SCHWELB, Associate Judge:
George E. Carter was convicted by a jury of two counts of armed robbery1 and one count of carrying a pistol without a license.2 His principal contention on appeal is that the trial judge committed reversible error by sustaining the invocation by his younger brother, Craig Carter, of the brother’s privilege against self-incrimination. We remand for further proceedings.
I.
THE TRIAL COURT PROCEEDINGS
A The Jaggers Inquiry.
The two complainants, Gregory Edmonds and Moses Williams, testified that George Carter and another man robbed them at gunpoint of jewelry, money, and other possessions. Both complainants were previously acquainted with George Carter.
Carter called his brother, Craig Carter, as a defense witness. He proffered that Craig would testify, among other things, that Ed-monds had acknowledged to Craig that he (Edmonds) did not know who had robbed him. Craig Carter was also expected to testify that Edmonds had sold Craig drugs and had been using drugs at the time of the alleged robbery. The prosecutor advised the [351]*351court that he would seek to cross-examine Craig Carter by inquiring about Craig’s own drug use, arguing that such drug use was relevant to Craig’s ability to perceive and relate his alleged encounter with Edmonds. The trial judge concluded, without defense objection, that such cross-examination would be permissible at least as to some of Craig Carter’s proposed testimony. The judge therefore appointed counsel to advise Craig Carter, who was incarcerated at the time, with respect to Craig’s rights under the Fifth Amendment. Craig Carter was then questioned outside the presence of the jury and, upon the advice of his attorney, stated that he would invoke his privilege against self-incrimination with respect to any questions that might be posed to him with regard to his use of unlawful drugs. Craig Carter’s counsel argued that his client’s answers to such questions could subject him to prosecution for unlawful possession of a controlled substance, and might also have adverse consequences for Craig’s prospects for parole.
The trial judge made a conscientious and comprehensive inquiry which was designed to determine “whether the risk of prosecution [was] substantial and real and not merely fanciful.” See Jaggers v. United States, 482 A.2d 786, 793 (D.C.1984) (per curiam) (citations omitted). During the course of that inquiry, George Carter’s counsel proffered that Craig Carter had been a resident of a halfway house both in May 1986 (when Ed-monds allegedly sold him heroin) and in July 1986 (when Edmonds allegedly admitted not knowing who robbed him), that test samples of Craig’s urine had been dirty in May, June and July of 1986, and that Craig’s work release privileges had been revoked and “that is why he is where he is.” Defense counsel argued that it would be “utterly unreasonable” to assume that Craig Carter would be prosecuted on the basis of historical evidence for simple possession of PCP a year and half before the trial, or that “the parole board will punish him twice.”
In response to the judge’s inquiry regarding the government’s position, the prosecutor stated that “we will not immunize [Craig Carter] in advance ... without [eliciting] from the witness all the facts underlying the witness’ alleged criminal activity.” Dissatisfied with this conclusory answer, the judge telephoned Alan Strasser, then the Chief of the Felony Trial Division of the United States Attorney’s office. With Mr. Strasser on the telephone, the judge described the government’s position as follows:
[Mr. Strasser] says, you know, that as a matter of fact it is rare — and I think that I would agree with the conclusion, that it is rare — extremely unusual for the government to ever prosecute misdemeanor drug possessions based on historical testimony.
He notes that on the other hand in an occasional narcotics conspiracy case where there is historical testimony or evidence that is relevant as part of the evidence in the conspiracy, he says to the extent that the government is being asked to predict what they will do in this matter, he says that he is unwilling to respond to that question, and he is also unwilling to say up front that he will grant immunity.
Through the office of the Superior Court’s legal adviser, the judge also contacted Gladys Mack, who was then the chairperson of the District of Columbia Board of Parole. Ms. Mack reported
that if the parole board learned that [Craig Carter] was using drugs over a longer period of time, or chronically or routinely rather than just the one or two times, three times — or whatever number of times his urine tested positive, that she could not rule out the possibility that he would ... be treated more harshly by the parole board if he were up for parole.
B. The Trial Judge’s Initial Ruling.
Recognizing the importance to the defense of the right to call potentially exculpatory witnesses, the judge was plainly troubled by the impact which Craig Carter’s exercise of his rights under the Fifth Amendment could have on his older brother’s rights under the Sixth Amendment. The judge repeatedly voiced skepticism that Craig Carter faced a realistic threat either of prosecution3 or of [352]*352deferment of parole.4 He asked rhetorically whether Craig’s sworn admission would be used against him “[e]ven where they already know about that drug use and have penalized him for it.” Following his conversation with Mr. Strasser, the judge opined that “while you can’t say that the risk of prosecution is fanciful ... the experiential element factored in here seems to me to suggest that it is not very realistic.” The judge stated that “with reference to the government’s position [as to] possible prosecution in this case, I find it a very, very difficult issue to resolve, if that were standing alone.” Noting that Ms. Mack had declined to rule out the possibility of harsher treatment on parole if chronic or routine drug use were shown, however, the judge found this to be a “serious concern.” He stated that potential adverse action by the parole authorities “is a valid consideration for [Craig Carter] in determining whether to assert the Fifth Amendment privilege.” The judge then concluded that, while the possibility of prosecution was not “fanciful,”
I am not sure on the other hand, that I can conclude that the risk of prosecution is substantial and real, standing alone. The court in Jaggers counter-posed those as opposites. It seems to me that there is some interim territory in there that we haven’t covered.
I guess I would say ... that the prospect of prosecution in this case is real, albeit not substantial.
But it does seem to me that the parole implications are not only real but significant in view of what I have learned from Mss Mack and through Mr. Erhardt.
Consequently, I am going to rule in this case that Mr. Craig Carter does have a Fifth Amendment right not to testify, and that it is being properly asserted in this case.
C. The Judge’s Revised Ruling.
Following the judge’s initial ruling, the trial continued. After the lunch break, however, the judge reopened the issue. He stated that he had read additional authorities, and that he “was not nearly as satisfied as he [had been] that the assertion of a Fifth Amendment privilege validly lies where the only danger sought to be protected against would be a revocation or denial of one’s parole or probation status.”
After extensive further discussion with counsel, however, the judge again sustained Craig Carter’s claim of privilege, now relying solely on the possibility that Craig would face further criminal prosecution if he were compelled to acknowledge past drug use:
THE COURT: Jackson v. United States, 490 A.2d 192 [D.C.1985], which is I think the most — probably the most recent case, at least in the four that I have looked at, the court talks about it being the legal possibility of subsequent prosecution rather than the practical likelihood of its coming to pass, as it determines the existence of the privilege so long as that legal possibility is real and not trifling.
I guess in view of everything that I have heard from Mr. Stern and in view of Jackson, I am inclined to abide by my ruling of this morning.
(Emphasis added.) During the course of his discussion of the issue, the judge made at least five separate references to the “legal possibility” of prosecution as the decisive factor with respect to Craig Carter’s claim of privilege.
The defense did not altogether lose the right to present Craig Carter’s testimony. The judge ruled that Craig would be permitted to testify regarding George Carter’s personal appearance at the time of the offense,5 and that if Craig’s direct testimony was confined to that subject, the prosecution would not be permitted to cross-examine him as to [353]*353his prior drug use.6 Craig Carter did testify to this limited extent. In light of the judge’s ruling on Craig’s invocation of his privilege, however, the defense was not able to present any testimony as to Edmonds’ alleged admission that he did not know the robber or as to Edmonds’ sale of unlawful drugs to Craig.
George Carter was convicted of all charges. This appeal followed.
II.
LEGAL DISCUSSION
A The Road Not Taken.
In (James) Harris v. United States, 614 A.2d 1277, 1281-82 (D.C.1992), we synopsized the basic legal principles applicable to the issue here presented as follows:
As the trial judge explicitly recognized, a criminal defendant’s right to present witnesses in his own defense is a fundamental one. Wilson v. United States, 558 A.2d 1135, 1140 (D.C.1989). Nevertheless, “in the crunch, when all else fails, the Fifth Amendment privilege of the witness prevails over the defendant’s right to compel him to testify.” Id. “Because both rights are so precious ... and because a forced election is so painful, it is the responsibility of the trial judge to take all reasonable steps to avoid a direct collision.” Id.
In situations of this kind the judge must attempt to accommodate not only the interests of the witness and of the defendant, but also the right of the prosecutor to cross-examine defense witnesses. “[T]he use of unlawful drugs is a proper subject of inquiry going to the credibility of a witness and his recollection of the events in question.” Coates v. United States, 558 A.2d 1148, 1152-53 (D.C.1989) (citations omitted). Thus, in the present case, the prosecutor was also asserting a legitimate interest.
Was this a ease in which all three of these interests could have been satisfactorily accommodated? With the incalculable benefit of hindsight, see Strickland v. Washington, 466 U.S. 668, 689, 104 S.Ct. 2052, 2065, 80 L.Ed.2d 674 (1984), and with a far more extensive opportunity to consider the question than was available to the trial judge and to trial counsel, we think that it may well have been. It was undisputed that Craig Carter’s urine had tested “dirty” at the relevant times. It was likewise undisputed that Craig’s halfway house privileges had been revoked on that account. If the parties had stipulated to these facts, or if the prosecutor had been limited to inquiring into the results of the drug tests (rather than into Craig’s possession and use of drugs), then the government’s interest in impeaching Craig’s recollection of events, as well as his credibility, might have been substantially vindicated without the preclusion of Craig’s testimony. The difference, for purposes of showing that Craig’s memory and judgment may have been impaired, between “dirty” drug tests on the one hand, and an admission by the witness that he used PCP on the other, would have been minimal.7
So far as we can discern, the prosecution would probably have lost very little, if anything, if its cross-examination of Craig Carter had been restricted to this limited degree. The scope and extent of cross-examination are committed to the trial court’s sound discretion. Roundtree v. United States, 581 A.2d 315, 323 (D.C.1990). It would therefore arguably have been at least permissible for the judge to confine the prosecutor to questions about the drug tests, to proscribe questions regarding Craig’s possession of PCP, and to avoid in this manner the loss to the defendant of his brother’s relevant testimony—a far more serious consequence than a comparatively minor restriction of one aspect of the prosecutor’s cross-examination.
Moreover, additional inquiry by the judge and by counsel might have revealed that the prosecutor’s proposed cross-exami[354]*354nation as to Craig Carter’s drug use at the relevant times, even if allowed, would not have required Craig to admit anything that he had not previously acknowledged. Craig’s halfway house privileges having been revoked, there may have been a revocation hearing in which the issue of his use of narcotics was adjudicated. Cf. Morrissey v. Brewer, 408 U.S. 471, 489-90, 92 S.Ct. 2593, 2604-05, 33 L.Ed.2d 484 (1972). The record does not disclose whether Craig testified at such a hearing or, if he did, whether he admitted or denied drug use at the relevant times. If he made admissions at a revocation hearing equivalent to those which would have been elicited from him at his brother’s trial, then his privilege would no longer have been available. See (James) Harris, supra, 614 A.2d at 1283. Here, as in (James) Harris,
[t]he focus of the inquiry should therefore have been on whether Martin would have been exposed to a substantial incremental risk of incrimination if he had been required to testify as a defense witness at the trial. The parties never addressed this question, and the trial judge never resolved it.
Id. (emphasis in original).
In this case, however, the judge and the attorneys were engaged in a thoughtful attempt to resolve a complex constitutional issue, and these possible solutions eluded them. Indeed, no such proposed resolution was presented to the trial court, or even to this court. The judge specifically asked counsel for George Carter if she had any basis for objecting to cross-examination by the prosecution with respect to Craig Carter’s alleged use of PCP in July 1986. Counsel responded that “[ojbviously, I don’t want him to be able to cross-examine, but I don’t legally see how I can prohibit it.” The defense likewise proffered nothing with respect to any testimony that Craig Carter may have given, or admissions that he may have made, during any revocation proceedings.
“Plain errors or defects affecting substantial rights may be noticed although they were not brought to the attention of the Court.” Super.Ct.Crim.R. 52(b).8 One might theoretically pose the question whether the trial judge committed plain error by not intervening, sua sponte, either to impose a restriction on the government’s cross-examination (when defense counsel never requested such a restriction) or to embark upon an investigation of what occurred at any prior revocation proceedings (when no party had asked that he do so). Cf. Irick v. United States, 565 A.2d 26, 33 (D.C.1989). We note, however, that even on appeal, defense counsel has not suggested that the judge should have taken either or both of these steps, or that he committed plain error by failing to do so.
To meet the rigorous standards of Rule 52(b), we must be satisfied that any alleged plain error was “obvious or readily apparent.” (Thomas) Harris v. United States, 602 A.2d 154, 159 (D.C.1992) (en bane); Foreman v. United States, 633 A.2d 792, 795 (D.C.1993). The error must be “clear under current law.” United States v. Olano, — U.S. -, -, 113 S.Ct. 1770, 1777, 123 L.Ed.2d 508 (1993); Foreman, supra, 633 A.2d at 795. We will not reverse for plain error unless a miscarriage of justice would otherwise result. (Thomas) Harris, supra, 602 A.2d at 159. Although the loss of a potential exculpatory witness is not a trivial matter, we are satisfied that there was no plain error here, if error at all; none of our hindsight-inspired observations could or should have been obvious to the trial judge.
B. The Scope of Review.
Although this court has frequently had occasion to consider trial court rulings regarding the invocation by defense witnesses of the privilege against self-incrimination, appellate counsel have not cited to us, nor has our own research identified, any decision by this court explicitly addressing the proper scope of appellate review. In the present case, which the trial judge correctly described as a very close one, the standard of [355]*355review is at least potentially decisive. Accordingly, we address that question directly.
In Hoffman v. United States, 341 U.S. 479, 71 S.Ct. 814, 95 L.Ed. 1118 (1951), the Supreme Court stated that
[t]o sustain the privilege, it need only be evident from the implications of the question, in the setting in which it is asked, that a responsive answer to the question or an explanation of why it cannot be answered might be dangerous because injurious disclosure could result. The trial judge in appraising the claim “must he governed as much by his personal perception of the peculiarities of the case as by the facts actually in evidence. ” See Taft, J., in Ex parte Irvine, 74 F. 954, 960 (C.C.S.D.Ohio, 1896).
Id. 341 U.S. at 486-87, 71 S.Ct. at 818 (emphasis added).9 The underscored language has been construed by some courts as investing the trial judge with “wide discretion” in resolving a self-incrimination claim. See, e.g., United States v. Melchor Moreno, 536 F.2d 1042, 1050 & n. 10 (5th Cir.1976) (abuse of discretion standard is implicit in Supreme Court’s allusion in Hoffman to judge’s “personal perception”); see also United States v. Van Deveen, 577 F.2d 1016, 1017 (5th Cir. 1978) (per curiam); State v. Parker, 79 Wash.2d 326, 485 P.2d 60, 63 (1971); State v. Kelly, 71 So.2d 887, 897 (Fla.1954); Commonwealth v. West, 321 Pa.Super. 329, 468 A.2d 503, 505 (1983).
Courts which apply the abuse of discretion standard have nevertheless reversed convictions where the defendant’s constitutional rights have been substantially undervalued. Melchor Moreno, supra, 536 F.2d at 1050; United States v. Chase, 281 F.2d 225, 227-29 (7th Cir.1960). In Chase, for example, the court held that the trial judge’s exercise of his discretion in rejecting the defendant’s invocation of the privilege against self-incrimination was undermined by his failure to apply the relevant law as declared by the Supreme Court.10
Although a number of courts have adopted the “abuse of discretion” standard, we question the soundness of such an analytical framework. “Discretion signifies choice.” Johnson v. United States, 398 A.2d 354, 361 (D.C.1979). A decision can fairly be viewed as having been entrusted to the trial court’s discretion if the judge is authorized to elect between a range of permissible alternatives. Where an exercise of discretion is called for, no single conclusion is preordained. Id. “[T]he decision maker, and not the law, decides.” Id.
This was not the kind of choice which the trial judge was called upon to make in the present case. If the risk that a witness will be prosecuted is real and not merely fanciful, the court must sustain the invocation of the privilege; the Fifth Amendment demands no less. See, e.g., Jaggers, supra, 482 A.2d at 793. If, on the other hand, the risk is merely fanciful, then the judge must reject the claim of privilege and permit the defendant to adduce the testimony of the witness; this is the command of the Sixth Amendment.
In order to determine whether the danger to the witness is sufficient to satisfy Joggers, the judge must assess what may plausibly happen in the future, both in terms of what the prosecutor will have the legal authority to do, and in relation to what he may in fact do. See Wilson, supra, 558 A.2d at 1141; Irby v. United States, 585 A.2d 759, 763-64 (D.C.1991). In making this inquiry, the judge is, in our view, confronted with identifiable questions of law and fact, and not with a discretionary choice between permissible alternatives. The scope of our review must be fashioned accordingly. See, e.g., [356]*356Griffin v. United States, 618 A.2d 114, 118 (D.C.1992).
Properly understood, Hoffman is not to the contrary. In observing that the judge must take into consideration “his personal perception of the peculiarities of the case,” 341 U.S. at 487, 71 S.Ct. at 818, the Supreme Court was not addressing the question whether appellate review is for “abuse of discretion” (as distinguished, say, from de novo review, or from review to determine whether a finding was clearly erroneous). Rather, the Court was implicitly recognizing the reality that the trial judge is present in the courtroom, in the midst of the action. The judge is therefore familiar with the atmospherics of the case and has a feel not only for what is and has been happening, but also for what may be about to happen. An appellate court, on the other hand, must view the proceeding from a lofty perch, far removed from the smells and sounds of battle. Appellate judges have more time than trial judges do to analyze and decide the difficult questions which arise in contested litigation, but they are compelled to act on the basis of a printed record rather than on what a trial judge observes from the perspective of an important participant who occupies an elevated seat in the front row. Given these differences in vantage point, the trial judge’s factual and predictive findings (as to the historical facts and as to what the prosecutor may or may not do) must be accorded considerable deference, and may be set aside by an appellate court only if they are clearly erroneous. Griffin, supra, 618 A.2d at 117; see D.C.Code § 17-305(a) (1989).
In considering the trial court’s legal conclusions, on the other hand, we apply the non-deferential de novo standard. Guadalupe v. United States, 585 A.2d 1348, 1352 n. 7 (D.C.1991); Griffin, supra, 618 A.2d at 117. In reviewing the trial court’s resolution of a mixed question of fact and law, we consider, among other things, whether the issue to be decided more closely resembles one of fact or of law, and whether the trial court or the appellate court is in a position to render the decision with the higher degree of accuracy. United States v. Felder, 548 A.2d 57, 61-63 (D.C.1988); Griffin, supra, 618 A.2d at 117-18. Finally, “where [as here] basic constitutional liberties are implicated, a more searching standard of review may be warranted.” Griffin, supra, 618 A.2d at 118 (citations omitted).
In the present case, the trial judge made a comprehensive and thoughtful inquiry into the facts, and we readily accept his determinations, both historical and predictive. The more difficult question is whether, as George Carter contends, the judge applied an incorrect legal standard.11
C. The Risk of Prosecution.
The basic question which the trial judge was called upon to resolve was whether the risk that Craig Carter would be prosecuted if he was cross-examined about his past drug use was “substantial and real and not merely fanciful.” Jaggers, supra, 482 A.2d at 793 (citations omitted). The judge understandably found this question to be a close and perplexing one. Initially, he expressed reservations regarding whether the risk of a criminal prosecution for unlawful possession of PCP, standing alone, was sufficient, but he upheld Craig Carter’s claim of privilege because Craig also faced what the judge viewed as a realistic prospect that his release on parole might be deferred. The judge eventually abandoned this position,12 and ruled in[357]*357stead that the result depended on “the legal possibility of subsequent prosecution rather than the practical likelihood of its coming to pass.” The judge obviously viewed the likelihood of a prosecution as very slight indeed— Mr. Strasser had conceded as much — but viewed the government’s authority to prosecute as decisive.
The judge cited Jackson, supra, 490 A.2d 192, as supporting his position. As the government concedes in its brief, however, Jackson restates the Joggers test, and there is nothing in that decision to suggest that the determination of Craig Carter’s Fifth Amendment claim could turn solely on the prosecutor’s legal authority to bring charges against the witness on the basis of the proposed testimony.13 In fact, the court in Jackson specifically noted that, under Joggers, the judge must determine whether “(1) the witness’ testimony would be incriminating,” and (2) “if so, whether the risk of prosecution is substantial and real and not merely fanciful.” Id. at 196 n. 8 (emphasis added; internal quotation marks omitted). The court added that the trial judge had sufficiently “considered the incriminatory nature of the proposed testimony and also the likelihood that criminal prosecution would result.” Id. (Emphasis added).
In any event, this court has conclusively rejected the proposition which the trial judge attributed to Jackson. In Jaggers, we stated that “[o]ne of the questions we must address to resolve the issues presented in this case is whether an absolute legal barrier to prosecution must exist before a witness can be compelled to testify in face of his claim of privilege. We think not.” 482 A.2d at 793. In Wilson, supra, we reiterated that
[w]here a witness might legally be prosecuted, but the threat of prosecution is not “real or appreciable,” this court has held that the privilege may not properly be invoked.
558 A.2d at 1141 (citing In re Neal, 475 A.2d 390, 392 (D.C.1984) (per curiam)). Similarly, in Irby, supra, we held that the trial judge must seek a commitment from the prosecutor that the potential defense witness will not be prosecuted, but that even if no commitment is forthcoming, “the trial judge must make an independent assessment, based on all of the circumstances, of the likelihood of prosecution.” Id., 585 A.2d at 763-64. In Irby, the trial judge made no inquiry regarding the second step, and we therefore remanded for further proceedings.14 Id. at 764.
Neal, Jaggers, Jackson, Wilson, Irby, and (James) Harris thus all address not only the authority of the government to prosecute the witness but also the existence vel non of a reasonable possibility that this will come to pass. Since the trial judge ultimately held that the first step alone was sufficient, we must return the case to him to determine, many years after the fact,15 whether there was a reasonable possibility at [358]*358the time of trial that Craig Carter' would be prosecuted, or to put it another way, whether the possibility of prosecution was “real” or merely “fanciful.”16 This determination must be made on the basis of all of the circumstances, including the fact that possession of PCP is itself a crime (and not simply a fact in a chain that might lead to the discovery of a crime), that Craig’s use of PCP had been known to correctional authorities for a year and a half, that his halfway house privileges had been revoked, and that, according to Mr. Strasser, misdemeanor prosecutions for unlawful drug possession, based on historical evidence, are most unusual.17 We are satisfied that if the judge concludes that Craig Carter’s claim of privilege was improperly sustained, the error was prejudicial, rather than harmless beyond a reasonable doubt.18 In that event, the judge must set aside George Carter’s convictions.
III.
CONCLUSION
For the foregoing reasons, the case is remanded to the trial court for further proceedings consistent with this opinion.
So ordered.
19