Opinion
CRIMINAL 1:22-CR-192
08-23-2023
UNITED STATES OF AMERICA v. JONATHAN LEE DEBUS, Defendant
MEMORANDUM
CHRISTOPHER C. CONNER UNITED STATES DISTRICT JUDGE
Defendant Jonathan Lee Debus pled guilty on September 12, 2022, to one count of receiving child pornography in violation of 18 U.S.C. § 2252A(a)(2). Presently before the court are the parties' memoranda in aid of sentencing. In addition to addressing the salient 18 U.S.C. § 3553(a) factors, Debus objects to application of a pair of five-level enhancements to his offense level. The first implicates defendants who “engaged in a pattern of activity involving the sexual abuse or exploitation of a minor,” see U.S.S.G. § 2G2.2(b)(5) (U.S. SENT'G COMM'N 2018); the second applies when the offense of conviction involves 600 or more images of child pornography, see Id. § 2G2.2(b)(7)(D). We will sustain Debus's objection to the pattern-of-activity enhancement and overrule his objection based upon the number of images attributed to him.
The following factual narrative derives from undisputed portions of Debus's presentence investigation report (“PSR”), (see Doc. 38), along with testimonial evidence adduced at an evidentiary hearing on April 19, 2023. The court reporter has provided the court with a rough transcript of that hearing. Citations to the rough transcript are abbreviated “4/19/23 Tr._,” and the government's hearing exhibits are cited as “Gov't Ex._.” Pagination of the rough transcript may vary from pagination of the official transcript. This recitation of facts reflects the court's credibility determinations. The court finds the testimony of Special Agent Angela Strause highly credible.
In the fall of 2008, when Debus was 29 years old, he had sexual intercourse with a 14-year-old girl in York County, Pennsylvania, approximately 15 times over several weeks. (See Doc. 38 ¶ 52). Debus pled guilty in state court to unlawful contact with a minor and statutory sexual assault, and he was sentenced to 18 to 60 months' imprisonment. (See Id. ¶¶ 20, 52). More than a decade later, in December 2021, a cloud-based service provider operating on Verizon's behalf informed federal investigators an account associated with Debus's telephone number uploaded suspected child pornography to its network. (See id. ¶¶ 5-6). Agents with the Federal Bureau of Investigation obtained a search warrant for Debus's account and found “numerous image and video files depicting illegal child sexual abuse material (CSAM).” (See id. ¶ 7). Special Agent Angela Strause-a 13-year veteran of the FBI specializing in human-trafficking cases and crimes against children-arrested Debus on May 12, 2022, and Mirandized him. (See id. ¶ 18; 4/19/23 Tr. 22:7-12). Debus admitted to viewing child pornography; his habits generally included accessing child pornography “every day or two,” watching it, and then deleting the files that night or the next day. (See 4/19/23 Tr. 22:13-18, 34:2-10).
We use “child pornography” and “CSAM” interchangeably.
Agents catalogued the contents of Debus's cloud-based account and “identified approximately 47 video files, 17 image files, and two computer-generated imagery (CGI) files depicting CSAM of prepubescent children.” (See Doc. 38 ¶ 9). A forensic examination of Debus's cell phone revealed an additional 70 unique images of child pornography he had received the evening before his arrest. (See id. ¶ 9). His laptop also contained four unique images of CSAM. (See id.).
Debus pled guilty pursuant to a written agreement. The United States Probation Office prepared a PSR that calculates his total offense level (35) using the five-level enhancements he now challenges. With respect to the image enhancement, the PSR attributes a total of 3,544 images to Debus, of which only 17 are traditional image files. (See Doc. 38 ¶ 34). To reach that number, the probation officer relies upon commentary to the Sentencing Guidelines providing “[e]ach video, video-clip, movie, or similar visual depiction shall be considered to have 75 images.” See U.S.S.G. § 2G2.2, cmt. n.6(B)(ii). Debus objects on the ground that the United States Sentencing Commission's 75:1 rule contravenes recent decisions from the United States Supreme Court and our court of appeals. (See Doc. 41 at 3-13 (citing, inter alia, Kisor v. Wilkie, 588 U.S., 139 S.Ct. 2400 (2019); United States v. Adair, 38 F.4th 341 (3d Cir. 2022))). He asserts he only “received 66 images” because each video should count as a single image, no matter its duration. (See Id. at 12). The government stands by the 75:1 rule. (See Doc. 49 at 8-24 (citing, inter alia, United States v. Phillips, 54 F.4th 374 (6th Cir. 2022))).
Our assessment of the parties' conflicting proposals revealed a third option- a frame-based standard-from Judge Larsen's concurring opinion in Phillips. See Phillips, 54 F.4th at 386-98 (Larsen, J., concurring in the judgment). That metric calculates an image number for each video by identifying the number of frames in which child pornography is depicted. See Id. at 398. We advised the parties of this potential alternative and scheduled an evidentiary hearing for the purpose of determining “the number of unique still-frame images depicting child pornography that may be derived from each video attributed to” Debus. (See Doc. 54). Agent Strause testified at the hearing. She prepared by examining the still-image and video files Debus received, watching each video in its entirety, and drafting a detailed report of her investigative methods and findings. (See 4/19/23 Tr. 1:25-3:13, 6:20-9:9; see also Gov't Ex. 1). She also compiled a spreadsheet listing the video files' properties, including file names, file types, frame rates (in frames per second (“fps”), if known), total duration and total CSAM duration (both in seconds), total frames of CSAM, and descriptions of the subject matter depicted. (See Gov't Ex. 2).
We summarize Agent Strause's findings as follows. Agent Strause examined 102 still-image files and 46 video files depicting child pornography recovered from Debus's account and devices. (See Gov't Ex. 1 at 2; see also 4/19/23 Tr. 16:7-14, 34:22-35:2). After excluding various duplicate, incalculable, or otherwise non-CSAM files, Agent Strause reports the total duration of the 37 remaining video files is 12,089 seconds (3 hours, 21 minutes). (See Gov't Ex. 1 at 4; see also 4/19/23 Tr. 18:1121). The total duration of child pornography depicted in those 37 videos is 10,760 seconds (2 hours, 59 minutes). (See Gov't Ex. 1 at 4; see also 4/19/23 Tr. 18:14-18).
There is a discrepancy between Agent Strause's still-image count and the estimate of still images set forth in the PSR. (See Doc. 38 ¶¶ 9, 19 (totaling 91 still images)). The parties do not dispute the accuracy of Agent Strause's tally. (Compare 4/19/23 Tr. 38:6-8, with id. at 37:10-12). We concur, and we attribute 102 still images of CSAM to Debus's devices.
Debus's video collection consisted of seven different file types: GIF, 3GP/3GPP, AVI, MOV, MP4, MPG, and WMV. (See Gov't Ex. 2 at 1-14). Of these, GIF, or “Graphics Interchange Format,” files present unique difficulties because they “can be used to create still images or a set of short animations which have the appearance of videos, as they repeat on a continuous loop or cycle.” (See Gov't Ex. 1 at 2). Debus possessed six GIF files (Files #1-6). Agent Strause used a stopwatch to determine the duration of each GIF's loop or cycle; they ranged between 1.5 and 14 seconds. She was unable to calculate a frame rate or frame total and therefore excluded these six files from her calculations. (See id. at 2-3; see also 4/19/23 Tr. 14:19-15:25). Agent Strause also excluded one video (File #9) from her tally as a duplicate, and two others (Files #34 and #35) as computer-generated imagery (“CGI”). (See Gov't Ex. 1 at 3; see also 4/19/23 Tr. 14:13-18).
Agent Strause found frame-rate information in the metadata of 34 video files. (See Gov't Ex. 1 at 2-4; see also 4/19/23 Tr. 7:22-8:4, 11:15-22). Frame rates ranged from 10 fps to 30 fps, with an average frame rate of 26.65 fps. (See Gov't Ex. 1 at 2, 4; see also 4/19/23 Tr. 18:22-19:4). Videos with lower frame rates “tended to be of poorer quality, i.e., movements appeared to be slower, and the content was a bit more pixelated.” (See Gov't Ex. 1 at 2; see also 4/19/23 Tr. 12:7-11). Three video files (Files #13, #45, and #46) had a total CSAM duration of 198 seconds but lacked frame-rate data. (See Gov't Ex. 1 at 2; see also 4/19/23 Tr. 14:8-12). Agent Strause multiplied the duration (in seconds) of child pornography depicted in each of the 34 video files for which frame-rate data was available by their respective frame rates and found a total of 306,111 individual frames of CSAM. (See Gov't Ex. 1 at 3-4; see also 4/19/23 Tr. 13:24-14:1, 16:21-18:4). Counting each of those 306,111 frames as images and adding the 102 still-image files recovered from Debus's devices equals 306,213 images of CSAM. (See Gov't Ex. 1 at 3-4; see also 4/19/23 Tr. 18:5-10).
II. Legal Standard
In sentencing, a court must engage in a three-step process pursuant to Gall v. United States, 552 U.S. 38 (2007). First, it must calculate the advisory Guidelines range. See United States v. Wright, 642 F.3d 148, 152 (3d Cir. 2011). Second, it must formally rule on any motions for departure and state the impact, if any, of such ruling on the Guidelines calculation. See id. Third, the court is required to exercise its discretion and consider the sentencing factors set forth in 18 U.S.C. § 3553(a), which may vary from the advisory Guidelines range. See id. The government generally bears the burden of proving any facts that may warrant a Guidelines enhancement by a preponderance of the evidence. See United States v. Ali, 508 F.3d 136, 143 (3d Cir. 2007).
III. Discussion
Whether the enhancements under Guidelines Sections 2G2.2(b)(5) and (b)(7)(D) apply to Debus's case turns upon what weight, if any, we must give the Sentencing Commission's commentary supplementing those provisions. The United States Supreme Court's decision in Kisor v. Wilkie, 588 U.S., 139 S.Ct. 2400 (2019), and our court of appeals' recent decisions in United States v. Nasir, 17 F.4th 459 (3d Cir. 2021) (en banc), and United States v. Adair, 38 F.4th 341 (3d Cir. 2022), guide this analysis.
Kisor marked a sea change in the Supreme Court's administrative law jurisprudence. Before Kisor, the Court counseled that the Sentencing Commission's commentary should control-even if it expanded the Guidelines, and “particularly when the commentary is ‘interpretive and explanatory'”-so long as the Commission's determinations were not “plainly erroneous or inconsistent with the regulation.” See Nasir, 17 F.4th at 470 (quoting Stinson v. United States, 508 U.S. 36, 45, 46-47 (1993)). Lower courts routinely misread that passage as sanctioning “uncritical and broad deference to agency interpretations of regulations.” See id. at 471. Kisor corrected that misreading, teaching “that Auer, or Seminole Rock, deference should only be applied when a regulation is genuinely ambiguous.” Id. (citing Kisor, 139 S.Ct. at 2414-15). Courts “must exhaust all the traditional tools of construction” to resolve latent ambiguities before even considering deference. See id. (quoting Kisor, 139 S.Ct. at 2415).
Auer v. Robbins, 519 U.S. 452 (1997); Bowles v. Seminole Rock & Sand Co., 325 U.S. 410 (1945).
This new paradigm, however, provided “only half of the framework for analyzing the Commission's interpretive commentary.” Adair, 38 F.4th at 348. As our court of appeals explained following Kisor, existence of a genuine ambiguity in the Sentencing Guidelines is merely a necessary condition for Auer deference, not a sufficient one. See id. Once genuine ambiguity is established, we must determine how much weight the Commission's commentary deserves. See id. To that end:
The agency's reading must be “reasonable[,]” as informed by “[t]he text, structure, history, and so forth[,]” which “establish the outer bounds of permissible interpretation.” A court “must make an independent
inquiry into whether the character and context of the agency interpretation entitles it to controlling weight[,]” including whether it is the agency's “official position[.]” Moreover, an agency's interpretation must “in some way implicate its substantive expertise” if it is to be given controlling weight, since “[s]ome interpretive issues may fall more naturally into a judge's bailiwick.” Finally, the reading must “reflect fair and considered judgment” and not simply be a “convenient litigating position.”Nasir, 17 F.4th at 471 (alterations in original) (quoting Kisor, 139 S.Ct. at 2415-17); see Adair, 38 F.4th at 349 (agency interpretations must “fall[] within the regulation's zone of ambiguity”). Context now drives “the degree of deference” we may afford agency interpretations. Nasir, 17 F.4th at 471.
Our court of appeals has underscored the practical implications of this seismic shift in administrative jurisprudence by cautioning district courts against reflexive reliance upon circuit precedents that predate Kisor's “reprised standard for Auer deference.” See Adair, 38 F.4th at 348-49. Past decisions that deferred to the Commission's interpretive comments “without engaging in the Kisor process” no longer “automatically retain . . . controlling force.” See id. at 349. Rather, “to remain binding, such a decision must have (presciently) complied with the Kisor process: a genuine-ambiguity analysis followed by an independent evaluation of the character and context of the agency's interpretation.” Id. (citing Nasir, 17 F.4th at 470-71). Recent precedents have charted a clear course for navigating challenges that implicate the Guidelines' commentary. We heed the directives of our court of appeals, and now turn to Debus's objections.
A. Pattern of Activity
Debus objects to paragraph 32 of the PSR, which adds five levels to his base offense level pursuant to U.S.S.G. § 2G2.2(b)(5). (See Doc. 38 ¶ 32). That section applies when a defendant “engaged in a pattern of activity involving the sexual abuse or exploitation of a minor,” defined by the Guideline's commentary as
any combination of two or more separate instances of the sexual abuse or sexual exploitation of a minor by the defendant, whether or not the abuse or exploitation (A) occurred during the course of the offense; (B) involved the same minor; or (C) resulted in a conviction for such conduct.U.S.S.G. § 2G2.2, cmt. n.1. Debus asserts this definition “impermissibly expands” the Guideline's scope and fails both steps of Kisor's interpretive framework. (See Doc. 41 at 13-19). The government does not engage Debus's argument in that respect, though it “in no way concedes that the commentary is not entitled to controlling weight.” (See Doc. 49 at 25). As the government sees it, Section 2G2.2(b)(5) applies to Debus under any “plausible definition of pattern” because he had sexual intercourse with a minor 15 times in one month. (See id. at 26). We agree with Debus that Application Note 1 improperly broadens the enhancement's reach, but we arrive at that conclusion from a different angle. Even if Debus's repeated sexual abuse of a minor over a monthlong period 15 years ago constituted a “pattern of activity,” Section 2G2.2(b)(5) does not apply to acts unrelated to the offense of conviction.
We begin with a brief history of the pattern-of-activity enhancement. In October 1991, Congress instructed the Sentencing Commission to amend Section 2G2.2 “to provide at least a 5 level increase for offenders who have engaged in a pattern of activity involving the sexual abuse or exploitation of a minor.” See Act of October 28, 1991, Pub. L. 102-141, tit. VI, § 632(1)(A), 105 Stat. 834, 876. The Commission obliged by adopting the language that now appears in
Section 2G2.2(b)(5). See U.S.S.G. App'x C, amend. 435 (U.S. SENT'G COMM'N Nov. 27, 1991). The Commission simultaneously amended its commentary to Section 2G2.2 by inserting the following:
‘Pattern of activity involving the sexual abuse or exploitation of a minor,' for the purposes of subsection (b)(4) [now (b)(5)], means any combination of two or more separate instances of the sexual abuse or the sexual exploitation of a minor, whether involving the same or different victims.See id.
Early challenges to the enhancement focused on its scope. After the First and Sixth Circuit Courts of Appeals held the pattern of activity in question must relate “to the offense of conviction,” see United States v. Chapman, 60 F.3d 894, 901 (1st Cir. 1995); United States v. Surratt, 87 F.3d 814, 819 (6th Cir. 1996), the Commission quickly amended its commentary to clarify no such relationship was required, see U.S.S.G. App'x C, amend. 537 (U.S. SENT'G COMM'N Nov. 1, 1996). Most federal appellate courts followed suit with varying degrees of uncritical deference to the amended commentary. See, e.g., United States v. Woodward, 277 F.3d 87, 91 n.2 (1st Cir. 2002); United States v. Reingold, 731 F.3d 204, 223-24 (2d Cir. 2013); United States v. Neilssen, 136 F.3d 965, 968-72 (4th Cir. 1998); United States v. Bacon, 646 F.3d 218, 221 (5th Cir. 2011); United States v. Gawthrop, 310 F.3d 405, 414 (6th Cir. 2002); United States v. Lovaas, 241 F.3d 900, 904-05 (7th Cir. 2001); United States v. Ashley, 342 F.3d 850, 852 (8th Cir. 2003); United States v. Garner, 490 F.3d 739, 742-43 (9th Cir. 2007); United States v. Groves, 369 F.3d 1178, 1185-88 (10th Cir. 2004); United States v. Anderton, 136 F.3d 747, 750 & n.2 (11th Cir. 1998).
Our court of appeals was no exception. In United States v. Olfano, 503 F.3d 240 (3d Cir. 2007), the Third Circuit confronted a defendant who, like Debus, pled guilty to receiving child pornography. See Olfano, 503 F.3d at 243. The district court imposed the pattern-of-activity enhancement based upon two acts of indecent assault Olfano committed as a juvenile in the mid- to late-1980s. See id. at 242. On appeal, the court looked to the Guidelines and concluded they placed no “explicit time limit on the previous activities that a court may consider in finding a ‘pattern of activity.'” Id. at 243. Nor could the court find any “case support for the proposition that previous events can be too remote in time to amount to a pattern.” Id.; see id. (agreeing with other circuits “that there is no temporal nexus necessary to establish a pattern of activity of sexual abuse or exploitation of a minor”). The “plain language” of the Guidelines posed “no problem” to treating Olfano's nearly two-decades-old juvenile offenses as a pattern of sexual misconduct. See id.
At first blush, Olfano appears to control the matter sub judice. But it runs headlong into Adair's admonition against hastily relying on past cases that failed to engage in “the Kisor process” before deferring to the Commission's interpretive commentary. See Adair, 38 F.4th at 348. The Olfano court did not conduct a genuine-ambiguity analysis or evaluate the character and context of the Guideline's commentary. Rather, the court began its discussion by quoting the commentary's definition of “pattern of activity,” see Olfano, 503 F.3d at 242, and then citing with approval the myriad circuit precedents collected above, see Id. at 244. The court said it relied upon the Guideline's “plain language,” but that brief discussion came at the very end of its analysis and was limited to finding the enhancement imposed no “similarity requirement” between prior crimes and the offense of conviction. See id. The court compared Olfano's past offenses-“inappropriate touching of juvenile females”-to his instant offense of receiving child pornography through a computer and deemed them “not so different in kind.” See id. The Supreme Court altered the landscape of administrative law in the decade and a half since Olfano issued, and our court of appeals has acknowledged Kisor upends any court of appeals' precedents that predated, and failed to engage in, the Court's “reprised standard for Auer deference.” See Adair, 38 F.4th at 349. We are constrained by Adair to conclude Olfano's circumscribed analysis retains no “controlling force.” See id.
Section 2G2.2(b)(5)'s plain language, read literally, admits of no temporal limitations; a defendant need only have “engaged” in certain activity. The enhancement's scope cannot be determined from that word alone, however-and therein lies ambiguity. Yet we need not determine anew whether the enhancement applies to offenses other than the offense of conviction. Chapman and Surratt “presciently” undertook the kind of context-driven textual analysis Kisor and Nasir now require, see Adair, 38 F.4th at 349, and concluded the enhancement could not apply to unrelated conduct. Like those earlier decisions, we have scrutinized the Sentencing Guidelines as a whole and find several contextual indicia supporting a more cabined construction. Because context resolves the enhancement's ambiguous scope, it is unnecessary to resort to the commentary. See Nasir, 17 F.4th at 471.
The strongest indicator the enhancement's application is confined to “past sexual abuse or exploitation activities related to the offense of conviction” is its placement in subsection (b) “under the listing of ‘Specific Offense Characteristics.'” See Surratt, 87 F.3d at 819 (emphasis added). Chapter Two of the Guidelines, where we find the enhancement, addresses “Offense Conduct,” while past acts unrelated to the offense of conviction are addressed in Chapter F our as part of the defendant's criminal history. See, e.g., U.S.S.G. § 4B1.5 (“Repeat and Dangerous Sex Offender Against Minors”). The general focus of each provision in Section 2G2.2 lends further support for an interpretation limited in scope. Eight of that section's various provisions address what “the offense involved.” See U.S.S.G. § 2G2.2(b)(3)(A), (C), (D), (E), (b)(4), (b)(6), (b)(7), (c)(1). One provision implicates “the material involved.” See id. § 2G2.2(b)(2). Another pertains to “the defendant's conduct.” See id. § 2G2.2(b)(1)(B). And three turn on “if the defendant” did something, e.g., “distributed in exchange for any valuable consideration”; “knowingly engaged in distribution, other than distribution described” in previous subdivisions; or “engaged in a pattern of activity.” See id. § 2G2.2(b)(3)(B), (F), (b)(5). It would be an unusual interpretive exercise to single out subsection (b)(5) as the sole provision applicable to conduct unrelated to the specific offense subject to enhanced punishment, as the commentary commands we must. We doubt Congress intended this result when it directed the Commission to adopt the pattern-of-activity enhancement.
Section 2G2.2's lengthy title refers to trafficking, receiving, or possessing “material involving the sexual exploitation of a minor,” not acts of sexual exploitation themselves. See Chapman, 60 F.3d at 897 (capitalization and emphasis modified) (quoting U.S.S.G. § 2G2.2, tit.). It is well settled that passively receiving child pornography does not constitute sexual abuse or exploitation of a minor. See id.; see also United States v. Ketcham, 80 F.3d 789, 794-95 (3d Cir. 1996). The Guidelines deal with sexual exploitation of minors separately in Section 2G2.1, as that provision's title “clearly indicates.” See Chapman, 60 F.3d at 897 (citing U.S.S.G. § 2G2.1 (“Sexually Exploiting A Minor . . .”)); see also Surratt, 87 F.3d at 819. Title 18 of the United States Code tracks that division. Compare 18 U.S.C. § 2251 (“Sexual exploitation of children”), with id. § 2252 (“Certain activities relating to material involving the sexual exploitation of minors”).
This transparent dichotomy evinces Congress and the Sentencing Commission's intent to partition the Guidelines to account for past acts of sexual exploitation unrelated to the offense of conviction somewhere other than Chapter Two. Indeed, the commentary advises “an upward departure may be warranted,” inter alia, “[i]f the defendant engaged in the sexual abuse or exploitation of a minor at any time (whether or not such abuse or exploitation occurred during the course of the offense or resulted in a conviction for such conduct).” See U.S.S.G. § 2G2.2, cmt. n.9 (emphasis added). The Commission “clearly knew how to” draft backward-looking provisions with expansive reach. See Chapman, 60 F.3d at 901 (citing then- existing iterations of U.S.S.G. § 2L1.2(1)-(2), which increased defendants' base offense level if they “previously” were deported after qualifying conviction). “The absence of similar language in subsection (b)[(5)],” and its conspicuous “classification] under the rubric of ‘Specific Offense Characteristics,'” compelled the Chapman court to conclude any patterns of activity must “relate to the offense of conviction.” Id.
We are persuaded by the First and Sixth Circuits' cogent analyses and adopt them as our own, as supplemented by the foregoing discussion. Section 2G2.2(b)(5) is not genuinely ambiguous when examined with “traditional tools of construction.” See Nasir, 17 F.4th at 471 (quoting Kisor, 139 S.Ct. at 2415). Context resolves any latent ambiguity about the enhancement's scope and forecloses consideration of the Sentencing Commission's more punitive position, which unreasonably expands the enhancement to reach conduct unrelated to a defendant's offense of conviction. Debus's past conduct lacks any logical or temporal connection to his admitted receipt of child pornography. He did not “engage” in a pattern-of-activity involving the sexual abuse of a minor when he received the images for which he has pled guilty. Thus, his previous conviction cannot be used to enhance his base offense level under Section 2G2.2(b)(5).
B. Images
Debus similarly invokes Adair in his challenge to paragraph 34 of the PSR, which employs the Sentencing Commission's 75:1 rule for images derived from video. (See Doc. 41 at 8-13). We agree with Debus that the term “image” is not genuinely ambiguous. An image is “a reproduction or imitation of the form of a person or thing”; “a visual representation of something,” as in “a picture produced on an electronic display (as a television or computer screen).” Image, MERRIAM-WEBSTER'S COLLEGIATE DICTIONARY (11th ed. 2003). No one disputes that a single photograph-whether analog or digital-constitutes an image. The parties disagree about how we should apply that term to circumstances involving a distinct visual medium: video. A video is “a recording of visual images and sound.” Videotape, MERRIAM-WEBSTER'S COLLEGIATE DICTIONARY (11th ed. 2003) (emphasis added).
Two decades ago, President George W. Bush signed into law the Prosecutorial Remedies and Other Tools to end the Exploitation of Children Today (“PROTECT”) Act of 2003, Pub. L. 108-21, tit. VI, § 401, 117 Stat. 650, 672-73. In promulgating the Act, Congress exercised its authority to directly amend Section 2G2.2(b) of the Sentencing Guidelines to create the “image table,” a fourtiered rubric for enhancing a defendant's base offense level for child pornography offenses. See U.S.S.G. § 2G2.2(b)(7). Then, as now, the image table provided as follows, If the offense involved
(A) at least 10 images, but fewer than 150, increase by 2 levels;
(B) at least 150 images, but fewer than 300, increase by 3 levels;
(C) at least 300 images, but fewer than 600, increase by 4 levels; and
(D) 600 or more images, increase by 5 levels.Id. The Commission adopted a conforming amendment effective the same day the PROTECT Act became law. See U.S.S.G. App'x C, amend. 649 (U.S. SENT'G COMM'N Apr. 30, 2003).
The PROTECT Act did not define “image,” so the Commission solicited public input to flesh out that term. See 68 Fed.Reg. 75340, 75342 (Dec. 30, 2003). The Commission proposed amending its commentary to define image as “any visual depiction described in 18 U.S.C. [§] 2256(5) and (8),” see id. at 75345, which, in turn, offered two complementary definitions:
(5) “visual depiction” includes undeveloped film and videotape, and data stored on computer disk or by electronic means which is capable of conversion into a visual image;
...
(8) “child pornography” means any visual depiction, including any photograph, film, video, picture, or computer-generated image or picture, whether made or produced by electronic, mechanical, or other means, of sexually explicit conduct, where-
(A) the production of such visual depiction involves the use of a minor engaging in sexually explicit conduct;
(B) such visual depiction is a digital image, computer image, or computer-generated image that is, or is indistinguishable from, that of a minor engaging in sexually explicit conduct; or
(C) such visual depiction has been created, adapted, or modified to appear that an identifiable minor is engaging in sexually explicit conduct.18 U.S.C. § 2256(5), (8) (effective Apr. 30, 2003, to Oct. 12, 2008).
The Commission also sought comment on the necessity of additional “instructions for counting images.” See 68 Fed.Reg. 75353 (Apr. 30, 2003). The Department of Justice urged the Commission to adopt a two- or three-level enhancement for crimes involving “video clips” due to the “increased harm caused by moving videos.” See Letter from Deborah J. Rhodes, Counselor to the Assistant Attorney General, to the U.S. Sentencing Commission, at 5 (Mar. 1, 2004). The Commission took a different approach. Instead of promulgating a separate enhancement for motion pictures, the Commission amended its commentary to create a special rule for video within its new definition of “images.” See U.S.S.G. App'x C, amend. 664 (U.S. SENT'G COMM'N Nov. 1, 2004). Application Note 6 now reads,
Application of Subsection (b)(7).-
(A) Definition of “Images”.-“ Images ” means any visual depiction, as defined in 18 U.S.C. § 2256(5), that constitutes child pornography, as defined in 18 U.S.C. § 2256(8).
(B) Determining the Number of Images.-For purposes of determining the number of images under subsection (b)(7):
(i) Each photograph, picture, computer or computer-generated image, or any similar visual depiction shall be considered to be one image. If the number of images substantially
underrepresents the number of minors depicted, an upward departure may be warranted.
(ii) Each video, video-clip, movie, or similar visual depiction shall be considered to have 75 images. If the length of the visual depiction is substantially more than 5 minutes, an upward departure may be warranted.Id. § 2G2.2, cmt. n.6. And thus the 75:1 rule was born.
The 75:1 rule faced few legal challenges in the federal appellate courts before the Supreme Court decided Kisor in 2019, and none of those cases involved a textual or historical analysis. Just two circuits have been asked to reexamine the rule in Kisor's wake. In United States v. Pratt, No. 20-10328, 2021 WL 5918003 (9th Cir. Dec. 15, 2021) (nonprecedential), a panel rejected a claim similar to Debus's principally because the Ninth Circuit continues to regard Guidelines commentary as “authoritative unless it violates the Constitution or a federal statute, or is inconsistent with, or a plainly erroneous reading of, that guideline.” See Pratt, 2021 WL 5918003 at *2 (quoting Stinson, 508 U.S. at 38). The panel found Section 2G2.2(b)(7) “genuinely ambiguous” in the alternative but without meaningful analysis, so it adds nothing to our understanding of the enhancement.
See, e.g., United States v. Leitenberger, 420 Fed.Appx. 188, 190 n.1 (3d Cir. 2011) (nonprecedential) (separation-of-powers challenge raised for first time on appeal forfeited); United States v. Garcia, 402 Fed.Appx. 768, 769 (4th Cir. 2010) (nonprecedential) (dismissing claim without analysis, finding no abuse of discretion); United States v. Miller, 665 F.3d 114, 122-23 (5th Cir. 2011) (rejecting challenge focused on fact enhancement applies in “large percentage” of cases, skewing sentences); United States v. Geerken, 506 F.3d 461, 464-66 (6th Cir. 2007) (dismissing arbitrariness challenge); United States v. Lynde, 926 F.3d 275, 279-82 (6th Cir. 2019) (same, when attack raised lack of empirical support for rule); United States v. Hanson, 693 Fed.Appx. 521, 522-23 (9th Cir. 2017) (nonprecedential) (district court free to rely on rule); see also United States v. Light, 753 Fed.Appx. 63, 65 (2d Cir. 2018) (nonprecedential) (expressing concern rule “lead[s] to many ‘run-of-the-mill' defendants receiving a full 5-level enhancement based on the arbitrary composition of that defendant's cache of pornography”).
The Sixth Circuit's decision in Phillips, by contrast, presents an extended historical discussion of the Sentencing Guidelines' treatment of child pornography offenses and the genesis of the image enhancement. See Phillips, 54 F.4th at 377-78 (citing U.S. SENT'G COMM'N, The History of the Child Pornography Guidelines 43-44 (2009)). The Phillips court purported to use an analytical framework like that adopted by our court of appeals in Nasir. See id. at 379-80. It found the term “image” to be ambiguous in the “wider” context of video notwithstanding the common definitions of those terms, see id. at 380, 382 (quoting Video, OXFORD ENGLISH DICTIONARY (2022) (“[A] recording of moving visual images . . . in a digital format.”)), and immediately turned to the Guideline's structure, history, and purpose-with a particular emphasis on the consequences of various methods of tallying images, see id. at 383.
One option the court considered was to treat each frame as an image. See Id. at 382 (citing Frame, OXFORD ENGLISH DICTIONARY (2022) (“[O]ne of the individual images on a strip of film; (later also) a single complete image in a series forming a television picture, film, or video sequence.”)). The majority quickly dismissed that approach on practical and penological grounds, suggesting parsing frames would be too “onerous and unrealistic” for courts, see id., and more punitive than Congress may have intended for defendants, see id. at 382-84 (opining frame-based approach “would nearly automatically vault” anyone who possesses video to maximum enhancement and offering series of hypotheticals to illustrate how approach could lead to “unreasonable results”). These concerns drove the court's conclusion that “image” is genuinely ambiguous and thus the Commission could clarify its meaning via supplemental commentary. Id. at 384. From there, the majority declared the rule “a reasonable interpretation” within the Guideline's zone of ambiguity, see id., and summarily found the character and context of the rule entitled it to controlling weight, see id. at 385-86 (finding rule set forth Commission's “official position,” as informed by public comments; implicates its substantive expertise in “promot[ing] sentencing goals”; and reflects its “fair and considered judgment” in responding directly to “challenges of applying the image table” (citing Kisor, 193 S.Ct. at 241617)). For those reasons, the court rejected Phillips' challenge and affirmed his conviction.
Judge Larsen saw things differently. The 75:1 rule satisfies none of Kisor's preconditions, in his view, and neither the government nor the majority even attempted to justify it as “a mere ‘interpretation' of the word ‘image.'” See id. at 387-88 (Larsen, J., concurring in the judgment). Among Judge Larsen's many sticking points was Application Note 6's declaration of what each video “shall be considered,” which is “the language of policy choice, not of interpretation.” See Id. at 388 (citing Sarmiento v. United States, 678 F.3d 147, 15 (2d Cir. 2012) (lawmakers use “shall be considered” to discard a term's “ordinary ‘plain English' meaning” in favor of “legal fiction” that “achieve[s] certain social policy goals”); Sturgeon v. Frost, 587 U.S., 139 S.Ct. 1066, 1076, 1081 (2019)). As well, he found the Commission's rule unreasonable, beyond the Guideline's plausible zone of ambiguity, and thus deserving of no deference. See id. at 389-90.
Judge Larsen set aside his colleagues' concern for fairness and consequences as improper considerations in a genuine-ambiguity analysis under Kisor. See Id. at 394-98. Carefully scrutinizing the Guideline's text, he identified three potential interpretations: an “image” may mean (1) a video, (2) a frame, or (3) “something like ‘imagery,' ‘impression,' or ‘scene.'” Id. at 390. Judge Larsen agreed with the majority that the first option-the preferred choice of both Phillips and Debus- failed out the gate because it was “contradicted by every dictionary definition and by common sense.” Id. at 392. He similarly believed the third option was nonsensical because it is unclear how a defendant could possess or distribute intangibles like “imagery,” “impressions,” or “scenes.” Id. at 393. Nor is it likely Congress expected courts to “comb through child-pornography videos” to identify “sequences of dramatic action” or “tally[] their ‘impressions' or the uses of ‘imagery.'” Id. All that remained was the frame-based definition, the most natural interpretation in Judge Larsen's mind. See id. at 391 (“An ‘image' means exactly what one would think: a ‘still representation' of something. And in the context of a video, that means a ‘frame.'”). Practically speaking, if Phillips paused one of his videos and printed a frame, or saved each frame individually in electronic format, no one could seriously dispute that the component part would constitute an “image.” Id. Adding up the frames in his 91 video undoubtedly exceeded 600 images, justifying the maximum enhancement. See id. at 398.
We concur in Judge Larsen's compelling, principled assessment. The term “image” is not genuinely ambiguous. Whatever uncertainty one might confront when attempting to apply Section 2G2.2(b)(5) to video is easily overcome by consulting any number of contemporary dictionaries, which confirm the operative words' ordinary meanings. See id. at 391-92 (citing Image, Video, and Frame, MERRIAM-WEBSTER'S COLLEGIATE DICTIONARY (11th ed. 2003); OXFORD ENGLISH DICTIONARY (3d ed. 2009); THE AMERICAN HERITAGE DICTIONARY OF THE ENGLISH LANGUAGE (5th ed. 2018); WEBSTER'S NEW WORLD COLLEGE DICTIONARY (5th ed. 2020); SHORTER OXFORD ENGLISH DICTIONARY (6th ed. 2007)). Videos, by definition, comprise a series of still images-i.e., frames-that, when played in sequence, create the illusion of a moving picture. Debus's preference for a 1:1 ratio plainly is a nonstarter because videos are a collection of images, and a single image does not a collection make. We need not consider the Guideline's commentary to reach that fundamental conclusion.
The Phillips majority, on the other hand, erroneously rejected the framebased standard because of the repercussions that option might entail. It feared such a standard would trigger the top-tier enhancement in most child pornography cases, never mind that the Sentencing Commission's own statistics show most offenders-73% in 2020-exceeded the 600-image threshold under the current rule. See U.S. Sent'g Comm'n, Use of Guidelines and Specific Offense Characteristics for Fiscal Year 2020, 43 (Apr. 2021). “Harsh outcomes” alone cannot justify setting aside a statute's natural construction, lest textualism gives way to consequentialism. See Dodd v. United States, 545 U.S. 353, 359 (2005); Lamie v. U.S. Tr., 540 U.S. 526, 538 (2004); see also Phillips, 54 F.4th at 394-95 (Larsen, J.) (collecting cases discussing absurdity doctrine). As Judge Larsen aptly observed, for every hypothetical concern the majority ascribed to the frame-based standard, the 75:1 rule offers its own problematic results. See, e.g., Phillips, 54 F.4th at 397 (Larsen, J.) (noting 75:1 rule “will be harsher ” for defendants who possess short videos and creates loophole for offenders who compile thousands of photographs into single video slideshow).
The 75:1 rule distorts the Guideline's plain meaning because it is a policy determination set by administrative fiat, not an effort at interpretation. The Commission arbitrarily treats all videos the same regardless of length or the duration of child pornography depicted therein. Arbitrariness, of course, is suggestive of a legislative act, as opposed to an interpretive one. See Catholic Health Initiatives v. Sebelius, 617 F.3d 490, 495 (D.C. Cir. 2010) (citing Hoctor v. U.S. Dep't of Agric., 82 F.3d 165, 171 (7th Cir. 1996)). An inflexible ratio cannot reasonably be considered “within the . . . zone of ambiguity” of the nonnumerical terms before us. See Adair, 38 F.4th at 349; cf. Catholic Health Initiatives, 617 F.3d at 495 (quoting HENRY J. FRIENDLY, Watchman, What of the Night?, in BENCHMARKS 144-45 (1967) (“[W]hen an agency wants to state a principle ‘in numerical terms,' . . . the agency is legislating and should act through rulemaking.”)). After all, we most certainly will not find “1/75th of a video” under any known dictionary definition of “image.” See Phillips, 54 F.4th at 389 (Larsen, J.).
Nor can we agree that setting a benchmark between two random, albeit legislatively promulgated, numbers in the image table “squarely implicates the Commission's ‘substantive expertise'” in “promot[ing] sentencing goals.” Cf. Phillips, 54 F.4th at 385 (quoting Kisor, 139 S.Ct. at 2417). Interpreting the meaning of a word (image) in a specific context (visual media) is an exercise that “fall[s] more naturally into a judge's bailiwick.” See Kisor, 139 S.Ct. at 2417. The 75:1 rule may be the Commission's “official position” and reflect its “fair and considered judgment,” see id. at 2416-17, but even sincerest defenses of agency commentary cannot overcome a lack of genuine ambiguity after resorting to traditional interpretive methods. If Congress or the Sentencing Commission wants to direct us to adopt the legal fiction that each video “shall be considered” 75 images, they can do so by formally amending the Guidelines. In the meantime, Section 2G2.2(b)(7)'s “text, structure, history, and purpose,” see id. at 2415, compel us to count each frame of a video as an image.
To be sure, a frame-based approach will require the government to change how it assesses contents of video files in child exploitation investigations, though we believe the added effort will be minimal. Agent Strause explained that reviewing evidence in these types of cases “[t]ypically” means watching a video using special categorization software, flagging the file as illegal as soon as CSAM appears, “and then proceeding]” to the next video. (See 4/19/23 Tr. 4:16-6:23). A frame-based standard will require the government to establish two criteria when it invokes the image enhancement for possessory offenses involving videos moving forward: the frame rate and the length of time child pornography is depicted.
The first task should be easy enough if Agent Strause's experience is any indication. She found frame rates for most video files in this case simply by “right clicking on each file” and selecting the “Details” tab under “Properties.” (See Id. at 11:15-22; Gov't Ex. 1 at 2). The second task, determining the length of time child pornography is depicted, will be a bit more involved than the typical approach Agent Strause described. Fortunately, investigators have tools at their disposal- like the fast-forward button-to speed up the viewing process. And to be clear, we do not expect the government to replicate Agent Strause's painstaking efforts in every future case. The government need only establish, for purposes of Section 2G2.2(b)(7)(D), that a defendant possessed a combination of 600 or more contraband photographs and still-video frames to surpass the threshold for the image table's five-level enhancement. Prosecutors may choose to describe a representative sample of video data or to present a comprehensive report; however, video frame analysis substantially in excess of 600 images is surplusage, so cataloguing even a handful of videos of sufficient length would ordinarily suffice to establish the enhancement. Obviously, the number and length of CSAM videos may factor into the sentencing court's Section 3553(a) analysis. See 18 U.S.C. § 3553(a)(1)-(2), (b)(1); see, e.g., United States v. Ziegler, 761 Fed.Appx. 104, 108-09 (3d Cir. 2019) (nonprecedential) (affirming upward variance based on, inter alia, “sheer volume of the recordings and reproductions” in defendant's possession); cf. United States v. Grober, 624 F.3d 592, 599-600 (3d Cir. 2010) (discussing justifications for varying from Guidelines in child pornography cases).
We understand that frame rates may not be readily available for every video file in a defendant's collection; however, we doubt metadata gaps will be an issue in many cases. In the unlikely event a defendant's collection consists primarily or exclusively of videos without locatable or manually derivable frame rates, we agree with the government's suggestion that it is reasonable to presume each video contains a frame rate of 10 fps. (See 4/19/23 Tr. 38:11-18). That is the minimum rate identified in the standard (non-GIF) video files at issue here, and it is representative of the videos produced on low quality cell phone cameras that appear in the majority of child pornography cases Agent Strause has investigated. (See id. at 12:2-4, 28:5-25). A defendant may rebut this presumption with appropriate evidence.
We have applied the frame-based standard to the undisputed facts of Debus's case, and we conclude the five-level enhancement for possessing more than 600 images of child pornography is appropriate. We find Agent Strause's calculations to be highly credible and persuasive. Agent Strause determined Debus possessed 102 still images and 306,111 frames depicting child pornography after excluding six GIFs, three videos lacking frame-rate data, two CGI files, and one duplicate video from her calculations. (See Gov't Ex. 1 at 3-4; see also 4/19/23 Tr. 18:5-10). Debus's stash of at least 306,213 contraband images easily exceeds the Guideline's 600-image threshold. His objection to the five-level image enhancement necessarily fails.
The commentary says “computer-generated image[s]” should be counted, see U.S.S.G. § 2G2.2, cmt. n.6(B)(i), but doing so arguably raises serious constitutional issues when no actual children are depicted, see Ashcroft v. Free Speech Coalition, 535 U.S. 234, 239-40, 256 (2002) (striking down on First Amendment grounds provision of federal statute prohibiting possession and distribution of “virtual child pornography” or CGI “that appear to depict minors but were produced without using any real children”). Because Debus's collection exceeds the maximum tier of the image table without including CGI files, we likewise exclude those images from our count, and we express no opinion on whether including them is constitutionally permissible.
We also agree with Agent Strause's exclusion of duplicate files. We note, however, that some circuits expressly decline to single out duplicates for exclusion, see, e.g., United States v. McNerny, 636 F.3d 772, 778-80 (6th Cir. 2011) (holding “duplicate digital images, like duplicate hard copy images, should be counted separately for purposes of calculating a sentence enhancement pursuant to § 2G2.2(b)(7)”); United States v. Sampson, 606 F.3d 505, 510 (8th Cir. 2010) (finding no “uniqueness” requirement when counting contraband digital images), though our court of appeals has observed (in dicta) the enhancement should account for “different images” in a defendant's possession, see United States v. Goff, 501 F.3d 250, 255 n.9 (3d Cir. 2007); see also United States v. Lacey, 569 F.3d 319, 322 & n.1 (7 th Cir. 2009) (not addressing propriety of district court's exclusion of duplicate images).
A single loop of the six GIFs, plus the three videos with missing frame rates, independently account for 231.5 seconds of CSAM, which equal 2,315 additional frames using a presumptive frame rate of 10 fps. Though our final tally may be more precise if we included these images, Debus has not had a meaningful opportunity to rebut the presumption's applicability to these files, nor do they affect the outcome of his objection, so we will leave the presumption for another day.
We offer a final observation. Technological advancements have rapidly improved the quality of footage that can be captured on even the cheapest cell phones and video cameras available today, and the ubiquity of these devices has, in a sense, made amateur cinematographers of us all. While 24 fps is still the standard frame rate for most filmmaking, some modern high-speed cameras can capture millions of frames per second. See Tim Moynihan, “A Slo-Mo Camera with an Insanely High Frame Rate (And Price Tag),” WIRED (July 16, 2014). Yet even the slowest video cameras will produce many frames that are indistinguishable to the human eye. We acknowledge that in a close case-one involving a handful of short videos with high frame rates, for example-a straightforward frame tally may have adverse “sentencing consequences” as the result of double counting virtually identical images. See Goff, 501 F.3d at 252 n.1.
This is not a close case. Debus possessed dozens of videos constituting nearly three hours of child pornography. The sheer amount of contraband footage he amassed contains enough frames of CSAM to exceed the 600-image threshold by an order of magnitude. We therefore find by a preponderance of the evidence Debus possessed more than 600 images of child pornography.
For example, we would reach the top of the image table in this case by counting one image for every 23 seconds of film in which child pornography is depicted. We calculated this charitable interval by first determining the maximum number of frames that could be counted as a single “unique” image while still satisfying the 600-image threshold when added to the 102 separate photographs Debus possessed. We divided the total number of frames depicting CSAM (306,111) by the number of additional frames needed to reach the threshold (498) and got an interval of approximately one image every 615 frames. We then found the average frame rate across the 34 videos included in Agent Strause's assessment (26.65 fps). Lastly, we divided our interval (1 image every 615 frames) by 26.65 fps, which equals an image every 23.08 seconds. Such a lengthy interval easily permits the naked eye to discern visual changes in the imagery depicted. See, e.g., Protocomm Corp. v. Fluent, Inc., No. 93-0518, 1994 WL 719673, at *2 (E.D. Pa. Dec. 17, 1994).
IV. Conclusion
For the foregoing reasons, we will sustain Debus's objection to the pattern-of-activity enhancement and overrule his objection to the image tally. An appropriate order shall issue.