In Opinion on 704(b), the Majority and Dissent Reveal Ignorance About How Mental States are Commonly Inferred (Guest post by Professors Teneille Brown and Emily Murphy)

[Editor's Note: On June 21, this blog featured an essay on Diaz v. United States criticizing the result but also the dissent in one important respect. Below is another perspective on Diaz, courtesy of Professors Teneille Brown and Emily Murphy, who submitted an evidence professors' amicus brief in the case.]

By now you’ve probably seen that the U.S. Supreme Court decided Diaz v. United States on June 21, 2024. The Court affirmed Ms. Diaz’s conviction 6-3, with Justice Thomas writing the majority opinion, Justice Jackson writing a separate concurrence, and Justice Gorsuch writing a dissent (joined by Justices Kagan and Sotomayor). 

We thought it might be helpful to share how we distill the reasoning of the majority, concurring, and dissenting opinions. One way to do this is by imagining a future case where a woman is charged with killing her abusive husband. The defendant claims it was done in self-defense. Now, suppose, she wants to introduce the testimony of an expert who will testify that her fear may have been reasonable, as a woman with battered woman syndrome.


According to the reasoning of the majority opinion written by Justice Thomas, an opinion by our hypothetical expert will not violate 704(b) if the expert limits herself to probabilistic statements about how “most” battered women think and behave. Probabilistic testimony, in Justice Thomas’s view, is not testimony about the defendant’s ultimate mental state, so it does not violate 704(b). Unsurprisingly, the opinion focuses on the textual analysis quite a bit (specifically, the use of the word “about”). The majority frames 704(b) as a narrow, later-added exception to the more permissive rule 704(a) that abolished the ultimate issue rule. 


Given the amount of time Justice Thomas spent including gratuitous facts that paint Ms. Diaz as guilty, one wonders whether he would come out the same in our hypothetical case. The reasoning appears to establish a clear-cut rule in favor of permitting probabilistic statements about how “most” people like the defendant behave: “most” does not equal “all;” therefore “most” does not violate 704(b). As with most black and white rules, however, this formalist interpretation does not grapple with the shades of gray that will inevitably appear in future cases. For example, what if an expert testifies as to how “nearly all” or “all” of her patients have behaved, and hints at the possibility that the defendant might be different in an important way that may exempt her from this class? As Justice Gorsuch points out in his dissent, if there is no direct link between the defendant and the expert’s opinion, the testimony might have a FRE 401/402 problem.


Moving to Justice Gorsuch’s dissenting opinion, he argues that the state should not be able to opine on the defendant’s mental state, even if the expert qualifies the opinion with “most people like defendant” as opposed to “all.” Agent Flood should not have been able to testify as to what “most” drug couriers know, because that gives prosecutors a “powerful new tool” (never mind that it has been the dominant Circuit interpretation for decades). But wait, what about if the testimony is raised by the criminal defendant, as in our hypothetical? The dissenting opinion has a confusing answer to this, as they say that the battered woman could present expert testimony as to her “symptoms” but not as to her mental state. The dissent does not elaborate on this, probably because “symptoms” can include common mental states (like fear), and also does not address the need for the symptoms to be relevant to this case. So, it is not clear how the dissent would treat this sort of defense-offered expert testimony.


Unfortunately, both the majority and dissent make armchair assumptions about how jurors read minds and draw inferences. Justice Thomas assumes that juries will be magically entranced by the word “all” in a way they would not be if the expert instead said “most.” If an expert said “in my opinion all battered women reasonably fear for their lives,” are we really to expect that juries will uncritically adopt this, without testing it against other evidence and their own read of the facts (and credibility of the expert)? Is the “most” really doing that much work? We don’t know; all that Justice Thomas offers is that the words “most” and “all” are sufficiently different. 


And in dismissing expert testimony on mental states as “telepathy” and “fortunetelling” Justice Gorsuch says the quiet part out loud: that while our criminal statutes require proving mens rea, it is incredibly hard to do so. The alternatives to 704(b) expert testimony Justice Gorsuch suggests (e.g., inferring Diaz’s guilty knowledge of the drugs simply from the fact that her story changed and she admitted to lying) may result in just convictions, but maybe not in circumstances that are unfamiliar to lay jurors (e.g., people lie to protect their friends).  Instead of engaging with how expert testimony on mental states might bridge the necessary inferential chain (and track how humans in their daily lives read the minds of others, using contextual and experience-based information), the dissent instead pivots to challenging the reliability of such law enforcement expert testimony and the concern it will be given too much weight. This argument is valid; we all know this issue should be addressed under FRE 702, something defense counsel failed to do here. Finally, if the mere process of inferring mental states from contextual information is such unreliable quackery, as Justice Gorsuch strongly implies, then this flouts the very purpose of the jury—to infer mental states by listening to the testimony and deciding whether they believe it is true.


It is only Justice Jackson’s concurring opinion that sheds some light on the nature of jury inference and what juries should hear in order to make those inferences. First, she writes separately to acknowledge the blind spots in both the majority and the dissent: 704(b) is party-agnostic, and prohibiting group mental state testimony will hurt criminal defendants. She also explains how the expert’s 704(b) testimony might help bridge the inferential chain—i.e., to move from the facts presented, to the expert’s group data, to the jury ultimately deciding whether those facts and expert data apply in this case: “the type of mental-state evidence that Rule 704(b) permits can be of critical assistance to lay factfinders tasked with determining a defendant’s mental state as an element of the alleged crime (or defense).” Critically, she also acknowledges that expert mental state evidence can provide important context for a defendant’s behavior or condition: “given the biases, stereotypes, and uneven knowledge that many people have about mental health conditions, such expert evidence could help jurors better understand a defendant’s condition and thereby call into question a mens rea that might otherwise be too easily assumed.” In sum, Justice Jackson - a former Federal Public Defender with eight years of district court judge experience - offered a theory of jury decision-making that is capable of nuance and benefits from having contextualizing information.


Acknowledging how the rules of evidence operate together, Justice Jackson described how unreliable cop expert testimony not excluded under Rule 704(b) nevertheless ought to be challenged under Rules 702, and/or 403, and if admitted, subjected to vigorous cross-examination. In a few places she cited the amicus brief we submitted on behalf of a group of distinguished evidence professors, including specifically adopting our suggestion of using jury instructions to distinguish the expert’s role from the jury’s. Alas, she did not explicitly adopt the “frameworks” evidence terminology our brief proposed, even though she endorsed the reasoning behind it.  But according to our seasoned counsel-of-record Matt Coles, having the brief cited in the opinion is apparently fairly rare. So, we are counting this as a win!


It was truly a fantastic experience to work on the brief, and to contribute to the interpretation of our evidence rules at such a high level. Thank you so very much to everyone who signed on and helped us be heard.