(dissenting). Few things are as vile as the sexual abuse of children. It is therefore understandable that we, as a society, would implement severe measures to try to prevent such abuse from occurring. But when we imprison someone for mere possession of a photograph of a naked child playing on a beach, we have lost all perspective. For the reasons detailed below, I respectfully dissent.
A. Sufficiency of the indictment. Because the defendant effectively conceded that he had adequate notice of the relevant charges against him, I am sympathetic to the majority’s efforts to find the indictment sufficient.1 After all, if the principal purpose that the indictment is designed to serve has been satis-*311fled, how can there be a problem? In addition, I have no quarrel with the idea that we may rely on the caption of an indictment to help inform the substance of the charges.
Nevertheless, the indictment here was so inexcusably mangled that I remain concerned about our deeming it sufficient.2 None of the cases on which the majority relies goes this far, and I am troubled by the standards we are setting. Moreover, assuming that the caption of the indictment can be taken to supply the key missing allegation that the photograph at issue was of the sort prohibited by the statute, the indictment would still fail to allege that the defendant had specific knowledge of the “nature and content” of the photograph. See G. L. c. 272, § 29C(vii). This would appear to be at odds with Commonwealth v. Palladino, 358 Mass. 28 (1970).3 Although some of Palladino’s “broad language has been limited by subsequent decisions,” Commonwealth v. Militello, 66 Mass. App. Ct. 325, 339 (2006), *312that case’s narrow holding still appears to stand. Perhaps the Supreme Judicial Court would reverse Palladino if asked to do so, but until that happens, it is our role to follow that case.4
B. Sufficiency of the evidence. 1. The photograph. The underlying facts are not in dispute. The debate is not over what can be seen in the photograph,5 but whether it makes out a crime when measured against the relevant legal standards.
The photograph shows an unclothed girl kneeling in the sand. A “flip-flop” sandal lies askew in the background, and there is no dispute that the setting is a beach. Although there is no way of knowing the girl’s exact age, she appears to be approximately eleven to twelve years old, as the Commonwealth maintained and as reasonable jurors could have concluded. The girl is not looking at the camera, and her gaze instead appears focused on the area of the beach immediately in front of her, where there is a small hole in the sand. Her sand-flecked arms are pointed in that direction and they appear to be reaching toward the hole. Her left hand, captured by the photograph in mid-air, and the hand’s shadow obscure a view of her genitals, although the area in the immediate vicinity of her genitals can be seen. Her partially developed breasts are also plainly visible.
The photograph has a distinct “snapshot” quality to it, and the Commonwealth never argued to the jury that the girl was posed. As the prosecutor acknowledged to the jury, “it seems like [the girl] is completely unaware that someone is taking her picture.”6 The jury had no information about who the girl is, what beach she was on, or who the photographer was. Thus, for example, the jury did not know if this was a vacation photograph taken by a member of the girl’s family, or if it was instead surreptitiously *313captured by a stranger. The defendant downloaded the photograph from a Russian photograph-sharing Web site. That Web site does not market itself as a pornographic site, and a Commonwealth witness acknowledged that it is a publicly accessible photograph-sharing site — similar to “Snapfish,” “Flickr,” and “Picasa” — that included what appeared to be “a lot of vacation photographs.”
2. The issue in dispute. The photograph found in the defendant’s possession unquestionably shows the girl’s “unclothed . . . partially developed breast[s].” G. L. c. 272, § 29C(vii). It is also fair to characterize the photograph as showing her unclothed “pubic area” (even if the anatomical boundaries of that statutory term remain unclear).7 However, for there to have been a violation of § 29C(vii), the photograph must additionally rise to the level of a “lewd exhibition” of her unclothed breasts or pubic area. G. L. c. 272, § 29C(vii).8 Indeed, as both the United States Supreme Court and the Supreme Judicial Court have long recognized, a statute that sought to punish the portrayal of mere nudity, even of children, would be constitutionally infirm. See Osborne v. Ohio, 495 U.S. 103, 112-114 (1990); Commonwealth v. Provost, 418 Mass. 416, 423 (1994); Commonwealth v. Bean, 435 Mass. 708, 711, 715 n.17 (2002) (Bean). The cases have long established that nudity, without more, does not establish “lewdness” or “lasciviousness.”9 See, e.g., Provost, supra; Bean, supra at 715 n.17; United States v. *314Arvin, 900 F.2d 1385, 1391 (9th Cir. 1990), cert, denied, 498 U.S. 1024 (1991); United States v. Amirault, 173 F.3d 28, 33, 35 (1st Cir. 1999) (Amirault); United States v. Russell, 662 F.3d 831, 843 (7th Cir. 2011), cert, denied, 132 S. Ct. 1816 (2012).
3. The Dost factors. As the majority accurately notes, in seeking to lend substance to the meaning of lewdness (or its linguistic twin, lasciviousness), courts have often looked to the so-called Dost factors. See ante at 302-303; United States v. Dost, 636 F. Supp. 828, 832 (S.D. Cal. 1986), aff’d sub nom. United States v. Wiegand, 812 F.2d 1239 (9th Cir.), cert. denied, 484 U.S. 856 (1987). For example, the United States Court of Appeals for the First Circuit relied on the Dost factors in determining that a photograph of a naked girl on a beach was not a lascivious exhibition, Amirault, supra, and the Supreme Judicial Court relied on them in determining that a photograph that revealed a fifteen year old girl’s breast did not establish the photographer’s “lascivious intent.” Bean, supra at 715-716. The photograph here is not lewd under a straightforward application of the Dost factors. This should hardly be surprising, because even a cursory examination of those factors and of the decades of case law applying them reveals that they are principally aimed at a particular species of lewdness not presented here: photographs of children deliberately posed in a manner or setting designed to sexualize them.10 In contrast, the photograph here is a candid, unposed snapshot of a girl apparently playing on a beach. “When a photographer selects and positions his subjects, it is quite a different matter from the peeking of a voyeur upon an unaware subject pursuing activities unrelated to sex.” United States v. Steen, 634 F.3d 822, 828 (5th Cir. 2011).11
In any event, the Dost factors do not support a conclusion *315that this photograph contains a “lewd exhibition.” As the Supreme Judicial Court has noted, a “beach setting is [a] ‘natural landscape’ that is not generally associated with sexual activity.” Bean, 435 Mass. at 716 n.18, citing Amirault, 173 F.3d at 33. The girl is not engaged in any sexual activity (as the prosecutor expressly acknowledged to the jury), nor is she posed to exhibit any willingness to engage in such activity. Indeed, as the prosecutor also acknowledged below, she does not appear even to be aware that her picture is being taken. Such a photograph can be considered “lewd” under the Dost factors only if those factors are stretched beyond recognition. See United States v. Steen, supra at 827 (questioning the applicability of the Dost factors “where a defendant’s conduct . . . proved to be no more than voyeurism”); United States vs. Johnson, U.S. Dist. Ct., No. 2:10-CR-71 -FtM-36DNF (M.D. Fla. June 15, 2011) (finding that images obtained by secretly recording a minor not engaged in sexual activity “constitute voyeurism and not child pornography”).
To be sure, some of the Dost factors are sufficiently malleable that some argument can be made that they apply. However, a closer examination of those factors demonstrates the inappropriateness of resting a determination of lewdness on them, at least in this context. A good example is the first Dost factor: “whether the focal point of the visual depiction is on” a forbidden anatomical area. 636 F. Supp. at 832. In her closing argument, the prosecutor repeatedly referenced the child’s developing breasts (and, to a lesser degree, her pubic area) as the “focal point” of the photograph. She made no effort to explain what made them so. The majority strives to supply what eluded the Commonwealth at trial: a critical assessment of how the visual elements that make up the snapshot’s “design” objectively accentuate the girl’s sexuality. But one hardly needs such analysis to explain that a viewer’s gaze may well be drawn to the girl’s developing breasts. To the extent that one’s eyes are drawn there, that fact is neither remarkable nor of significant import. *316As the Supreme Judicial Court emphasized in Bean, whether a viewer sees a naked breast or other body part as the focal point of a photograph is often “a subjective assessment of little value.” Bean, supra at 716.12 To some extent, a viewer’s attention may be drawn to the image’s portrayal of the girl’s breasts in part precisely because her evolving sexuality invokes cultural taboos. See United States v. Frabizio, 459 F.3d 80, 93 (1st Cir. 2006) (Torruella, J., concurring) (“The images are jarring, more because of the cultural taboo regarding nude pictures of prepubescent girls than because of the specific portrayal of the girls”). In fact, here as with the beach photograph in Amirault, the “truly striking aspects of the photograph [are] the girl’s nakedness and her youth.” 173 F.3d at 35. But “[t]hese factors alone are not enough to render the photo ‘lascivious.’ ” Ibid.13
Also problematic is the sixth Dost factor, which asks “whether the photograph is intended or designed to elicit a sexual response in the viewer.” 636 F. Supp. at 832. The First Circuit has identified this factor as “the most confusing and contentious of the Dost factors.” Amirault, 173 F.3d at 34. Applying this factor makes some sense in the context of a prosecution against the photographer. In that context, “[i]t [is] logical ... to hold that *317the picture[] [is] ‘a lascivious exhibition because the photographer arrayed it to suit his peculiar lust.’ ” United States v. Rivera, 546 F.3d 245, 252 (2d Cir. 2008), cert. denied, 555 U.S. 1204 (2009), quoting from United States v. Wiegand, 812 F.2d at 1244. Even in that context, some courts have noted that the sixth factor adds little, if any, independent value to a factors-based analysis.14 See, e.g., United States vs. Clark, U.S. Ct. App., No. 10-4520, slip op. at 4-5 (3d Cir. Nov. 22, 2011), cert. denied, 132 S. Ct. 1778 (2012), quoting from United States v. Villard, 885 F.2d 117, 125 (3d Cir. 1989) (“the sixth factor, ‘rather than being a separate substantive inquiry about the [depictions], is useful as another way of inquiring into whether any of the other five . . . factors [is] met’ ”).15 The problematic role played by the sixth factor is heightened in a possession case given that the possessor of a photograph is not necessarily the one who “intended or designed” the composition. See United States v. Rivera, supra at 252 (“The sixth Dost factor is not easily adapted to a possession case”).16 Even to the extent that *318the underlying considerations that animate the sixth factor are relevant in a possession case, it makes little sense to try to get at these considerations by shoe-homing them into one “factor” in a six-part factor-based analysis. Put differently, the test of time has demonstrated that the Dost factors cannot bear the full analytical weight that has at times been placed on them. Rather than treating the Dost factors as providing the overarching doctrinal framework under which lewdness issues should be analyzed, it makes far more sense to treat them as helpful considerations especially designed to ferret out a particular type of lewdness not present here.
4. Otherwise lewd. Of course, a photograph that is not lewd under the Dost factors may still be lewd for other reasons. See United States v. Frabizio, 459 F.3d at 87 (sharply criticizing a District Court judge for resting on an “unqualified application” of the Dost factors and for failing to consider whether the beach photographs at issue there could be considered lascivious for other reasons). In Frabizio, the First Circuit emphasized that the Dost factors were never intended to be exclusive, and it aptly pointed out that, in upholding the original use of these factors, the United States Court of Appeals for the Ninth Circuit described them as “over-generous to the defendant.” Id. at 88, quoting from United States v. Wiegand, 812 F.2d at 1244. See United States v. Steen, 634 F.3d at 829 (Higginbotham, 1, concurring) (Dost “factors often create more confusion than clarity”).
Such observations are well taken as far as they go; while Dost was a well-meaning effort to lend some objectivity to the elusive question where the boundary between mere nudity and lewdness lies, its many shortcomings have been well documented.17 However, once one leaves the Dost factors behind, the question remains: against what standard should “lewdness” be assessed? In the context of considering the analogous Federal statute, the First Circuit flatly declined to address this question, commenting that lasciviousness “needs no adornment.” United *319States v. Frabizio, 459 F.3d at 85-86 & n.9.18 The trial judge in this case followed this hands-off approach,19 as does the majority.
This approach is untenable. One cannot hope to address whether a visual image of a naked child is “lewd” without knowing what it means for something to be “lewd.” Appellate courts have a responsibility to provide clarity on what demarcates the boundary between the exercise of free expression and a lengthy prison term.20 Failing to provide such guidance threatens to leave the issue to unbridled subjectivity in an area ruled by emotions. That in turn raises both First Amendment and due process concerns. See Commonwealth v. Kenney, 449 Mass. 840, 850 (2007), quoting from Commonwealth v. Orlando, 371 Mass. 732, 734 (1977) (“It is well established that due process requires criminal statutes that are not ‘sufficiently explicit to give clear warning as to proscribed activities’ to be declared unconstitutional. . . . Prohibited imprecision includes ambiguity in the terms of a criminal statute, especially one implicating the right to freedom of expression, that might give rise to arbitrary enforcement”).21 Simply put, someone charged with the heinous *320crime of possessing child pornography is entitled to something more than a “trial by Rorschach test.”
5. What makes something “lewd”? A visual image of a naked child cannot be considered a “lewd exhibition” unless it presents the child in an overtly “sexualized” manner that is tantamount to sexual abuse or exploitation of the child.22 As discussed below, at least three sources support that conclusion. I begin with the statute’s stated purpose.23
“The Legislature’s purpose in enacting the statute could not be clearer: ‘[T]o protect children from sexual exploitation . . . [by] prohibiting] the production of material which involves or is derived from such exploitation and to exclude all such material from the channels of trade and commerce.’ ” Commonwealth v. Kenney, 449 Mass. at 853, quoting from St. 1997, c. 181, § 1(1). Moreover, with regard to the criminalization of the mere possession of child pornography, “the Legislature enacted G. L. c. 272, § 29C, in recognition that the ‘mere possession or control of any sexually exploitative material results in continuing victimization of children [because] such material is a permanent record of an act or acts of sexual abuse or exploitation of a child and that each time such material is viewed the child is harmed.’ ” Ibid., quoting from St. 1997, c. 181, § 1(2).
That G. L. c. 272, § 29C, criminalizes behavior that amounts to the “sexual exploitation” and “sexual abuse” of children is *321reinforced by the fact that it is this critical governmental interest that allows a direct content-based restriction on speech. As the United States Supreme Court has recognized, the production of child pornography may be criminalized because that production “is itself the crime of child abuse.” Ashcroft v. Free Speech Coalition, 535 U.S. 234, 254 (2002). See United States v. Stevens, 130 S. Ct. 1577, 1586 (2010), citing New York v. Ferber, 458 U.S. at 759 (States have “a compelling interest in protecting children from abuse,” and child pornography constitutes a permanent record of that abuse).24 “The production of the work, not its content,” is the target of valid child pornography statutes. Ashcroft v. Free Speech Coalition, supra at 249.25 The dissemination of child pornography may also be criminalized, but that is because it provides “an economic motive for and [is] thus an integral part of the production of such materials.” New York v. Ferber, supra at 761. In turn, in order to prevent the abuse inherent in the production of child pornography, States may even seek to “stamp out this vice at all levels in the distribution chain” by criminalizing its mere possession. Osborne v. Ohio, 495 U.S. 103, 110 (1990). Contrast Stanley v. Georgia, 394 U.S. 557 (1969) (holding that the State may not criminalize mere possession of obscenity).
A necessary corollary is that “where the speech is neither obscene nor the product of sexual abuse, it does not fall outside the protection of the First Amendment.” Ashcroft v. Free Speech Coalition, supra at 251, citing New York v. Ferber, 458 U.S. at 764-765. In other words, the class of speech that constitutionally may be restricted is limited to those images whose suppression actually advances the State’s compelling interest in preventing the sexual abuse of children — namely, images whose production involved the sexual abuse or exploitation of actual children. See id. at 252-254. To the extent that a prohibition on possessing images of naked children fails to further such interests, it is unconstitutional. Id. at 244, 254, 256 (“virtual *322child pornography” that does not depict actual children may not be criminalized).
Finally, insight can be drawn from the principle of ejusdem generis. “When elements are listed in a series, the rules of statutory construction require the general phrase to be construed as restricted to elements similar to the specific elements listed.” Commonwealth v. Zubiel, 456 Mass. 27, 31 (2010). The reference in G. L. c. 272, § 29C(vii), to “lewd exhibition[s]” appears at the end of a list of specific sexual acts that children cannot be shown doing.26 This strongly suggests that for a visual depiction of a naked child to rise to the level of a “lewd exhibition,” it has to sexually exploit the child in a manner akin to that done by a photograph of the child engaged in the prohibited sex acts listed in the prior six subsections. It follows that, at a minimum, the photograph must “sexualize” its portrayal of the child’s nudity in some material way. See, e.g., United States v. Wallenfang, 568 F.3d 649, 657 (8th Cir.), cert, denied, 130 S. Ct. 566 (2009), quoting from United States v. Kemmerling, 285 F.3d 644, 646 (8th Cir.), cert. denied, 537 U.S. 860 (2002) (“A picture is ‘lascivious’ only if it is sexual in nature”).27
6. Is this photograph “lewd”?28 As the prosecutor affirmatively conceded below, “there is certainly no sexual activity in this *323picture.” In contending that the photograph is nevertheless lewd, the Commonwealth now focuses especially on the proximity of the girl’s left hand to her pubic area. Although no one can say with certainty what exactly the girl is doing with her hands, she appears to be playing in the sand (as the majority acknowledges). However, according to the Commonwealth, the happenstance of where the girl’s hand ended up frozen in time allowed a viewer to fantasize about the girl masturbating.29 In other words, the Commonwealth’s new theory is that the photograph is accidentally lewd. The majority appears to accept this argument, concluding that “the direction and curve of the left hand toward [the girl’s] exposed pubic area could be viewed as sexually suggestive.” Ante at 305.
Perhaps this defendant selected the photograph for the very reason that the Commonwealth now highlights. However, whether the defendant found this photograph lewd is not the test. “If [the defendant’s] subjective reaction were relevant, a sexual deviant’s quirks could turn a Sears catalog into pornography.”30 Amirault, 173 F.3d at 34. How a platonic observer would view the placement of the hand is somewhat imponderable. Some reasonable viewers will view the girl’s hand simply in motion on its way to (or back from) the hole in the sand in front of her. Other reasonable viewers will no doubt agree with the majority’s view that the placement of the hand “could be *324viewed as sexually suggestive.” Ante at 305. However, the applicable test is whether the photograph sexualizes its portrayal of the girl’s nudity in an objective and sufficiently material way to warrant a criminal conviction merely for possessing it. While the photograph need not be “obscene” (at least in the sense that that word is used in the context of adult pornography), it still has to be so noxious that one commits a felony merely by holding the photograph in one’s hand. This photograph does not come even close to meeting that test. Indeed, it is not clear how any photograph that is at most subtly sexually suggestive could ever be deemed “lewd” beyond a reasonable doubt.
This conclusion does not undercut the purposes that G. L. c. 272, § 29C, was enacted to serve. The photograph cannot fairly be described as a permanent record of the girl being subjected to sexual abuse or exploitation; rather, it is a permanent record of her playing in the sand.31 Of course, this is not to say that the girl expected or desired that her photograph would be taken that day on the beach, and needless to say, she might well be mortified if she knew that this photograph were available on the Internet. But the question before us is not whether whoever took the photograph (or placed it on the photograph-sharing Web site) broke the law in whatever jurisdiction this occurred. Nor is it whether the girl might have other remedies to address the invasion of her privacy. The question instead is whether the harm inflicted by the continued existence of the photograph is so substantial as to support treating its mere possession as a crime.
7. The majority’s methodology. Some additional comment is warranted as to the methodology through which the majority determined the evidence sufficient. After having preliminarily determined that the photograph is “lewd,” the majority goes on to examine whether “other evidence or characteristics of the photograph indicate that it is not lewd.” Ante at 306. Specifi*325cally, the majority examines whether the photograph “constitutes a form of artistic, educational, or other personal exhibition or expression protected by the First Amendment.”32 Ibid. Finding such content lacking, the majority suggests that First Amendment values have been protected.
The majority’s mode of analysis is problematic. Whether an image has artistic or other social value is irrelevant to the question whether it may constitutionally be proscribed as “child pornography.”33 If a photograph truly constitutes a “lewd exhibition” of a naked child, then it thereby falls outside First Amendment protection, and no amount of redeeming social content can protect it. New York v. Ferber, 458 U.S. at 773. Id. at 774-775 (O’Connor, J., concurring). This makes sense. If a visual image involves a child’s having been subjected to sexual abuse or exploitation, then whether there might be artistic or scientific value in the image is beside the point. Id. at 761, quoting from Memorandum of Assemblyman Lasher in support of N.Y. Penal Law, Art. 263, § 263.15 (“It is irrelevant to the child [who has been abused] whether or not the material . . . has a literary, artistic, political or social value”).34
By the same token, a judicial determination that a work lacks any redeeming value cannot justify a less rigorous inquiry into whether that work qualifies as child pornography. There is no sliding scale upon which a work deemed less valuable or artistic *326is therefore more likely to be lewd than a more valuable work containing the same objective depiction.35 Judges are in any event ill-equipped to make such a determination. See, e.g., Bleistein v. Donaldson Lithographing Co., 188 U.S. 239, 251 (1903) (“It would be a dangerous undertaking for persons trained only to the law to constitute themselves final judges of the worth of pictorial illustrations, outside of the narrowest and most obvious limits”); Pope v. Illinois, 481 U.S. 497, 505 (1987) (Scalia, J., concurring) (“For the law courts to decide ‘What is Beauty’ is a novelty even by today’s standards”).
Rather than judging the value of the work at issue, it is the role of the courts to guard the boundaries of free expression, permitting speech to be suppressed based upon its content only if it falls within one of the “well-defined and narrowly limited classes” that lie outside the First Amendment’s protection. United States v. Stevens, 130 S. Ct. at 1584, quoting from Chap-linsky v. New Hampshire, 315 U.S. 568, 571 (1942). Speech that fits comfortably within accepted social norms requires little judicial enforcement of First Amendment protections. Instead, the need for judicial intervention is implicated where the expression the government is challenging is unpopular or offensive. *327“Indeed, ‘the point of all speech protection . . . is to shield just those choices of content that in someone’s eyes are misguided, or even hurtful.’ ” Snyder v. Phelps, 131 S. Ct. 1207, 1219 (2011), quoting from Hurley v. Irish-American Gay, Lesbian & Bisexual Group of Boston, Inc., 515 U.S. 557, 574 (1995). While judges have an important role in making an independent review whether the expression at issue has crossed the boundary into unprotected speech, if that boundary has not been crossed, then judges have no proper role in determining whether particular speech merits protection.
The need for an independent judicial role is particularly acute in cases where the expression at issue involves a photograph of a naked child. Even where the child is not engaged in any of the sex acts specifically prohibited by the statute, and even though “mere nudity” is not (indeed, cannot be) prohibited by the statute, many members of the public would no doubt find any portrayal of a child’s nakedness offensive (particularly where, as here, it depicts an adolescent’s budding sexuality).36 That becomes increasingly likely to the extent one considers the possessor’s motives for acquiring such a photograph. Given that the defendant had no other obvious reason for downloading the photograph, the jury readily could have inferred that he derived sexual excitement from it.37 Indeed, defense counsel came close to conceding the point when she acknowledged to the jury that they probably found her client’s conduct “weird” and that it “probably makes you uncomfortable.” An adult’s sexual attraction toward young children is highly disturbing. Needless to say, when an individual actively pursues such urges to the point of *328crossing the boundary into objectively criminal behavior that exploits children, such conduct can and should warrant severe criminal sanctions. However, where the individual has not crossed that line, there is a risk that a jury will nevertheless seek to punish him or her based on the revulsion they feel for what is inside his or her head. Cf. Stanley v. Georgia, 394 U.S. 557, 566 (1969) (recognizing that the government “cannot constitutionally premise legislation on the desirability of controlling a person’s private thoughts”). In this area perhaps more than in any other, “a jury is ‘unlikely to be neutral with respect to the content of [the] speech,’ posing ‘a real danger of becoming an instrument for the suppression of . . . unpleasant’ expression.” Snyder v. Phelps, 131 S. Ct. at 1219, quoting from Bose Corp. v. Consumers Union of U.S., Inc., 466 U.S. 485, 510 (1984). It is the responsibility of judges to try to ensure that this does not happen.
C. Jury instructions. Even if the evidence had been sufficient, the jury instructions were wholly inadequate. Notably, the principal theory of lewdness that the Commonwealth presented to the jury was that a photograph that shows an adolescent girl in a state of nudity is for that reason lewd. Specifically, the prosecutor urged the jurors to ignore “[w]hat is acceptable or appropriate somewhere else,” and to focus their attention on the question: “Doesn’t a child reach an age where her nakedness is private?” According to the prosecutor, the girl “is too old to be depicted in a picture with her developing breasts and her pubic area on display,” and “given the age of this girl and her physical development and the nakedness of her breasts and her pubic area in this picture, isn’t that what makes this lewd?” In this manner, the Commonwealth urged the jury to convict the defendant on a theory that is untenable on both statutory and constitutional grounds.
The defendant objected to the prosecutor’s closing and specifically requested a curative instruction that nudity alone cannot make a photograph “lewd.” When the judge declined to provide such an instruction, the defendant specifically objected to its absence. Because the principle that mere nudity is insufficient is well established by the case law, the requested instruction should have been given.38 Regardless of whether the defendant in the *329end waived the issue (through his counsel’s stating that she was “content” with the instructions), the absence of adequate instruction on lewdness created a substantial risk of a miscarriage of justice. Indeed, the import of the inadequate instruction is self-evident given that the Commonwealth urged the jury to conclude that mere nudity of an adolescent girl was by itself sufficient. In my view, it is not only possible, but likely that adequate instruction would have made a difference here.39 And that leaves aside the separate problem, recognized by the majority, that “the judge could have stated more clearly that the defendant’s subjective reaction to the photograph was not relevant to the jury’s determination of the lewdness of the photograph itself.” Ante at 309.
D. Introduction of the other photographs. Finally, while I agree with the majority that the judge did not abuse her discretion in allowing in evidence the other images that the defendant viewed, I arrive at that conclusion by a different route. There is little merit to the Commonwealth’s protests about its need to use the other photographs, and its entitlement to do so. That is because the Commonwealth’s evidence that the defendant knowingly possessed the photograph was not only strong, it was uncontested.40 At the same time, the other photographs are not, in the end, significantly prejudicial. In fact, these photographs had some exculpatory value that lay unexploited. Most of them simply reinforced the “day at the beach” quality of the images the defendant was viewing.41 While a few of the photographs showed other young girls in undeniably lewd (but clothed) poses *330(see note 10, supra), those photographs could have been used to underscore the absence of lewdness in the one photograph for which the defendant was charged.42
Conclusion. In my view, the photograph of a naked girl playing on a beach does not rise to the level of a “lewd exhibition.” Therefore, even if the indictment adequately charged the defendant with conduct amounting to a crime, the Commonwealth’s proof of such conduct failed as a matter of law.
Of course, the fact that the case has now been fully tried (twice) creates an : .ntive to find the indictment sufficient. But as to this issue, I note that once *311the trial judge sua sponte raised the problem about the wording of the indictment during the defendant’s first trial, defense counsel sought dismissal of the indictment on this basis but volunteered to the judge that the defendant could be retried on a proper indictment without violating double jeopardy. Had the problem with the indictment been addressed at that point, all that might have been lost was one day of jury selection.
The body of the indictment here charged the defendant only with conduct that is indisputably not a crime: possessing a photograph of a child.
In Palladino, the court found that several complaints charging possession of obscene materials with intent to distribute them, G. L. c. 272, § 28A, repealed by St. 1974, c. 430, § 2, were insufficient because they failed to allege the defendant’s “knowledge of the nature of the material.” 358 Mass. at 32. That element, which is distinct from the general scienter for possession offenses, is “knowledge of a more specific kind,” Commonwealth v. Bacon, 374 Mass. 358, 361 (1978), and proof of it is constitutionally required to convict a defendant of possession of either obscenity or child pornography. Smith v. California, 361 U.S. 147, 152-154 (1959). New York v. Ferber, 458 U.S. 747, 765 (1982). Section 29C(vii) enumerates each scienter separately, requiring proof both that a defendant “knowingly purchase[d] or possesse[d]” the material and that he did so “with knowledge of the nature or content thereof.” The indictment in this case, like the complaints in Palladino, supra, failed to allege the “knowledge of the nature or content” element. Subsequent Massachusetts cases have narrowed the thrust of Palladino without overruling it. Together, Palladino and the cases that distinguish it establish that while the general knowledge element need not be alleged in a complaint or indictment, the “knowledge of a more specific kind” at issue in speech cases must be. Commonwealth v. Bacon, supra at 360-361. See Commonwealth v. Green, 399 Mass. 565, 567 (1987); Commonwealth v. Militello, 66 Mass. App. Ct. 325, 339 (2006).
I agree with the majority’s analysis of the O’Dell issue. See Commonwealth v. O’Dell, 392 Mass. 445 (1984).
The photograph was admitted in evidence and it shows whatever it shows. The only possible debate regarding the photograph has to do with whether the girl’s genitals are visible. The photograph does not support the Commonwealth’s position, asserted for the first time on appeal, that “[a] portion of her genitalia is exposed.”
The majority emphasizes that the girl “is not smiling.” Ante at 301. True enough. Neither is she frowning. In fact, her expression is utterly unremarkable, and no one can plausibly claim to know what she may have been thinking or feeling when this moment was captured. The Commonwealth’s claim in its brief that the girl looks “sad and defeated” amounts to pure fancy.
As noted, the girl’s genitals cannot be seen. Whether the photograph shows her “pubic area” depends on how you define where that area begins and ends. Neither the majority nor the parties has focused on this issue. For a discussion of the issue in anatomical detail, see United States v. Knox, 977 F.2d 815, 819-820 (3d Cir. 1992), vacated on other grounds, 510 U.S. 939 (1993), S.C., 32 F.3d 733 (3d Cir. 1994), cert. denied, 513 U.S. 1109 (1995).
Otherwise, the term “lewd exhibition” would be rendered mere surplusage. See, e.g., Commonwealth v. Disler, 451 Mass. 216, 227 (2008) (“Every word in a statute should be given meaning . . . , and no word is considered superfluous”). Under the express words of the statute, there needs to be a “lewd exhibition” even of “partially developed” breasts. This directly undercuts the suggestion that a photograph is “lewd” by definition if it shows a naked girl “on the cusp of puberty.” Ante at 305, quoting from United States v. Frabizio, 459 F.3d 80, 86 (1st Cir. 2006).
The parallel subsection of the Federal child pornography statute criminalizes images containing a “lascivious exhibition of the genitals.” 18 U.S.C. § 2256(2)(A)(v) (2006). The Federal courts have found that the terms “lewd” and “lascivious” are “virtually interchangeable” in this context. United States v. Wiegand, 812 F.2d 1239, 1243-1244 (9th Cir.), cert. denied, 484 U.S. 856 *314(1987). We thus may look to Federal constructions of the term “lascivious exhibition of the genitals” for instruction in our interpretation of G. L. c. 272, § 29C(vii).
Examples of this type of lewdness can be found among the other photographs that the defendant viewed (but for which he was not charged). For example, one such photograph shows a young girl sitting in an armchair, looking at the camera, with her legs spread as wide as possible to reveal a view of her underpants. In this manner, the girl is shown, albeit clothed, in an unnatural pose self-evidently designed to sexualize her.
In Steen, the defendant had secretly filmed a girl in a private room at a tanning salon. Even though the film revealed the girl’s naked pubic area, the *315United States Court of Appeals for the Fifth Circuit concluded that — as a matter of law — the film did not constitute a “lascivious exhibition.” 634 F.3d at 828. In the case at bar, the girl was not only photographed seemingly unaware as she “pursu[ed] activities unrelated to sex,” ibid.., she was doing so in what was, from all appearances, a fully public setting.
The limited import of evaluating whether a child’s nakedness is the focus of a visual depiction is underscored by both the majority and dissenting opinions in Osborne v. Ohio, 495 U.S. 103. That case involved an Ohio statute that on its face punished the possession of images of naked children. The Ohio Supreme Court had ruled that only “lewd exhibitions” or those that had a “graphic focus” on the child’s genitalia could be subject to prosecution. Id. at 107, 113. In upholding the Ohio statute against a claim that it was facially overbroad, the majority opinion applied its own further limiting gloss: “The crucial question is whether the depiction is lewd, not whether the depiction happens to focus on the genitals or the buttocks.” Id. at 114 n.11. In dissent, in reference to the Ohio Supreme Court’s “graphic focus” language, Justice Brennan emphasized: “Not only is this factor dependent on the perspective and idiosyncrasies of the observer, it is also unconnected to whether the material at issue merits constitutional protection. Simple nudity, no matter how prominent or ‘graphic,’ is within the bounds of the First Amendment. Michelangelo’s ‘David’ might be said to have a ‘graphic focus’ on the genitals, for it plainly portrays them in a manner unavoidable to even a casual observer.” Id. at 138 (Brennan, J., dissenting).
The fourth Dost factor simply asks whether the child is nude, a factor that alone cannot support a finding of “lewdness.” 636 F. Supp. at 832. See Steen, 634 F.3d at 827 (“Surreptitiously filming a nude tanner, on its own, does not meet the standard for producing child pornography”).
In defining “lascivious intent,” the Legislature incorporated a version of the first five Dost factors but omitted the sixth. See G. L. c. 272, § 31. Thus, the Supreme Judicial Court had no occasion to consider the sixth Dost factor in Bean (a prosecution pursuant to G. L. c. 272, § 29A, in which the Commonwealth had to prove “lascivious intent”). However, the omission of the sixth factor from the definition of “lascivious intent” appears to reflect the Legislature’s recognition that the sixth Dost factor largely duplicates the overarching inquiry into whether a photographer acted with “lascivious intent” in producing a photograph. Compare Doe v. Chamberlin, 299 F.3d 192, 196 (3d Cir. 2002) (“The final Dost factor simply puts again the underlying question: Is the exhibition lascivious?”).
For its part, the majority engages in no independent analysis of the sixth factor, instead concluding that the image is “designed to elicit a sexual response” because “the over-all design of the visual depiction is to direct the viewer to the girl’s developing breasts and pubic area” and “the pose could be seen as sexually suggestive.” Ante at 305. In other words, the majority asserts that because in its view the first and third Dost factors are satisfied, the sixth factor is as well.
Given that Dost itself was a prosecution for production of child pornography, it is unsurprising that the Dost factors are tailored to such cases. The failure to distinguish between production and possession cases helps explain the frustration and confusion expressed by some courts when attempting to apply the sixth factor. See, e.g., United States v. Villard, 885 F.2d 117, 125 (3d Cir. 1989); Amirault, 173 F.3d at 35; United States v. Hill, 322 F. Supp. 2d 1081, 1086 (C.D. Cal. 2004), aff’d, 459 F.3d 966 (9th Cir. 2006), cert, denied, 549 U.S. 1299 (2007).
See, e.g., United States v. Rivera, 546 F.3d at 251-252, and cases cited. See also Adler, The Perverse Law of Child Pornography, 101 Colum. L. Rev. 209, 261-265 (2001) (criticizing the Dost factors for encouraging courts to view images of children the way a pedophile would, allowing “everything [to] become[] child pornography in the eyes of the law”).
In fact, the majority opinion in United States v. Frabizio eschewed any efforts to explain what the word means, lest the court inadvertently make the scope of the statute’s prohibitions narrower than Congress intended. However, the United States Supreme Court has made clear that it is the opposite concern that is paramount. Where a defendant’s speech is alleged to fall within a category unprotected by the First Amendment, the role of the courts is “to confine the perimeters of any unprotected category within acceptably narrow limits in an effort to ensure that protected expression will not be inhibited.” Bose Corp. v. Consumers Union of U.S., Inc., 466 U.S. 485, 505 (1984). In such cases, “[providing triers of fact with a general description of the type of communication whose content is unworthy of protection has not, in and of itself, served sufficiently to narrow the category, nor served to eliminate the danger that decisions by triers of fact may inhibit the expression of protected ideas.” Ibid. See United States v. Frabizio, 459 F.3d at 94 (Torruella, J., concurring) (criticizing the majority opinion for not providing further guidance).
The judge told the jury that “as far as defining lewd exhibition, the courts say those are words that are within the ken of ordinary jurors and that they should be given the meaning that ordinarily attach to those words.”
Here, having been convicted on a prior occasion of ten counts of the same or similar charges, the defendant received a mandatory prison sentence of ten to fifteen years (to be served concurrently with a sentence he was already serving).
Commonwealth v. Kenney upheld G. L. c. 272, § 29C, against a claim that it was facially overbroad or void for vagueness. However, the specific arguments that the court rejected involved the term “actually or by simulation,” which appears in the statute’s first six subsections. The last subsection, *320proscribing “lewd exhibitions,” is referenced only in passing in a footnote. The United States Supreme Court did specifically reject an argument that the use of the term “lewd exhibition of the genitals” made a New York child pornography statute constitutionally overbroad on its face, but it did so because “in view of the examples of ‘sexual conduct’ provided by the statute, [the Court was] willing to assume that the New York courts would not ‘widen the possibly invalid reach of the statute by giving an expansive construction to the proscription on “lewd exhibition^] of the genitals.” ’ ” Osborne, 495 U.S. at 131 (Brennan, J., dissenting), quoting from New York v. Ferber, 458 U.S. at 773.
The Dost factors are best understood as one means of trying to measure whether a photograph meets this standard.
Typically, the ordinary usage of the language of a statute provides its principal source of meaning. However, one must look to other sources here, because “lewd” has an elastic meaning. Dictionary definitions confirm that the word is commonly used to refer to things ranging from the merely “vulgar” or “indecent” (at one end of the spectrum) to the downright “obscene” (at the other end). See, e.g., Oxford American Dictionary 381 (1980); Merriam-Webster’s Collegiate Dictionary 715 (11th ed. 2005).
Cf. Commonwealth v. Oakes, 407 Mass. 92, 97 (1990) (Oakes II) (“posing” a minor in order to photograph her is “expressive conduct” entitled to less protection than the “pure speech” of disseminating any photographs produced).
See United States v. Stevens, 130 S. Ct. at 1586 (noting that child pornography presents a “special case” where the speech is “an integral part” of the independently criminal act of producing child pornography).
Section 29C of G. L. c. 272 criminalizes the possession of visual material depicting children engaged in seven different categories of activity. The first six involve specific sexual acts (real or simulated), such as sexual intercourse, masturbation, or “lewd fondling, touching, or caressing.” These six categories establish fairly bright-line rules. For example, whether a particular photograph depicts a child “actually or by simulation engaged in any act of sexual intercourse with any person or animal” is unlikely to be subject to extended debate. G. L. c. 272, § 29C(i). Nor can there be any reasonable debate that photographs of children engaged in such acts lie outside First Amendment protection.
Cf. Manuel Enterprises, Inc. v. Day, 370 U.S. 478, 482-483 (1962) (“While in common usage the words [obscene, lewd, lascivious, indecent, filthy or vile] have different shades of meaning, the [Comstock Act] since its inception has always been taken as aimed at obnoxiously debasing portrayals of sex” [footnote omitted]).
The analysis that follows focuses on the term “lewd” and assumes arguendo that an unposed snapshot taken in public could amount to a lewd “exhibition.” However, even this is in doubt. The term “exhibition” suggests that the Legislature had in mind portrayals of children that had been posed (or otherwise actively engaged in the taking of the picture), not passive subjects who may well be unaware of the photograph’s existence. To the same effect is *323the characterization in G. L. c. 272, § 31, of “lewd exhibition” as a form of sexual “conduct.” Strictly speaking, the definition of “sexual conduct” in § 31 expressly applies only to sections other than § 29C.
In its brief, the Commonwealth suggested that the unknown photographer actively posed the girl to simulate a sex act: “Her crooked hand points at her genitalia. She is either deliberately covering her genitalia in a posed manner, or is posed to make it appear that she is about to touch herself there or masturbate.” At oral argument, the Commonwealth wisely distanced itself from that suggestion.
Notably, the literature suggests that “[pjedophiles may prefer ‘innocent’ pictures. . . . According to certain theorists, the stimulation derived from a picture can be inversely proportional to its overtly sexualized nature. It is often the very innocence —• the sexual naiveté — of the child subject that is sexually stimulating.” (Footnotes omitted.) Adler, Inverting the First Amendment, 149 U. Pa. L. Rev. 921, 944-945 (2001). An inquiry into whether the defendant in a child pornography case found the images lewd could thus potentially turn on factors diametrically opposed to those that support depriving the image of First Amendment protection.
The majority acknowledges that the taking of the photograph did not amount to any sexual abuse of the girl, but it declares that her “developing sexuality is being exploited by the taking and sharing of this revealing snapshot of her breasts and pubic area.” Ante at 307. Although the majority does not explain what it means by the term “exploited” or how that standard is met here, it appears that the majority rests on the fact that the girl has been shown “at a vulnerable stage in her sexual and emotional development [and that] her privacy is being invaded.” Ibid. But, as already discussed, mere nudity, even of a developing child, is not enough to make an image “lewd.”
The majority engages in this analysis even while recognizing that a photograph may still be lewd even if it has social value, and may still be subject to First Amendment protection even if it does not. Ante at 306.
A work’s literary, artistic, political, or scientific value is, of course, relevant to determining whether that work is obscene. See Miller v. California, 413 U.S. 15, 24 (1973). The Commonwealth has never suggested that the photograph at issue here is obscene, nor could it.
If the photograph here is swept within the statute’s prohibitions, then it is difficult to see how other nonsexual depictions of naked adolescents that society might not want to ban (such as those in medical textbooks) would not be swept in as well. When pressed on this point at oral argument, the Commonwealth responded simply that it would use its enforcement discretion not to prosecute cases involving such materials (even if they were used by a pedophile for sexual excitement). “But the First Amendment protects against the Government; it does not leave us at the mercy of noblesse oblige.” United States v. Stevens, 130 S. Ct. at 1591. A statute incorporating such a broad interpretation of “lewdness” would be intolerable, even if “the Government promised to use it responsibly.” Ibid.
Bean, 435 Mass. 708, is not to the contrary. That was not a possession case. Instead, the defendant there (a photographer) was charged with posing a minor in the nude, and the Commonwealth was required to prove (unlike here) that the defendant was acting with “lascivious intent.” Id. at 709. Compare G. L. c. 272, § 29A, with G. L. c. 272, § 29C. Thus, whether the photographer was pursuing artistic aims was directly relevant to the court’s analysis. Significantly, a photographer may be convicted under § 29A for posing a minor in a state of nudity even if the photographs produced are indisputably not child pornography, and would therefore be legal to possess. Oakes II, 407 Mass, at 97-98 & n.8. It is also undoubtedly the case that a photographer, though not acting with lascivious intent (and thus not subject to prosecution under § 29A), could nevertheless produce photographs which would fall outside the First Amendment’s protection. In addition, while the defendant here does not claim the photograph has artistic value (and none readily appears to my eye), I note that any distinction suggested by Bean between art photographs and “mere snapshots,” 435 Mass, at 715, may be a difficult one to draw. In recent years, the art world has seen a burgeoning appreciation for the artistic value of snapshots and other “found art” that freezes random moments in time. See, for example, “The Art of the American Snapshot, 1888-1978,” an exhibition of photographs selected from the collection of snapshot collector Robert E. Jackson, which were displayed at the National Gallery of Art in Washington, D.C., from October 7 to December 31, 2007, and memorialized in a bound volume of the same name.
See Osborne v. Ohio, 495 U.S. at 132 (Brennan, J., dissenting) (“Many would characterize a photograph of a seductive fashion model or alluringly posed adolescent on a topless European beach as ‘lewd,’ although such pictures indisputably enjoy constitutional protection”).
Although the prosecutor acknowledged at sidebar that whether the defendant himself found the photograph lewd was irrelevant, she nevertheless urged the jury to focus on the image of a middle-aged man “trolling” photographs of naked children on a Russian Web site until he found the one he wanted. We recognize that the judge instructed the jury that what the defendant “was going to do with the material is not for you to consider in this case,” and we ordinarily presume that the jury followed a judge’s instructions. In the context of this case, such an instruction has as realistic a chance of successfully getting the jury to put the defendant’s thoughts out of their minds as would a plea to “stop thinking about the elephant in the room.”
The majority acknowledges that this is the state of the law but provides no convincing explanation for why such an instruction was not warranted.
This conclusion is supported by the different outcomes of the two trials. The only significant difference in the presentation of the case at the two trials had to do with the jury instructions. Although the judge did not provide the defendant his requested “mere nudity is not enough” instruction in the first trial either, she did instruct the first jury that whether the defendant’s conduct was appropriate “has no place here,” because “the fact that it’s inappropriate does not alone or necessarily mean it violates the statute.”
Although there was no stipulation to that effect, defense counsel informed the jury during her opening statement that the defendant was not contesting possession, and she stuck by that statement throughout the trial (as she did in the first trial).
Most of the photographs are other candid snapshots of people at beaches, some walking along or lying on the sand, some in the water. The people, most but not all of whom are naked, include adults, children, and people of indeterminate age (either because their backs are to the camera or the photograph *330is too blurry). In the setting of a nude beach, it is hardly “unnatural” for a child to be nude, regardless of whether the child passes the majority’s new “Coppertone-age” test. At least one of the additional beach photographs shows the same girl depicted in the photograph for which the defendant was charged. This second photograph reinforces the candid, snapshot quality of the first. A second naked child, a toddler, has wandered into the camera’s field of view, and a portion of a leg of someone lying in the sand can also be seen. To the extent that the majority suggests that the girl is trying to shield her breasts from the photographer in that second photograph, this is belied by the positioning of her forearm and the fact that her fingers are spread apart.
I also note that the defendant did not move to bifurcate resolution of the question whether the photograph was lewd from that of whether he knowingly possessed it. Having the jury initially consider whether the photograph was lewd would have presented an obvious means of allowing that issue to be tried without any extraneous considerations presented to the jury.