Reset A A Font size: Print

United States Court of Appeals,Fourth Circuit.

William Thomas BAUBERGER, Petitioner-Appellee, v. Grady J. HAYNES, Respondent-Appellant.

No. 09-8111.

Decided: February 11, 2011

Before WILKINSON and MOTZ, Circuit Judges, and DamonJ. KEITH, Senior Circuit Judge of the United States Court of Appeals for the Sixth Circuit, sitting by designation. ARGUED:Clarence Joe DelForge, III, North Carolina Department of Justice, Raleigh, North Carolina, for Appellant. David L. Neal, Hillsborough, North Carolina, for Appellee. ON BRIEF:Roy Cooper, Attorney General of the State of North Carolina, Raleigh, North Carolina, for Appellant.


William Thomas Bauberger was convicted of second-degree murder and assault with a deadly weapon inflicting serious injury after he drove his car the wrong way down an exit ramp, killing one person and wounding another. Bauberger unsuccessfully challenged his murder conviction in state court after learning that the jurors read dictionary definitions of several words in the judge's instructions.

Bauberger then sought federal habeas relief under 28 U.S.C. § 2254, arguing that the jurors' dictionary use violated his federal constitutional rights and prejudiced his verdict. The district court agreed and granted the writ. Given that the dictionary definitions did not materially alter the instruction as a whole and that the government presented significant evidence of malice, any misconduct the jurors may have committed did not exert a “substantial and injurious effect ․ in determining the jury's verdict.” Brecht v. Abrahamson, 507 U.S. 619, 623 (1993) (quoting Kotteakos v. United States, 328 U.S. 750, 776 (1946)) (internal quotation marks omitted). Accordingly, the district court's order granting Bauberger the writ is reversed, and the case is remanded with directions to dismiss the petition.


On February 3, 2002, William Bauberger attended a Super Bowl party at a friend's house. Over the course of the five or so hours he was there, he drank more than ten beers. After the game Bauberger, despite his intoxicated condition, decided to drive to another friend's house. He never made it. Instead, he drove his Cadillac-at a speed somewhere around 45 to 55 miles per hour-the wrong direction down an exit ramp off U.S. Highway 421 in Winston-Salem, North Carolina. Disregarding “Do Not Enter” and “Wrong Way” signs, as well as the honks and swerves of several cars traveling in the proper direction, Bauberger finally crashed his car into that of William and Carol Foy. William sustained several broken bones, and Carol died within minutes.

The government charged Bauberger with second-degree murder and assault with a deadly weapon inflicting serious injury. At trial the government introduced Bauberger's troubled driving record. He had two prior driving-while-impaired (“DWI”) convictions, as well as a reckless driving conviction and other driving offenses. He also had disregarded prior court orders not to drive and was driving on a revoked license the night of the collision. Bauberger admitted that his blood-alcohol content that night was 0.20 and that he was aware of the dangers of driving while intoxicated. In light of this evidence, Bauberger conceded guilt to the lesser-included offense of involuntary manslaughter at trial but insisted that he lacked the malice necessary for a second-degree murder conviction under North Carolina law.

The jury convicted Bauberger of second-degree murder and assault with a deadly weapon inflicting serious injury, and he was sentenced to 189 to 236 months in prison. Shortly after the verdict came down, however, the parties and the court learned that the jury may have used a dictionary during its deliberations. Bauberger filed a postconviction Motion for Appropriate Relief (“MAR”), arguing that the dictionary was an impermissible extraneous influence on the jurors and that the dictionary definitions lowered the government's burden of proof regarding malice. The MAR court determined from the jurors' affidavits that the jury's foreperson left the courthouse during a break in deliberations, went to a public library, and brought back the 1953 edition of Webster's New Collegiate Dictionary. He read to the other jurors the dictionary's definition of several terms in the judge's malice instruction but not the definition of “malice” itself.

The MAR court denied Bauberger's requested relief, reasoning that the jurors' actions, though improper, were harmless. The North Carolina Court of Appeals affirmed, reasoning that Bauberger's federal constitutional rights were not violated because the definitions “concerned legal terminology, not evidence developed at trial.” State v. Bauberger, 626 S.E.2d 700, 706 (N.C.Ct.App.2006). The North Carolina Supreme Court affirmed by an equally divided vote, leaving Bauberger's convictions in place but stripping the lower court's decision of precedential effect. See State v. Bauberger, 637 S.E.2d 536 (N.C.2006).

Bauberger filed for federal habeas relief under 28 U.S.C. § 2254. The district court granted his petition, holding that the jurors' dictionary use violated Bauberger's clearly established Sixth Amendment rights and that the error prejudiced him because the dictionary's definitions of “recklessly” and “wantonly” may have lowered the government's burden of proof regarding malice.


We review the district court's decision to grant the writ de novo. Bell v. Ozmint, 332 F.3d 229, 233 (4th Cir.2003).

We shall assume without deciding that the North Carolina Court of Appeals's rejection of Bauberger's Sixth Amendment claims was “contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States,” satisfying the threshold demands for habeas relief under the Antiterrorism and Effective Death Penalty Act (“AEDPA”). 28 U.S.C. § 2254(d)(1); see also Golphin v. Branker, 519 F.3d 168, 189-90 (4th Cir.2008) (leaving unresolved whether the state court unreasonably applied federal law because any error did not have prejudicial impact under Brecht.) By doing so, we skirt the problems long associated with unnecessary constitutional decisionmaking: we avoid wasting the parties' and the courts' limited resources on “questions that have no effect on the outcome of the case,” Pearson v. Callahan, 129 S.Ct. 808, 818 (2009), and we minimize the threat to good adjudication that arises when courts enter into thorny constitutional areas with inadequate briefing or in fact-bound dispositions, see id. at 819-20.

Assuming arguendo that the state court erred in rejecting Bauberger's claim does not end our inquiry, however. “[M]ost constitutional errors can be harmless,” including those of the kind we shall assume occurred at Bauberger's trial. Arizona v. Fulminante, 499 U.S. 279, 306 (1991); see, e.g., Fullwood v. Lee, 290 F.3d 663, 678-83 (4th Cir.2002) (third-party influence on juror and jurors' consideration of extraneous evidence subject to harmless error review). On direct review, the government has the burden of proving that a constitutional error was “harmless beyond a reasonable doubt.” Chapman v. California, 386 U.S. 18, 24 (1967).

On collateral review, however, the calculus changes. Because of the threat collateral attacks pose to “finality, comity, and federalism,” Fry v. Pliler, 551 U.S. 112, 116 (2007), habeas petitioners may secure the writ only if the error “actual[ly] prejudice[d]” them, Brecht, 507 U.S. at 637 (quoting United States v. Lane, 474 U.S. 438, 449 (1986)) (internal quotation marks omitted). In making this determination, courts ask whether the error had a “substantial and injurious effect or influence in determining the jury's verdict.” Id. (quoting Kotteakos, 328 U.S. at 776) (internal quotation marks omitted). In the “unusual” situation where “the matter is so evenly balanced that [the habeas judge] feels himself in virtual equipoise as to the harmlessness of the error [under Brecht ]”-that is, where the judge is in “grave doubt”-the court must grant the writ. O'Neal v. McAninch, 513 U.S. 432, 435 (1995).

The Supreme Court recently addressed the relationship between the Chapman direct review standard, AEDPA's deference requirement, and the Brecht collateral review standard. In Fry, the question before the Court was whether habeas courts should apply the AEDPA/Chapman test (asking whether a state court unreasonably applied Chapman's direct review standard), the Brecht test (asking whether a constitutional error had a substantial and injurious effect), or both. See Fry, 551 U.S. at 119-20.

The Court made two critical determinations. First, it held that “in § 2254 proceedings a court must assess the prejudicial impact of constitutional error in a state-court criminal trial under the [Brecht ] standard ․, whether or not the state appellate court recognized the error and reviewed it for harmlessness under [Chapman ].” Fry, 551 U.S. at 121-22 (emphasis added) (citations omitted). Second, the Court stated that “it certainly makes no sense to require formal application of both tests (AEDPA/Chapman and Brecht ) when the latter obviously subsumes the former.” Id. at 120. That is to say, where an error is harmful under Brecht, any state court decision declaring it harmless must have unreasonably applied Chapman. As a result, any error satisfying Brecht will also satisfy AEDPA's deference requirements. See Ruelas v. Wolfenbarger, 580 F.3d 403, 411-13 (6th Cir.2009).

Thus the Court took a somewhat convoluted debate over the appropriate standard in harmless error habeas cases and opted for simplicity itself. Federal habeas courts must always review constitutional errors in state trials under Brecht, but they need not debate whether a state court's harmless error determination also unreasonably applied Chapman, as most circuits since Fry have explicitly or implicitly recognized. See Welch v. Workman, 607 F.3d 674, 686 (10th Cir.2010); Wesbrook v. Thaler, 585 F.3d 245, 255-56 (5th Cir.2009); Ruelas v. Wolfenbarger, 580 F.3d 403, 411-13 (6th Cir.2009); Moses v. Payne, 555 F.3d 742, 755 (9th Cir.2009); Farley v. Bissonnette, 544 F.3d 344, 347-48 (1st Cir.2008); Bond v. Beard, 539 F.3d 256, 275-76 (3d Cir.2008). But see Perkins v. Herbert, 596 F.3d 161, 175-77 (2d Cir.2010) (leaving undecided whether both AEDPA/Chapman and Brecht must be used); Johnson v. Acevedo, 572 F.3d 398, 403-04 (7th Cir.2009) (requiring AEDPA/Chapman and Brecht where a state court addressed harmlessness).

Of course, most successful habeas petitioners must still go through “AEDPA['s] mandate[d] ․ two-step analysis” by demonstrating that the state court's resolution of their constitutional claim was contrary to or unreasonably applied clearly established federal law and that the error was prejudicial under Brecht. Baum v. Rushton, 572 F.3d 198, 205 (4th Cir.2009); see also Barbe v. McBride, 521 F.3d 443, 453 (4th Cir .2008). But this two-step process does not require us to address the first prong where the petitioner's claims fail on the second, see Golphin, 519 F.3d at 189-90, and as explained above, Fry absolves us of any need to consider both AEDPA/Chapman unreasonableness and Brecht prejudice in the harmless error context. Accordingly, we need only determine whether the jury's dictionary use prejudiced Bauberger under the Brecht standard.1


In analyzing whether the prejudice from the jury's dictionary use meets the Brecht standard, we begin by laying out the jury's actions in greater detail.


The trial judge instructed the jurors on the meaning of “malice,” the only element of second-degree murder that Bauberger disputed, as follows:

Malice is a necessary element which distinguishes second degree murder from manslaughter. Malice arises when an act which is inherently dangerous to human life is intentionally done so recklessly and wantonly as to manifest a mind utterly without regard for human life and social duty and deliberately bent on mischief.

Shortly after beginning its deliberations, the jury requested a copy of the elements of second-degree murder and manslaughter. In response, the judge re-read the instructions to the jurors and returned them to their deliberations. However, they soon sent another note to the judge asking for a copy of the instruction regarding malice and any other contested element, stating that “[m]any of us are visual people.” The judge informed them he would prepare a copy during their upcoming lunch recess and dismissed them for that recess.

As previously indicated, during lunch the jury's foreperson retrieved a dictionary and read some of its definitions to the other jurors when deliberations resumed. Though he did not read the definition for “malice” itself to the jury, he did read them the definitions for some of the words in the judge's instruction, including “recklessly” and “wantonly.”2 The dictionary defined the former as “lack of due caution” and the latter as “arrogant recklessness of justice or the feelings of others.” Two hours into their resumed deliberations the jury informed the judge that they had resolved one count but were split seven to five on the other. The judge instructed them to do their best to “reconcile [their] differences ․ without the surrender of conscientious convictions.” Within an hour of this instruction the jury stood at ten to two, and within two hours they had convicted Bauberger of second-degree murder and assault with a deadly weapon inflicting serious injury.


Bauberger argues that the dictionary definitions of “recklessly” and “wantonly” lowered the government's burden of proof on the issue of malice, an issue with which he contends the jury apparently struggled. As a result, Bauberger claims that the erroneous use of those dictionaries meets the Brecht standard: one or more of the jurors may have applied an unacceptably low malice standard in convicting Bauberger of second-degree murder, and therefore the dictionary use had a substantial and injurious effect upon his verdict. We disagree: the dictionary definitions did not materially alter the meaning of the instruction as a whole, and the government introduced significant evidence of malice.


In assessing the harm from jurors' use of a dictionary we look in part to the difference between the dictionary definition and the legal definition. See Henderson v. Kibbe, 431 U.S. 145, 154 (1977) (looking to the difference between the actual instruction given and the correct instruction to determine prejudice); McNeill v. Polk, 476 F.3d 206, 226-27 (4th Cir.2007) (King, J., concurring in part and concurring in the judgment) (examining difference between dictionary and legal definitions); McNeill, 476 F.3d at 229-30 (Gregory, J., dissenting in part and concurring in part) (same). Under North Carolina law the “distinction between ‘recklessness' indicative of murder and ‘recklessness' associated with manslaughter ‘is one of degree rather than kind.’ “ State v.. Rich, 527 S.E.2d 299, 303 (N.C.2000) (quoting United States v. Fleming, 739 F.2d 945, 948 (4th Cir.1984)). North Carolina courts have indicated, however, that the difference in degree is large. See, e.g., id.; State v. Wilkerson, 247 S.E.2d 905, 918 (N.C.1978). Nonetheless, they have upheld second-degree murder convictions even where a portion of the malice instruction veered toward culpable negligence in defining recklessness, provided the instruction as a whole conveyed the difference. See Rich, 527 S .E.2d at 301-03 (upholding a second-degree murder conviction where a portion of the recklessness instruction suggested that “recklessness of consequences” sufficed).

If Bauberger's jury had only the dictionary's “lack of due caution” definition to go on in determining the recklessness portion of the malice element, he would have more of a case. But as the Supreme Court has long emphasized in the analogous context of jury instructions, “a single instruction to a jury may not be judged in artificial isolation, but must be viewed in the context of the overall charge.” Cupp v. Naughten, 414 U.S. 141, 146-47 (1973). Viewed in this light, the instruction, even as possibly modified by the definitions the jurors consulted, fully conveyed the essence of North Carolina law concerning malice. The modified version still referred to an “intentionally done,” “inherently dangerous” act. It still described the “lack of due caution” as of such a nature as to “show a mind fully or totally without respect or consideration for human life” or “social duty.” And it still spoke of an individual whose acts were so “arrogant[ly] reckless[ ] of justice or the feelings of others” as to indicate a mind “deliberately bent on mischief.”

This is not the language of mere culpable negligence. Phrases like these impose a much higher standard than the “thoughtless disregard of consequences” and “heedless indifference to the safety and rights of others” culpable negligence requires. State v. Mack, 697 S.E.2d 490, 494 (N.C.Ct.App.2010) (quoting State v. Wade, 589 S.E.2d 379, 382 (N.C.Ct.App.2003)) (internal quotation marks omitted). Bauberger's verdict was not substantially and injuriously affected by the dictionary's definition of “recklessly” because the altered instruction as a whole remained materially equivalent to the one given by the judge.

So too with “wantonly.” Under North Carolina law, “wantonness” describes “intentional wrongdoing,” conduct undertaken in “conscious and intentional disregard of and indifference to the rights and safety of others.” State v. Williams, 199 S.E.2d 409, 412 (N.C.1973); see also State v. Young, 559 S.E.2d 814, 818 (N.C.Ct.App.2002). Again, it would be one thing if the jurors had only considered “arrogant recklessness of justice or the feelings of others” in determining whether Bauberger's conduct was wanton. But the possibly modified instruction in its entirety still conveyed the knowing disregard of others' safety central to wantonness under North Carolina law. It spoke of “intentionally done” acts, ones performed with such “arrogant recklessness” toward others as “to show a mind fully or totally without respect or consideration for human life ․ and deliberately bent on mischief.” (emphases added). Any modification of the instruction that came about by virtue of the dictionary's definition of “wantonly” did not materially affect that instruction's malice standard.

The other jury instructions also indicate that the jurors could not have read the isolated definitions so as to alter the malice instruction. “[T]he challenged instruction [is often] but one of many such instructions․” Henderson, 431 U.S. at 152 n. 10 (quoting Cupp, 414 U.S. at 147) (internal quotation marks omitted). Here, those other instructions illustrate that the jurors were well aware of the difference between malice and culpable negligence. The jurors were instructed that if they acquitted Bauberger of second-degree murder they must consider whether he was guilty of involuntary manslaughter, which requires a showing of culpable negligence. To find culpable negligence, the trial court instructed the jurors that they must find that Bauberger drove while impaired, that he “willful[ly,] wanton[ly,] or intentional[ly]” violated the law, or that his “inadvertent or unintentional violation of the law ․ [was] accompanied by recklessness of probable consequences of a dangerous nature ․ amounting altogether to a thoughtless disregard of consequences or a heedless indifference to the safety of others.” The jurors also knew that Bauberger had conceded guilt to involuntary manslaughter, including its lower culpable negligence standard, making malice the central disputed issue before them.

Given these circumstances, it is unlikely the jurors seized on isolated dictionary definitions to transmute the malice standard-which they knew to be the disputed issue in the case-into the culpable negligence standard, to which Bauberger had already conceded. In other words, Bauberger's verdict was not substantially affected because the dictionary definitions, viewed both in light of the instruction as a whole and in light of the entire trial, did not materially alter the malice standard.


We also look to the strength of the evidence in assessing whether the dictionary use substantially and injuriously affected Bauberger's verdict. See McNeill, 476 F.3d at 226 (King, J., concurring in part and concurring in the judgment); id. at 229 (Gregory, J., dissenting in part and concurring in part). This approach makes sense: if the evidence presented was such that the issue of malice was not likely a close one, it is less likely that the error impacted the jury's decision.

On this score, Bauberger's claim to prejudice is particularly weak. His is the paradigmatic case of second-degree murder via drunk driving. As detailed above, he had several prior drunk driving convictions, prior court orders not to drive, and a revoked license on the night in question. He stipulated to a high blood-alcohol content and admitted knowledge of alcohol's dangerous effects upon drivers. He also drove at a high speed, into oncoming traffic, and in disregard of several signs. North Carolina courts have routinely accepted evidence less formidable than this in upholding second-degree murder convictions. See, e.g., Rich, 527 S.E.2d at 304 (finding sufficient evidence of malice where the defendant “drove his vehicle at a high rate of speed while impaired, on the wrong side of the road, in a no-passing zone and in violation of right-of-way rules”); State v. Westbrook, 623 S.E.2d 73, 78-79 (N.C.Ct.App.2005) (finding sufficient evidence of malice where the defendant had a prior DWI conviction and sped through a traffic light on the wrong side of the road). While we are not analyzing Bauberger's conviction for the sufficiency of the evidence, such compelling evidence of malice makes it considerably less likely that the dictionary use affected the jury's ultimate decision.

Bauberger discounts the importance of this evidence by pointing to the struggles the jury apparently had in reaching its decision. He contends that the actual jury's decisionmaking process-its requests for printed instructions, its foreperson's decision to retrieve the dictionary, and its initially divided votes-is what counts under Brecht, not the actions of a hypothetical jury looking at the evidence. We agree in principle and have accordingly looked to the jury's difficulty in reaching a decision in assessing prejudice in prior dictionary use cases. See McNeill, 476 F.3d at 226 (King, J., concurring in part and concurring in the judgment); id. at 229 (Gregory, J., dissenting in part and concurring in part).

However, as applied here, this principle cannot do the work Bauberger sets out for it because the record of the jury's deliberations cannot bear this much weight. The jurors heard the definitions only an hour into their deliberations. And yet despite Bauberger's contentions that the definitions lowered the malice standard to a level to which Bauberger had essentially conceded, the jury still needed an additional four hours before reaching a decision. This timeline tells us little about the issues around which the jurors' struggles revolved. We are unwilling ultimately to put dispositive weight on a jury's ambiguous actions. That would be altogether too speculative. Where the modified instruction fully conveyed the gist of North Carolina law and where the evidence presented was so strong, it takes clearer evidence of jury difficulty than this to satisfy Brecht.


“The principle that collateral review is different from direct review resounds throughout [the Supreme Court's] habeas jurisprudence,” Brecht, 507 U.S. at 633, and understandably so. Casually upending state convictions in federal court threatens “finality, comity, and federalism,” Fry, 551 U.S. at 116, words which singly and together convey a sense of respect toward state judiciaries and the significant responsibilities assigned them. Thus we have respected state judgments where an error's impact, considered in the context of the trial as a whole, was not significant. See, e.g., Golphin, 519 F.3d at 190-92 (overwhelming evidence rendered any error from admitting a confession harmless under Brecht ); Wilson v. Ozmint, 357 F.3d 461, 469 (4th Cir.2004) (sentencing record rendered any error from not admitting one mitigating statement harmless under Brecht ).

The error we assume arguendo occurred at Bauberger's trial does not justify the relief he now seeks. Neither the court nor the government had any role in bringing about the erroneous dictionary use. The potentially modified instruction, taken as a whole, conveyed the essence of North Carolina law regarding malice. And Bauberger's disturbing driving record, combined with his activities on the night of the accident, would more than permit a jury to find a “mind utterly without regard for” the life-threatening and ultimately life-ending risks created by his conduct. Given these circumstances, there was no substantial and injurious effect on his verdict from the jurors' dictionary reading.3 The district court's order granting Bauberger the writ is therefore reversed and the case remanded with directions to dismiss the petition.


I respectfully dissent from the majority's opinion. It is axiomatic that the Constitution prohibits jurors from going outside of the record to independently determine the standards that are to be used when deciding an accused person's guilt or innocence. This is based on many sound principles; most importantly, the reality that such standards frequently differ from the governing rule and, thus, undermine the basic precept that all persons are guaranteed due process and equal protection under the law. Such concerns are not to be taken lightly. Were we to treat such violations cavalierly, it would open the floodgates to jurors ignoring the instructions of the court and, in essence, the responsibilities entrusted them by the Constitution.

I. Analysis

As the majority notes, the appropriate question when considering a defendant's petition for writ of habeas corpus is whether the state court's decision was contrary to, or involved an unreasonable application of clearly established federal law. See 28 U.S.C.S. § 2254(d)(1). “Clearly established federal law” under § 2254(d)(1) refers to the governing legal principle or principles set forth by the Supreme Court at the time the state court rendered its decision. Bell v. Cone, 535 U.S. 685, 698 (2002); Williams v. Taylor, 529 U.S. 362, 405, 413 (2000).

Even if the state court proceedings violated clearly established law, the court may not grant a defendant's petition for relief if the error was harmless. Jones v. Polk, 401 F.3d 257, 265 (4th Cir.2005). The habeas petitioner will be entitled to relief if a habeas court is “in grave doubt about whether a trial error of federal law had substantial and injurious effect or influence in determining the jury's verdict.” O'Neal v. McAninch, 513 U.S. 432, 436 (1995) (internal quotation marks and citation omitted). “[G]rave doubt exists when, in the relevant circumstances, the question is so evenly balanced that the reviewing court finds itself in virtual equipoise on the harmlessness issue.” Barbe v. McBride, 521 F.3d 443, 461 (4th Cir.2008) (internal quotation marks and citations omitted). The test is whether it can be said with fair assurance that not a single juror's decision was swayed by resort to the extrinsic influence. Parker v. Gladden, 385 U.S. 363, 366 (1966) (A defendant is “entitled to be tried by 12, not 9 or even 10, impartial and unprejudiced jurors.”); Fullwood v. Lee, 290 F.3d 663, 678 (4th Cir.2002) (“[I]f even a single juror's impartiality is overcome by an improper extraneous influence, the accused has been deprived of the right to an impartial jury.”); Lawson v. Borg, 60 F.3d 608, 613 (9th Cir.1995) (noting that if even one juror was improperly influenced the verdict must be reversed).

A. The Trial Court Unreasonably Applied Established Supreme Court Precedent.

The Sixth Amendment provides, in relevant part, that “the accused shall enjoy the right to a ․ trial[ ] by an impartial jury ․ [and to] be confronted with the witnesses against him.” U.S. Const. Amend. VI. The right to trial by an impartial jury “guarantees ․ a fair trial by a panel of impartial, indifferent jurors.” Irvin v. Dowd, 366 U.S. 717, 722 (1961) (internal quotation marks omitted). This right prohibits “any private communication, contact, or tampering directly or indirectly, with a juror during trial about the matter pending before the jury.” Remmer v. United States, 347 U.S. 227, 229 (1954).

As the Fourth Circuit has interpreted the relevant Supreme Court precedent, all influences, outside the record, on a juror's decisions are not necessarily prohibited. The Supreme Court has clearly established that while external influences on a jury's deliberations are prohibited, internal ones are permissible. Robinson v. Polk, 438 F.3d 350, 362 (4th Cir.2006).

Under established Supreme Court case law, an influence is external if it (1) is extraneous prejudicial information; i.e., information that was not admitted into evidence but nevertheless bears on a fact at issue in the case or (2) is an outside influence upon the partiality of the jury, such as private communication, contact, or tampering ․ with a juror.

Id. at 363 (internal quotation marks and citations omitted). Examples of internal influences include alcohol or drugs taken by a juror, Turner, 379 U.S. at 466, or Bible readings which a juror relies upon for the purpose of “examining his or her own conscience from within,” Robinson, 438 F.3d at 363-64.

It is clear under the Supreme Court's precedents that when the jury relies on a source outside of its own knowledge or beliefs, not presented at trial or by the trial judge as part of his or her instructions, to determine what relevant law to apply, the jury has been subject to an “external influence” in violation of the Sixth Amendment to the United States Constitution. Rogers v. United States, 422 U.S. 35 (1975) (finding violation of the Sixth Amendment where the court, without consulting the defendant, provided further instruction to the jury which the jury then relied on in convicting the defendant). See also Tanner v. United States, 483 U.S. 107, 117-18 (1987) (noting the distinction between “external” influences, such as a juror reading a newspaper or hearing prejudicial statements from others, and “internal” influences); Parker, 385 U.S. at 364-66 (finding that a bailiff's statement to jurors that the defendant was a “wicked fellow” and “guilty” constituted an “outside influence” that violated the defendant's Sixth Amendment right to fair trial and confrontation because “the evidence developed against a defendant shall come from the witness stand in a public courtroom where there is full judicial protection of the defendant's right of confrontation, of cross-examination, and of counsel”); Turner v. Louisiana, 379 U.S. 466, 473 (1965) (finding a violation of defendant's Sixth Amendment rights where two deputies who testified against him were assigned to guard, and fraternized with, the jury); Remmer v. United States, 347 U.S. 227, 229 (1954) (stating that “private communication, contact, or tampering” with the jury is presumptively prejudicial); Mattox v. United States, 146 U.S. 140, 149 (1892) (stating that “in capital cases ․ the jury should pass upon the case free from external causes tending to disturb the exercise of deliberate and unbiased judgment”).

Unlike ingesting alcohol or drugs or reading a Bible to settle one's conscience, looking up legal terms to apply in the decision-making process is not merely an “internal” matter that merely affects how one feels or facilitates an examination of what one already thinks or believes. Rather, in this case, the jury in considering the dictionary's definitions was consulting an external source-the dictionary-specifically because they found their internal knowledge to be insufficient. See Robinson, 438 F.3d at 364 (external influences impart pressure or knowledge on “a juror apart from the juror himself, the reading of Bible passages invites the listener to examine his or her own conscience from within.”). Had the jury had the knowledge within itself, it would not have found it necessary to consult the dictionary in the first place.

Likewise, unlike reading the Bible or ingesting alcohol or drugs, consulting the dictionary to determine the applicable legal principle was not merely incidental to the issues before the jury, but was directly relevant as the jury looked up words it was instructed to apply. See id. at 363 (finding that juror's reading of the Bible did not require reversal of the conviction because “the Bible had no bearing on any fact relevant to sentencing, and was therefore not tantamount to ‘evidence’ that was used against him at sentencing.”). Any doubt concerning the correctness of this conclusion is all but vitiated in a situation such as this where the defendant has conceded all but one legal element and the jury acquired information about this one legal element at issue.

The cases the state provides, in response, are clearly distinguishable. In none of the cases did the jurors rely on an influence to resolve a legal question or factual dispute relevant to the case before it. Wolfe v. Johnson, 565 F.3d 140 (4th Cir.2009) (finding no error where juror brought pictures of his son into jury room and another juror spoke with his wife); Robinson, 438 F.3d at 364 (finding no clearly established law prohibiting a juror from relying on Bible not to garner any external fact or legal principle merely to reflect on his own conscience); Lynch v. Polk, 204 F. App'x. 167 (4th Cir.2006) (same). The state's reliance on the aforementioned cases suffers from the flawed assumption that merely because the influence of some materials may not violate the Sixth Amendment, the consideration of any and all external material does not violate the Sixth Amendment.

Beyond citing these cases, the state provides no explanation as to why it believes that a dictionary used to define the element at the center of trial constitutes a permissible internal influence as opposed to an impermissible external influence. Rather, the state simply asserts that no case has specifically found that a jury's use of a dictionary to define the specific legal terms at issue in this case violates the Sixth Amendment. Accordingly, it concludes that the issue in this case is not clearly established.

This line of argument misconstrues the relevant legal principle. “[T]he relevant Supreme Court precedent need not be directly on point, but must provide a ‘governing legal principle’ and articulate specific considerations for lower courts to follow when applying the [relevant] precedent.” Quinn v. Hayes, 234 F.3d 837, 844 (4th Cir.2000) (citing Williams v. Taylor, 529 U.S. 362, 413 (2000)); see Panetti v. Quarter-man, 551 U.S. 930, 953, (2007) ( “That the standard is stated in general terms does not mean the application was reasonable. [The statute] does not ‘require state and federal courts to wait for some nearly identical factual pattern before a legal rule must be applied.’ ”); Lockyer v. Andrade, 538 U.S. 63, 76, (2003) (“Section 2254(d)(1) permits a federal court to grant habeas relief based on the application of a governing legal principle to a set of facts different from those of the case in which the principle was announced.”).1

Finally, this conclusion is buttressed by the decisions of other courts which similarly have found that the reliance of a jury on a dictionary in defining the applicable legal principles violates the Sixth Amendment. See United States v. Duncan, 598 F.2d 839 (4th Cir.1979); McNeill v. Polk, 476 F.3d 206 (4th Cir.2007) (two of three judges on panel finding that juror relied on improper external influence by consulting dictionary for definition of mitigate) (King, J., concurring in judgment) (Gregory, J., dissenting in judgment); Marino v. Vasquez, 812 F.2d 499, 505 (9th Cir.1987) ( “[U]nauthorized reference to dictionary definitions constitutes reversible error which the State must prove harmless beyond a reasonable doubt.”); United States v. Kupau, 781 F.2d 740, 744 (9th Cir.1986), cert. denied, 479 U.S. 823 (1986) (same).2

Accordingly, while the majority assumes, without deciding that the jury's conduct violated the Sixth amendment, no such assumptions are necessary. Under both the Supreme Court's and this Circuit's jurisprudence, such conduct is unconstitutional.

B. The Error had a Substantial and Injurious Effect on the Defendant.

Upon assuming that the jurors' conduct violated the Sixth amendment, the majority erroneously rejects Bauberger's petition on the grounds that any such error was not prejudicial to him.

In Mayhue v. St. Francis Hosp. of Wichita, Inc., 969 F.2d 919 (10th Cir.1992), the Tenth Circuit set out a five part test to use when determining whether a lower court's error is prejudicial. The test assesses:

(1) The importance of the word or phrase being defined to the resolution of the case; (2) The extent to which the dictionary definition differs from the jury instructions or from the proper legal definition; (3) The extent to which the jury discussed and emphasized the definition; (4) The strength of the evidence and whether the jury had difficulty reaching a verdict prior to introduction of the dictionary definition; and (5) Any other factors that relate to a determination of prejudice.

Id. at 924. The test was subsequently adopted by the majority of the judges of a panel of this circuit. McNeill, 476 F.3d at 226, 229. However, the history of this circuit's consideration of these factors is not so nascent. The Supreme Court and the Fourth Circuit have long considered each of the factors set out in Mayhue and adopted in McNeill when considering the prejudicial effect of constitutional violations. Henderson v. Kibbe, 431 U.S. 145, 154 (1977)(examining the difference between the instruction given and the instruction that should have been given in determining prejudice); Barbe, 521 F.3d at 459-60 (explaining that the trial court's error prevented the defendant from cross-examining the prosecution's expert as to a matter “crucial to his presentation of an effective defense”); Fitzgerald v. Greene, 150 F.3d 357, 366 (4th Cir.1998) (noting that a potentially biased juror's impact was minimal as the jury rejected his sentencing suggestion and the jury had made its decision before he made potentially prejudicial statements); Sherman v. Smith, 89 F.3d 1134, 1143 (4th Cir.1996) (concluding that “in light of all the evidence presented at trial, [the court] harbor[ed] no grave doubt” as to the minimal prejudicial effect of the juror's “site visit.”) (internal quotation marks and citations omitted); Stockton v. Virginia, 852 F.2d 740, 746 (4th Cir.1988) (finding comment to jury by restaurant owner prejudicial where it “bore on the exact issue ․ that the jurors were deliberating on at the time”); Duncan, 598 F.2d at 866 (finding error was not prejudicial as the foreman immediately squelched any discussion of dictionary definition); Cairns v. Johnson, 267 F. App'x 240, 247 (4th Cir.2008) (noting that potentially erroneously excluded journals were relevant to the central issue in the case). Given the well-founded basis for the application of the factors the test lists, the parties, likewise, agree on its applicability.

Nonetheless, the majority focuses its analysis primarily on the extent to which the dictionary definition differed from the jury instruction-which in and of itself was substantial-and the strength of the evidence supporting Bauberger's conviction. It meanwhile attaches no weight to the centrality of the issue for which the jury sought assistance and the length of time for which the jury was exposed to the erroneous definitions, both of which past courts of this circuit have deemed relevant. Consideration of such in addition to a careful weighing of the important differences between the dictionary and legal definitions of the examined terms establishes that both the magistrate and district judges correctly concluded “grave doubt” existed as to the prejudicial effect of the constitutional error.

As to the first factor, the parties do not dispute that the words the jurors looked up in the dictionary were of supreme importance in this case. As noted above, Bauberger conceded every other element, but whether he had acted with malice-the precise term the jury sought additional help in defining.

The parties, as the majority notes, do disagree as to the extent the definitions provided in the dictionary for the words “recklessly” and “wantonly” differ from their legal counterparts. The dictionary defined “recklessly” as “lack of due caution.” App. 157. The dictionary defined “wantonly” as an “arrogant recklessness of justice or the feelings of others.” Id. Bauberger argues that the jury's use of these definitions effectively lowered the standard for malice to something more equivalent to negligence. The state argues, and the majority agrees, that the standard imposed by these terms is virtually identical to their legal counterpart.

To understand the effect these definitions had on the jury, one must examine, as the majority does, the trial judge's charge to the jury on malice, of which the disputed words were part. The judge's instruction to the jury on malice provided that “[m]alice arises when an act which is inherently dangerous to human life is intentionally done so recklessly and wantonly as to manifest a mind utterly without regard for human life and social duty and deliberately bent on mischief.” Tr. Vol. IV at 38 (Emphasis added). The trial judge contrasted this standard with the more lenient one applicable to involuntary manslaughter, with which Bauberger was also charged. To be guilty of involuntary manslaughter, the defendant must have acted with “culpable negligence.” “Culpable negligence,” the court explained, requires either (1) a “willful, wanton, or intentional” violation of law governing the operation of a motor vehicle, or (2) an “inadvertent or unintentional violation of the law” that is “accompanied by recklessness of probable consequences of a dangerous nature when tested by the rule of reasonable foresight amounting altogether to a thoughtless disregard of consequences or a heedless indifference to the safety of others.” Trial Tr. Vol. VI at 21-22. The trial judge also instructed the jury on the meaning of “reckless” in the context of reckless driving, which required that the jury find Bauberger “acted carelessly and heedlessly in willful or wanton disregard of the rights or safety of others.” Id.

The North Carolina courts have clearly stated, as the majority acknowledges, “malice” for the purposes of murder requires “a high degree of recklessness.” State v. Rich, 527 S.E.2d 299, 303 (N.C.2000). However, even in contrast with the trial court's instruction regarding the meaning of “reckless” in the context of reckless driving, the dictionary definition-lack of due caution-sets a low standard. It seems clear that a person may have acted without due caution, but not necessarily in willful or wanton disregard of the rights or safety of others. Stated otherwise, there may be many circumstances in which a person may not have exercised proper caution (thereby acting recklessly as defined by the dictionary) but may not necessarily have acted with willful or wanton disregard for the rights or safety of others. Bauberger correctly points out that the dictionary's definition of recklessness, because it does not require that the disregard for others be willful or wanton, resembles the legal standard for negligence.

North Carolina courts have similarly interpreted wantonness, in the criminal context, as requiring more than a mere unintentional disregard of others. “Wantonness ․ connotes intentional wrongdoing․ Conduct is wanton when in conscious and intentional disregard of and indifference to the rights and safety of others.” State v. Williams, 199 S.E.2d 409, 412 (N.C.1973) (internal quotation marks and citation omitted) (emphasis added). Furthermore, “[t]he words ‘willful’ and ‘wanton’ have substantially the same meaning when used in reference to the requisite state of mind for a violation of a criminal statute.” State v. Davis, 356 S.E.2d 607, 610 (N.C.1987)(citing Williams 199 S.E.2d at 412). “ ‘Willful’ as used in criminal statutes means the wrongful doing of an act without justification or excuse, or the commission of an act purposely and deliberately in violation of the law.” Id. at 610. The dictionary in contrast defined “wanton” as the “arrogant recklessness of justice or the feelings of others.” App. 157. Again, the dictionary definition imposed a more lenient standard. Not every act which is in arrogant disregard of others will necessarily involve a conscious and intentional disregard of others. Rather, as noted above, the dictionary's definition, because it does not require that the disregard be conscious or intentional, resembles the lower standard of negligence.

The state argues, and the majority agrees, that even if the jury attached the lower standard associated with each of the term's dictionary definitions, when the terms are read in the context of the overall instruction for malice their insertion could only have had a minimal effect. The majority emphasizes that a reading of the instruction as a whole shows that the criminal act must still have been “intentionally done” and “manifest[ed] a mind utterly without regard for human life and social duty and deliberately bent on mischief.” In actuality, the majority by simply emphasizing that these words are present in the instruction, regardless of their placement, views them in isolation and out of context. In doing so, it ignores both the instruction's text and the facts of this case.

The instruction, on its face, has two separate parts. Under the instruction's terms, for an individual to have acted with malice, he must not only have committed the act intentionally, but also acted in willful or wanton disregard of the risk his actions posed, i.e., he must have been aware of the risk his actions entailed and disregarded them nonetheless. As noted, the dictionary definition of reckless-lack of due caution-requires no such awareness. Accordingly, as the instruction read to the jury, Bauberger must a) have committed an intentional act; and b) done so in spite of the fact a grave danger existed-whether or not he was aware of it. To simply conflate the instruction's two parts, and assume the jury likewise did so, ignores the instruction's plain construction.

Importantly, this conclusion is further undercut by the jury's actions. Had the jury imputed “intentionally” to the whole instruction, as the majority suggests, it is not clear why the foreman would have so felt the need to define the following scienter terms that he, on his lunch break, would have gone out of his way to obtain further assistance.

Likewise, pursuant to the third factor, there is strong reason to believe that the jury placed significant emphasis on the terms for which they garnered definitions. The relevant considerations under this factor include the number of jurors exposed to the potentially prejudicial information, at what point in their deliberations they received the violative material and the length of time they considered it. See, e.g., Fitzgerald, 150 F.3d at 366 (noting that a potentially biased member's impact was minimal as jury rejected her sentencing suggestion and they had made their decision before she made potentially prejudicial statements).

It is undisputed that several members of the jury requested further instruction from the judge on the definition of malice, of which the terms they looked up are part. Likewise, the parties do not dispute that once the foreman retrieved the dictionary, the definitions for the disputed terms were shared with all twelve of the jurors. It is relevant that it was the foreman who was responsible for obtaining the dictionary and sharing it with other jurors. That he was the party providing the definitions suggests that the other jurors may have attached additional weight to the external information. See Mayhue, 969 F.2d at 925. Finally, it is notable that the foreman obtained the definitions relatively early in the jury's deliberative process.

The state argues, and the majority agrees, the fact that the jury deliberated for more than four hours after acquiring the dictionary definitions before reaching a verdict indicates that the dictionary had little or no effect. While this is relevant, it is not as conclusive as the state presumes. The state misconstrues the appropriate question before the court. As noted above, the relevant question is merely whether one individual was affected by the dictionary. Accordingly, even if one vote shifted as a result of the dictionary's use, the violation had a prejudicial effect.

The fourth factor-the strength of evidence and the difficulty of the jury in reaching a verdict-weighs in favor of the state. As the majority notes, there is substantial evidence that the defendant acted with malice.

Nonetheless, there is evidence that the jury, despite the aforementioned evidence, struggled to decide whether malice existed. The jury requested additional clarification as to the meaning of the term “malice.” During lunch, the foreman took the unusual step of going to the library to acquire a dictionary to aid the jury in determining whether the evidence was sufficient to support a finding of malice. However, even after returning from lunch, receiving the additional materials requested and reviewing the dictionary, the jury needed multiple votes and an Allen charge before it found Bauberger guilty.

Considering all these factors, the question of whether the jury's consideration of the dictionary was prejudicial is a close question. Factors weighing in favor of a finding of prejudice are that the jury sought evidence directly relevant to the sole disputed legal issue, and the substantial difference between the terms as defined in the dictionary and as they are defined by North Carolina Courts. However, the substantial evidence justifying a finding of malice weighs against a finding of prejudice. Ultimately, my conclusion that the constitutional violation was prejudicial is based on two realities. First, while, as a legal matter, there may have been substantial evidence of malice, the jury obviously did not see the situation as such. As noted, the foreman felt it necessary to go to the library, and the other members of the jury likewise felt it necessary to request additional information regarding the legal rules to apply. Even once these pieces of information had been provided, only after an Allen charge and multiple votes did the jury reach its conclusion. Second, as noted above, the defendant, to garner a new trial, need not prove that the result was a proximate result of the impropriety, but merely that the evidence is in virtual equipoise that his trial may have been affected.

Given the centrality of the issue to the case, the substantial difference in the standard applied and that which was appropriate and the jury's struggles to resolve the issue, I, like the district court, find that Bauberger has raised grave doubt as to whether the jury's decision was in fact unbiased.

II. Conclusion

The Constitution's protections exist not only because they ensure that all persons are treated fairly and equally by our criminal justice system, but because they speak to the integrity and values of the nation and its residents. Therefore, when these rights are infringed, it hurts us all.

Today's majority opinion concerns me deeply. With it, we embark down a dangerous road which permits jurors to go outside of the record to determine guilt or innocence. Such conduct threatens to open a Pandora's box of unconstitutional behavior. It creates the potential for a system in which any juror may determine another human being's guilt on nothing more than a whim; a system more worthy of Kafka than the Constitution. With these thoughts in mind, I respectfully DISSENT from today's majority opinion.


1.  We therefore need not resolve the parties' dispute about whether we should “look through” the North Carolina Court of Appeals's decision (which did not conduct a harmless error analysis) to the MAR Court's decision (which did) to perform AEDPA/Chapman analysis. See Ylst v. Nunnemaker, 501 U.S. 797, 803 (1991) (habeas courts should look through summary affirmances to the “last reasoned decision” to determine whether a state procedural rule bars habeas review).

2.  The jurors also heard the definitions of “manifest” (“show”), “utterly” (“fully, totally”), and “regard” (“respect or consideration for”), and earlier in Bauberger's trial one juror looked up the definition of “malice” in a dictionary at his home but did not copy it down, recall it, or share it with others. Following the district court, Bauberger focuses on the prejudicial impact of the dictionary definitions for “recklessly” and “wantonly” on appeal.

3.  To the extent the dissent wishes to ascribe some larger significance to our holding, we should note simply that we have assumed for purposes of argument the existence of the error and rested decision upon the application of the Brecht standard to the facts and circumstances of the case. The deference due state court judgments of conviction upon collateral attack has of course not absolved the court of the need to conduct a careful review of the matter, and we thank our distinguished colleague in dissent for his willingness to do likewise.

1.  The state posits in its reply brief that it is not arguing that the Supreme Court must have decided an identical issue before a federal court may grant a petition for habeas relief, merely that the Supreme Court has not decided that looking up the words which the jurors looked up in this case violated the Sixth Amendment. This argument is obviously internally inconsistent. The state at once argues that it is not arguing that there must be an identical case and then complains that there is not an identical case.

2.  As the district court noted, “[t]he unreasonableness of a state court's application of ․ [the relevant] Supreme Court jurisprudence cannot be established by decisions of lower federal courts or state courts. [However], the analysis of these decisions is persuasive in the objective inquiry before the court.” Bauberger v. Haynes, 666 F.Supp.2d 558, 563 (M.D.N.C.2003) (internal citation omitted).

Reversed and remanded by published opinion. Judge WILKINSON wrote the majority opinion, in which Judge MOTZ joined. Senior Judge KEITH wrote a dissenting opinion.

Copied to clipboard