Bui v. DiPaolo

SELYA, Circuit Judge. This habeas case presents both procedural and substantive quandaries. Procedurally, we must determine the effect of a district court’s grant of a certificate of appealability on some, but not all, of the issues that a habeas petitioner seeks to pursue. Substantively, we must determine whether the petitioner has shown legal cause for us to set aside his state court convictions. I. BACKGROUND In November 1989, police discovered the stabbed bodies of a mother and daughter, Ngoc Le and Dixie Poulin. Evidence at the crime scene (the victims’ apartment) suggested that Dixie also had been bludgeoned with a blunt instrument, probably a gun, and that the perpetrator(s) likely had absconded with jewelry and cash. Several months later, Thinh Trinh, an acquaintance of the victims, informed the authorities that the petitioner, Tam Bui, had boasted of wielding a gun at the apartment around the time of the murders. Thinh’s wife, Linh Nguyen, substantiated the story, recalling that the petitioner told her that he had participated in the crime. As additional corroboration, Thinh and Linh produced two pieces of jewelry that they claimed the petitioner had given to Linh. The jewelry appeared to have belonged to Ngoc Le. In due course, the police obtained warrants for the petitioner and two other suspects. They first sought to find Bui at his parents’ fiat. He was not there, but police seized a .38 caliber handgun that Bui’s father stated belonged to his son. Shortly thereafter, the authorities apprehended the petitioner and charged him (and the other suspects). At trial in a Massachusetts state court, the Commonwealth’s most incriminating evidence consisted of (i) certain statements that the petitioner had made to police after his arrest, (ii) the testimony of Thinh and Linh, and (iii) the handgun (which the Commonwealth hypothesized had been used to bludgeon Dixie Poulin). The jury convicted the petitioner on two counts of first-degree murder and one count of armed robbery. The court sentenced him to serve two consecutive terms of life imprisonment. The Massachusetts Supreme Judicial Court (SJC) affirmed the convictions. See Commonwealth v. Bui 419 Mass. 392, 645 N.E.2d 689 (1995) (Bui I). The petitioner then sought habeas corpus relief in the federal district court, see 28 U.S.C. § 2254, naming as respondents two state officials. His petition enumerated several purported constitutional faults. The district court found the claims of error unpersuasive and declined to disturb the state court judgment. See Bui v. DiPaolo, 985 F.Supp. 219 (D.Mass.1997) (Bui II). Because Bui filed his application for habe-as corpus relief in December 1996, the Anti-terrorism and Effective Death Penalty Act (AEDPA), Pub.L. No. 104-132, 110 Stat. 1214, 1217-18 (Apr. 24, 1996) (codified in scattered sections of 28 U.S.C.), governs this case. See Lindh v. Murphy, 521 U.S. 320, 117 S.Ct. 2059, 2068, 138 L.Ed.2d 481 (1997) (holding that, in general, AEDPA’s provisions apply to cases filed after its effective date). Under the AEDPA amendments, a habeas petitioner can prosecute an appeal only if he first obtains a certificate of appeal-ability (COA). See 28 U.S.C. § 2253(c). In this instance the district court, after denying the application for a writ of habeas corpus, certified one issue for appeal — the petitioner’s Sixth Amendment claim that the trial court impermissibly thwarted his lawyer’s attempts to cross-examine Thinh — but declined *236 to certify the petitioner’s Fifth Amendment claim. Undaunted, the petitioner sought to proceed in this venue on both claims. 1 Because we had not had occasion to resolve the question of whether COAs are to be treated as case-specific or issue-specific, we permitted the petitioner to brief his Fifth Amendment claim on the merits, but required that he simultaneously brief the antecedent procedural question. Consequently, we begin by erecting a procedural framework for the handling of COAs in multi-issue cases. We then discuss the petitioner’s two substantive contentions. II. CERTIFICATES OF APPEALABILITY Bui and the amici argue that we are obligated to mull the merits of his Fifth Amendment claim because, in their view, COAs issued under the AEDPA-spawned habeas amendments should be treated as case-specific rather than issue-specific. In other words, they contend that the grant of a COA on any one issue opens all issues in the case to full appellate review. Although this result would have been consistent with practice as it existed before the AEDPA amendments, see Magouirk v. Phillips, 144 F.3d 348, 356 (5th Cir.1998); Tejeda v. Dubois, 142 F.3d 18, 22 n. 4 (1st Cir.1998), we think it is now outmoded. Congress, in enacting the AEDPA, meant to change prior practice and succeeded in doing so: the language and structure of the amended habeas statute pretermit appellate consideration of claims not properly certified for appeal. The AEDPA predicates the very issuance of a COA — without which “an appeal may not be taken to the court of appeals,” 28 U.S.C. § 2253(c)(1) — on whether an “applicant has made a substantial showing of the denial of a constitutional right.” Id. at § 2253(c)(2). A habeas petitioner who fails to demonstrate that his claims satisfy the substantial showing standard may not appeal the denial of habeas corpus at all. This rule is easily applied in situations in which a habeas court’s decisions on the merits and on the availability of a COA are congruent. Thus, if a habeas application raises only issues that, in the district court’s opinion, warrant neither relief nor a COA no appeal lies. Conversely, if the application raises only issues that, according to the district court, uniformly pass muster under the substantial showing standard and, thus, are certifiable (even though in the district court’s view they do not justify relief), then a free-ranging appeal lies. In these examples, the COA itself is but a variable that depends wholly on the existence of a substantial showing that the petitioner’s constitutional rights have been abridged. The question at hand is whether this relationship is altered in a situation that lacks essential congruence, that is, when the district court, having denied a habeas application, deems some, but fewer than all, of the petitioner’s claims certifiable. We believe that the necessity for a substantial showing extends independently to each and every issue raised by a habeas petitioner. Had the mere existence of a COA been sufficient to qualify all the issues in a habeas case for appellate review, it would have been pointless for Congress to go on to state, in section 2253(e)(3), that a COA “shall indicate which specific issue or issues satisfy” the substantial showing standard of section 2253(c)(2). This language was absent from section 2253’s previous incarnation, and *237 the fact that Congress added it in the AED-PA signals an intent that the courts should accord some significance to it. The petitioner’s reading of the statute would render this phrase superfluous, for as long as one issue were certifiable, all others would be amenable to review. We customarily read statutes in a manner that gives effect to all words and phrases, see Walters v. Metropolitan Educ. Enters., Inc., 519 U.S. 202, 117 S.Ct. 660, 664, 136 L.Ed.2d 644 (1997); United States v. Ven-Fuel, Inc., 758 F.2d 741, 751-52 (1st Cir.1985), and we see no valid reason to deviate from this salutary practice here. Common sense also suggests that such an interpretation should be preferred. Were we to adopt the petitioner’s construction of the statute, we would establish a paradigm under which tenuous habeas claims could not be appealed if asserted on their own, but could command appellate attention if accompanied by a solitary claim that met the substantial showing standard. This dichotomy lacks rhyme or reason, and there is no indication that Congress intended to create so curious a structure. The better reading of the habeas amendments is one that links section 2253(e)(3)’s insistence on an issue by issue enumeration of what has been certified for appeal with the substantial showing requirement of section 2253(c)(2). We hold, therefore, that a court of appeals should not consider the merits of an issue advanced by a habeas petitioner unless a COA first has been obtained with respect to that issue. Accord Ramsey v. Bowersox, 149 F.3d 749, 759 (8th Cir.1998), cert. denied, — U.S. -, 119 S.Ct. 1083, 143 L.Ed.2d 85 (1999); Murray v. United States, 145 F.3d 1249, 1250-51 (11th Cir.1998); Sylvester v. Hanks, 140 F.3d 713, 715 (7th Cir.1998); Lackey v. Johnson, 116 F.3d 149, 151 (5th Cir.1997); In re Certificates of Appealability, 106 F.3d 1306, 1308 (6th Cir.1997). Were the power to issue a COA exclusively a prerogative of the district courts, this holding might answer the procedural question that confronts us. Under the AED-PA amendments, however, a habeas petitioner who fails in his effort to obtain a COA in the district court may then beseech the court of appeals to issue one. See Fed. R.App. P. 22(b)(1) (“If the district judge has denied the certificate, the applicant for the writ may then request issuance of the certificate by a circuit judge.”). The procedure for cases in which the district court does not grant a COA at all is uncontroversial: even if there is no express request for a COA, the court of appeals must deem a notice of appeal to represent a request for a COA on all issues raised. See Fed. R.App. P. 22(b)(l)-(2). The rules are silent, however, with respect to the appropriate procedure in hybrid eases. Thus, we must decide what course a habeas petitioner should follow when the district court certifies some, but not all, of the issues that he wishes to vent on appeal. If the district court issues a limited COA (i.e., a COA that permits the petitioner to go forward on fewer than all of the issues raised in his petition), the petitioner must present promptly to the court of appeals an express request for a complementary COA. Although this court’s local rules do not definitively resolve the timing issue, it is our present practice to notify each habeas petitioner of an outside date by which his request for a COA must be filed in the court of appeals. Failure to file within that interval may be deemed to constitute a waiver. The request for a complementary COA must be explicit as to the issues which the petitioner wishes us to consider and, in such circumstances, we will not treat an inexplicit notice of appeal, without more, as an invitation to review the district court’s denial of certification with regard to the petitioner’s remaining claims. In taking this stance, we adopt the rule stated in United States v. Kimler, 150 F.3d 429, 430 (5th Cir.1998) (declaring that a notice of appeal in a multi-issue habeas case will not be treated as a constructive request for expanding a limited COA issued by the district court), and reject the rule stated in Kincade v. Sparkman, 117 F.3d 949, 953 (6th Cir.1997) (treating a notice of appeal in a hybrid case as “an informal request with this Court to review the district Court’s denial as to [the petitioner’s] remaining claims”). We make one related point: the request for a complementary COA, *238 whether contained in a separate document or incorporated into the notice of appeal, should be accompanied by a copy of the district court’s order and a memorandum giving specific and substantial reasons (not mere generalizations) why a certificate should be granted. In the interests of fairness and judicial economy, we also rule that, when the district court grants a limited COA and the petitioner seasonably seeks to expand it, his appeal on the certified issue(s) should be held in abeyance, and full briefing deferred, until this court determines the appealability of the issues that the district court deemed unworthy of appellate scrutiny (and, thus, whether we will issue a complementary COA). Accord Kinder, 150 F.3d at 431 n. 1; Kincade, 117 F.3d at 953; United States v. Simmonds, 111 F.3d 737, 740-41 (10th Cir.1997). As an administrative measure, we advise litigants that, to the extent practicable, the panel that determines whether to issue a complementary COA also will be the panel that adjudicates the appeal on the merits. 2 We are aware that a few cases are pending in which, as here, a motions judge has adopted an ad hoc framework for handling situations in which a district court, over a habeas petitioner’s objection, issued a limited COA. Such actions were entirely appropriate in the absence of a definitive resolution of the attendant procedural questions. We will, of course, allow those appeals to proceed in the ordinary course, under the terms of the provisional orders previously entered. In future cases, however, we will require habeas petitioners to walk the procedural path that we set in place today. We regard the instant case as grandfathered. III. THE FIFTH AMENDMENT CLAIM When the district court issued a COA restricted to the Sixth Amendment issue, the petitioner’s able counsel timely filed a notice of appeal and contemporaneously submitted an application to the district court to enlarge the COA to encompass the Fifth Amendment claim. When no amended certificate emerged, counsel sought a complementary COA in this court. We entered an order, previously described, directing the petitioner, inter alia, to brief his Fifth Amendment claim on the merits. Reading the new habe-as amendments literally, we next should consider whether the petitioner has made a “substantial showing of the denial of a constitutional right” in respect to this Fifth Amendment point. 28 U.S.C. § 2253(c)(2). Here, however, under the terms of our interim order, the parties have fully briefed and argued the merits of that claim. Since we decline to take two steps where one will do, we skip the preliminary “substantial showing” inquiry and turn directly to the merits. We caution, however, that this expedient departs from the norm, and we employ it here solely because of the terms of the interim order. The petitioner’s Fifth Amendment claim has its genesis in a 30-minute conversation between him and the police conducted shortly after his arrest. The facts surrounding the interview, derived primarily from a police report, are not in dispute. When the authorities arrested the petitioner on August 18, 1990, they gave him Miranda warnings in his native language (Vietnamese). The petitioner acknowledged that he understood his rights. Before interviewing the petitioner, the police again advised him of his rights, this time in both English and Vietnamese. In response, the petitioner proclaimed that he knew his rights, that “my Constitution will protect me,” and that “you have nothing.” When an officer asked if he had anything to say about why he was being arrested, the petitioner replied in the negative. Then— before the officer posed another question— he inquired: “Who said I did this?” The petitioner thereafter answered some questions, while brushing others aside. The in *239 terview ended when he declined to respond to a query asking what he had “to say about what happened around [the time of the murders].” After the state trial court had refused to suppress the interview, a number of the petitioner’s statements were admitted into evidence. For example, the Commonwealth introduced the petitioner’s comment about the gun that had been found in his parents’ abode and his remark to the police that “you have nothing.” On direct appeal, the SJC upheld the admissibility of these statements. See Bui I, 419 Mass, at 396-97, 645 N.E.2d at 692-93. In considering the petitioner’s ensuing habeas application, the district court found no unequivocal assertion of the right to remain silent and dismissed the Fifth Amendment claim on the authority of Davis v. United States, 512 U.S. 452, 114 S.Ct. 2350, 129 L.Ed.2d 362 (1994). See Bui II, 985 F.Supp. at 226-28. 3 Seizing on the fact that Davis involved only the right to counsel—in Davis, 512 U.S. at 458-59, 114 S.Ct. 2350, the Court held that if an accused wishes to assert a right to counsel, he must so state unambiguously — the petitioner posits that Davis is inapplicable where, as here, a defendant anchors a Fifth Amendment claim upon the right to remain silent. This means, he says, that we must grant habeas relief because Davis does not comprise “clearly established Federal law, as determined by the Supreme Court of the United States,” within the purview of 28 U.S.C. § 2254(d)(1). The petitioner’s attempt to characterize Davis as something other than “clearly established Federal law” rests on a mistaken premise: the idea that, merely because Davis does not directly compel the particular outcome reached by the state court, a federal habeas court perforce must issue the writ. This is a non sequitur. Under the AEDPA amendments, federal courts review state court judgments to determine whether those judgments construe or apply federal law in a manner that is “contrary to” or an “unreasonable application of’ the Supreme Court’s “clearly established” jurisprudence. See O’Brien v. Dubois, 145 F.3d 16, 24 (1st Cir.1998) (citing statutory language). The question, then, is whether the Massachusetts courts’ rulings in this case contravene the Court’s precedents or represent an unreasonable application of them. If Davis is the relevant precedent, we do not see how we can disturb the judgment. To be sure, Davis concerned only the right to counsel recognized in Miranda v. Arizona, 384 U.S. 436, 86 S.Ct. 1602, 16 L.Ed.2d 694 (1966). Nonetheless, every circuit that has addressed the issue squarely has concluded that Davis applies to both components of Miranda: the right to counsel and the right to remain silent. See United States v. Mills, 122 F.3d 346, 350-51 (7th Cir.) (citing United States v. Banks, 78 F.3d 1190, 1196-97 (7th Cir.1996)), cert. denied, - U.S. -, 118 S.Ct. 637, 139 L.Ed.2d 615 (1997); Medina v. Singletary, 59 F.3d 1095, 1100-01 (11th Cir.1995); United States v. Johnson, 56 F.3d 947, 955 (8th Cir.1995). 4 For the purposes of habeas corpus review, we simply cannot deem unreasonable a conclusion by the Massachusetts courts that is consistent with the approach taken by so many respected tribunals. In an effort to avoid the force of this reasoning, the petitioner maintains that Davis is inapposite to Miranda’s right to remain silent. Although the similarity of the analyses requisite for assessing claims anent Miranda’s right to counsel and its right to remain silent suggests that Davis constitutes strong evidence of how the Supreme Court likely would decide this right to remain silent question, we acknowledge that Davis does not “authoritatively” answer the question in the narrow, technical sense of that term. We therefore assume for argument’s sake that *240 Davis is not the relevant precedent. 5 Even on this assumption, the petitioner’s challenge to the state court judgment fares poorly. As the petitioner sees it, his claim implicates a line of authority which states that, as long as a criminal defendant articulates his Fifth Amendment rights in even an equivocal manner, police interrogation must cease immediately. See Michigan v. Mosley, 423 U.S. 96, 103-04, 96 S.Ct. 321, 46 L.Ed.2d 313 (1975); Quinn v. United States, 349 U.S. 155, 162-64, 75 S.Ct. 668, 99 L.Ed. 964 (1955); United States v. Barone, 968 F.2d 1378, 1384 (1st Cir.1992); see also Wayne R. LaFave & Jerold H. Israel, 1 Criminal Procedure § 6.9(e), 531-33 (1984 & Supp.1991) (discussing implied assertion of Miranda rights). Under this legal rubric, the petitioner argues, his statement to the effect that “my Constitution will protect me” and his negative response to the officer’s query anent whether “he had anything to say about what he was being arrested for” were, at the very least, equivocal assertions of the right to remain silent, and, therefore, precluded the police from doing more than following up with clarifying questions. See United States v. March, 999 F.2d 456, 461 (10th Cir.1993) (collecting cases). Since the police went further, the petitioner reasons, the trial court should have suppressed all the statements. Setting to one side the question of whether the doctrinal underpinnings of the petitioner’s argument retain much integrity after Davis, we believe that the facts of this case are controlled by a different line of authority. In North Carolina v. Butler, 441 U.S. 369, 373, 99 S.Ct. 1755, 60 L.Ed.2d 286 (1979), the Court made it pellucid that an express statement is not invariably necessary to support a finding that the defendant waived either the right to remain silent or the right to counsel. We, too, have employed this implied waiver doctrine, see United States v. Garcia, 983 F.2d 1160, 1169 (1st Cir.1993), and it appears to be the primary ground on which the SJC resolved the petitioner’s Fifth Amendment claim, see Bui I, 419 Mass, at 397, 645 N.E.2d at 692-93. Precisely when waiver may be implied depends on the circumstances. Even so, there are certain types of cases in which courts routinely conclude that a defendant who has professed an understanding of his right to remain silent has waived that right. For example, a defendant will be held to have effected a waiver when, after receiving warnings and asserting (equivocally or unequivocally) a right to remain silent, he spontaneously recommences the dialogue with his interviewers. See Edwards v. Arizona, 451 U.S. 477, 484-85, 101 S.Ct. 1880, 68 L.Ed.2d 378 (1981); United States v. Conley, 156 F.3d 78, 83 (1st Cir.1998). So, too, if a defendant’s incriminating statements were made either as part of a “steady stream” of speech, Bradley v. Meachum, 918 F.2d 338, 342 (2d Cir. 1990), or as part of a baek-and-forth conversation with the police, Baskin v. Clark, 956 F.2d 142, 146 (7th Cir.1992), courts regularly have found waivers. A waiver of Miranda rights also may be implied when, after having received Miranda warnings, a criminal defendant responds selectively to questions posed to him. See United States v. Soliz, 129 F.3d 499, 503 (9th Cir.1997); United States v. Eaton, 890 F.2d 511, 513-14 (1st Cir.1989) (Breyer, J.); United States v. Chong, 829 F.2d 1572,1574 (11th Cir.1987). The petitioner’s interview displays characteristics that fit the implied waiver profile. In his initial statements, uttered immediately after the second round of Miranda warnings, the petitioner proclaimed that he knew his rights and that the Constitution would protect him. The petitioner’s claim to the contrary notwithstanding, these comments cannot plausibly be construed to suggest that the speaker intended to assert Miranda rights. In all events, the spontaneous utterance “you have nothing,” made immediately after those comments and without any semblance of police provocation, is not protected by the Fifth Amendment. See *241 Conley, 156 F.3d at 83; Baskin, 956 F.2d at 146. Up to this point, the dialogue has the distinct flavor of a baek-and-forth conversation between the petitioner and his interviewer. See, e.g., Baskin, 956 F.2d at 146. An examination of the remainder of the interview lends credence to this impression. Once the petitioner boasted that the police had nothing, the interviewer legitimately followed up by asking him whether he had something to say about the reasons for his arrest. The petitioner answered “no”—and this is a point at which he now avers the police should have stopped their questioning. He might have had a stronger argument had he not added anything to his response, see Conley, 156 F.3d at 83, but he then immediately asked the police: “who said I did this?” The petitioner’s continuation of the conversation by means of this question both negated any effect his prior response may have had as an assertion of his right to remain silent and confirmed that the interview was indeed a back-and-forth exchange. See Baskin, 956 F.2d at 146. To say more on this point would be supererogatory. The short of it is that the petitioner did not unequivocally assert his right to remain silent, and, moreover, selectively responded to questions during what otherwise appears to have been a fairly standard two-way conversation with the authorities. The prosecution thereupon introduced those responses at trial. If the relevant precedent here is Davis, the petitioner cannot prevail because he made no clear articulation of a desire to stand mute. If, however, Davis does not apply, the relevant precedent is Butler, and federal courts have on numerous occasions employed its implied waiver principle to uphold the admissibility of evidence under circumstances similar to those presented in this case. See, e.g., Ramirez, 79 F.3d at 305; Baskin, 956 F.2d at 146; Bradley, 918 F.2d at 342; see generally LaFave & Israel, supra § 6.9(d), at 530-31 (discussing implied waivers). Either way, the SJC’s application of federal law (and, for that matter, the state trial court’s determination of subsidiary facts) does not fall outside the sphere of acceptability. See O’Brien, 145 F.3d at 25. IV. THE SIXTH AMENDMENT CLAIM We pause to set the petitioner’s Sixth Amendment claim into perspective. Thinh was one of two persons who testified that the petitioner admitted involvement in the murders, and one of three who described him brandishing a knife and suggesting that he had used it to commit a crime. In addition, Thinh recounted important information concerning the petitioner’s possession of certain items of jewelry ostensibly belonging to Ngoc Le. See Bui I, 419 Mass, at 398-99, 645 N.E.2d at 693 (summarizing Thinh’s testimony). The petitioner had hoped to impeach Thinh’s credibility by developing, inter alia, a rather intricate theory of bias. In the petitioner’s scenario, Thinh and Linh were coerced into testifying falsely by Linh’s father, Mong Nguyen, a drug kingpin, who took this action because the petitioner had refused, in the presence of Mong’s peers (equally notorious drug lords), to transport drugs. According to the petitioner, extreme retributive tendencies characterize interpersonal relations in Vietnamese society, and Thinh’s rejection so affronted Mong that he exacted revenge by directing his daughter and quondam son-in-law to frame Bui. At trial, the petitioner’s counsel first attempted to gather support for this hypothesis by questioning Linh on voir dire, outside the jury’s presence. He drew a blank. He revisited the subject with Thinh, but the trial judge sustained the prosecution’s objections to his initial round of questions. Counsel thereupon made an offer of proof. The judge remained steadfast and refused, on relevancy grounds, to allow the attorney to ask the proposed questions. After the jury convicted the petitioner, the SJC affirmed this evidentiary ruling, finding the petitioner’s theory of bias too speculative to justify cross-examination. See id. at 401, 645 N.E.2d at 694-95. The Confrontation Clause places a high value on the right to cross-examination. See Delaware v. Van Arsdall, 475 U.S. 673, 678-79, 106 S.Ct. 1431, 89 L.Ed.2d 674 (1986). That right is witness-specific. See id. at 680, 106 S.Ct. 1431. Furthermore, a criminal defendant’s entitlement to cross-examine a witness increases in sensi *242 tivity in direct proportion to the witness’s importance to the prosecution’s case. See United States v. Mizell, 88 F.3d 288, 292-93 (5th Cir.1996); United States v. Taylor, 17 F.3d 333, 340 (11th Cir.1994); see also United States v. Tracey, 675 F.2d 433, 437-38 (1st Cir.1982) (collecting cases). Given Thinh’s status as a key figure in the prosecution’s case, it is readily apparent that the petitioner’s claim demands close attention. The petitioner exhorts us to hold that the Massachusetts courts violated his rights under the Confrontation Clause by prohibiting all inquiry into a cogent bias theory. Citing eases such as United States v. Lynn, 856 F.2d 430, 433 (1st Cir.1988), he maintains that the trial court’s actions violated the general rule requiring courts to permit a threshold level of cross-examination on bias and, thus, the SJC’s holding was “contrary to” clearly established Supreme Court precedent. We reject this plaint. Speaking in the large, the jurisprudence of the Confrontation Clause rarely lends itself to Procrustean application, see O’Brien, 145 F.3d at 26, and the case law simply does not state the “threshold level of inquiry” rule in terms as absolute as the petitioner would have it. Indeed, the Van Arsdall Court explicitly recognized that, under certain circumstances, trial judges possess broad discretion to limit cross-examination. See Van Arsdall, 475 U.S. at 679, 106 5.Ct. 1431 (enumerating various factors that may justify circumscription of cross-examination). Thus, to the extent that the petitioner is suggesting that a criminal defendant has license to cross-question a prosecution witness concerning every conceivable theory of bias, regardless of the prevailing circumstances, he is plainly wrong. The threshold requirement imposed by the Confrontation Clause is satisfied as long as the defendant is given a fair chance to inquire into a witness’s bias. See United States v. Sinclair, 109 F.3d 1527, 1537 (10th Cir.1997); United States v. Guthrie, 931 F.2d 564, 568-69 (9th Cir.1991); United States v. Boylan, 898 F.2d 230, 254-55 (1st Cir.1990). A criminal defendant is not regarded as being deprived of that chance if a trial court legitimately determines that his cross-examination is inappropriate. See Van Arsdall, 475 U.S. at 679, 106 S.Ct. 1431. With these principles in mind, we cannot say that the Massachusetts courts acted contrary to established Supreme Court case law in regard to this matter. The petitioner cross-examined Thinh at length with respect to bias, and the trial court based its restriction of further cross-examination on relevancy grounds — something Van Arsdall, in principle, expressly permits. Thus, we must test the propriety of the Massachusetts courts’ conclusions by asking whether forbidding further cross-questioning represented an unreasonable application of the Supreme Court’s Confrontation Clause precedents. 6 See 28 U.S.C. § 2254(d)(1). In itself, this is not a surprising development: when a constitutional claim requires balancing rather than applying a strict, one-dimensional legal rule, that ordinarily is an indication that a habeas court, in the end, will have to gauge the reasonableness of the state court’s application of the rule. See, e.g., Holman v. Gilmore, 126 F.3d 876, 881-82 (7th Cir.1997) (stating that “when the constitutional question is a matter of degree, rather than of concrete entitlements, a ‘reasonable’ decision by the state court must be honored”), cert. denied, - U.S. -, 118 S.Ct. 1169, 140 L.Ed.2d 179 (1998). Before turning to specifics, we reiterate the narrow circumference of appellate review in connection with the “reasonable application” standard. Under this formulation, a federal court may not “grant the writ merely because it disagrees with the state court’s decision, or because, left to its own devices, it would have reached a different result.” O’Brien, 145 F.3d at 25; accord Corwin v. Johnson, 150 F.3d 467, 471-72 (5th Cir.), cert. denied, — U.S. -, 119 S.Ct. *243 613, 142 L.Ed.2d 548 (1998); Neelley v. Nagle, 138 F.3d 917, 924 (11th Cir.1998), cert. denied, — U.S. -, 119 S.Ct. 811, 142 L.Ed.2d 671 (1999); Hennon v. Cooper, 109 F.3d 330, 334-35 (7th Cir.), cert. denied, — U.S. -, 118 S.Ct. 72, 139 L.Ed.2d 32 (1997). The question, rather, is whether the state court has applied federal law in a manner that is “so offensive to existing precedent” or so lacking in foundation that its conclusion is “outside the universe of plausible, credible outcomes.” O’Brien, 145 F.3d at 25. As suggested by our discussion of the Fifth Amendment claim, what this standard implies is that we cannot deem a state court’s application of federal law unreasonable when federal courts — typically, the circuit courts of appeals — regularly resolve cases in a similar fashion. See Williams v. Taylor, 163 F.3d 860, 869 (4th Cir.1998); Davis v. Johnson, 158 F.3d 806, 813-14 (5th Cir.1998), petition for cert. filed (U.S. Feb. 23, 1999) (No. 98-8209); O’Brien, 145 F.3d at 25; cf. Lambrix v. Singletary, 520 U.S. 518, 117 S.Ct. 1517, 1530, 137 L.Ed.2d 771 (1997) (engaging in a similar analysis in a pre-AEDPA habeas proceeding). Another dimension of the problem deserves mention. In order for a federal court to conclude that a state court has applied federal law reasonably, it is not necessary that the federal court agree with every last detail of the state court’s analysis. By like token, state courts are not required to supply the specific reasons that a federal court thinks are most persuasive for upholding the judgment. Rather, it suffices that the state court generally articulates and applies tenets that can reasonably sustain its judgment. Thus, the question on habeas corpus review is whether the state courts’ invocation and reliance on enunciated legal principles are reasonably consistent with the facts, circumstances, and outcome of the case. See, e.g., Hennon, 109 F.3d at 335 (“It doesn’t follow that the criterion of a reasonable determination is whether it is well reasoned. It is not. It is whether the determination is at least minimally consistent with the facts and circumstances of the case.”). Were we to insist on perfect jurisprudential harmony, not only would we risk defying Congress’s will by transforming the AEDPA standard into one of plenary review, but we also would place ourselves in a position in which we might invalidate convictions not because a “defendant may have been the victim [ ] of [a] constitutional error but merely of a failure of judicial articulateness.” Id. The logic of the AEDPA is flatly inconsistent with such an approach. Mindful of these background principles, we turn to the SJC’s determination that the petitioner was not entitled to inquire along the line that he proposed. As we already have observed, cross-examination is subject to reasonable restrictions. One well-established basis for circumscribing cross-examination is a party’s inability to lay a proper evidentiary foundation for the questions that he wishes to pose. See, e.g., United States v. Ovalle-Marquez, 36 F.3d 212, 218-19 (1st Cir. 1994); United States v. Carty, 993 F.2d 1005, 1010 (1st Cir.1993). The SJC relied upon this principle to conclude that the petitioner’s Confrontation Clause claim lacked merit. See Bui I, 419 Mass, at 401, 645 N.E.2d at 694-95 (holding that the petitioner’s offer of proof was too tenuous to justify the desired inquiry). This brings us to the offer of proof. It prophesied that Thinh, if allowed to answer the proposed questions, would acknowledge that the petitioner rebuffed Mong’s entreaty to ferry narcotics from California to Boston. Defense counsel purposed to argue from this response that the repudiation so angered Mong that he directed Thinh to frame the petitioner. But to that point in the trial, there had been no evidence even remotely supporting the petitioner’s offer of proof. When the court inquired into the predicate for the offer, counsel did not cite any evidence, but replied instead that he had a good-faith basis founded on attorney-client privilege. 7 On this exiguous record, the *244 mere invocation of the petitioner as the source was not enough to negate the inherent speculativeness of the offer, especially since the petitioner neither testified at trial nor submitted an affidavit as part of the proffer. Even were we to assume, as did the state trial court, that petitioner had made a good-faith proffer, we still could not conclude that the SJC’s determination that the offer of proof was unsatisfactory so offended existing Supreme Court precedent as to be unreasonable. Apart from its foundational infirmities, the petitioner’s offer of proof was inherently defective because it did not include crucial elements of his bias theory. He sought to ask Thinh three questions: (1) Was Mong present in California while Thinh and the petitioner were there? (2) Did Mong ever ask the petitioner to transport drugs to Boston? (3) If so, did the petitioner refuse? However, we agree with the SJC that the petitioner’s theory of bias had more components than could be addressed by responsive answers to these three questions. See Bui I, 419 Mass, at 401-02, 645 N.E.2d at 695. It ultimately depended on demonstrating facts relating to matters such as the relationship between Mong and the petitioner, the relationship between Mong and Thinh, the presence (and identity) of others in the course of the alleged repudiation, and the tenor of Mong’s reaction. Neither the proposed questions nor the petitioner’s offer of proof touched upon these issues. On this sparse record, both the trial court’s limitation on cross-examination and the SJC’s affirmance on the ground that the offer of proof was too tenuous were in conformity with federal court precedents requiring a satisfactory evi-dentiary foundation for cross-questioning, see, e.g., Ovalle-Marquez, 36 F.3d at 218-19; Catty, 993 F.2d at 1010, and thus neither decision implicated the AEDPA standard. United States v. Lynn, 856 F.2d 430, much relied upon by the petitioner, does not demand a different result. If anything, the facts of Lynn highlight one of the main shortcomings of the petitioner’s offer of proof. Lynn exemplifies those cases in which courts have disapproved restrictions on questioning about bias when cross-examination seeks to probe whether a prosecution witness will get preferential treatment from law enforcement in return for testifying. In these instances, courts have been especially open to permitting cross-examination about agreements between government and witness. In the same vein, courts have recognized, albeit less frequently, that similarly distorting influences may arise from other innately coercive relationships. See, e.g., United States v. Abel, 469 U.S. 45, 52, 105 S.Ct. 465, 83 L.Ed.2d 450 (1984) (discussing witness’s membership in a gang or secret group as productive of coercion). Whatever the source of the alleged coercion, however, courts traditionally have required a solid showing that coercion is present. See Catty, 993 F.2d at 1010. Where there is no evidence of a coercive dimension in the relationship between a witness and an organization, trial courts cast well within their discretion in prohibiting cross-examiners from mounting fishing expeditions. See Sinclair, 109 F.3d at 1537; United States v. Elkins, 70 F.3d 81, 83-84 (10th Cir.1995); United States v. Warren, 18 F.3d 602, 603 (8th Cir.1994); United States v. Piche, 981 F.2d 706, 715-16 (4th Cir.1992). Here, the singularly essential element of the petitioner’s theory—a relationship of coercion between a person purportedly interested in the outcome of the trial (Mong) and the witness (Thinh)—utterly lacked an evidentiary foundation. 8 Indeed, the petitioner made no attempt to address this element in his offer of proof or contemporaneously in some other fashion. 9 Under these circumstances, we do *245 not believe the petitioner can rely on Lynn to undo the SJC’s conclusions. In performing our Confrontation Clause analysis in this case, we find the opinion in United States v. Lin, 101 F.3d 760 (D.C.Cir. 1996), to be particularly instructive because that case presented circumstances comparable to those that surround this case. There, the defendant sought to assail the government’s witnesses by suggesting that they were testifying against him because they worked for a competing gambling parlor and wanted him out of the way. When Lin attempted to cross-examine one witness by inquiring into his gambling activities, the government objected and the district court requested a proffer. Defense counsel made an offer of proof that limned his theory of bias and identified his client as the source of the allegations. The trial court eventually sustained the government’s objection. The court of appeals affirmed, upholding the trial court’s decision not to allow the initiation of “a highly prejudicial line of cross-examination,” in circumstances in which the questioner did not show that he possessed facts sufficient to “support a genuine belief that the witness committed the offense or the degrading act to which the questioning relates.” Id. at 768 (citation and internal quotations marks omitted). Like Bui, Lin elected not to testify, and the court viewed his proffer as “too ambiguous” to indicate that the proposed questioning had any realistic potential for showing bias. Many of the principles that animated Lin are pertinent to the case at hand. First and foremost, Lin, like so many of the precedents we have culled, holds that cross-examination presupposes the existence of a proper foundation. Offers of proof that do not carry the promise of demonstrating bias do not satisfy this requirement; indeed, such offers may be considered irrelevant. See Carty, 993 F.2d at 1010. Furthermore, in Lin as in this case, the trial court left open the possibility that witnesses could be recalled and the line of inquiry reopened if the defense subsequently made a better proffer. Like Lin, Bui never returned with a beefed-up presentation. Finally, as Lin suggests, questions that implicate a witness’s involvement in elaborate conspiracies to commit illegal acts have the potential to be prejudicial and misleading. And when the potential for this kind of prejudice exists, the need to insist that cross-examination be based on a solid foundation is heightened. In an effort to deflect this logic, the petitioner seizes upon the trial judge’s comment that the questions his lawyer sought to ask were irrelevant (and thus inadmissible). This assessment was misguided, the petitioner says, because the general issue of bias is always relevant. See Davis v. Alaska, 415 U.S. 308, 316, 94 S.Ct. 1105, 39 L.Ed.2d 347 (1974). We are unimpressed by this wordplay. In context, we interpret the trial judge’s relevancy ruling to mean that, given the scantiness of the record evidence and the incomplete offer of proof, the petitioner’s proposed questions were not sufficiently probative of his particular theory of bias, and, therefore, not relevant to that issue. Let us be perfectly clear. We do not mean to suggest that a criminal .defendant must testify in his own behalf, or alternatively, that he must arm himself with irrefutable proof before he may cross-examine witnesses about a particular bias. We do believe, however, that a defendant in such circumstances must present a satisfactory foundation for the critical elements on which his hypothesis of bias depends. In this instance, the petitioner, having been accorded a fair opportunity to inquire into Thinh’s bias, failed to *246 meet that criterion. Hence, we cannot fault the trial judge’s decision to bar the petitioner from asking Thinh all the specified questions. See Delaware v. Fensterer, 474 U.S. 15, 20, 106 S.Ct. 292, 88 L.Ed.2d 15 (1985) (stating that the “Confrontation Clause guarantees an opportunity for effective cross-examination, not cross-examination that is effective in whatever way, and to whatever extent, the defense might wish”); Boylan, 898 F.2d at 254 (same). We add a coda. Even under the regime of Teague v. Lane, 489 U.S. 288, 109 S.Ct. 1060, 103 L.Ed.2d 334 (1989), constitutional law was not to be considered the “exclusive province” of inferior federal courts, Caspari v. Bohlen, 510 U.S. 383, 395, 114 S.Ct. 948, 127 L.Ed.2d 236 (1994), and such courts, in the absence of a contrary, established rule announced by the Supreme Court, were obliged to “validate[ ] reasonable, good-faith interpretations of existing precedents made by state courts even though they [were] shown to be contrary to later decisions.” Butler v. McKellar, 494 U.S. 407, 414, 110 S.Ct. 1212, 108 L.Ed.2d 347 (1990). That approach has been underscored by the AEDPA standard of review. See O’Brien, 145 F.3d at 25; see also Hennon, 109 F.3d at 335 (stating that, under AEDPA, state courts are no longer in a “tutelary relation” with federal courts). Thus, absent a controlling Supreme Court precedent, mere disagreement with a state court’s legal analysis is not enough to permit a federal court to vacate a state conviction; rather, it must be shown that the state court’s application of federal law was unreasonable. This sets the bar quite high. There is, of course, some doubt as to the degree of elevation necessary to separate reasonable and unreasonable applications of federal law. Here, however, we need not pinpoint that location precisely. The SJC’s analysis regarding the petitioner’s Confrontation Clause claim clears the bar because it comports with the approach taken by the courts of appeals in numerous cases raising comparable scenarios. When, as now, a petitioner can show only that rational minds might differ over how to apply certain general constitutional principles to the specific circumstances of his case, the current habeas corpus standard of review does not allow a federal court to invalidate a state conviction. V. CONCLUSION We need go no further. The petitioner fails to convince us that the SJC’s holdings are contrary to, or an unreasonable application of, Supreme Court precedent. Accordingly, the judgment below must be Affirmed. Footnotes [1] The petitioner also has raised a third contention, in the nature of a generalized due process claim based on the Constitution’s guarantee to criminal defendants of "a meaningful opportunity to present a complete defense." Crane v. Kentucky, 476 U.S. 683, 690, 106 S.Ct. 2142, 90 L.Ed.2d 636 (1986). The protections of the Sixth Amendment are themselves specific means through which the Constitution realizes the aforementioned guarantee. See id..; see also Strickland v. Washington, 466 U.S. 668, 684-685, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984). If the petitioner’s due process claim rested on a factual predicate that was not amenable to conventional Sixth Amendment analysis, such as the circumstances presented in Crane or in California v. Trombetta, 467 U.S. 479, 104 S.Ct. 2528, 81 L.Ed.2d 413 (1984), we would not hesitate to consider this claim separately. Here, however, the due process claim relies entirely on the factual predicate of the petitioner’s Sixth Amendment claim. Because there is a virtually complete overlap between the Sixth Amendment and due process claims, we need only discuss the former. [2] The Federal Rules of Appellate Procedure authorize COA determinations to be made by two, or even one, circuit judges. See Fed. R.App. P. 22(b)(2) (stating that COA requests addressed to the court of appeals "may be considered by a circuit judge or judges, as the court prescribes”). Some courts of appeals have exercised that authority. See, e.g., In re Certificates of Appealability, 106 F.3d at 1307. This circuit has opted, at least for the time being, to commit all such determinations to three-judge panels. [3] Although the SJC did not cite to Davis, its reference to the “ambiguous” nature of the petitioner’s assertion of his Fifth Amendment rights, see Bui I, 419 Mass, at 397 n. 4, 645 N.E.2d at 693 n. 4, meshes neatly with the district court's rationale. [4] Two courts, recognizing a possible theoretical distinction between the right to counsel and the right to remain silent, have left the issue open. See Evans v. Demosthenes, 98 F.3d 1174, 1176 (9th Cir.1996); United States v. Ramirez, 79 F.3d 298, 305 (2d Cir.1996). [5] There are limits to this assumption. To the extent that the petitioner calumnizes the district court’s reliance on Davis as improperly importing Sixth Amendment principles into the Fifth Amendment context, he is barking up the wrong tree. Davis did not deal with the Sixth Amendment, but, rather, with the Fifth Amendment right to counsel articulated in Miranda. See Davis, 512 U.S. at 457-58, 114 S.Ct. 2350. [6] To be sure, the petitioner argues that the trial court’s conclusion contravened established Supreme Court precedent holding that the issue of bias is always relevant. See Davis v. Alaska, 415 U.S. 308, 316, 94 S.Ct. 1105, 39 L.Ed.2d 347 (1974). As we explain below, our review of the record convinces us that the petitioner misconstrues the trial court’s ruling and that the “contrary to” prong of the AEDPA standard is not implicated here. [7] In his principal brief on appeal, the petitioner now suggests that he had supported his theory by evidence that Mong was Thinh’s father-in-law; that Mong, Thinh, and Bui had been together in California shortly after the murders; that Thinh had previously been convicted of a drug offense; and that autopsy reports showed that one of the victims had cocaine in her blood. This evidence *244 says nothing about Mong's interest or role in framing the petitioner. Moreover, under the AEDPA standard of review, such a shaky eviden-tiary foundation would not permit us to discredit the reasonableness of the SJC’s conclusions. [8] The trial judge's earlier refusal to permit the petitioner to cross-examine one of the police officers who had investigated the murders about Mong’s criminal history does not fill this void. For one thing, the petitioner does not challenge this ruling on appeal. For another thing, the attempted cross-examination was not in any way tied either to Thinh’s cross-examination or to the petitioner’s offer of proof. [9] The petitioner suggests that two items of evidence sufficed to close this gap. He says, first, that the idea that Mong would not endanger his daughter and Thinh becomes "laughable” when *245 we consider that the latter sometimes sold drugs on Mong’s behalf, and second, that in order to understand the reactions of an irate Vietnamese drug lord, we need only refer to a Massachusetts state trooper’s testimony that the Vietnamese community is "closed” and "in part controlled by intimidation and retribution." We are unpersuaded by these suggestions. With respect to the first matter, even if we assume, contrary to the established record, that there was evidence that Mong and Thinh were collaborators in the drug trade, the mere fact that Thinh at one unspecified time may have sold drugs for Mong says little about the existence of coercion, let alone its nature or whether that coercion was continuing. As for the police officer’s testimony (which the petitioner thoroughly mischaracterizes), we decline this invitation to arrive at a legal conclusion based on nothing more than an invidious stereotype of the members of an ethnic community.