375 Mass. 510

February 13, 1978 - June 26, 1978

Suffolk County


The judge at a criminal trial erred in allowing a one-on-one voice identification of the defendant by the victim in open court under unnecessarily suggestive conditions [516-519]; such error was not harmless where the evidence apart from the voice test, while sufficient to warrant a guilty verdict, was not overwhelming [519-521].

INDICTMENTS found and returned in the Superior Court on September 24, 1975.

The cases were tried before Roy, J.

After review by the Appeals Court, the Supreme Judicial Court granted leave to obtain further appellate review.

Dyanne Klein Polatin for the defendant.

Michael J. Traft, Special Assistant District Attorney, for the Commonwealth.

KAPLAN, J. After a trial centering on the question of identification, a jury found the defendant guilty on three indictments for rape by unnatural sexual intercourse (G. L. c. 265, Section 22), and indictments for assault with a dangerous weapon (c. 265, Section 15B) and for armed robbery (c. 265, Section 17), all arising out of a criminal event on the evening of Wednesday, August 13, 1975. As will be seen, there was substantial evidence pointing to the defendant as the culprit. But as the trial drew to a close the prosecutor was permitted

Page 511

over repeated objections to conduct a "one-on-one" voice identification in open court, and it is claimed that the proceedings were thereby infected with serious error. The Appeals Court affirmed the convictions. Commonwealth v. Marini, 5 Mass. App. Ct. 834 (1977). This court granted further appellate review.

We describe the trial in some detail and then deal with the claimed error. [Note 1]

As its case in chief the Commonwealth brought out the following through testimony of the complaining witness (whom we shall refer to as the victim) regarding the details of the crime, and through testimony of the victim and police officers regarding the efforts at identification. On the date mentioned the victim, aged twenty-three, and her girl friend were sitting on a bench on the platform of the elevated station in Boston called Science Park, awaiting the trolley. The only other person on the platform was a man appearing to be in his late teens or early twenties standing nearby carrying a shopping bag. The man walked behind the girl friend and, drawing a gun and warning the two not to scream, forced them along the platform and into a small, squalid utility room lit by an unshaded electric bulb. He demanded and took from the victim all the money in her wallet (but refused the wallet) and dropped the money into his shopping bag. Still menacing the women with his gun, he ordered them to take off their clothes. He let down his pants and in vulgar words ordered the victim and the girl friend to perform acts of fellatio on him and to submit themselves to acts of cunnilingus. They did so. After threatening to take their clothes, the assailant relented, and left, warning them to remain in the room for a half hour without summoning help. The episode had taken fifteen to twenty minutes. The women stayed for perhaps a further fifteen minutes and then returned to the victim's apartment in the neighborhood.

Next morning the victim went to a station house of the Metropolitan District Commission and told her story to Officer

Page 512

Flynn. [Note 2] She was then taken to another office where she and Officer Morris produced cooperatively (by the "Ident-a-kit" method) a composite picture of the assailant as she recalled him. A week later Officer DeGuilio of the police department of the Massachusetts Bay Transit Authority called at the victim's apartment and asked her to look through some sixty-six "mug" photographs of white males around twenty years old. [Note 3] On the first run through she selected a picture, then paused at another, but soon rejected it, and on a second examination of the deck rather firmly identified the picture she first chose. This was a picture of the defendant. It showed a clean-shaven young man whereas the composite was with moustache and goatee, but otherwise the facial resemblance was close. The defendant was arrested a few days after this identification. There was no lineup or showup. At the probable cause hearing held at the Municipal Court of the City of Boston the victim, so she testified at trial, recognized the defendant at sight, but it is fair to say he was so placed in the court room that any onlooker would probably have perceived he was the subject of the hearing. The victim identified the defendant at trial which occurred in May, 1976.

The verbal descriptions of the assailant given by the victim to the three officers, and at a voir dire hearing, [Note 4] and at

Page 513

trial, were consistent in depicting a "very thin," or "very, very thin" white male in his late teens or early twenties with a narrow face, unkempt dark blond or light brown hair, blue eyes, and moustache and goatee. [Note 5] Some differences, however, did appear. As to height: Officer Flynn said the victim gave the assailant's height as six feet two inches. [Note 6] On the composite, height was stated as six feet. Officer DeGuilio said the victim had estimated five feet ten inches to six feet. [Note 7] At trial, the victim put the assailant's height at approximately six feet, figuring that he seemed about four inches taller than her own height of five feet seven inches. As to weight: The composite gave weight as 165 pounds, but the victim testified that she believed she had given estimates from 150 to 165 pounds to Officers Flynn and DeGuilio. [Note 8] This was the Commonwealth's case in chief.

The defense called Mrs. Jean Ratliff. She testified that on the evening of August 13, around 8 P.M., she was visiting with the defendant and his mother at their home. As supervisor of the mailing department of Winthrop Printing Company, she had gotten a job for the defendant and he was to report the next day. There was discussion of the nature of the job and possibilities of advancement because the defendant had a driver's license. The defendant was "going back and forth" getting dressed during the conversation, and left the house about 9 P.M. On cross-examination Mrs. Ratliff said the defendant started work the next morning (a Thursday) and did some day work (i.e., work on call) the next few days. Pressed as to when she learned of the importance of the defendant's whereabouts at 8 P.M., she said she learned

Page 514

of his arrest through a newspaper report on Monday, August 25, but that did not mention when the crime occurred. She did not recall when she became informed of this. She said she had been a friend of the defendant's mother for sixteen years. On redirect, the witness said she was first in touch with defense counsel in January, 1976. At that time she told counsel she had visited at the defendant's home on August 13. [Note 9]

The defendant took the stand in his own behalf and on direct examination testified consistently with Mrs. Ratliff as to the visit on August 13 around 8 P.M.

The Commonwealth opened cross-examination by eliciting from the defendant that he was nineteen years old and five feet eight inches tall (as indicated, he said, on his driver's license). He said he left the house on August 13 about 9 P.M. but he could not say where he went that night; nor could he identify the job he claimed he was doing earlier that day as a helper to his father, an electrician. He said he was at work at Winthrop's the next morning and worked perhaps two days before he was arrested, apparently on Saturday, August 23. Some time after the arrest, it occurred to him to have his mother look into his employment at Winthrop's in order to help fix his actions on August 13.

About this stage of the cross-examination the prosecutor at a bench conference asked the judge to admit the victim to the court room -- thus modifying his previous order sequestering witnesses -- and to direct the defendant to stand so that the victim could undertake a further identification of

Page 515

the defendant with particular reference to his height. The judge refused the request.

On the opening of court the following morning, the prosecutor at a further bench conference proposed to the judge that the victim be admitted to hear the defendant speak. He thought the defendant's voice had a guttural quality. He had spoken with the victim and she thought she might be able to recognize the voice. Over objection, the judge granted the request.

After some continuation of the substantive cross-examination, the prosecutor, over further objection, had the defendant repeat after him in conversational tones, "Don't scream or I'll kill you"; the vulgar words -- which need not be set out here -- commanding fellatio (twice repeated); and, "I don't want to see your immigration papers. I don't care if you are a foreigner" (the response made by the assailant when the victim, apparently by way of appeal to him, said the girl friend was a foreigner).

On short redirect, the defense introduced Winthrop's voucher stubs showing that during the week ending August 16 (a Saturday) the defendant had worked fifteen hours; and during the week ending August 23, nineteen hours. But, as the prosecutor pointed out, these records did not establish that the defendant had started work, or worked, on August 14. Here the defense rested. For rebuttal, the Commonwealth recalled Officer DeGuilio. Earlier the defendant, on cross-examination, when asked whether he had not told DeGuilio in September, 1975, that he was on Cape Cod on August 13, had replied that he did not recall whether that date had been mentioned. DeGuilio now testified that the defendant had made the statement. Cross-examined, DeGuilio indicated that he had been actually questioning the defendant about three dates, and that he had made no notes of the conversation with the defendant.

The last rebuttal witness was the victim, recalled. She said the voice she heard in the court room was the assailant's, "[t]he voice of the man who held my girlfriend

Page 516

and I at gunpoint." On cross-examination defense brought out that there had been a lapse of some nine months since the victim heard the voice to be identified. Thus the trial proper ended.

In the closing arguments to the jury, counsel on both sides referred to the voice identification. Defense counsel said that when the victim testified initially about the voice she said it was undistinguished, not unusual; hearing the voice after nine months, she now claimed to recognize it; the jury were to judge whether there was anything in the voice that would make it "stick out." We must interpolate here that, in testifying before the jury, the victim had not said anything about the assailant's voice. She had, however, testified to it at voir dire in the absence of a jury: COUNSEL FOR THE DEFENDANT: "Could you describe his voice?" THE WITNESS: "Well, there was nothing really distinctive about it. It was not low, it was not high, no noticeable accent."

The prosecutor, in his closing argument, followed defense counsel's apparent mistake or confusion in misplacing the victim's original testimony. He said the victim had testified on direct and cross-examination that there was "nothing special about the voice." But, said the prosecutor, "there is a certain sort of guttural quality" about it, hard to describe, and he left it to the jury whether one would be able to remember it on hearing it a second time. The judge in his charge to the jury referred to the voice identification but did not give it particular prominence.

1. We have no hesitation in saying that the prosecutor was wrong to attempt to stage the voice identification as he did, and that it was error of constitutional dimension for the judge to permit it. We notice the error even though the defendant might have been clearer in describing the exact grounds of his objections.

Identification by voice alone has long been thought to involve "grave dangers of prejudice to the suspect." Palmer v. Peyton, 359 F.2d 199, 201 (4th Cir. 1966). It has indeed been suggested that, at least where a witness has a basis for an identification of the culprit by sight, he should not be

Page 517

asked to make a voice identification unless he himself suggests it. Sobel, Assailing the Impermissible Suggestion: Evolving Limitations on the Abuse of Pre-Trial Criminal Identification Methods, 38 Brooklyn L. Rev. 261, 305 (1971). At all events, police and prosecutors are warned to take particular pains to avoid suggestive conditions in making arrangements for out-of-court tests where a witness tries to match live voices with his recollections of a voice heard in the usually stressful original setting. One-on-one auditions ought certainly to be avoided (see Commonwealth v. Torres, 367 Mass. 737, 740 [1975]); there should be the best approach to a "lineup" with precautions against directing undue attention to any participant. See Model Code of Pre-Arraignment Procedure Section 160.2(3), Commentary at 444-445 (1975). Preferably the witness should not be viewing the participants as he listens to the words spoken by them in turn. Sobel, supra. Because emotion confuses discernment and discrimination, the words chosen for repetition by the lineup should not be those heard by the witness at the scene. Note, 45 Miss. L.J. 489, 502-503 (1974); Note, 55 Minn. L. Rev. 779, 819 (1971); Murray, The Criminal Lineup at Home and Abroad, 1966 Utah L. Rev. 610, 628. As reliability is negatively affected not only by the degree of suggestiveness of the conditions of the test but also by natural loss of memory over time, the earlier the test, the better, with recognition that there is likely to be a greater rate of loss of memory in the hours immediately following the original perception than afterwards. Marshall, J., dissenting, in Manson v. Brathwaite, 432 U.S. 98, 118, 131 (1977). (To avoid misunderstanding, we should say here that, as with sight identifications, there will be occasions when standard precautions may well be overlooked, for example, when a suspect is apprehended very soon after the criminal event and is brought promptly to the victim for direct voice identification, as in Wise v. United States, 383 F.2d 206 [D.C. Cir. 1967], cert. denied, 390 U.S. 964 [1968]. Cf. Commonwealth v. Torres, supra; Roper v. Beto, 454 F.2d 499 [5th Cir. 1972]; Model Code, supra at 444.)

Page 518

Now it will be observed that the in-court voice test in the present case was conducted in breach of the procedures just mentioned. It was one-on-one, with the defendant in full sight of the victim. The "damning words" (Palmer v. Peyton, supra at 201), not neutral expressions, were utilized. High suggestiveness was injected not only by the circumstance that the speaker was the very accused on trial, but also by the fact that the person attempting the voice identification had already committed herself by having made separate positive identifications: these would be put in question or nullified by a negative identification as to voice, and perhaps shaken even by an avowed inability to make a voice identification. [Note 10] Hearing the precise words uttered by a man the victim had already identified could be expected to renew in her the horror and disgust of the heinous original encounter, and all this under the pressure and excitement of actual trial. If emotion can distort recollection we seem to have here a setting definitely favorable to such a reaction. The final point to be noted is that there was no need to indulge in an attempt at voice identification which contained all these suggestive elements and was to be carried out before the jury. [Note 11] If a need was felt for a voice test, it could have been conducted under standard lineup conditions out-of-court or at a voir dire hearing, and then testified to at trial.

Page 519

The law has not taken the position that a jury can always be relied on to discount the value of an identification by a proper appraisal of the unsatisfactory circumstances in which it may have been made. On the contrary, this court, like others, has read the Constitution to require that where the conditions are shown to have been highly and unnecessarily suggestive, the identification should not be brought to the attention of the jury. See, e.g., Commonwealth v. Kazonis, 356 Mass. 649 (1970); Stovall v. Denno, 388 U.S. 293, 302 (1967); Simmons v. United States, 390 U.S. 377, 384 (1968). It is true that in Manson v. Brathwaite, supra, referred to in Commonwealth v. Nolin, 373 Mass. 45, 51 (1977), it is suggested that there are situations in which evidence of an identification may be admitted although the identification was carried out under needlessly suggestive conditions. Manson would allow this to be done when special factors attending the identification show it to be reliable. Here, however, the indicia of reliability set out by the Supreme Court in Manson (at 114) were lacking; indeed, there was a distance of nine months between the criminal incident and the voice identification and a seeming discordance between the identification and the victim's prior testimony at voir dire.

2. We have next to say whether the error in permitting the in-court identification can be passed over as "harmless." The question is not simple, and is a matter of close judgment. As we said at the outset, the case against the defendant, apart from the voice identification, was substantial. There was opportunity to observe the assailant, and a pretty definite verbal description of him, followed by the making of the composite and the selection out of many photographs of one which conformed to the composite. The testimony of Mrs. Ratliff was by no means certain, and her credibility was subject to doubt. Similarly the defendant's testimony was infirm, as his insistence on the August 13 alibi was not matched by a solid recollection of what else he had done that night, and the documentary evidence of his employment

Page 520

at Winthrop's was not conclusive as to the day on which he started work there.

On the other hand, there remains some probability that the victim's verbal description and consequent graphic reconstruction were inaccurate. The better test of a live lineup for sight identification (P. Wall, Eye-Witness Identification in Criminal Cases 70 [1965]; Marshall, J., dissenting, in Manson v. Brathwaite, supra at 118, 132-133) was not employed. Whereas the divergences in the statements of height and weight were not strikingly great, the repeated characterizations of the assailant as "very thin" were of dubious compatibility with a youth who might be five feet eight inches in height even if his weight was as low as 150 pounds. The testimony of Mrs. Ratliff and the defendant cannot be wholly disregarded for the present purpose, and Officer DeGuilio's recollections in his rebuttal testimony about the defendant's statements to him do not inspire conplete confidence.

We have had earlier occasion to remark on the "somewhat variant definitions of the concept [of harmless error] propounded by the Supreme Court." Commonwealth v. LeBlanc, 364 Mass. 1, 10 (1973). See Commonwealth v. Graves, 363 Mass. 863, 865-866 (1973). The Court has been swinging from one to another polar position which may be represented, respectively, by Chapman v. California, 386 U.S. 18 (1967), and Milton v. Wainwright, 407 U.S. 371 (1972). Under Chapman, "the beneficiary of a constitutional error [is required] to prove beyond a reasonable doubt that the error complained of did not contribute to the verdict obtained." Id. at 24. Under Milton, harmlessness must still be established beyond a reasonable doubt, but the demonstration may be made by pointing to "overwhelming evidence" (at 373) of guilt, distinct from the tainted evidence which for this purpose is to be disregarded.

By the light of Chapman, it would be hard to say that the in-court voice identification, amounting almost to a compelled reenactment of the crime, was without effect on the jury and did not contribute to the verdict. It is true that the

Page 521

closing arguments of counsel called attention to the fact that the victim had stated earlier that there was nothing exceptional about the voice. This (although muddled by the mistake shared by both counsel) could have minimized the importance of the voice identification in the minds of the jury, or possibly suggested to them that the identification meant merely that the quality of the voice in the court room was not inconsistent with that recalled by the victim from the actual event. Still a significant doubt remains whether the jury were not actuated to some material degree by the display before them at the close of trial. If the Milton test is applied, a similar doubt remains, for while the evidence apart from the voice test would certainly have been enough to warrant a guilty verdict, it cannot be said to have been "overwhelming," at least if we take the facts of the Milton case itself as indicative of the meaning of that word. [Note 12] (We need not consider in detail whether the Milton standard itself can survive fine analysis. [Note 13])

We note that the evidence improperly received in the present case was not addressed to a collateral issue in the case but ran to the center, the question of identification. We are mindful, too, that, if the concept of harmless error is loosely applied, it can serve too readily as a bridge for a procession of mistakes and injustices.

On full consideration, we feel obliged to set aside the judgments of conviction. A new trial will be in order. We

Page 522

can but express our regret at the waste of time and resources invested in the trial, as well as at the delay in reaching a final disposition of the charges, whatever that may be, all occasioned by a procedure that we hold to have been prejudicial.

Judgments of the Superior Court reversed.

Verdicts set aside.


[Note 1] Another claimed error is dealt with at n.9 below.

[Note 2] The girl friend, visiting here from Belgium, declined to go to the police, returned to her country, and did not appear at any stage of the proceedings. At first, to save the girl friend from embarrassment, the victim did not mention her in speaking to the police.

[Note 3] The victim testified on direct examination, without objection, that the photographs were mug shots, as was indeed obvious. They were received in evidence and passed to the jury with all but the pictures masked by tape. See the remarks of Tuttle, J. (sitting by designation), on the handling of mug shots in United States v. Fosher, 568 F.2d 207 (1st Cir. 1978).

[Note 4] At voir dire (which immediately preceded the trial proper) the judge suppressed one identification. Some days after the photographic identification at the victim's apartment, perhaps on the day of the probable cause hearing, Officer DeGuilio showed the victim the photograph she had identified but with a moustache and goatee pencilled in. She recognized this doctored picture as that of the assailant. The judge ruled this identification to be suggestive. He did not so rule the identification at the probable cause hearing.

[Note 5] Other elements: The assailant, according to the victim, wore sunglasses with gold colored rims (he removed these glasses on reaching the utility room), faded jeans with holes at the knees, sneakers, and on the ring finger of one hand a gold ring with a dark stone.

[Note 6] The victim thought she had said six feet, or five feet ten inches to six feet.

[Note 7] The victim thought she had said five feet ten inches to six feet one inch.

[Note 8] At voir dire the victim seemed to put the assailant's height at six feet or up to six feet two inches, and his weight at 145 to 150 pounds.

[Note 9] However, the judge on his own motion excluded a further question on redirect asking "exactly what you told him [counsel]." Exception was taken to the exclusion. The defense argues that the Commonwealth in cross-examining Mrs. Ratliff had attacked her testimony about the meeting of August 13 on grounds amounting to bias and recent contrivance, and that the defense was therefore entitled to prove she had made prior consistent statements to counsel (see Commonwealth v. Zukoski, 370 Mass. 23, 26-28 [1976]), and should have been allowed to make the proof in greater detail. Even on the defendant's assumption as to the thrust of the cross-examination, we are not prepared to say that the judge was wrong to draw in the reins on the redirect as he did.

[Note 10] We can only speculate how much suggestiveness was involved in the prosecutor's approach to the victim about getting her relieved of the sequestration order and having her listen to the defendant. The judge, be it noted, had not considered what, if any, submission to the jury the prosecutor might be required to make if, after hearing the defendant, the victim told the prosecutor she was unable to make a voice identification or was actually prepared to say this was not the voice she remembered. Incidentally, the prosecutor, in applying to the judge to lift the sequestration order, did not state that he intended to ask the defendant to repeat the "damning words."

[Note 11] Compelling the defendant to repeat the exact words (including the vulgar ones) before the jury may be thought of, independently, as an abuse of the judge's discretion.

[Note 12] In the Milton case the Court remarked that "[t]he jury, in addition to hearing the challenged testimony, was presented with overwhelming evidence of petitioner's guilt, including no less than three full confessions that were made by petitioner prior to his indictment." 407 U.S. at 372-373.

[Note 13] A meticulous examination and critique of the Court's decisions on harmless error lead Professor Martha A. Field to a demonstration of the inadequacies of the Milton formula and to espousal of a rule which would allow a finding of harmlessness under either the Chapman formula or a "cumulative evidence" test; she finds support for the latter in some cases, especially Harrington v. California, 395 U.S. 250 (1969). We think the error in the present case would not be termed harmless under the "cumulative evidence" analysis. Field, Assessing the Harmlessness of Federal Constitutional Error -- A Process in Need of a Rationale, 125 U.Pa. L. Rev. 15 (1976).