B. Tyson's Objections
Tyson objects to the Magistrate's conclusion that even if the trial court erred in denying funding for an expert, such error was harmless. According to the Report, the most important ground for this conclusion was the fact that Tyson had not met his burden of showing that the "linguistic discourse analysis" his expert provided would have been admissible at trial. See Report at 24-27. Tyson does not dispute the inadmissibility of this evidence; instead, he argues that if the proposed expert had been hired, his attorney would have known that the tape was authentic and therefore would have used a different, and perhaps more successful, strategy at trial.
When a federal court conducts a habeas review of a conviction, the standard it applies is less stringent than it would be if the court were conducting a direct appellate review. See Brecht v. Abrahamson, 507 U.S. 619, 123 L. Ed. 2d 353, 113 S. Ct. 1710 (1993). An appellate court will set aside a conviction if there is constitutional error that is not "harmless beyond a reasonable doubt." Chapman v. California, 386 U.S. 18, 24, 17 L. Ed. 2d 705, 87 S. Ct. 824 (1967); Peck v. United States, 106 F.3d 450, 454 (2d Cir. 1997). On the other hand, absent a fundamental, "structural" error,
habeas relief is only granted when the error "had substantial and injurious effect or influence in determining the jury's verdict." Brecht, 507 U.S. 619, 637-38, 123 L. Ed. 2d 353, 113 S. Ct. 1710 (1993) (quoting Kotteakos v. United States, 328 U.S. 750, 776, 90 L. Ed. 1557, 66 S. Ct. 1239 (1946)). If "grave doubt" exists as to the harmlessness of the error, habeas relief must be granted. Agard v. Portuondo, 117 F.3d 696, 714 (2d Cir. 1997) (quoting O'Neal v. McAninch, 513 U.S. 432, 437, 130 L. Ed. 2d 947, 115 S. Ct. 992 (1995)).
The trial court's purported error was harmless under this standard. Tyson suggests no reason why his now-preferred defense -- that the tape was authentic, but his "admissions" were not truly inculpatory because the complainant was effectively putting words in his mouth -- would have been more successful at trial than the defense he actually presented. While Tyson's conviction proves that the defense raised at trial failed, this does not mean that it could not have worked, nor that the new defense would have been more likely to. I accept the proposition that the trial court's decision had an "effect" on the trial in the sense that the issues presented would have been different had the expert testimony been available; however, to have an " injurious effect," as the habeas standard requires, the erroneous decision must have led to a trial that was not only different, but less favorable to the defendant than the one he otherwise would have had. Brecht, 507 U.S. at 637 (emphasis added). If anything, however, it appears that the trial court's purported error may have improved Tyson's chance for an acquittal at trial: Under his now-preferred defense, Tyson would have to explain, for example, why his taped statement to the complainant, "I didn't mean to rape you," was not truly probative of his guilt. At first blush, at least, a stout denial that he ever made such a comment would appear to be significantly more credible. For aught that appears, then, the trial court's purported error may have had a beneficial rather than an injurious effect on the trial, from Tyson's point of view.
The absence of any injury to Tyson is made still clearer by the fact that nothing prevented him from raising his now-preferred defense at trial, even without the benefit of the expert. As Tyson himself points out:
It is possible that defense counsel, even in the absence of any . . . suggestion by the expert . . . would have decided to admit at trial that it was petitioner's voice on the tape, but to minimize the impact of the tape by arguing to the jury that the complainant was dominating the conversation, was putting words in [Tyson's] mouth, and petitioner was therefore not admitting that he had committed any of the crimes charged. That argument is apparent from review of the transcript. . . .