Trial lawyers are routinely confronted with recantations and inconsistent witness statements before and during trial. Attorneys decide which version advances their trial theory and then use their trial skills and other evidence to shore up that version and discredit the other. Jurors are tasked with evaluating the credibility of the inconsistent witness. While each advocate argues in summation that their preferred version represents the truth and should be credited, the law does not require jurors as fact-finders to presume any particular version true. Indeed, the entire legal system presupposes jurors capable of determining which version–if either–is worthy of their belief and which should be discredited. The jury is also free to conclude that they simply do not know which version to believe or disbelieve.
It is through this adversarial process that we believe (hope) the truth will emerge. What happens, though, when the truth arrives to the party a little late, as in the form of a post-conviction recantation?
The Troy Davis case brought national attention to the challenges created by post-conviction recantations. Legal, political, and social advocates for Davis assigned great significance to the number of recanting witnesses and were astonished to discover that the courts lacked the desire or means to effectively assess the reliability of post-conviction recantations. State and federal courts, it turns out, are extremely distrustful of recantation evidence and always have been.
The Georgia Supreme Court has been dismissive of recantations as far back as 1876. In Felton v. State, 56 Ga. 84, the court noted:
[I]t would be easy, for love or money, to get some witness to say something contradictory to his evidence on the trial.
The Court of Appeals of New York’s 1916 assessment of recantation evidence remains the prevailing view in many jurisdictions:
There is no form of proof so unreliable as recanting testimony. In the popular mind it is often regarded as of great importance. Those experienced in the administration of the criminal law know well its untrustworthy character.
It is this untested presumption of unreliability that enables courts to quickly deny relief for those relying on recantations as proof of innocence. But what is it about recantations that cause courts to presume them inherently unreliable? Does the fact that some recantations are unreliable mean that all recantations are? Does the truth come with an expiration date? Is a liar no less a liar simply because the witness and/or the lawyer convinced the jurors to believe the liar? How is it possible to know, merely from the occurrence of a recantation, which version–if either–is true? As to the latter question, I think the Missouri Supreme Court, in State ex.re. Amrine v. Roper, 102 S.W.3d 541 (Mo. 2003), astutely characterized the problem when a recantation directly undermines the only evidence at trial.
With the witnesses’ recantations, we do not know whether [the man convicted] is actually innocent. We similarly do not know whether he is guilty, despite the final judgment in his case. The question is: which time were these [recanting] witnesses lying? When they testified [at trial], or when they recanted?.
For obvious reasons, the very act of recanting calls into question the credibility of the witness. Nevertheless, no reasonable person would suggest that an untested recantation is sufficient to undermine the verdict. By the same token, a system premised on the categorical distrust of all recantations does not serve the truth and is often fatal to the claims of innocence and sometimes, as in the case of Troy Davis, to the innocent themselves. Context is everything.
It is irrefutable that some unknown number of people are convicted on the word of liars. As of this writing, there are 280 DNA exonerations listed on the Innocence Project’s site, thanks to the tireless work of members of the Innocence Network. Moreover, as documented here and here, hundreds of other innocent people have had their questionable convictions reversed without the benefit of DNA, including 121 non-DNA exonerations from death row. Unfortunately, because biological evidence is generally not present in the majority of crimes, DNA testing is often of no use to innocent people convicted on perjured testimony. Therefore, if we are determined to follow the truth, we need a framework for assessing the reliability of recantation evidence in a case-by-case basis.
In 2007, we filed an amicus brief in the Georgia Supreme Court on behalf of Troy Davis, wherein we suggested a contextual analysis:
There are a number of common sense factors that amici believe would assist the Court in differentiating between untrustworthy recantations and those that may well signal a wrongful conviction. The factors, suggested by a review of those cases with known reliable recantations, are: 1) the recanting witness’s motivation, including any relationship the witness may have to the defendant and the circumstances surrounding the decision to recant; 2) the reliability of the original trial testimony and whether the recantation is corroborated by other facts known at trial; 3) the number of independent recantations and the degree of consistency among them; 4) the existence or absence of other material evidence of guilt.
The court would assess the reliability of the recantation by considering it in light of each of the factors. While some recantations may be easily and consistently categorized as either reliable or unreliable as each factor is considered, undoubtedly, there will be other recantations that perform well in some categories and not so well in others. Those recantations will require the court to “balance” the results from each factor and use its discretion to determine whether there are sufficient indicia of reliability to warrant a new trial.
Certainly our judges should be as capable as our jurors when it comes to determining which version–if either–is to be trusted.