Scipio v. State

928 So. 2d 1138 (2006)

Stephen SCIPIO, Petitioner,
v.
STATE of Florida, Respondent.

No. SC04-647.

Supreme Court of Florida.

February 16, 2006. Rehearing Denied April 24, 2006.

James S. Purdy, Public Defender, Noel A. Pelella and Larry B. Henderson, Assistant Public Defenders, Seventh Judicial Circuit, Daytona Beach, FL, for Petitioner.

Charles J. Crist, Jr., Attorney General, Tallahassee, FL, and Kellie A. Nielan, Assistant Attorney General, Daytona Beach, FL, for Respondent.

ANSTEAD, J.

Stephen J. Scipio seeks review of the Fifth District Court of Appeal's decision in Scipio v. State, 867 So. 2d 427 (Fla. 5th DCA 2004), on the basis of conflict with numerous prior decisions of other district courts and this Court, including the decisions in Pender v. State, 700 So. 2d 664 (Fla.1997); Williams v. State, 863 So. 2d 1189 (Fla.2003); and State v. Schopp, 653 So. 2d 1016 (Fla.1995). This Court has jurisdiction pursuant to article V, section 3(b)(3) of the Florida Constitution. Because we find the State has not demonstrated the harmlessness of a discovery violation, we quash the decision of the Fifth District and remand with instructions that a new trial be conducted. We also clarify our holding in Schopp.

PROCEEDINGS TO DATE

Scipio was convicted of first-degree murder and sentenced to life in prison without the possibility of parole.^[1] On appeal, he claimed a substantial State discovery violation tainted his conviction. Scipio v. State, 867 So. 2d 427, 429-30 (Fla. 5th DCA 2004). The Fifth District described the proceedings and the facts presented at trial in its opinion:

At Scipio's trial, four witnesses identified him as Smith's assailant, but there was no physical evidence connecting Scipio to the crime. Smith was shot four times, and two projectiles were recovered from his body. They were possibly fired from the same pistol, most likely a revolver, but the gun was never recovered.

The testimony adduced at trial was that immediately prior to his death, Smith was shooting pool with Miles Garey in the pool room of the Southside Inn, a Daytona Beach nightclub, which is located in a high crime area. Garey testified that he and Smith had been at the Southside Inn only a short while, and that he was bending over shooting pool, when he heard Smith say "Oh, s—t, he's got a gun." Garey felt a bullet whiz by his ear and he ran outside the bar. Garey could not identify the shooter.

When police arrived, Smith was lying in the parking lot and bystanders were trying to help him. Smith was pronounced dead at the scene. Although police questioned potential witnesses at that time, no one claimed to have seen what happened. Scipio was not questioned by police because he was not in the area. However, three people (Harper, White and Gaines) subsequently gave police information about the crime, and testified at Scipio's trial that they saw him shoot Smith. A fourth witness testified that Scipio admitted shooting Smith to him.

The state's theory of the case was that Smith's murder was in retaliation for his testimony against an individual named Robert Allen. Smith had been the victim of a robbery perpetrated by Allen about a year earlier. Allen had been convicted of the robbery and sentenced to fifteen years imprisonment, based in part on Smith's testimony.

....

Scipio's defense was that another person had committed the crime. The defense intended to rely, in part, on the testimony of Robert Burch, an investigator for the Medical Examiner's Office. During Burch's deposition, defense showed him a photograph of the crime scene outside the Inn, in the parking lot. Based upon this photograph, Burch said there had been a semi-automatic pistol under Smith's body, and that he had turned it over to law enforcement. Defense believed that this evidence could have created doubt in the mind of the jury as to whether Scipio was the assailant.

Defense counsel met with Burch on the morning of the trial and asked him to review his deposition testimony. Burch did so and said he agreed with it. However, Burch was subsequently approached by the prosecutor and asked about the weapon. Burch told the prosecutor he thought there was a photograph of a weapon.

The prosecutor brought Burch to the State Attorney's Office to review the photographs. Defense asked to accompany them but the prosecutor refused. The prosecutor produced the crime scene photos. When Burch reviewed them, he decided the FDLE photo showed the object he had believed to be a weapon, was in fact a pager. Burch realized that he had been mistaken in his deposition testimony, but this information was not given to defense prior to the commencement of the trial.

When Burch testified at trial that he had not seen a gun at the scene, but rather he had seen only a pager, the defense was completely surprised. Defense counsel made a motion for a mistrial based on Burch's changed testimony, but it was denied. The defense was permitted to impeach Burch, but Burch stuck to his story, explaining that he had made a mistake.

Scipio, 867 So.2d at 428-30. As noted, Scipio was convicted and sentenced to life without the possibility of parole.

On appeal, the Fifth District Court of Appeal rejected all issues raised except for Scipio's claim that a harmful discovery violation had occurred during trial by reason of the State's failure to disclose Burch's changed testimony to the defense. The district court held that a "failure to disclose a significant change in a witness's testimony is as much a discovery violation as a complete failure to disclose a witness." Id. at 430 (citing State v. Evans, 770 So. 2d 1174, 1182 (Fla.2000)). The court concluded that the State had "actively procured the defense witness's recantation" while "knowing that defense counsel intended to rely on [the witness's earlier] deposition." Id. The court declared that this "type of `dirty pool'" is "exactly the type of conduct the discovery rules were designed to prevent." Id. The Fifth District nevertheless held that the discovery violation was not sufficiently harmful to require reversal. Id. at 431. Scipio now seeks review, claiming that the Fifth District misapplied the harmless error standard enunciated by this Court in Schopp.

DISCOVERY VIOLATION

The State first asserts that the Fifth District erred in finding any misconduct by the State, let alone a discovery violation, thereby making any harmless error assessment moot. However, we conclude that the Fifth District was correct in its determination that the State's discovery violation was a flagrant case of "dirty pool" that is "exactly the type of conduct the discovery rules were designed to prevent." Scipio, 867 So.2d at 430.

Not surprisingly, Scipio agrees with the Fifth District's assessment. Further, he asserts that the prosecutor's actions constituted a violation of the spirit and purpose of the discovery rules, as well as a violation of the prosecutor's ethical and professional responsibilities. He claims that the prosecutor's action constituted the "trial by ambush" that this Court has consistently condemned.

The Fifth District relied upon case law concerning discovery issues from this Court and the district courts, as well as an analysis of Florida's criminal discovery rules, to demonstrate the prosecution's discovery violation for failing to disclose the substantial change in testimony by the medical examiner's investigator. The criminal rules that codify the prosecutor's obligation to provide discovery include Florida Rule of Criminal Procedure 3.220(b) (prosecutor's discovery obligation) and 3.220(j) (continuing duty to disclose). These rules of criminal procedure provide:

(b) Prosecutor's Discovery Obligation.

(1) Within 15 days after service of the Notice of Discovery, the prosecutor shall serve a written Discovery Exhibit which shall disclose to the Defendant and permit the defendant to inspect, copy, test, and photograph the following information and material within the state's possession and control:

. . . .

(B) The statement of any person whose name is furnished in compliance with the preceding subdivision. The term "statement" as used herein includes a written statement made by the person and signed or otherwise adopted or approved by the person and also includes any statement of any kind or manner made by the person and written or recorded or summarized in any writing or recording. . . .

. . . .

(j) Continuing Duty to Disclose. If, subsequent to compliance with the rules, a party discovers additional witnesses or material that the party would have been under a duty to disclose or produce at the time of the previous compliance, the party shall promptly disclose or produce the witnesses or material in the same manner as required under these rules for initial discovery.

Fla. R.Crim. P. 3.220(b)(1)(B), (j). The "preceding subdivision" referred to in rule 3.220(b)(1)(B) requires the State to provide the defense "a list of names and addresses of all persons known to the prosecutor to have information that may be relevant to any offense charged or any defense thereto." Fla. R.Crim. P. 3.220(b)(1)(A).

Under these rules, it is apparent that the medical examiner's investigator, who examined the crime scene and the body of the victim, was a person whom the State had an obligation to disclose as having information relevant to the case. That obligation is not in dispute here, and is comparable to the State's obligation to disclose the identity of an investigating police officer who has gathered evidence of the crime. Further, since the medical examiner's investigator actually gave a statement by deposition that he possessed relevant information of the crime scene, the State also had an obligation to disclose any material change in that statement. See Fla. R.Crim. P. 3.220(j). The form in which the statement was provided, i.e., a recorded deposition versus a written statement, does not affect that obligation. Of course, we now know by hindsight that the investigator first gave a statement helpful to the defense, and then, at trial, gave testimony contrary to the initial statement and harmful to the defense while helpful to the State.

In establishing a discovery violation by the State, the Fifth District also relied upon this Court's analysis of a similar discovery issue in State v. Evans, 770 So. 2d 1174 (Fla.2000). See Scipio, 867 So.2d at 430. The witness in Evans provided a statement to police and a pretrial deposition, and in both statements asserted that she "didn't see anything" pertaining to the crime. Evans, 770 So.2d at 1176. The same witness subsequently testified as a witness for the State at trial, asserting, contrary to her police deposition and statement, that she actually saw the defendant shoot the victim. Id. On cross-examination the witness disclosed that she had provided her changed statement to the State a month before trial. Id. However, the State did not notify the defendant of the change and that it would be relying on this important new evidence at trial. Id. This Court found that the State had committed a discovery violation by failing to disclose that the witness had changed her original statement "to such an extent that the witness is transformed from a witness who `didn't see anything' into an eyewitness who observed the material aspects of the crime charged" and that this was "tantamount to failing to name a witness at all." Id. at 1182.^[2] This Court concluded:

Finally, we determine that the State's failure in this case to disclose both the substance of the oral statement allegedly made by Evans and the transformation of Green into an eyewitness was harmful because we cannot say beyond a reasonable doubt that the defense was not procedurally prejudiced by the violation. See State v. Schopp, 653 So. 2d 1016, 1021 (Fla.1995) ("[O]nly if the appellate court can say beyond a reasonable doubt that the defense was not procedurally prejudiced by the discovery violation can the error be considered harmless.").

Id. at 1183 (modification in original).

Like the Fifth District, other district courts have also found discovery violations when a witness's subsequent statement to the State materially alters an existing written or recorded statement provided by the State to the defense. In Jones v. State, 514 So. 2d 432 (Fla. 4th DCA 1987), for example, the Fourth District Court of Appeal concluded that when a witness informs the State of his intention to materially alter information provided in a sworn statement, rule 3.220(f) (the predecessor to rule 3.220(j)) imposes a continuing duty upon the State to disclose such information to the defense. Id. at 435. Similarly, the Fourth District has held that the State committed a discovery violation by furnishing a defendant with misleading and inaccurate information regarding the testing of a victim's clothing, when in fact the laboratory report provided by the State actually reflected the testing of the defendant's clothing. McArthur v. State, 671 So. 2d 867, 870 (Fla. 4th DCA 1996). The court concluded that furnishing such discovery was "tantamount to providing no discovery at all." Id.; see also Neimeyer v. State, 378 So. 2d 818, 821 (Fla. 2d DCA 1979) (State had duty to disclose change in testimony by medical examiner who had given prior statement and deposition); Alfaro v. State, 471 So. 2d 1345, 1345-46 (Fla. 4th DCA 1985) (State violated discovery rules by not informing the defense that the medical examiner performed last-minute reconstruction analysis).

PURPOSE OF DISCOVERY

Importantly, this Court has consistently held that "Florida's criminal discovery rules are designed to prevent surprise by either the prosecution or the defense. Their purpose is to facilitate a truthful fact-finding process." Kilpatrick v. State, 376 So. 2d 386, 388 (Fla.1979). In Kilpatrick we explained:

Florida's criminal discovery rules are designed to prevent surprise by either the prosecution or the defense. Their purpose is to facilitate a truthful fact-finding process. Discovery under Florida Rule of Criminal Procedure 3.220 is commenced by service of a demand for discovery by the defense on the State. The rule imposes a continuing mandatory duty on the prosecution to disclose certain specifics, including the names of prospective witnesses. Once having invoked this procedure, the defense must also affirmatively respond by disclosing certain information to the prosecution including the names of prospective witnesses. Both sides are entitled to rely on full and fair compliance with the rule in preparing their cases for trial.

Id. This Court has held that the chief purpose of our discovery rules is to assist the truth-finding function of our justice system and to avoid trial by surprise or ambush. See, e.g., Evans, 770 So.2d at 1182.

Because full and fair discovery is essential to these important goals, we have repeatedly emphasized not only compliance with the technical provisions of the discovery rules, but also adherence to the purpose and spirit of those rules in both the criminal and civil context. See Binger v. King Pest Control, 401 So. 2d 1310, 1314 (Fla.1981). This Court has explained that the rules of discovery are intended to avoid surprise and "trial by `ambush.'" Binger, 401 So.2d at 1314. In Binger, we emphasized that the "search for truth and justice can be accomplished only when all relevant facts are before the judicial tribunal. Those relevant facts should be the determining factor rather than gamesmanship, surprise, or superior trial tactics." Id. at 1313 (quoting Dodson v. Persell, 390 So. 2d 704, 707 (Fla.1980)). Moreover, in Binger, this Court declared:

The argument that the possibility of being ambushed by an unlisted impeachment witness encourages truthful testimony reflects an outdated methodology for ascertaining the truth at trials. It lacks the solid foundation on which courts should alone build exceptions to general discovery principles. Moreover, the logic of the proposition is fairly disputable. The listing of contradictory witnesses whose impeachment testimony can be discovered by interrogatory or deposition provides equal assurance that trial testimony will be accurate and truthful.

Id. at 1314. Following our lead, the district courts of appeal have also interpreted Florida Rule of Civil Procedure 1.360(b), which requires disclosure of the reports of expert witnesses, to also require the disclosure of a substantial reversal of such an opinion. For example, even when no explicit duty to supplement discovery existed under the rules of civil procedure, this Court and the district courts have repeatedly emphasized that "litigation should no longer proceed as a game of `blind man's bluff.'" Jones v. Seaboard Coast Line R.R. Co., 297 So. 2d 861, 863 (Fla. 2d DCA 1974). As the Second District noted in Jones:

[I]t's now rudimentary that the primary purpose of pretrial discovery is twofold: (1) to "discover" evidence relevant and pertinent to the triable issues pending before the court, and (2) if in written form to serve, of itself, as evidence at trial if otherwise admissible.

Jones, 297 So.2d at 863.

Similarly, as this Court concluded in Evans, the Fourth District has pointed out, in finding error in a litigant's failure to disclose a change in an expert's opinion, "[a] party can hardly prepare for an opinion that it doesn't know about, much less one that is a complete reversal of the opinion it has been provided." Office Depot, Inc. v. Miller, 584 So. 2d 587, 590 (Fla. 4th DCA 1991). The Fourth District concluded:

While judicial economy may have been served by adhering to the decision made at trial to deal with the problem, justice is ultimately better served even if a new trial must be held to insure fairness to the litigants. The trial court's action here sends out a strong message to those who do not adhere to the code of fair play advanced by Binger. Serious violations of the pretrial disclosure rules may result in the exclusion of important evidence, and may, in extreme circumstances, lead to the grant of a new trial.

Id. at 591. Moreover, the Fourth District in Office Depot emphasized, relying on Binger, "[p]arties who fail to make such disclosure do so at their peril, depending on the circumstances of the particular case." Id.

Of course, the policy of avoiding trial by ambush or surprise has even greater application in the criminal context, where the stakes are much higher and the obligation of the State to see that justice is done is much greater than that of the private litigants in a civil dispute.^[3] Indeed, the United States Supreme Court long ago made clear:

The [prosecutor] is the representative not of an ordinary party to a controversy, but of a sovereignty whose obligation to govern impartially is as compelling as its obligation to govern at all; and whose interest, therefore . . . is not that it shall win a case, but that justice shall be done. . . . He may prosecute with earnestness and vigor — indeed, he should do so. But, while he may strike hard blows, he is not at liberty to strike foul ones.

Berger v. United States, 295 U.S. 78, 88, 55 S. Ct. 629, 79 L. Ed. 1314 (1935). These words seem especially appropriate to the circumstances before us today.

Under our discovery rules and our case law, we conclude that the Fifth District was correct in its determination that the State committed a discovery violation when it failed to disclose to Scipio a material change in the State investigator's deposition statement. Further, the Fifth District's reliance on our decision in Evans is consistent with our case law stressing disapproval of trial by ambush.

The State's calculated failure to inform the defense of the important and dramatic change in testimony of its medical examiner's investigator not only violated the prosecutor's duty not to strike "foul" blows, but undermined the very purpose of the discovery rules as set out by this Court in Kilpatrick and Evans, since the State was fully aware that the defense intended to rely heavily on the testimony of the State's investigator and would be completely surprised by the witness's changed testimony at trial.

HARMLESS ERROR

Although we agree with the Fifth District's assessment of the discovery issue, we find the court's decision on harmless error in conflict with this Court's decision in Schopp. However, it is also apparent that there is some conflicting language in our Schopp opinion that was responsible for the district court's holding and that requires clarification.

Initially, it is important to note that prior to this Court's decision in Schopp, a trial court's failure to conduct a Richardson^[4] inquiry as to discovery violations was considered to constitute such substantial harm as to mandate automatic reversal without any consideration of harmless error. See Smith v. State, 500 So. 2d 125, 126 (Fla.1986). However, while we decided in Schopp that it was possible that a Richardson violation could be harmless, we retained the strict procedural prejudice standard set out in Smith as the standard to be utilized in determining prejudice or harm. Schopp, 653 So.2d at 1021.^[5]

In Schopp this Court went to great lengths to explain and repeatedly reaffirm our adherence to the procedural prejudice standard:

We reasoned [in Smith] that an appellate determination as to whether a Richardson violation is harmless is impossible in light of the fact that "[t]he purpose of a Richardson inquiry is to ferret out procedural, rather than substantive, prejudice." 500 So.2d at 126 (quoting Wilcox [v. State], 367 So.2d [1020] at 1023 [(Fla.1979)]). We explained:

The certified question in [Smith] misapprehends the very purpose of a Richardson hearing, which is precisely to determine if a violation is, in fact, harmless. One cannot determine whether the state's transgression of the discovery rules has prejudiced the defendant (or has been harmless) without giving the defendant the opportunity to speak to the question. We repeat what the court made clear in Wilcox. A reviewing court cannot determine whether the error is harmless without giving the defendant the opportunity to show prejudice or harm. 367 So.2d at 1023. In Wilcox,. . . this Court explained that the question of "prejudice" in a discovery context is not dependent upon the potential impact of the undisclosed evidence on the fact finder but rather upon its impact on the defendant's ability to prepare for trial [.]

500 So.2d at 126.

This distinction between substantive and procedural prejudice in the context of discovery violations continues to be a valid one. However, this case demonstrates that there are cases in which a reviewing court can determine that a discovery violation is harmless beyond a reasonable doubt, absent an adequate Richardson inquiry at the trial level. Here, we can say beyond a reasonable doubt that neither the discovery violation nor the trial court's failure to inquire into whether corrective sanctions were warranted materially hindered the defendant's trial preparation or strategy.

Schopp, 653 So.2d at 1019-20 (emphasis supplied) (footnote omitted). Indeed, we repeatedly emphasized throughout our Schopp opinion that when reviewing a claim of error based upon a Richardson violation, the reviewing court's focus should be on procedural and not substantive prejudice:

In determining whether a Richardson violation is harmless, the appellate court must consider whether there is a reasonable possibility that the discovery violation procedurally prejudiced the defense. As used in this context, the defense is procedurally prejudiced if there is a reasonable possibility that the defendant's trial preparation or strategy would have been materially different had the violation not occurred. Trial preparation or strategy should be considered materially different if it reasonably could have benefited the defendant. In making this determination every conceivable course of action must be considered. If the reviewing court finds that there is a reasonable possibility that the discovery violation prejudiced the defense or if the record is insufficient to determine that the defense was not materially affected, the error must be considered harmful. In other words, only if the appellate court can say beyond a reasonable doubt that the defense was not procedurally prejudiced by the discovery violation can the error be considered harmless.

. . . .

We recognize that in the vast majority of cases it will be readily apparent that the record is insufficient to support a finding of harmless error. We also recognize that where the defendant's trial preparation or strategy reasonably could have been affected by the discovery violation it will be difficult to determine whether the verdict could have differed had the violation not occurred or had the trial court acted to avert the prejudice. However, the mere fact that there is a high probability that a given error will be found harmful does not justify categorizing the error as per se reversible. [State v.] DiGuilio, 491 So.2d [1129] at 1137 [(Fla.1986)]. Thus, we recede from Smith, Cumbie [v. State, 345 So. 2d 1061 (Fla.1977)], Wilcox, and their progeny to the extent that they require per se reversal, and we hold that the harmless error analysis set forth above should be applied where a trial court fails to conduct an adequate Richardson inquiry.

Schopp, 653 So.2d at 1020-21 (emphasis supplied).

Even the dissent in Schopp, authored by Justice Harding, recognized the Court's continued adherence to the procedural prejudice standard:

While the majority opinion correctly notes that the prejudice at issue in the discovery context is procedural rather than substantive [653 So.2d at 1019-20], it focuses solely upon a single aspect of the prejudice, whether the discovery violation prevented the defendant from properly preparing for trial. As this Court explained in Wilcox, there are two separate but interrelated aspects of prejudice in the discovery context: 1) whether the discovery violation prevented the defendant from properly preparing for trial; and 2) what is the proper sanction to invoke for the discovery violation. 367 So.2d at 1023. Granted, if a court determines that a defendant was not prejudiced in trial preparation or strategy by a discovery violation, it may also determine that no sanction is warranted. Smith, 500 So.2d at 126. But without an adequate Richardson inquiry the court cannot determine if a discovery violation was willful and warrants a sanction on that basis.

Schopp, 653 So.2d at 1023 (Harding, J., dissenting) (footnote omitted).

Hence, our decision in Schopp formulated a strict procedural standard for proving harmless error in place of an even stricter per se rule of reversal. Id. at 1020. It placed the burden on the State and emphasized that a finding of harmless error should be "the exception rather than the rule." Id. The decision further noted that "the vast majority of cases" will not have a record sufficient to support a finding of harmless error and that there is a "high probability" that any given error will be found harmful. Id. at 1021.

Unfortunately, however, included among all of our statements in Schopp affirming the rule of procedural prejudice, there is one statement that appears inconsistent with that rule:

This analysis recognizes the procedural prejudice inherent in discovery violations. It also takes into account the fact that errors that reasonably could affect trial preparation or strategy are "prejudicial," and therefore harmful for appellate purposes, only when a change in trial tactics reasonably could have benefited the defendant by resulting in a favorable verdict.

Id. (emphasis supplied). In this case, the Fifth District focused on this brief and inconsistent passage from Schopp in concluding the State's discovery violation was harmless in this case. See Scipio, 867 So.2d at 431. We now conclude, however, that this isolated statement is inconsistent with the extensive analysis in Schopp preceding the statement, adopting a procedural, not substantive prejudice analysis. See Schopp, 653 So.2d at 1020. We now take this opportunity to clarify our opinion in Schopp and recede from this inconsistent language. One cannot consistently say, as we did clearly and repeatedly in Schopp, that the test for prejudice does not concern the impact of the undisclosed evidence on the fact-finder, and yet apply a test that does consider that impact.

In contrast to the Fifth District's opinion below, this Court in Evans correctly applied the Schopp standard of procedural prejudice when concluding:

Finally, we determine that the State's failure in this case to disclose both the substance of the oral statement allegedly made by Evans and the transformation of Green into an eyewitness was harmful because we cannot say beyond a reasonable doubt that the defense was not procedurally prejudiced by the violation. See State v. Schopp, 653 So. 2d 1016, 1021 (Fla.1995) ("[O]nly if the appellate court can say beyond a reasonable doubt that the defense was not procedurally prejudiced by the discovery violation can the error be considered harmless.").

Evans, 770 So.2d at 1183. Further, in Pender v. State, 700 So. 2d 664 (Fla.1997), this Court also noted that the district court's analysis of harmless error in that case was in direct contravention of our holding in Schopp because it had considered the impact of a discovery violation on the fact-finder instead of the impact on the defendant's trial preparation. Id. at 666. Similarly, in Cox v. State, 819 So. 2d 705 (Fla.2002), this Court began its harmless error analysis by stating, "[W]here the State commits a discovery violation, the standard for deeming the violation harmless is extraordinarily high." Id. at 712. Then, in analyzing whether the discovery violation was harmless, this Court reiterated the procedural prejudice analysis of Schopp: "[O]nly if the appellate court can say beyond a reasonable doubt that the defense was not procedurally prejudiced by the discovery violation can the error be considered harmless." Id. (quoting Pomeranz v. State, 703 So. 2d 465, 468 (Fla. 1997)); see also Irish v. State, 889 So. 2d 979, 981 (Fla. 4th DCA 2004) (holding that a prior statement made by the defendant to the police that was withheld from the defense attorney undermined the defense's theory and therefore was harmful); Suda v. State, 838 So. 2d 665, 666 (Fla. 1st DCA 2003) (concluding that the State's nondisclosure of its intent to call a witness prior to trial was harmful because it may have affected the defense's trial strategy); Portner v. State, 802 So. 2d 442, 446 (Fla. 4th DCA 2001) (holding that the prosecutor's unannounced use of a prior deposition by the defendant was harmful).

On the other hand, and in contrast to the procedural prejudice analysis discussed above, this Court in Pomeranz v. State found no harmful error because we concluded that the potential change in trial tactics would not have "resulted in a more favorable outcome." 703 So.2d at 469.

It is apparent from the Fifth District's decision and this Court's opinions, including the opinion in Pomeranz, that there is a need to clarify our opinion in Schopp. Importantly, in Schopp, we approved the following language from Smith:

[T]he question of "prejudice" in a discovery context is not dependent upon the potential impact of the undisclosed evidence on the fact finder but rather upon its impact on the defendant's ability to prepare for trial[.]

Schopp, 653 So.2d at 1019 (quoting Smith, 500 So.2d at 126) (emphasis added). As this Court explained at great length in its opinions in Smith and Schopp, although related, procedural prejudice and substantive prejudice are not the same. See Schopp, 653 So.2d at 1019-1021; Smith, 500 So.2d at 126. An analysis of procedural prejudice does not ask how the undisclosed piece of evidence affected the case as it was actually presented to the jury. Rather, it considers how the defense might have responded had it known about the undisclosed piece of evidence and contemplates the possibility that the defense could have acted to counter the harmful effects of the discovery violation. Evans, 721 So.2d at 1210. By contrast, an analysis of substantive prejudice would ask whether it was possible that the error affected the jury's verdict. Schopp, 653 So.2d at 1021. While this standard itself is a high one for the State to overcome, it is of an entirely different quality than a procedural prejudice analysis. Id. at 1019.

We now clarify that Schopp's harmless error standard does not focus on whether the discovery violation would have made a difference in the verdict. Such an analysis would make the standard for procedural prejudice identical to substantive prejudice. Instead, we reaffirm our statements in Schopp that the inquiry is whether there is a reasonable possibility that the discovery violation "materially hindered the defendant's trial preparation or strategy." Schopp, 653 So.2d at 1020. Under Schopp, only if the appellate court can determine beyond a reasonable doubt that the defense was not procedurally prejudiced by the discovery violation can the error be considered harmless. Id. at 1021.

THIS CASE

This case stands in sharp contrast to the facts of Schopp, both as to the nature of the discovery violation and its impact on the defense. Here, the prosecution committed a discovery violation by its purposeful failure to inform the defense that the medical examiner's investigator had recanted his prior statement about the gun under the body. In the final portion of his closing statements, the prosecutor essentially ridiculed the attempt at a defense presented by defense counsel based on the investigator's now-changed testimony. The prosecutor asserted that the defense theory was based on "imaginary things" and called it "ridiculous," reminding the jury that the investigator testified there was no gun under the body. The jury found Scipio guilty of first-degree murder, and the court sentenced him to life in prison without the possibility of parole.

Of course, because of the State's misconduct, we are essentially left to speculate as to what strategy the defense would have adopted had it been made aware of the State investigator's recantation of testimony. If the defense attorney had known about the recantation, the defense could have sought a continuance since the medical investigator's testimony was critical to the defense's case, or the defense could have chosen not to call the witness and to avoid the surprise and humiliation that took place. It is even possible that the defense would have chosen to call and impeach the investigator and to present the jury with his prior inconsistent testimony, although that would appear to have been highly unlikely. What is certain is that the State's misconduct directly prevented the defense from considering these options and instead forced the defense into a humiliating scenario of having its witness turn against the defense. It is precisely because we are left to speculate that we conclude here that the State has not demonstrated harmless error under the procedural prejudice analysis of Schopp.

What we do know is that the State not only committed the discovery violation, but also completely turned the tables on the defense by this surprise "gotcha" tactic. This is precisely the kind of procedural harm of which we were fearful in Schopp, placing the defense in the impossible position of simply being unable to prepare for such a devastating event. Under these circumstances, we simply cannot conclude that the State carried its heavy burden under Schopp to demonstrate the procedural harmlessness of its misconduct. As the Fourth District concluded in Office Depot, Inc.:

While judicial economy may have been served by adhering to the decision made at trial to deal with the problem, justice is ultimately better served even if a new trial must be held to insure fairness to the litigants. The trial court's action here sends out a strong message to those who do not adhere to the code of fair play advanced by Binger. Serious violations of the pretrial disclosure rules may result in the exclusion of important evidence, and may, in extreme circumstances, lead to the grant of a new trial. 584 So.2d at 591. We reach a similar conclusion today.

Indeed, as emphasized by the Fifth District, the circumstances presented here constitute a classic case of "trial by ambush." See Binger, 401 So.2d at 1314. When the State's investigator provided the defense with testimony that he had not only observed a gun under the victim's body, but had also taken possession of the gun and turned it over to the police, he provided the defense with a substantial basis to dispute the State's theory of the case, i.e., that the defendant had shot an unarmed victim without provocation. As the Fifth District emphasized, when the State deliberately prevented the defense from learning of the changed testimony, the rug was completely pulled out from underneath the defense's theory of the case. See Scipio, 867 So.2d at 430. Not only was the only available defense evidence removed, in the process the defense was made to look utterly foolish, as later pointed out and emphasized by the State in its closing argument to the jury. Hence, not only did the State improperly ambush or pull a "gotcha" on the defense, it then used its improper ambush as a hammer to humiliate the defense before the jury.^[6]

CONCLUSION

Accordingly, we approve the district court's holding that the State committed a discovery violation, clarify our holding in Schopp, and disapprove of the district court's opinion to the extent that it conflicts with our analysis of harmless error under Schopp. We quash and remand for further proceedings consistent herewith.

It is so ordered.

PARIENTE, C.J., and LEWIS and QUINCE, JJ., concur.

CANTERO, J., dissents with an opinion, in which WELLS and BELL, JJ., concur.

CANTERO, J., dissenting.

I share the majority's indignation at the prosecutor's conduct in this case. No less than the majority, I firmly believe in promoting the highest standards of professional conduct. See, e.g., Boca Burger, Inc. v. Forum, 912 So. 2d 561, 572 (Fla. 2005) (recognizing "the basic principle that a lawyer's duty to his calling and to the administration of justice far outweighs — and must outweigh — even his obligation to his client, and, surely what we suspect really motivates many such inappropriate actions, his interest in his personal aggrandizement") (quoting Rapid Credit Corp. v. Sunset Park Centre, Ltd., 566 So. 2d 810, 812 n. 1 (Fla. 3d DCA 1990) (Schwartz, C.J., specially concurring)). I agree that the prosecutor's conduct here falls short of those standards. But when it comes to deciding the discrete issues before us, I must part company. Those issues are: (A) does the district court's decision expressly and directly conflict with one of ours? (B) did the prosecution violate a specific discovery rule? and (C) if so, was it harmless error?

I disagree with the majority's resolution of each of these issues. The district court's decision does not conflict with any of our cases. To the contrary, the district court faithfully applied them. We cannot base our jurisdiction on allegedly "conflicting language in our [own] opinion." Majority op. at 1146. Even if we had jurisdiction, however, I would dissent on the merits. In failing to inform Scipio of a defense witness's last-minute change in testimony, the State did not violate any specific rule of criminal procedure, and violating the nebulous "spirit" of the rules is not enough. Finally, even assuming a violation, it did not procedurally prejudice the defense. I will explain each of my disagreements in turn.

A. Jurisdiction

The majority's first obstacle is jurisdiction. There is none. The majority concludes that the district court applied the wrong standard for procedural prejudice and thereby created conflict with two of our decisions — State v. Schopp, 653 So. 2d 1016 (Fla.1995), and Pender v. State, 700 So. 2d 664 (Fla.1997). To the contrary, the district court followed them as closely as it possibly could have.

The majority quotes our decision in Schopp at great length, noting its emphasis on procedural, as opposed to substantive, prejudice. See majority op. at 1146-48. As the majority notes, Schopp held that in determining whether the State's discovery violation was harmless, the issue is whether the undisclosed information would have affected the defendant's trial strategy (procedural prejudice). Specifically, we stated, "[t]he defense is procedurally prejudiced if there is a reasonable possibility that the defendant's trial preparation or strategy would have been materially different had the violation not occurred. Trial preparation or strategy should be considered materially different if it reasonably could have benefited the defendant." Schopp, 653 So.2d at 1020.

That is precisely the standard the district court applied. The court wrote:

A violation is harmful if the defense is procedurally prejudiced. Pender [, 700 So.2d at 666]; Schopp [, 653 So.2d at 1016]. The defense is procedurally prejudiced if there is a reasonable possibility that the defendant's trial preparation or strategy would have been materially different had the violation not occurred. Pender; Schopp. Trial preparation or strategy will be materially different if it reasonably could have benefitted the defendant. Id. If the reviewing court finds that there is a reasonable possibility that the discovery violation prejudiced the defense, the error must be considered harmful. Id. Put another way, if the appellate court can say, beyond a reasonable doubt, that [the] defense was not procedurally prejudiced by the violation, then the error is harmless. Id.

Scipio v. State, 867 So. 2d 427, 430 (Fla. 5th DCA 2004) (emphasis omitted). As is evident from the five citations to both Schopp and Pender, the district court applied the same standard we announced in those cases.

In applying this standard, the district court concluded that, had the defense known about the witness's change in testimony,

its trial strategy could only have been one of two things. One, it would not have called the witness. Clearly, the outcome of this case would not have been affected given the multiple eyewitness testimonies concerning Scipio's shooting Smith inside the Inn. Or second, the defense would have called the witness and sought to impeach him, to present the jury with his prior testimony. That is, in effect, what happened in this case, which if anything, was more favorable to Scipio than not having called the witness.

Id. at 431. As this quote demonstrates, the district court, applying Schopp and Pender, analyzed whether the State's discovery error materially hindered the defendant's trial preparation or strategy. Its conclusion was "no." Id. at 430. However, it cautioned that "[i]n a different case where the evidence is less overwhelming, or a recanting witness more material," the same sort of discovery error might mandate reversal for a new trial. Id. at 431.

The majority takes issue with the district court's consideration of "the outcome of this case," claiming it expressly and directly conflicts with Schopp. As the majority acknowledges, however, see majority op. at 1146, language in Schopp directly supports the district court's reference to the verdict. The majority minimizes its importance, calling the passage "brief and inconsistent." Id. at 1148. But it is not so brief, and not necessarily inconsistent. The passage from Schopp reads in full:

This analysis recognizes the procedural prejudice inherent in discovery violations. It also takes into account the fact that errors that reasonably could affect trial preparation or strategy are "prejudicial," and therefore harmful for appellate purposes, only when a change in trial tactics reasonably could have benefited the defendant by resulting in a favorable verdict.

We recognize that in the vast majority of cases it will be readily apparent that the record is insufficient to support a finding of harmless error. We also recognize that where the defendant's trial preparation or strategy reasonably could have been affected by the discovery violation it will be difficult to determine whether the verdict could have differed had the violation not occurred or had the trial court acted to avert the prejudice. However, the mere fact that there is a high probability that a given error will be found harmful does not justify categorizing the error as per se reversible.

653 So.2d at 1021 (emphases added). As if these two paragraphs were not clear enough, we then explained in applying the law to the facts that "because Schopp effectively `won' his case [by persuading the jury to convict on a lesser included offense], there is no reasonable possibility that a change in trial tactics could have benefited him." Id. at 1022.

These repeated references to the verdict cannot be dismissed as accidental. Schopp went to great pains to clarify that the procedural prejudice standard requires a reasonable possibility both that (a) the State's discovery error materially hindered the defendant's trial preparation or strategy, and that (b) preventing the error might have benefited the defendant by enabling a more favorable outcome. We derived the second prong from the harmless error analysis in State v. DiGuilio, 491 So. 2d 1129, 1135 (Fla.1986), which requires the State "to prove beyond a reasonable doubt that the error complained of did not contribute to the verdict or, alternatively stated, that there is no reasonable possibility that the error contributed to the conviction." See Schopp, 653 So.2d at 1020 (citing DiGuilio and requiring "[a]pplication of a harmless error analysis in this context").

The district court in this case applied Schopp's harmless error standard and concluded that neither prong was satisfied because the State's discovery error did not "materially hinder[ ] Scipio's trial preparation or strategy," and "the outcome of this case would not have been affected." Scipio, 867 So.2d at 430-31.

Notwithstanding the fact that Schopp offers clear textual support for the district court's analysis, the majority contends that the two decisions conflict because of another sentence in Schopp, which the majority interprets as forbidding any consideration of the outcome. The sentence is actually a quotation from Smith v. State, 500 So. 2d 125 (Fla.1986), the case from which Schopp receded. Smith explained that "the question of `prejudice' in a discovery context is not dependent upon the potential impact of the undisclosed evidence on the fact finder but rather upon its impact on the defendant's ability to prepare for trial." Id. at 126. We acknowledged in Schopp that, although Smith no longer reflects the current state of the law, "[t]his distinction between substantive and procedural prejudice in the context of discovery violations continues to be a valid one." 653 So.2d at 1019.

Based on this language, the majority deems Schopp internally "inconsistent" and "conflicting," majority op. at 1146, 1148, and concludes that the district court's decision conflicts with Schopp by favoring the outcome-considering part of it over the outcome-ignoring part. Majority op. at 1144. The strange result of this position is that the majority resolves the "conflict" by simultaneously receding from Schopp and disapproving the district court's decision as conflicting with it.

I cannot join in this reasoning. Schopp is not internally inconsistent. It merely recognizes, as did DiGuilio, which Schopp cites, that the harmless error standard necessarily involves determining the possible effect of the error on the outcome. Although Schopp rejects an exclusive focus on "the potential impact of the undisclosed evidence on the fact finder," it permits inquiry into the outcome as part of the analysis. Thus, in determining prejudice the issue is whether the discovery violation prevented the defense from adjusting its trial preparation or strategy in a way that might have benefited the defendant by affecting the outcome. As Schopp explains, under this approach "every conceivable course of action must be considered" — including, for example, the possibility that the defense may have decided to pursue other evidence, other theories of the case, or other styles of presentation. 653 So.2d at 1020. In contrast, the approach Schopp rejected would focus narrowly on the undisclosed evidence. We did not contradict ourselves by embracing the broader approach.

Even if Schopp were internally inconsistent, as the majority claims, it would not create a basis for jurisdiction. We have jurisdiction to review district court decisions that expressly and directly conflict with ours, not those that faithfully apply the allegedly "conflicting language in our [own] opinion." Majority op. at 1146. We must wait for a district court to reject or misapply Schopp, or at least to interpret it in a way that conflicts with the analysis of another district court. None of that happened here.

The majority also contends that the district court's decision conflicts with Pender, a more recent decision of ours. In Pender, we disapproved a misapplication of the procedural prejudice standard, stating:

[T]he [district] court did not indicate that in conducting a harmless error analysis it considered the impact of the State's discovery violation or the trial court's failure to conduct a Richardson hearing on the defendants' trial preparation. Instead, it appears that the district court, in direct contravention of Schopp, considered the impact of these errors on the fact finder. The opinion... seems to indicate that the district court based its holding on a finding that the defendants did not experience any substantive prejudice.

700 So.2d at 666.

The district court in this case did not commit the same error as did the district court in Pender, which utterly failed to consider how the discovery error affected the defense's trial preparation. To the contrary, the decision below "turn[ed] on whether the discovery error . . . materially hindered Scipio's trial preparation or strategy." Scipio, 867 So.2d at 430 (citing Pender and Schopp). Because the error "had no such effect," the district court found no procedural prejudice. Id. This analysis complied with Pender.

The majority portrays Pender as not only demanding a focus on trial preparation and strategy, but also prohibiting consideration of whether the strategic changes would have affected the verdict. This reading of Pender is flawed. Pender only disapproved of decisions "in direct contravention of Schopp," 700 So.2d at 666, and thus cannot be construed as receding from the language in Schopp that expressly permitted consideration of the outcome. Moreover, only two months after Pender, we decided another case in which, as part of our prejudice analysis, we again considered the outcome. See Pomeranz v. State, 703 So. 2d 465, 469 (Fla.1997) (holding that a discovery error was harmless because "[w]e cannot see how these potential changes in trial tactics would have resulted in a more favorable outcome"). Thus, the proposition that Pender revised the Schopp analysis is untenable.

Because the district court's decision in this case closely followed Schopp and did not commit the error disapproved by Pender, there is no express and direct conflict with either case. I would therefore discharge review as improvidently granted.

B. The State Did Not Violate a Discovery Rule

If the majority were correct that the district court misapplied our decisions in Schopp and Pender, then, as we have in many other cases, we should simply quash the decision and remand for the district court to correctly apply our precedents. See, e.g., Williams v. State, 863 So. 2d 1189, 1190 (Fla.2003) (quashing a decision that misapplied the DiGuilio harmless error standard and remanding for reconsideration); Knowles v. State, 848 So. 2d 1055, 1058-59 (Fla.2003) (same). There is no reason in this case to go further and to analyze the merits, unless it is to recede from Schopp, which only proves that the district court's decision did not conflict with it.

Because the majority does address the merits, I must dissent from its conclusion that the State's conduct — unprofessional though it was — violated a specific discovery rule. The majority concludes that the State violated the "chief purpose," the "spirit," and the "policy" behind the discovery rules by engaging in conduct "tantamount" to what the rules forbid. Majority op. at 1144, 1145. While this may be true, and while I agree that the State's conduct falls well short of the prosecutorial ideal, I cannot identify any specific rule of criminal procedure the State violated. Thus, unlike the majority, I must accept its conduct as lawful and call for an amendment to the rules to expressly prohibit its conduct.

Scipio alleges that the State violated the discovery rules by failing to inform the defense that a forensic investigator from the medical examiner's office changed his testimony. In a pretrial deposition, the investigator had asserted that he found a semiautomatic firearm under the victim's body. The defense intended to call him as a witness to cast doubt on the State's theory of the case. But when, just before trial, the prosecutor showed the investigator a photograph of the victim as he had been discovered at the crime scene, the investigator acknowledged that the object was not a firearm but a pager. Defense counsel was not privy to the conversation. To the defense's great surprise, the investigator testified to that effect at trial. The district court concluded that "the state committed a discovery violation because it failed to advise the defense of the changed testimony prior to the commencement of the trial." Scipio, 867 So.2d at 430.

The majority agrees with the district court that the State violated its continuing duty to disclose information under Florida Rule of Criminal Procedure 3.220(j). That rule provides:

If, subsequent to compliance with the rules [of discovery], a party discovers additional witnesses or material that the party would have been under a duty to disclose or produce at the time of the previous compliance, the party shall promptly disclose or produce the witnesses or material in the same manner as required under these rules for initial discovery.

This rule refers to other discovery rules. It is violated only when the State withholds "additional witnesses or material" that, if discovered earlier, would have been subject to mandatory disclosure.

According to the majority, the State would have been required to disclose the investigator's oral change in testimony under the other discovery rules that require it to give the defendant

(A) a list of the names and addresses of all persons known to the prosecutor to have information that may be relevant to any offense charged or any defense thereto, or to any similar fact evidence to be presented at trial . . .

(B) the statement of any person whose name is furnished in compliance with the preceding subdivision. The term "statement" as used herein includes a written statement made by the person and signed or otherwise adopted or approved by the person and also includes any statement of any kind or manner made by the person and written or recorded or summarized in any writing or recording.

Fla. R.Crim. P. 3.220(b)(1).

Neither of these rules required the State to disclose the investigator's change in testimony. The State had already provided the defense with the investigator's name and address, as required by subdivision (A), listing him as a "Category A" eyewitness. See Fla. R.Crim. P. 3.220(b)(1)(A)(i). It was the State's identification of the investigator as an eyewitness that led the defense to depose him. The State's categorization of the investigator never changed. What changed was the investigator's testimony. But his new testimony was not "written or recorded or summarized in any writing or recording," which is how subdivision (B) defines a "statement." The exclusion of oral statements from subdivision (B) is confirmed by their express inclusion in subdivisions (C) and (D), which mandate the disclosure of "any written or recorded statements and the substance of any oral statements" made by a defendant or codefendant. Fla. R.Crim. P. 3.220(b)(1)(C)-(D).

Thus, the majority simply cannot escape the fact that, under the plain meaning of the rules, the State did not commit a discovery violation.^[7] Maybe the rules should be changed. I would agree to refer the issue to the Florida Bar Criminal Rules Committee, or any other committee the majority deems appropriate. But the rules as written do not prohibit this conduct. See, e.g., Fennell v. State, 544 So. 2d 1017, 1018 (Fla.1989) ("The language of the [criminal procedure] rule is unambiguous and therefore must be construed according to its plain meaning.").

The majority, however, concludes from our decision in State v. Evans, 770 So. 2d 1174 (Fla.2000), that the discovery rules should not be interpreted according to their plain meaning. That case does involve similar circumstances. But they are not identical, and we did not in that case recede from prior opinions holding that failure to disclose oral changes in testimony generally will not violate the rules. In Evans, a woman had informed the police shortly after a shooting that she "didn't see anything." Id. at 1176. About one month before trial, she told a new story: she saw the defendant shoot the victim, and did not see any physical provocation by the victim. The State failed to disclose this new story. We held that, "[i]n essence, the State's nondisclosure of the changes in [the woman's] testimony from her original police statement was tantamount to failing to name a witness at all" and therefore violated rule 3.220(b)(1)(A). Id. at 1182.

This case differs from Evans in important respects. Here, the investigator was recognized as a relevant eyewitness both before and after his change in testimony. Also, unlike the situation in Evans, where the witness retracted a statement that the State gave the defense pursuant to rule 3.220(b)(1)(B), here the State never gave the defense any inaccurate statements. The mistaken testimony occurred in a deposition conducted by the defense itself. Despite these differences, the majority now interprets Evans as standing for a much broader proposition: that prosecutors have "an obligation to disclose any material change" in testimony, majority op. at 1142, or at least any "important and dramatic change." Majority op. at 1145.

The majority's expansion of Evans dramatically changes our caselaw, which, at least since our decision in Bush v. State, 461 So. 2d 936 (Fla.1984), generally permitted the nondisclosure of material oral changes in testimony. In Bush, an investigator had stated in a deposition that a particular witness did not identify any photographs of the defendant. At trial, however, the investigator testified that the witness did identify the defendant's photograph during the photo lineup. We held that "[t]he prosecutor's failure to inform the defense of this change of testimony is not a discovery violation." Id. at 938. We reasoned:

When testimonial discrepancies appear, the witness' trial and deposition testimony can be laid side-by-side for the jury to consider. This would serve to discredit the witness and should be favorable to the defense. Therefore, unlike failure to name a witness, changed testimony does not rise to the level of a discovery violation . . . .

Id. at 938 (emphasis added).

We have never abrogated that rule. Rather, in Evans, we tempered Bush's sweeping language by clarifying that certain extreme changes in a witness's testimony must be disclosed by the State. As we explained,

our statements in Bush do not control in situations where the State provides the defendant with a witness's "statement" — as that term is defined in rule 3.220(b)(1)(B) — and thereafter fails to disclose to the defendant that the witness intends to change that statement to such an extent that the witness is transformed from a witness who "didn't see anything" into an eyewitness who observed the material aspects of the crime charged.

770 So.2d at 1182. In other words, when a statement disclosed by the State asserts that a person has no recollection of the crime, and the person suddenly announces that he or she is an eyewitness to it, the State must also disclose that information to the defense.

Ignoring for the moment Justice Wells's objection to the Evans exception as "creat[ing] uncertainty," Evans, 770 So.2d at 1183 (Wells, C.J., joined by Quince, J., dissenting), this case clearly does not fall within it. Scipio suggests it is "the converse of Evans," meaning that the investigator went from being an important eyewitness to one who "didn't see anything." Petitioner's Reply Brief at 9-10. That characterization is not entirely accurate, because the investigator did see something under the victim's body, and he remained a relevant eyewitness after his change in testimony. But even if one accepts that premise, the logic of Evans does not run smoothly in reverse. In Evans, we recognized that a witness's metamorphosis into an eyewitness needed to be disclosed because the prosecutor has a continuing duty to disclose "additional witnesses" under rule 3.220(j). The witness in Evans went from a witness to nothing (because she did not see anything), and therefore not a witness at all, to an eyewitness to the crime.

In this case, however, the investigator cannot plausibly be characterized as an additional witness. To the contrary, Scipio essentially wishes the investigator had been subtracted from the witness list, because his testimony became less relevant. Subtraction is not a requirement of the discovery rules. They encourage the prosecutor to err on the side of over-inclusion, listing "all persons known . . . to have information that may be relevant to the offense." Fla. R.Crim. P. 3.220(b)(1)(A) (emphasis added). The defense then has the opportunity to test the relevancy and reliability of those witnesses' accounts.

Another crucial distinction between this case and Evans is that here the State never gave the defense an inaccurate statement under rule 3.220(b)(1)(B). Rather, the mistaken testimony was elicited in a deposition conducted by the defense. By its own terms, Evans did not extend this far. It merely held that "our statements in Bush do not control in situations where the State provides the defendant with a witness's `statement' — as that term is defined in rule 3.220(b)(1)(B) — and thereafter fails to disclose to the defendant that the witness intends to change that statement." Evans, 770 So.2d at 1182 (emphasis added). This is not one of those situations.

Nevertheless, the majority argues that "since the medical examiner's investigator actually gave a statement by deposition that he possessed relevant information of the crime scene, the State also had an obligation to disclose any material change in that statement." Majority op. at 1142. This argument implies that a deposition is a "statement" as defined in rule 3.220(b)(1)(B). But the definition cannot bear such an interpretation. The definition clearly refers to written or recorded statements known to the State and unknown to the defendant. No court has ever interpreted the rule as requiring the State to "disclose" depositions that the defense attended. Rather, depositions offer the defense its own opportunity to force disclosure directly from witnesses. As Justice Wells has explained:

The continuing duty to disclose pursuant to Florida Rule of Criminal Procedure 3.220(j) has to be afforded a reading which takes into account the right to depositions under our criminal procedures. The right to depositions has to have the concomitant obligation that in the deposition the party taking the deposition will seek information which is available at the time of the deposition. Otherwise, the omission of questions about which information was available at the time of deposition can be used as a strategic depository of discovery violations.

Reese v. State, 694 So. 2d 678, 686 (Fla. 1997) (Wells, J., concurring in part and dissenting in part). Because deposition testimony is not a "statement" that the State must disclose in the first place, rule 3.220 does not require the State to disclose oral changes to it. Thus, contrary to the majority's assertion that "[t]he form in which the statement was provided, i.e., a recorded deposition versus a written statement, does not affect [the State's] obligation," majority op. at 1142, both the language and logic of the discovery rules suggest that form is fundamentally important.

The majority, perhaps recognizing that its decision stretches the discovery rules beyond their textual capacity and even beyond our interpretation of them in Evans, emphasizes that the State's conduct also defies "the purpose and spirit of those rules," majority op. at 1144, which is the "policy of avoiding trial by ambush or surprise." Id. at 1145. I concede as much. But the whole point of adopting specific discovery rules is to give prosecutors more concrete guidance than an abstract statement of policy — such as "no trials by ambush or surprise." By allowing the policy behind the rules to justify a result contrary to their plain meaning, the majority leaves prosecutors without firm boundaries.

There were good reasons for establishing a firm boundary between oral changes in testimony and changes "written or recorded or summarized in any writing or recording." Fla. R.Crim. P. 3.220(b)(1)(B). When a statement is written or recorded, the prosecutor can easily copy and forward it to the defense. But oral statements pose more of a problem, because they tend to be more frequent, less formal, and not as easily verifiable. As the dissenters in Evans cautioned, mandating the disclosure of material oral changes in testimony

creates uncertainty as to what the trial judge is to do when a witness's testimony at trial is not precisely as given in an earlier deposition or written statement, i.e., how much and what quality of deviation requires a mistrial, how long must the State be alleged to have known the changes in testimony, and what could be impeached by the prior statements and not require a mistrial. The variations in a witness's trial testimony from an oral statement made by the witness to the prosecutor minutes before going on the witness stand will more likely come in all shades of gray — not in striking contrasts.

770 So.2d at 1183 (Wells, C.J., dissenting, joined by Quince, J.).

Perhaps we could develop a rule that would assuage these concerns, while also protecting defendants from the sort of surprise that occurred here. But in fairness to prosecutors, we should do so through a prospective rule amendment, not through a retroactive reinterpretation of the discovery rules. I disagree with the majority's decision to take the latter course.

C. Any Violation Did Not Prejudice the Defendant

Having found that the State violated the discovery rules, the majority — and this is the real purpose of the opinion, since the district court found a violation as well — concludes that the defendant was prejudiced. Under the standards enunciated in Schopp, Pender, and Pomeranz, however, the violation did not procedurally prejudice the defense. Apparently, the majority agrees, since it has to recede from Schopp to reach its conclusion.

Scipio maintains that, if the investigator's change in testimony had been disclosed before trial, then "the defense clearly would not have put on any evidence" at all. At trial, the defense called only one other witness. That witness arrived at the crime scene after the shooting and could not recall seeing any of the State's eyewitnesses present. Scipio claims that, had he known the object under the victim's body was just a pager, he would have foregone not only the investigator's embarrassing testimony, but also the other witness's mildly helpful testimony in exchange for securing the opening and closing portions of closing argument. See Fla. R.Crim. P. 3.250 ("[A] defendant offering no testimony in his or her own behalf, except the defendant's own, shall be entitled to the concluding argument before the jury.").

To satisfy the procedural prejudice standard, this alternative strategy must be materially different from the strategy that Scipio actually used at trial. We explained in Schopp that "[t]rial preparation or strategy should be considered materially different if it reasonably could have benefited the defendant" by "resulting in a favorable verdict." 653 So.2d at 1020-21. We adapted this standard from the harmless error analysis in DiGuilio, 491 So.2d at 1135, which requires the State to "prove beyond a reasonable doubt that the error complained of did not contribute to the verdict or, alternatively stated, that there is no reasonable possibility that the error contributed to the conviction." See Schopp, 653 So.2d at 1020 (citing DiGuilio and requiring "[a]pplication of a harmless error analysis in this context"). Schopp's definition of procedural prejudice has never been overruled. In fact, we applied it only two years later, cementing its status as established law. See Pomeranz, 703 So.2d at 469 (deeming a discovery error harmless because its prevention would not "have resulted in a more favorable outcome").

The district court concluded that any change in Scipio's trial strategy would not have benefited him because impeaching the investigator "was more favorable to Scipio than not having called the witness." Scipio, 867 So.2d at 431. According to the district court, if the defense had changed its strategy to exclude the investigator's testimony, "the outcome of this case would not have been affected given the multiple eyewitness testimonies." Id.

I agree with the district court's analysis. The State presented four main witnesses. Three claimed to be eyewitnesses. One saw Scipio start to shoot and then ran away in fear. The other two watched as the victim crawled out of the bar with Scipio following behind, screaming obscenities and firing additional shots until bystanders intervened. The fourth witness testified that Scipio came to his home on the night of the shooting and, with a gun protruding from his coat, admitted that he had just killed a man who testified against his "homeboy." (The victim had recently testified against a man convicted of robbery.) Although Scipio challenged the credibility of these witnesses at trial and pointed out minor inconsistencies in their stories, the jury nevertheless convicted him based largely on their testimony.

Given this overwhelming evidence of guilt, I cannot imagine how it would have benefited Scipio to withdraw two witnesses who, at the very least, raised concerns about how carefully the crime scene investigation had been handled. The investigator, for example, admitted to the jury that he believed (until just before trial) that he had discovered a semiautomatic firearm beneath the victim's body. Yet he never followed up on that finding, which happened to be mistaken. This oversight could have reduced the jury's confidence in the quality of the investigation. The other defense witness testified that the State's eyewitnesses were not present when he arrived after the shooting. This testimony also exposed a possible gap in the record.

Although neither defense witness offered strong, exculpatory testimony, they were better than Scipio's stated alternative — no evidence at all. Both witnesses at least challenged the State's investigation. Neither undermined Scipio's effort to discredit the State's eyewitnesses, who were the centerpiece of the case against him. See Cox v. State, 819 So. 2d 705, 713 (Fla.2002) (holding a discovery error harmless "[s]ince the appellant's theory of defending himself was just as viable after [the investigator's] testimony as it was prior to the introduction of this evidence"). Exchanging both defense witnesses for the right to sandwich the State during closing argument could not reasonably have benefited the defendant, much less affected the outcome. Thus, the State's failure to disclose the investigator's change in testimony did not prejudice the defense. It was harmless beyond a reasonable doubt.

The majority now recedes from Schopp's harmless error standard and bars any consideration of the verdict in a procedural prejudice analysis. I see no basis for doing so. We depart from our precedents only "when an established rule of law has proven unacceptable or unworkable in practice," Allstate Indem. Co. v. Ruiz, 899 So. 2d 1121, 1131 (Fla.2005) (citing Brown v. State, 719 So. 2d 882, 890 (Fla.1998) (Wells, J., dissenting)), or when "necessary. . . to remedy continued injustice." Id. (quoting Haag v. State, 591 So. 2d 614, 618 (Fla.1992)). The Schopp standard has not proven unworkable. To the contrary, it has prevented scarce judicial resources from being expended on new trials that undoubtedly would have resulted in the same verdict. The majority's approach will have the opposite effect, mandating reversal so that defendants can follow a new path to the same destination.

Ignoring the verdict in our procedural prejudice analysis will not "remedy continued injustice." If anything, it will cause injustice by making it easier for defendants to obtain retrials for discovery violations than for other equally (if not more) serious trial defects. Most of our other prejudice standards consider the possible effect on the verdict. For example, newly discovered evidence must be "of such a character that it would probably produce an acquittal on retrial." Johnson v. State, 904 So. 2d 400, 404 (Fla.2005) (quoting Mills v. State, 786 So. 2d 547, 549 (Fla. 2001)). We also consider the verdict when the State violates Brady v. Maryland, 373 U.S. 83, 83 S. Ct. 1194, 10 L. Ed. 2d 215 (1963), by willfully or inadvertently suppressing exculpatory evidence. Specifically, we ask whether "there is a reasonable probability that had the information been disclosed to the defendant, the result of the proceeding would have been different." Young v. State, 739 So. 2d 553, 559 (Fla. 1999). Even the standard that applies when the State knowingly allows a witness to give false testimony, in violation of Giglio v. United States, 405 U.S. 150, 92 S. Ct. 763, 31 L. Ed. 2d 104 (1972), requires that the State prove beyond a reasonable doubt that there is no "reasonable likelihood that the false testimony could have affected the judgment of the jury." Mordenti v. State, 894 So. 2d 161, 175 (Fla.2004) (quoting Guzman v. State, 868 So. 2d 498, 506 (Fla. 2003)). Thus, under the majority's redefinition of procedural prejudice, retrials will now be harder to obtain when the State knows that a witness is testifying falsely at trial than when the State knows (as it did here) that a witness will testify truthfully at trial but fails to warn the defense of a change in testimony. This is an arbitrary and unjust distinction. Schopp avoided such distinctions by expressly incorporating the harmless error analysis into the procedural prejudice standard.

The majority's approach eliminates the clear boundaries of the Schopp standard and replaces them with an unmarked demilitarized zone. The majority now focuses on whether the discovery violation "materially hindered the defendant's trial preparation or strategy." But how does one evaluate whether the hindrance was "material" without determining whether a different strategy would have affected the outcome? Every discovery error has some effect on trial preparation and strategy. That, of course, was the whole point of the Schopp standard. "Material" meant "affecting the outcome." Under the majority's reasoning, "material" means whatever a particular judge thinks it should mean. The majority also fails to analyze whether the trial court must accept at face value the defendant's assertions that his trial strategy would have changed. The majority fails to offer any guidance on these issues.

The majority's application of the new rule to this case demonstrates just how malleable it is. The majority considers the hindrance to the defense "material" because, had the defendant known of the investigator's changed testimony, it either would not have presented him as a witness, would have sought a continuance, or would have impeached him. Majority op. at 1150. If the defendant had not presented the investigator, then it would have been left with little if anything as a defense theory. I would hardly consider any hindrance to the defense "material." The second alternative, according to the majority, was seeking a continuance. In that event, the defendant, after days or weeks to think about it, would have been left with the same choices — present the witness and impeach him or leave him out. The third alternative, to present the witness but impeach him with the prior inconsistent testimony, is essentially what happened at trial. The majority, without discussion, deems it "highly unlikely" that the defense would have employed the same strategy if given time to reflect. Id. But at least that strategy gave the defendant the chance to introduce the deposition testimony and the possibility that the object found under the victim was a firearm and not a pager. I agree that such testimony benefited the defendant only marginally. But that simply demonstrates that even the investigator's original testimony, and thus its effect on the trial, was itself marginal, and therefore the State's failure to disclose the investigator's change in testimony could not have been material.

Rather than adopt the majority's malleable new rule, I would continue to apply the comparatively precise and predictable Schopp standard, which focuses on whether a reasonable possibility exists that the defendant's trial preparation or strategy would have been materially different such that it could have benefited the defendant by resulting in a more favorable outcome. In this case, the answer to that question is clearly no.

Conclusion

As explained above, I disagree with the majority's assertion of jurisdiction over this case. I also disagree with the majority's conclusions that the State committed a discovery violation and that it procedurally prejudiced the defense. I find neither a violation nor prejudice. Therefore, I respectfully dissent.

WELLS and BELL, JJ., concur.

NOTES

[1] At the outset, it is important to note that this case concerns the most serious crime and the most severe penalty, short of death, provided for under Florida's criminal laws.

[2] In Evans, we also discussed and qualified our earlier holding in Bush v. State, 461 So. 2d 936 (Fla.1984), finding no discovery violation by the State in a similar situation. We explained:

To the extent that our determination here may be interpreted as being inconsistent with our "changed testimony" statements in Bush, we clarify that our statements in Bush do not control in situations where the State provides the defendant with a witness's "statement" — as that term is defined in rule 3.220(b)(1)(B) — and thereafter fails to disclose to the defendant that the witness intends to change that statement....

Evans, 770 So.2d at 1182. In Bush, a state investigator stated in his deposition that a witness had not identified a photograph of the defendant in a photographic lineup, but at trial testified that a witness had identified the photograph during the lineup. 461 So.2d at 938. The investigator testified that the inconsistency in his testimony was explained by the fact that defense counsel had asked two different questions. Id. This Court found that under those circumstances the prosecution's failure to inform the defense of the change in the investigator's testimony did not constitute a discovery violation. Id.

Neither Evans nor Bush presents circumstances identical to those presented here. Whereas the witness who changed his testimony in Bush was a designated State witness and was subject to impeachment by the defense on the basis of inconsistent statements, the witness in the instant case was a State investigator but was called by the defense in the reasonable and good faith belief that the investigator would provide substantive evidence helpful to the defense. Furthermore, while the changed testimony in Bush may have resulted from differences in questioning by the defense rather than a mistake or clear change in testimony by the witness, the changed testimony in the instant case was directly contrary to the testimony given earlier, and occurred after "the state actively procured the defense witness's recantation of his earlier deposition testimony." Scipio, 867 So.2d at 430.

[3] See, e.g., Brady v. Maryland, 373 U.S. 83, 87, 83 S. Ct. 1194, 10 L. Ed. 2d 215 (1963) (holding that the State must disclose material information within the State's possession or control that tends to negate the guilt of the defendant); Giglio v. United States, 405 U.S. 150, 154, 92 S. Ct. 763, 31 L. Ed. 2d 104 (1972) (concluding that where the prosecutor knowingly uses perjured testimony, or fails to correct what the prosecutor later learns is false testimony, the false evidence is material if there is any reasonable likelihood that the false testimony could have affected the judgment of the jury).

[4] Richardson v. State, 246 So. 2d 771 (Fla. 1971).

[5] In Schopp, this Court held that in applying the Smith standard of procedural prejudice, there could be instances where some discovery violations could be found harmless beyond a reasonable doubt. Schopp, 653 So.2d at 1021. Indeed, we held that Schopp represented that rare instance of harmless error since, in essence, Schopp had "won" his case because his defense strategy was to admit commission of a lesser offense while denying guilt of the charged crime. Id. at 1022. The jury apparently accepted this defense in acquitting him of the charged crime but finding him guilty of the lesser offense;

Because it is clear in this case that the State's failure to include the officer on its original witness list could not have materially hindered the defense, the trial court's failure to adequately inquire into what, if any, corrective measures should have been taken was harmless beyond a reasonable doubt. During discovery, defense counsel was given the police report, which was consistent with the officer's testimony at trial. The challenged testimony was cumulative to other testimony, including that of the defendant. And, even if the officer had been listed as a witness and the defense had deposed him, there would have been nothing in the officer's testimony that could have supported a strategy different from that taken — to admit that Schopp committed burglary but maintain that he was unarmed and did not know that he was stealing a firearm. Moreover, because Schopp effectively "won" his case, there is no reasonable possibility that a change in trial tactics could have benefited him.

Schopp, 653 So.2d at 1021-22.

[6] Viewing such conduct by the State, whose obligation is not only to prosecute but to see that justice is done, might make this Court wonder whether all its efforts to stress ethical and professional conduct are in vain. We are confident, however, that this is not typical of Florida's prosecutors.

[7] Because the change in testimony did not "tend[ ] to negate the guilt of the defendant as to any offense charged," it also did not fall within the prosecution's obligation to disclose all exculpatory information. Fla. R.Crim. P. 3.220(b)(4).