Florida Bar Journal

Nullius in Verba? A Multidisciplinary Assessment of the Assertion, “Eyewitness Testimony Is Inherently Unreliable”

Nullius in verba, which is Latin for “take nobody’s word for it,” is the motto adopted in 1662 by The Royal Society of London for Improving Natural Knowledge.^[1] While this motto may be applicable to the realm of the natural sciences where the reliance on assertions unverified by the scientific method engenders skepticism, strict application of the motto to trial practice would cause a seismic shift in how truth is reached in the jury trial process. After all, witness testimony is the primary form of evidence used in the courtroom to tell the story of what happened in a case.

As a trial judge, I was recently confronted with this issue after different attorneys proclaimed in two successive murder trials that “eyewitness testimony is inherently unreliable.” This assertion was placed before the jury pool in both cases during jury selection and put forward in one case during opening statements. However, is it appropriate for a party to make such an assertion to jurors or potential jurors? As a matter of caselaw, science, and epistemology, the assertion that “eyewitness testimony is inherently unreliable” seems to be overly broad. It also appears to be unsuitable to be placed before jurors as a matter of axiom.

This article examines the increasingly common courtroom assertion that “eyewitness testimony is inherently unreliable.” While this claim sounds sophisticated, a closer examination demonstrates its vulnerability across three fields of inquiry. Legally, it ignores existing safeguards for testing witness credibility and potentially usurps the jury’s traditional role. Scientifically, the broad statement is unsupported and much of the psychological research attacking eyewitness testimony has been discredited by well-documented replication failures. Philosophically, accepting it would undermine most human knowledge acquisition and destabilize the truth-seeking function of our judicial system. Rather than blanket skepticism toward witnesses, practitioners should evaluate eyewitness testimony contextually. In those rare cases in which expert testimony on witness reliability proves necessary, Florida law provides a framework for its admission. After exploring why the assertion fails in each field, this article offers guidance for Florida practitioners on when such expert testimony might be appropriate.

The Legal Issues With the Statement and Why Expert Testimony Is Required

As philosopher David Hume noted, “there is no species of reasoning more common, more useful, and even necessary to human life, than that which is derived from the testimony of men, and the reports of eyewitnesses and spectators.”^[2] Varied sources of evidence suggest that eyewitness testimony is the most persuasive form of evidence presented in court.^[3] Unlike a video recording of an event, which is objective in its presentation, eyewitness accounts of an event can provide context and substance through the questioning of the witnesses that a video recording cannot provide. But as with all subjective forms of presentation, eyewitness testimony can be fallible for a variety of reasons. The evidence code and the structure of evidentiary presentation offer ways to test subjective eyewitness testimony. For example, during cross examination, a skilled trial attorney can tease out bias,^[4] lack of sensory acuity,^[5] or dishonesty.^[6] Such techniques are well known among legal movie aficionados as portrayed in entertaining fashion in My Cousin Vinny.^[7]

However, the statement made in my courtroom was not attacking an individual witness. It was attacking all witnesses. Since the statement, “eyewitness testimony is inherently unreliable,” is not a commonly accepted fact, and no expert was going to testify regarding the reliability of all eyewitness testimony, I instructed counsel under F.S. §90.104(2),^[8] to not make the statement again. But could a party call an expert to testify about the reliability of eyewitness testimony? In Florida, it depends on whether you are litigating in federal or state court.

In United States v. Thevis, 665 F.2d 616, 641 (5th Cir. 1982), the former^[9] federal Fifth Circuit Court of Appeals reviewed a criminal appeal in which the defendant was challenging a trial court’s decision to not allow expert testimony to discuss problems with eyewitness identification. Specifically, the expert was going to explain why aircraft pilots, as a group, were not better equipped than ordinary witnesses to make identifications. After a proffer, the trial judge did not allow the expert to testify as he concluded that the probative value of the evidence was substantially outweighed by the possibility of prejudice emanating from this “expert” testimony. The trial court concluded that the subjects of perception, memory, and identification were within the expertise of the jury. Upon review, the appellate court concluded that the trial court had not abused its discretion by disallowing the testimony. The Thevis court noted, “To admit such testimony in effect would permit the proponent’s witness to comment on the weight and credibility of opponents’ witnesses and open the door to a barrage of marginally relevant psychological evidence.”^[10] The court ultimately agreed with the trial court in that “the problems of perception and memory can be adequately addressed in cross-examination and that the jury can adequately weigh these problems through common-sense evaluation.”^[11]
The 11th Circuit Court of Appeals has reviewed its per se rule prohibiting expert testimony on the reliability of eyewitness identification several times. In both United States v. Benitez, 741 F.2d 1312, 1315 (11th Cir. 1984), and United States v. Holloway, 971 F.2d 675, 679 (11th Cir. 1992), the court has stood by its conclusion that “[t]he established rule of this circuit is that such testimony is not admissible.”^[12] In United States v. Fred Smith, 122 F.3d 1355 (11th Cir. 1997), the court once again reviewed its rule prohibiting expert testimony regarding eyewitness identification, this time in light of the U.S. Supreme Court’s decision in Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993). While the 11th Circuit determined that trial courts may continue to exclude expert testimony on eyewitness identification, even under Daubert, it did note that a request for “jury instructions that highlight particular problems in eyewitness recollection” was a potential tool for defense counsel in federal criminal trials as was the “powerful tool of cross-examination.”^[13]
More recently, this issue has been raised twice at the 11th Circuit, once in a dissent regarding the denial of an en banc review, United States v. Owens, 682 F.3d 1358, 1359 (11th Cir. 2012) (Barkett, J., dissenting from the denial of rehearing en banc), and once in a concurrence, United States v. Daniels, 91 F.4th 1083, 1103 (11th Cir. 2024) (Jordan, J., concurring). In Owens, Judge Barkett argued that the 11th Circuit should reconsider Thevis and its progeny. Specifically, she contended the court should change its “position that appellate courts are never permitted to review for abuse of discretion the exclusion of expert testimony regarding the reliability of eyewitness identifications.”^[14] The main thrust of her argument was that the field of psychology over the 30 years since Thevis had produced an abundance of evidence concerning the unreliability of eyewitness testimony. These psychological studies seemed to demonstrate that variables such as high stress, limited time viewing the suspect, poor police line-up protocols, among others, may impact the reliability of an eyewitness.^[15]
In Daniels, Judge Jordan discussed eyewitness unreliability in his call for the “Eleventh Circuit Committee on Pattern Jury Instructions to revise the pattern instruction on identification to allow juries to consider, in appropriate cases, that the witness and the person identified are of different races.”^[16] In his concurrence, Judge Jordan cites to Judge Barkett’s dissent in Owens wherein she asserted that “unreliability of eyewitness testimony is now widely recognized in the psychological literature and by law enforcement.”^[17] Judge Jordan also repeats an assertion Justice Sotomayor made in a dissent she authored in 2012 in the case of Perry v. New Hampshire, 565 U.S. 228, 249 (2012) (Sotomayor, J., dissenting). Quoting the New Jersey Supreme Court, Justice Sotomayor proclaimed that the psychological research regarding eyewitness testimony “represents the ‘gold standard in terms of the applicability of social science research to the law.’”^[18]

Unlike the federal courts in Florida, which have adopted a prohibitory view, the state courts have adopted the discretionary view regarding expert testimony about eyewitness reliability.^[19] In one of the earliest cases in Florida to explore the issue, the Florida Supreme Court reviewed a trial court’s refusal to allow a professor of psychology to testify about the reliability of eyewitness identification. The court in Johnson v. State, 438 So. 2d 774, 777 (Fla. 1983), stated:

Expert testimony should be excluded when the facts testified to are of such nature as not to require any special knowledge or experience in order for the jury to form its conclusions. Johnson. We hold that a jury is fully capable of assessing a witness’ ability to perceive and remember, given the assistance of cross-examination and cautionary instructions, without the aid of expert testimony. We find no abuse of discretion in the trial court’s refusal to allow this witness to testify about the reliability of eyewitness identification.

Fifteen years later, in McMullen v. State, 714 So. 2d 368 (Fla. 1998), the Florida Supreme Court performed an in-depth review of its precedent on this matter and reaffirmed its holding in Johnson. However, the court took McMullen as an opportunity to clarify a misunderstanding that had arisen in the precedent. As one district court judge commented in the case below, “The opinion of the supreme court in [Johnson] is to my mind a rather categorical holding that such evidence is inadmissible.”^[20] The statement from McMullen that caused the confusion was the following: “We hold that a jury is fully capable of assessing a witness’ ability to perceive and remember, given the assistance of cross-examination and cautionary instructions, without the aid of expert testimony.”^[21] The court in McMullen definitively dispelled the confusion over whether expert testimony regarding the reliability of eyewitness identification was admissible in state courts. It concluded, “We hereby reaffirm our holding in Johnson by concluding that the admissibility of expert testimony regarding the reliability of eyewitness testimony is left to the sound discretion of the trial judge. By so holding, we are continuing to align ourselves with a majority of other jurisdictions.”^[22] With this issue cleared up, the state courts of Florida were firmly in the discretionary view camp.

Similar to contrarian voices in the federal system, Florida Supreme Court Justice Pariente authored two notable concurring opinions — first in Simmons v. State, 934 So. 2d 1100, 1123 (Fla. 2006) (Pariente, J., specially concurring), encouraging trial courts to exercise discretion in admitting eyewitness expert testimony, and later in Peterson v. State, 154 So. 2d 275, 285 (Fla. 2014) (Pariente, J., concurring), more forcefully advocating that such testimony “should be generally admissible” and urging her colleagues to abandon their restrictive precedent.

In Simmons, Justice Pariente emphasized that research over the past 23 years had revealed that common-sense assumptions about eyewitness testimony are often wrong. For example, while intuition suggests that a more confident witness is more accurate, contemporaneous research showed confidence correlates weakly with accuracy.^[23] She cited research demonstrating other counter-intuitive findings, such as difficulties in cross-racial identification.^[24] She argued that expert testimony can serve as a “powerful tool” to help achieve the criminal justice system’s goal of convicting the guilty while acquitting the innocent. Rather than mandating admission of such testimony, Justice Pariente encouraged trial courts to genuinely exercise their discretion under §90.702, which allows expert testimony when specialized knowledge would assist the jury. She urged judges to consider whether expert testimony would introduce relevant accuracy considerations that could not be addressed through cross-examination, jury instructions,^[25] or common sense alone. Her concurrence in Simmons essentially argued for a more flexible approach that would allow expert testimony on eyewitness identification in appropriate cases, particularly when such testimony is the primary evidence against a defendant.

In Peterson, Justice Pariente concurred with the majority’s decision but wrote separately to clarify that expert testimony on eyewitness identification should generally be admissible, particularly in cases relying substantially or entirely on eyewitness testimony. She expressed concern that Florida’s precedent has created the false impression that such testimony is generally inadmissible. Justice Pariente again critiqued the Johnson precedent as being out of step with the national trend and asserted that Florida’s approach is out of step with widespread judicial recognition that eyewitness identifications are unreliable in ways unknown to average jurors.

However, the acclaimed studies from the fields of social and cognitive psychology, many of which are from the era before the replication crisis came to light, may not be as reliable as Justice Pariente, Judge Jordan, and Judge Barkett were led to believe. As discussed next, the tumult in the field of psychology over the last two decades may call into question whether an expert would even have a valid scientific basis to make an assertion about the reliability of eyewitness testimony. Some of the assertions could run into issues under the Daubert standard.^[26]

The Scientific Problems With the Statement and Why Recent Developments May Undermine Expert Opinion

Despite seemingly compelling arguments and the growing judicial acceptance of eyewitness expert testimony based on decades of psychology research, recent developments in the field have introduced new complexities to this debate. As has been recently discovered, the social and cognitive psychology research that has driven this judicial shift toward admitting eyewitness expert testimony rests on increasingly shaky scientific ground. The replication crisis that has plagued social and cognitive psychology since at least 2008 has revealed that many foundational studies in the field — including key research on memory, perception, and cognition — cannot be replicated under rigorous scientific conditions.^[27] Rather than representing the “gold standard” of social science application to law, much of this research may very well suffer from methodological flaws, publication bias, and inflated effect sizes that call into question the very premises underlying expert testimony on eyewitness identification. As courts rush to embrace this supposed scientific evidence, they may be building legal precedent on a foundation of irreproducible research that fails to meet basic standards of scientific reliability.

For example, a study published in 2015 titled, Estimating the Reproducibility of Psychological Science, by the Open Science Collaboration addressed the replication crisis in psychology.^[28] This article reported the results of a large-scale project designed to evaluate the reproducibility of 100 psychological studies originally published in 2008 across three prominent journals: Psychological Science; Journal of Personality and Social Psychology; and Journal of Experimental Psychology: Learning, Memory, and Cognition. The results revealed a significant gap between original and replicated outcomes. Only 36% of replications achieved statistically significant results, compared to 97% of the original studies. Subjective assessments deemed 39% of replications successful. Interestingly, social psychology had a much lower subjective replication rate (28%) as compared to cognitive psychology (54%). These findings suggested that many published psychological results, particularly in the sub-field of social psychology, may be unreliable and potentially inflated by publication bias, selective reporting, or methodological flaws.

In another article, Replicability Crisis in Social Psychology: Looking at the Past to Find New Pathways for the Future, the authors recognized the replicability crisis in social psychology was particularly acute and represented a fundamental challenge to the field’s scientific credibility.^[29] Citing the 2015 Open Science Collaboration study, the authors identified various methodological issues that have contributed to the crisis, including problems with statistical power, publication standards, and research practices that may have inflated the likelihood of false positive results. The frequent inability to replicate findings in psychology, particularly social psychology, has led many scientists to question the reliability of established psychological knowledge.

Even prior to the revelation of a replication crisis in the field of psychology, conscientious researchers acknowledged some shortcomings in the study of eyewitness testimony. In a 2003 article titled, Eyewitness Testimony, the authors provided an overview of what findings in psychology concerning eyewitness testimony were compelling, what areas needed more research, and what limitations the evidence exhibited.^[30] The authors noted the two sets of variables that affect eyewitness identification accuracy: system and estimator. System variables are those under the control of the criminal justice system, such as lineup instructions and functional lineup size. Estimator variables are those that cannot be controlled, such as lighting conditions during witnessing and whether the witness and perpetrator are of the same race. The scientific literature regarding system variables has had the most practical impact, as attested to by the Department of Justice’s 1999 publication of Eyewitness Evidence: A Guide for Law Enforcement.^[31] However, despite that influence, the authors of Eyewitness Testimony acknowledged the deficiencies in the literature, including that much of the data is laboratory driven and lacks backing in real world data. Further, laboratory driven experiments tend to focus on a specific issue impacting eyewitness reliability and cannot accurately replicate the complexity of actual criminal encounters and investigations. This is a common admission even of the most strenuous advocates for the admission of eyewitness expert testimony.^[32]
Recent scholarship continues to expose fundamental problems with eyewitness research. In a recent article by Kaja Głomb, she asserts that despite more than 40 years of research on eyewitness testimony, there are fundamental methodological and theoretical shortcomings that have led to contradictory findings regarding some aspects of the research.^[33] For example, while there is a supposed “consensus” that emotions negatively affect memory, the research shows numerous studies supporting both positive and negative effects of emotions on memory. The critique suggests that much of the existing research on emotions and eyewitness testimony may be methodologically flawed, explaining the contradictory findings in the field.

From this brief overview of the latest literature in psychology, the evidentiary and empirical bases for some of the psychological findings in recent years lack a solid foundation in science. Though the findings related to system variables seem to have a more solid foundation, the findings regarding estimator variables, especially in social psychology, appear uncertain.^[34] This could call into question whether some of the underlying claims made by psychologists over the last 30 years are indeed the “gold standard.”^[35] While these methodological concerns do not invalidate all eyewitness research, they do complicate claims that psychology provides definitive answers about witness reliability.^[36] In any event, while there may be some issues with witness reliability under certain circumstances, the expansive assertion that “eyewitness testimony is inherently unreliable” is certainly not confirmed by any scientific research.

The Epistemological Problems With the Statement Vis-À-Vis the Truth-Seeking Function of the Judicial System

This brings us to a third analytical framework to consider beyond legal precedent and empirical research. We must examine epistemologically whether categorical skepticism toward witness testimony undermines the very foundations of knowledge itself. Epistemology, the philosophical study of knowledge, truth, and justified belief, reveals inherent contradictions in the claim that “eyewitness testimony is inherently unreliable.” If we accept this blanket dismissal, we face a paradox. Much of human knowledge, including the scientific research used to question a testimony’s reliability, ultimately depends on forms of witnessing and reporting. Given that our justice system fundamentally relies on distinguishing truth from falsehood, epistemological analysis exposes the logical inconsistencies in the wholesale rejection of testimonial evidence.

The adversarial trial system is set up to search for the truth.^[37] During a trial, a trial judge is “presumed to have drawn on his talents, his knowledge and his experience to keep the search for the truth in a proper channel.”^[38] But as Pontius Pilate asked Jesus, “What is truth?” In our justice system, truth is determined through evidence and the collective wisdom of juries.^[39] Pervasive skepticism toward the most prevalent form of evidence — eyewitness testimony — would radically undercut our system. Juries could not find truth if they could take nobody’s word as fact. Doubt based on some identifiable fault with a particular witness is valid, and our evidentiary code recognizes that. However, the blanket assertion that “eyewitness testimony is inherently unreliable” represents the kind of universalizing claim that even postmodern thinkers critiqued.^[40] While knowledge via eyewitness testimony may be contextual, and the trial system rightly tests veracity and accuracy, knowledge through that medium is not unattainable.

If witnesses to criminal acts can never be trusted to convey what they saw, the trial system collapses and criminals escape accountability. Successful prosecutions would require video evidence with clear facial identification or DNA that perfectly matches the defendant (as a matter of statistical probability) without contamination from other sources. Yet even these evidentiary sources crumble when subjected to excessive skepticism: DNA results require expert interpretation and testimony, while video evidence faces the growing threat of AI-generated deepfakes. So, what sources of evidence survive this deep-seated skepticism? Perhaps circumstantial evidence? However, in my experience on the bench, juries often view circumstantial evidence with reservation. Thus, if we accept that eyewitness testimony is inherently unreliable, we would effectively dismantle the jury trial system.

This brings us to the first epistemological problem with the statement itself. The word “inherently” suggests testimony is unreliable by its very nature, but this seems too broad. While testimony can be unreliable in some circumstances, it is reliable enough in others to serve as a foundation for knowledge. For example, we successfully navigate daily life largely through testimonial knowledge — from simple facts like store hours to complex scientific theories we accept from experts.

The next issue is the baseline problem. What standard of reliability is being used when one asserts that “eyewitness testimony is inherently unreliable?” All sources of knowledge, including perception and reasoning, involve some degree of fallibility. If we reject testimony for being unreliable, consistency might require rejecting other knowledge sources that are similarly imperfect, such as the findings in the field of psychology. Rejecting all testimony as unreliable because certain testimony is unreliable some of the time is a form of nihilism.^[41]
The broad statement also runs afoul of the context-independence problem. The statement ignores how reliability varies with context. Testimony about simple, directly observable events may be quite reliable, while testimony about complex, emotionally charged, or temporally distant events may be less so in some cases. Treating all testimony as equally unreliable overlooks these important distinctions.

As a matter of common sense, people vary dramatically in their observational skills, memory capacity, attention to detail, and honesty. Training and experience may also impact one’s reliability. For example, a trained forensic observer, an experienced pilot, or a professional wine taster (sommelier) may have vastly superior reliability in their domains of expertise compared to untrained individuals. Also, research shows substantial variation in memory accuracy, attention span, perceptual acuity, and susceptibility to suggestion across individuals.^[42] Some people are simply better witnesses than others due to cognitive differences, personality traits, or life experiences. Other individuals possess extraordinary memory abilities — a photographic or eidetic memory — such as the capacity to recall visual details in great detail after only a brief exposure. The statement ignores these meaningful differences in testimonial competence.

These various concerns suggest that while witness testimony certainly has limitations and context-dependent reliability issues, the blanket claim of inherent unreliability is epistemologically problematic. In other words, just because any individual eyewitness may be fallible, that does not equate to the unreliability of all eyewitnesses. This individual variation problem suggests that rather than making blanket judgments about testimonial reliability, advocates should focus on witness competence, context, and individual differences. The successful advocate will understand the limitations of eyewitness testimony generally, and the shortcomings of an individual witness specifically.

A Framework for Admission of Eyewitness Expert Testimony in the State Courts of Florida

Having examined the broad assertion regarding eyewitness testimony across legal, scientific, and philosophical dimensions, the analysis now turns to practical guidance for Florida practitioners confronted with the need for eyewitness expert testimony regarding a specific witness. While the phrase, “eyewitness testimony is inherently unreliable,” is certainly overly broad, there are limited circumstances in which eyewitness expert testimony may be useful and appropriate. Justice Bell, in his concurrence in Simmons, provided a beneficial framework outlining when such testimony could be considered in Florida courts.^[43] He provided the following guiding principles:

1. Whether to allow expert testimony on the issue of eyewitness identification is left to the sound discretion of the trial judge.

2. As with other witnesses, evaluating the credibility and reliability of eyewitnesses is generally for the jury to determine, without the help of eyewitness experts. Effective cross-examination, presentation of contradictory evidence, sound argument of counsel, and jury instructions are all effective, efficient, and generally sufficient means of assisting the jury in properly fulfilling its critical function as factfinder.

3. Expert testimony on eyewitness identification will be necessary in very limited circumstances.

4. If eyewitness expert testimony is deemed necessary, it must be “Frye tested.” The expert’s opinion must rest upon research and principles that are generally accepted in the relevant scientific community.

His fourth principle must substitute the Daubert standard and should now state: “If eyewitness expert testimony is deemed necessary, it must meet the Daubert standard. The expert’s opinion must rest upon research and principles that are reliable, e.g., able to be replicated, and generally accepted in the relevant scientific community.”

If the proposed expert testimony is suitable under these four principles, Justice Bell also provides five factors to be considered for admissibility:

(a) Identification of the defendant is a critical, disputed issue in the case; and

(b) The state’s case essentially rests upon the credibility of eyewitness testimony; and

(c) There is little or no other evidence to independently establish that the defendant is the person who committed the crime; and

(d) The expert testimony addresses a specific circumstance of the eyewitness identification and provides a particularized explanation which will assist the jury in sorting out the facts; and

(e) The defendant maintains that he was not present when the offense was committed and advances an alibi.

These guiding principles and admissibility factors are exceptionally helpful in assisting the bench and bar navigate when eyewitness expert testimony may be useful and appropriate in a given case.^[44] Importantly, they ensure the scales of justice are balanced. After all, as the former Fifth Circuit stated in United States v. Garza, 608 F.2d 659, 662 (5th Cir. 1979):

A criminal trial provides a neutral arena for the presentation of evidence upon which alone the jury must base its determination of a defendant’s innocence or guilt. Attorneys for both sides, following rules of evidence and procedure designed to protect the neutrality and fairness of the trial, must stage their versions of the truth within that arena. That which has gone before cannot be considered by the jury except to the extent it can be properly presented at the trial and those things that cannot properly be presented must not be considered at all.

A fair and impartial forum that admits relevant and reliable evidence creates the optimal conditions for discovering truth and achieving justice. Such conditions are compromised when courts admit expert testimony grounded in methodologically flawed research or when jury decisionmaking is influenced by radical skepticism toward the pursuit of truth. If we allow the search for truth to become inconsequential in a court of law, the legal profession devolves into a mere power struggle — precisely what American philosopher Charles Sanders Peirce predicted more than a century ago:

When society is broken into bands, now warring, now allied, now for a time subordinated one to another, man loses his conceptions of truth and of reason. If he sees one man assert what another denies, he will, if he is concerned, choose his side and set to work by all means in his power to silence his adversaries. The truth for him is that for which he fights.^[45]

Attacking all eyewitness testimony as unreliable is a symptom of this decline.^[46]

Conclusion

This analysis across legal, scientific, and philosophical dimensions demonstrates that responsible advocacy requires abandoning oversimplified claims about witness reliability in favor of nuanced evaluation. Florida practitioners must recognize that attacking all witness testimony as inherently unreliable not only lacks evidentiary support but threatens the foundational assumptions underlying jury trials. When eyewitness expert testimony seems necessary, it should address specific circumstances with reliable research rather than broad generalizations or methodologically flawed studies. By maintaining rigorous standards for both witness evaluation and expert testimony, Florida courts can preserve the integrity of the truth-seeking process while avoiding the extremes of either uncritical acceptance or systematic skepticism.