. Christie stresses that, after Mr. Wulf consented to the computer’s seizure, she asked for copies of schoolwork she had stored on the computer—a request that, by all accounts, FBI Special Agent Chad Oakes refused. But this fact does not compel the inference Ms. Christie wanted the computer itself or objected to Mr. Wulf's decision to give it to the authorities. In fact, during their conversation Agent Oakes told Ms. Christie he had already obtained Mr. Wulf’s permission for the seizure and Ms. Christie offered no objection in response.

2 . Ms. Christie presses two other suppression arguments, but neither merits extended discussion. First, in contesting the second warrant, she suggests the government lacked probable cause to think the computer contained evidence related to the charges against her. But by the time of the second warrant the government had amassed a great deal of evidence suggesting not only that BW died as a result of Ms. Christie’s neglectful treatment, but also that Ms. Christie’s computer would yield further valuable evidence. Second, Ms. Christie contends the temporal scope of the second warrant was overbroad, authorizing as it did a search for materials dated from BW’s birth to after her death. But evidence during this entire period—again, so long as it ”relate[d] to the murder, neglect, and abuse of” BW—was relevant to the case. The government, for example, had to prepare to rebut the defense’s suggestion that BW's death was caused by her preexisting medical problems rather than Ms. Christie’s neglect, and the government could do so only by showing that BW was (relatively) healthy before Ms. Christie assumed sole responsibility for her care.

3 . Ms. Christie raises two other complaints about her trial and sentencing but both are significantly less compelling.

First, she argues that the trial record contains insufficient evidence to sustain her second-degree murder charge. Everyone agrees that to prevail on that charge the government had to show "conduct which is reckless and wanton, and a gross deviation from a reasonable standard of care, of such a nature that a jury is warranted in inferring that defendant was aware of a serious risk of death or serious bodily harm." United States v. Wood, 207 F.3d 1222, 1228 (10th Cir.2000). Ms. Christie says the government failed to prove she possessed the requisite state of mind in light of (among other things) her testimony professing ignorance of BW’s condition. But on a sufficiency challenge, an appellant must prove that no "rational trier of fact”- could think the government met its burden. United States v. Hutchinson, 573 F.3d 1011, 1033 (10th Cir.2009). And on the facts as we've outlined them, we have no doubt a rational juror could have disbelieved Ms. Christie’s account and concluded that she saw and disregarded BW's glaring and¹ dire symptoms. Second and for similar reasons, we reject Ms. Christie's challenge to the district court’s application of a sentencing enhancement (U.S.S.G. § 3C1.1) for testifying falsely. Ms. Christie insists the district court's finding— that she lied about being unaware of the severity of BW's condition—was wrong. But to unseat the district court’s factual finding, she must show the court clearly erred. And just as we cannot say any rational juror had to believe her story over the government’s competing evidence, neither can we say the district court had to do so.

1 . The federal murder statute has been amended since Lewis was decided. Federal law now classifies as first-degree murder "[e]very murder ... committed in the perpetration of, or attempt to perpetrate ... child abuse.” 18 U.S.C. § 1111(a). Child abuse is defined as “intentionally or knowingly causing death or serious bodily injury to a child.” § 1111(c)(3).