Carlos Escobar Gomez v. Merrick Garland ( 2021 )


Menu:
  •                                     UNPUBLISHED
    UNITED STATES COURT OF APPEALS
    FOR THE FOURTH CIRCUIT
    No. 20-1654
    CARLOS ALEXANDER ESCOBAR GOMEZ,
    Petitioner,
    v.
    MERRICK B. GARLAND, Attorney General,
    Respondent.
    On Petition for Review of an Order of the Board of Immigration Appeals.
    Argued: September 22, 2021                                  Decided: December 10, 2021
    Before WILKINSON, WYNN, and FLOYD, Circuit Judges.
    Petition for review granted and remanded for further proceedings by unpublished opinion.
    Judge Floyd wrote the opinion in which Judge Wynn joined. Judge Wynn wrote a separate
    concurring opinion. Judge Wilkinson wrote a dissenting opinion.
    ARGUED: Nathan Randal Bogart, BOGART, SMALL + NAYLOR, Fayetteville,
    Arkansas, for Petitioner. Paul Fiorino, UNITED STATES DEPARTMENT OF JUSTICE,
    Washington, D.C., for Respondent. ON BRIEF: Jeffrey C. Bossert, Acting Assistant
    Attorney General, Carl H. McIntyre, Jr., Assistant Director, Office of Immigration
    Litigation, Civil Division, UNITED STATES DEPARTMENT OF JUSTICE, Washington,
    D.C., for Respondent.
    Unpublished opinions are not binding precedent in this circuit.
    FLOYD, Circuit Judge:
    Carlos Escobar Gomez seeks review of the Board of Immigration Appeals’ (BIA)
    dismissal of his application for asylum. The BIA determined that Escobar Gomez was
    ineligible for asylum because he failed to establish membership in a particular social group
    defined with sufficient particularity. Because this ruling is not supported by a reasoned
    explanation, we grant the petition for review and remand to the BIA for further
    proceedings.
    I.
    Escobar Gomez is a thirty-year-old native and citizen of El Salvador. He departed
    El Salvador on November 23, 2013, and arrived in the United States on January 19, 2014,
    at or near Hidalgo, Texas, entering without inspection. Escobar Gomez was placed into
    removal proceedings pursuant to 8 U.S.C. § 1229a following service of a Notice to Appear
    (NTA) on February 12, 2014. Escobar Gomez conceded his removability, but sought
    asylum, withholding of removal, and relief under the Convention Against Torture (CAT),
    claiming that he would be unsafe in El Salvador because he had witnessed a murder and
    subsequently was threatened by gang members.
    Per Escobar Gomez’s testimony, on October 24, 2013, while returning home from
    a soccer practice in his hometown in El Salvador, Escobar Gomez and two of his friends
    witnessed two members of an unspecified gang shoot and kill a known acquaintance. On
    November 13, 2013, one of the gang members involved in the shooting, Costello,
    approached Escobar Gomez and warned him that if he went to the authorities or told anyone
    2
    else what he had seen, he would meet the same fate as the murder victim. Costello left
    Escobar Gomez with his cell phone number so Escobar Gomez could inform him if the
    police asked any questions, warning him they had further things to discuss. Neither
    Escobar Gomez nor his two friends reported either witnessing the murder or the subsequent
    threat to legal authorities. He believed that the police would be unable to protect him,
    fearing that even if Costello were arrested, other members of the gang would seek
    retribution by following through on Costello’s threat. Ten days after receiving the threat
    against his life, on November 23, 2013, Escobar Gomez fled the country.
    At the merits hearing held June 19, 2018, the Immigration Judge (IJ) denied Escobar
    Gomez’s application for asylum based on his finding that Escobar Gomez’s proposed
    particular social group, “witnesses to a crime,” was not cognizable. * The IJ evaluated the
    proposed particular social group to determine whether it was (i) immutable, (ii) socially
    visible, and (iii) defined with sufficient particularity.    A.R. 50 (citing Matter of
    M-E-V-G-, 
    26 I. & N. Dec. 227
    , 237 (BIA 2014)). The IJ indicated concerns about the
    mutability of the group but did not enumerate them. He found that the particular social
    group lacked social distinction, in large part because Escobar Gomez did not tell anyone
    except his mother that he had witnessed the murder and thus would not be recognized by
    others in the community as a member of that social group. On the issue of particularity, he
    determined that the particular social group was too amorphous and did not lend itself to a
    benchmark for group membership. Having determined that Escobar Gomez did not belong
    *
    Escobar Gomez did not press his CAT claim on appeal and so we focus only on
    his asylum and withholding-of-removal claims.
    3
    to a particular social group, the IJ did not analyze the other requirements for asylum
    eligibility. Additionally, the IJ did not assess Escobar Gomez’s other proposed particular
    social group—“witnesses to a murder in El Salvador.”
    Escobar Gomez filed a timely Notice of Appeal with the BIA on July 13, 2018. On
    May 13, 2020, the BIA affirmed the IJ’s decision and dismissed Escobar Gomez’s appeal.
    The BIA found that Escobar Gomez’s proposed particular social group “witnesses to a
    crime” is not particularly defined and thus does not constitute a protected ground because
    “it includes both people who report crime, and those who . . . do not.” A.R. 5. Regarding
    the second proposed particular social group “witnesses to a murder in El Salvador,” the
    BIA said only that it is not particularly defined “[f]or the same reason,” and thus the IJ’s
    failure to evaluate the cognizability of that group constituted harmless error. A.R. 5.
    Escobar Gomez petitions for review of the BIA’s order denying his application for asylum.
    II.
    “When, as here, the BIA adopts and affirms the IJ’s decision and supplements it
    with its own opinion, we review both decisions.” Cordova v. Holder, 
    759 F.3d 332
    , 337
    (4th Cir. 2014). Both the IJ and the BIA based their denials on an analysis of the issue of
    particularity. That is to say neither addressed the other elements of an asylum claim—past
    persecution or well-founded fear of future persecution, nexus between the particular social
    group and persecution suffered, or the government’s ability or willingness to protect the
    individual. See Perez Vasquez v. Garland, 
    4 F.4th 213
    , 221 (4th Cir. 2021). In order to
    make a successful asylum claim, the petitioner must establish each of these individual
    4
    elements, in addition to meeting the separate standard for establishing a particular social
    group.     Canales-Rivera v. Barr, 
    948 F.3d 649
    , 660 n.1 (4th Cir. 2020) (Agee, J.,
    concurring) (quoting Crespin-Valladares, 
    632 F.3d 117
    , 124–126 (4th Cir. 2011)) (stating
    that although the INA does not define “particular social group,” we have “long deferred to
    the BIA’s interpretation . . . , which provides that to qualify as a particular social group, a
    group’s members (1) must share common, immutable characteristics that (2) give the group
    social visibility and that (3) define the group with sufficient particularity to have well-
    defined boundaries, i.e., constitute a discrete class of persons”) (cleaned up). As the BIA
    in this case only addressed particularity on appeal, we do not, and jurisdictionally cannot,
    evaluate any other elements of the asylum test. Amaya v. Rosen, 
    986 F.3d 424
    , 429 (4th
    Cir. 2021). Our review is limited to the grounds relied upon by the agency. Cordova, 759
    F.3d at 337.
    The Court is obliged to uphold a decision of the BIA unless it is “manifestly contrary
    to the law and an abuse of discretion.” See Tairou v. Whitaker, 
    909 F.3d 702
    , 706 (4th Cir.
    2018). The Court may find that the BIA abused its discretion if it “failed to offer a reasoned
    explanation for its decision, or if it distorted or disregarded important aspects of an
    applicant’s claim.” 
    Id.
     The BIA also “abuses its discretion in making an error of law.” 
    Id.
    Although deference to the BIA’s determinations is substantial, it is not unlimited. Essohou
    v. Gonzales, 
    471 F.3d 518
    , 520 (4th Cir. 2006); Lopez-Soto v. Ashcroft, 
    383 F.3d 228
    , 233
    (4th Cir. 2004).
    Given those grounds for review, we find the agency abused its discretion by failing
    to provide a reasoned explanation and, in the limited explanation it did provide, making an
    5
    error of law. In its analysis of Escobar Gomez’s claims, the agency relied on its finding
    that the proposed particular social group lacked limiting features because “it includes both
    people who report crime, and those who . . . do not.” A.R. 5. It applied the same reasoning
    to both proposed particular social groups, failing to explain why the broader group of
    “witnesses to a crime” should be synonymous with the narrower group “witnesses to a
    murder in El Salvador” as a matter of particularity. Its finding on particularity as it pertains
    to the second proposed social group requires further analysis.
    Additionally, the stated grounds on which the agency relies—the failure to
    distinguish between those who report crime and those who do not—is insufficient to
    support its determination on particularity. “[T]he focus of the particularity requirement is
    whether the group is discrete or is, instead, amorphous.” Matter of W-G-R-, 
    26 I. & N. Dec. 208
    , 214 (BIA 2014); see also Matter of M-E-V-G-, 26 I. & N. Dec. at 239 (stating
    that to be particular a social group must “be discrete and have definable boundaries—it
    must not be amorphous, overbroad, diffuse, or subjective”). But we have stated that
    whether the proposed social group may be divided into smaller subgroups is not dispositive
    of a particularity finding. Amaya, 986 F.3d at 434. A particular social group need not be
    made up of homogenous members, nor does the existence of smaller parts within the whole
    automatically discount the existence of a particular social group. See id. “What matters is
    not whether the group can be subdivided based on some arbitrary characteristic but whether
    the group itself has clear boundaries.” Id. Large and internally diverse social groups are
    cognizable without further qualification, id. at 433 n.6, so long as boundaries drawn around
    the group are clear. Other statutory requirements, including nexus, operate to further limit
    6
    the number of members within the groups who qualify for asylum, but do not bear on an
    assessment as to whether the proposed group provides a clear, objective benchmark for
    who is in the group.
    The agency’s limited justification in this instance, applied cursorily to both
    proposed groups, is neither thorough nor properly reasoned. By relying exclusively on
    such reasoning, the agency abused its discretion. “We may not affirm the [BIA]’s decision
    on any conceivable basis, but rather only if ‘the grounds upon which the agency acted . . .
    were those upon which its action can be sustained.’” Cordova, 759 F.3d at 337 (quoting
    Nken v. Holder, 
    585 F.3d 818
    , 822 (4th Cir. 2009)). And we may not consider new
    arguments raised before this Court. Portillo Flores v. Garland, 
    3 F.4th 615
    , 637 (4th Cir.
    2021). Under this scope of review, we cannot uphold the agency’s decision.
    III.
    The agency’s action in this case cannot be sustained on its proffered grounds and
    we must thus remand Escobar Gomez’s asylum and withholding-of-removal claims for
    further proceedings.
    PETITION FOR REVIEW GRANTED
    AND REMANDED FOR FURTHER PROCEEDINGS
    7
    WYNN, Circuit Judge, concurring:
    I fully agree with the majority opinion that the Board of Immigration Appeals
    legally erred by failing to adequately support its decision, and therefore that remand is
    appropriate. Accordingly, I readily concur. I write separately, however, because I would
    go further and hold that Escobar Gomez’s proposed particular social group is, in fact,
    sufficiently particular. 1 I also write separately to address the dissenting opinion’s erroneous
    application of both principles of immigration law and, specifically, our precedent related
    to particularity.
    I.
    There is no need to reinvent the wheel in this case. In fact, our limited reviewing
    role precludes us from doing so. The question before us is an extraordinarily narrow one:
    is the particular social group “witnesses to a murder in El Salvador” sufficiently particular?
    I first explain why the answer, according to our case law, is yes. I then turn to the
    dissent and explain how it departs from binding precedent.
    A.
    “[P]articularity chiefly addresses the outer limits of a group’s boundaries and is
    definitional in nature.” Matter of M-E-V-G-, 
    26 I. & N. Dec. 227
    , 241 (B.I.A. 2014)
    (emphasis added) (internal quotation marks omitted). Particularity requires a group to have
    Our dissenting colleague is wrong to charge the majority with finding Escobar
    1
    Gomez’s group to be particular. Dissenting Op. at 34. The majority does not do so; I,
    however, would.
    8
    “concrete, identifiable boundaries that allow an observer to distinguish members of a group
    from non-members.” Temu v. Holder, 
    740 F.3d 887
    , 892 (4th Cir. 2014). Thus, “[a]
    particular social group must be defined by characteristics that provide a clear benchmark
    for determining who falls within the group.” Matter of M-E-V-G-, 26 I. & N. Dec. at 239.
    We recently took a close look at the meaning of “particularity” in Amaya v. Rosen,
    
    986 F.3d 424
     (4th Cir. 2021). We emphasized that “the purpose of the particularity
    requirement is ‘to avoid indeterminacy’” and that a particular social group must have
    “‘definable boundaries’ so that it is sufficiently clear who is in and out of the group.” Id.
    at 429 (first quoting Zelaya v. Holder, 
    668 F.3d 159
    , 165 (4th Cir. 2012); then quoting
    Matter of M-E-V-G-, 26 I. & N. Dec. at 239). We further noted that the particularity prong
    of the analysis must not be conflated with the social distinction prong, because “the
    requirements serve distinct purposes, and it is important to consider them separately and
    with integrity to their purposes.” Id. at 433. “[P]articularity is a definitional question—an
    inquiry meant to ensure there is an ‘adequate benchmark’ for setting the boundaries of the
    group. . . . Put another way, is it evident from the group’s description who is in and who is
    not?” Id. at 433–34 (quoting Lizama v. Holder, 
    629 F.3d 440
    , 447 (4th Cir. 2011)). By
    contrast, the social distinction requirement “asks whether the ‘home society actually does
    recognize that group as being a “distinct” and identifiable group.’” Id. at 433 (emphasis
    added) (quoting S.E.R.L. v. Att’y Gen. U.S., 
    894 F.3d 535
    , 553 (3d Cir. 2018)); see also
    Matter of M-E-V-G-, 26 I. & N. Dec. at 241.
    Thus, wealth, Americanization, opposition to gangs, resisting gang recruitment,
    being a returning migrant, and criminal history are not sufficiently particular. See Amaya,
    9
    986 F.3d at 429 (citing Lizama, 
    629 F.3d at 447
    ); Zelaya, 668 F.3d at 166; Moreno-Osorio
    v. Garland, 
    2 F.4th 245
    , 255 (4th Cir. 2021). This is because such groups “change[]
    dramatically based on who defines [them]”; the proposed benchmarks are subjective rather
    than objective. Temu, 740 F.3d at 895. By contrast, groups consisting of the family unit,
    individuals with bipolar disorder who exhibit erratic behavior, and prosecution witnesses
    against gangs are each “self-limiting” and particular. Amaya, 986 F.3d at 429 (citing
    Crespin-Valladares v. Holder, 
    632 F.3d 117
    , 125 (4th Cir. 2011)); see Temu, 740 F.3d at
    896; Zelaya, 668 F.3d at 169 (Floyd, J., concurring). 2
    Further, “the size of the group has no bearing on the clarity of the group’s
    boundaries.” 3 Amaya, 986 F.3d at 433 n.6. So even the broad group “landowners” is
    sufficiently particular, since whether someone is or is not a landowner is objectively
    determinable. Id. at 433–34. Whether a group of such size would be socially distinct is a
    different question; in some societies, like the modern United States, so many people are
    2
    My citation to Zelaya here is to Judge Floyd’s concurrence. However, that
    concurrence was joined by a second judge on the three-judge panel. Accordingly, the
    concurrence is best understood as a second majority opinion with precedential weight.
    3
    We have previously noted that the Board of Immigration Appeals “has explained”
    that “[s]ize is relevant” but is “not dispositive.” Alvarez Lagos v. Barr, 
    927 F.3d 236
    , 253
    (4th Cir. 2019) (citing Matter of S-E-G-, 
    24 I. & N. Dec. 579
    , 584 (B.I.A. 2008)). But the
    statement in Matter of S-E-G- that size is relevant to particularity comes in the context of
    an example of the precise problem identified in Amaya—that the agency has at times
    conflated social distinction with particularity. Matter of S-E-G- notes that “the size of the
    proposed group may be an important factor in determining whether the group can be”
    “described in a manner sufficiently distinct that the group would be recognized, in the
    society in question, as a discrete class of persons.” Matter of S-E-G-, 24 I. & N. Dec. at
    584. That inquiry, as we clarified in Amaya, pertains to social distinction. Amaya, 986 F.3d
    at 433–34. Accordingly, after Amaya, size is not relevant to the particularity prong in this
    circuit. See Majority Op. at 6 (citing Amaya, 986 F.3d at 433 n.6).
    10
    landowners that the group may not be distinct, while in others, land ownership may be
    relatively rare and thus distinctive. Id. But when it comes to particularity, “[o]wning land
    is no less clear a boundary for group membership in the United States than in any other
    country.” Id. at 433.
    Nor is the fact that a group may have difficult applications dispositive of whether it
    is defined with particularity: “[a] difficult application does not turn a clear rule into a vague
    standard.” Id. at 435. We illustrated this point in Amaya using a tennis court as an analogy.
    A tennis court, of course, “has clear lines that indicate when a ball is in or out.” Id. Though
    it can sometimes be “hard to determine whether the ball landed inside or outside the line,”
    the “difficulty comes from an inability to see where the ball landed, not from an inability
    to see the line.” Id. (emphasis added). We concluded that “[p]articularity is no different.”
    Id. “Whether an applicant is a member of the group is a factual question for the fact-finder
    to determine on a case-by-case basis,” a decision that “may sometimes be difficult based
    on the available evidence.” Id. at 435–36 (emphases added). “[W]hat matters for
    particularity,” however, “is that there are clear lines delineating the boundaries of the
    group.” Id. at 436 (emphasis added).
    And, relatedly, the fact that there may be gradations within the group, or different
    ways of being a part of the group, does not eliminate particularity. See Majority Op. at 6–
    7 (“Large and internally diverse social groups are cognizable without further qualification,
    so long as boundaries drawn around the group are clear.” (citation omitted)). In Amaya, we
    used Catholicism as an example to illustrate this point. We noted that Catholics “vary
    widely in the time they have been part of that faith as well as in their level of commitment
    11
    and involvement. Those differences may make it difficult to establish the required nexus
    of persecution or even whether a petitioner is or is not Catholic. But they have nothing to
    do with particularity.” Amaya, 986 F.3d at 435.
    To put the matter simply: as the majority correctly recognizes, Majority Op. at 6–7,
    the only question, in determining particularity, is whether “there are clear lines delineating
    the boundaries of the group,” Amaya, 986 F.3d at 436. The size of the group, the difficulty
    of its application, and any gradations within the group are all irrelevant for particularity—
    even if they ultimately are relevant to other aspects of the asylum or withholding-of-
    removal analysis, such as other elements of the three-part particular-social-group test that
    the Board has implemented to define the statutory term “particular.” When it comes to the
    particularity prong of that three-part test, however, all we care about is whether there is a
    boundary line between the in-group and out-group that is not simply a matter of subjective
    interpretation. Id.
    B.
    Under the standard articulated in Amaya, the proposed particular social group of
    “witnesses to a murder in El Salvador” is particular. There is a clear delineation of who is
    in and who is out. There is no subjective component, like “wealth.” There may be difficult
    applications—does a murderer himself count?—but that does not eliminate the clarity of
    the boundary itself, as we stated, explicitly, in Amaya.
    In Amaya, we evaluated the particularity of the proposed particular social group of
    “former Salvadoran MS-13 members.” Id. at 426. We held that group to be sufficiently
    12
    particular—even assuming Chevron deference applied, which it does not here. 4 Id. at 432.
    And we did so over the Government’s protestations that it was too “difficult to ascertain
    who exactly is a former gang member.” Id. at 429. Rather, this Court found more persuasive
    the petitioner’s argument: that “‘former Salvadoran MS-13 members’ is sufficiently
    particular because it refers to a discrete class of persons and has self-limiting components—
    it excludes all current members of MS-13, all non-Salvadorans, and anyone who has never
    been a member of MS-13.” Id. Thus, the group “provides a clear benchmark for who is in
    the group, therefore satisfying the particularity requirement.” Id.
    So too here. At bottom, the proposed particular social group in this case “provide[s]
    objective goalposts delineating the boundaries to the group”: someone is only in the group
    if they have (1) witnessed (2) a murder (3) in El Salvador. Id. at 435. This group includes
    “self-limiting features baked into the definition”; it does not include those who have never
    witnessed direct violence at all, or who have witnessed violence that did not end in murder,
    or who have witnessed a murder somewhere other than El Salvador. Id. at 434.
    To be sure, there is some ambiguity surrounding the phrase “witnesses to a murder”:
    What exactly must have been seen or heard to qualify? What about a killing in self-defense?
    4
    The Board’s rejection of Escobar Gomez’s claim is not entitled to Chevron
    deference. See Moreno-Osorio, 2 F.4th at 255 n.4. Rather, we may give the decision
    Skidmore deference, but “that modest deference depends upon ‘the thoroughness evident
    in [the Board’s] consideration, the validity of its reasoning, its consistency with earlier and
    later pronouncements, and all those factors which give it power to persuade.’” Martinez v.
    Holder, 
    740 F.3d 902
    , 910 (4th Cir. 2014) (quoting A.T. Massey Coal Co. v. Holland, 
    472 F.3d 148
    , 169 (4th Cir. 2006)). I see no reason to give even Skidmore deference to the
    decision here, given the very limited nature of the Board’s analysis and the fact that even
    that limited analysis cannot be reconciled with our decision in Amaya.
    13
    But it is hard to imagine a particular social group that would not have some ambiguity.
    Even the family unit, which we have found to be particular time and again, suffers from
    some level of vagueness: How distant of a relative counts?
    More importantly, we rejected the argument in Amaya that “it is difficult to ascertain
    who exactly is a former gang member.” Id. at 429. We brushed aside concerns that “former”
    and “member” both include too much variability for the group to be particular. Id. at 434–
    35. And we came to that conclusion even though the difficulty in determining who is a
    former member of a gang is substantial. Id. at 429 (acknowledging that “Amaya does not
    remember a moment when he officially left the gang, and he never told any gang leader
    that he was leaving”).
    For those reasons, I would hold that Escobar Gomez’s proposed particular social
    group satisfies particularity.
    II.
    Our dissenting colleague departs from these and other key principles and comes to
    a conclusion that cannot be reconciled with Amaya. I agree with the majority’s articulation
    of the central guiding principles in this case, and I first explain how the dissent departs
    from them. I then also explain why the dissent’s view of the particularity issue in this case
    is at odds with our precedent.
    A.
    Our friend in dissent purports to start from first principles. But in doing so, he
    fundamentally misconstrues the framework that applies in immigration cases like this one.
    14
    “We generally give Chevron deference to interpretations of statutes the [Board of
    Immigration Appeals] administers, namely the Immigration and Nationality Act . . . , if the
    [Board]’s interpretation carries the force of law.” Portillo Flores v. Garland, 
    3 F.4th 615
    ,
    625 (4th Cir. 2021) (footnote omitted). The Board issued an interpretation of “membership
    in a particular social group” in Matter of M-E-V-G-, 26 I. & N. Dec. at 237. Under the
    Board’s test, “an applicant for asylum or withholding of removal seeking relief based on
    ‘membership in a particular social group’ must establish that the group is (1) composed of
    members who share a common immutable characteristic, (2) defined with particularity, and
    (3) socially distinct within the society in question.” Id. We have always treated the
    overarching three-prong test described in Matter of M-E-V-G- as binding, 5 e.g., Nolasco v.
    Garland, 
    7 F.4th 180
    , 187 (4th Cir. 2021), and it must guide our analysis here, as the
    majority properly recognizes. Majority Op. at 5. Yet, inexplicably, our dissenting friend
    nowhere refers to it, instead only citing obliquely to “the particularity requirement.”
    Dissenting Op. at 29.
    This leads to a fundamental error in his opinion. Our friend purports to directly
    interpret the statutory term “particular social group.” See 
    id.
     at 27–29. But not only was
    this question unbriefed by the parties, our role in this case is not to interpret “particular
    social group” as that term is used in the statute. Rather, it is to apply the definition from
    5
    That said, we have at times rejected some of the more detailed aspects of the
    descriptions of the elements given by Matter of M-E-V-G- and other Board opinions. See,
    e.g., Amaya, 986 F.3d at 433–34 (concluding that the Board’s “landowners” example could
    be relevant for social distinction but not for particularity, even though the Board relied on
    this example in the context of particularity in Matter of M-E-V-G-, 26 I. & N. Dec. at 241,
    and Matter of W-G-R-, 
    26 I. & N. Dec. 208
    , 214–15 (B.I.A. 2014)).
    15
    Matter of M-E-V-G-, as understood in our case law. Put differently, we certainly must
    consider the meaning of “particularity.” But our role is to consider the meaning of that term
    as used as one element of the three-part test provided by the agency—not in the statutory
    phrase “particular social group.” The three-part test in Matter of M-E-V-G- already
    represents the agency’s interpretation of the statutory phrase, and we must defer to that
    understanding.
    B.
    Our dissenting colleague next fundamentally errs by failing to account for the
    jurisdictionally limited nature of our inquiry here. See Majority Op. at 5. For Escobar
    Gomez to ultimately receive asylum or withholding of removal, he will have to prove at
    least seven different aspects of his claim. But only a single aspect—particularity—is before
    us now. We have jurisdiction to consider only that aspect.
    For an applicant to succeed in obtaining asylum or withholding of removal on the
    basis of membership in a particular social group, the applicant must show (1) persecution;
    (2) with a nexus to a group that is (3) immutable, (4) socially distinct, and (5) particular;
    (6) of which the applicant is a member; (7) by an organization that the government is unable
    or unwilling to control. See id. at 4; Hernandez-Avalos v. Lynch, 
    784 F.3d 944
    , 949 (4th
    Cir. 2015); Amaya, 986 F.3d at 435 (noting that “whether [the applicant] met his burden
    that he was actually in the” particular social group is an inquiry distinct from the group’s
    particularity). These are separate elements that may not be conflated. E.g., Amaya, 986
    F.3d at 433 (rejecting the “conflation of the particularity requirement with the social
    distinction requirement”).
    16
    Importantly, we “lack jurisdiction to consider” any grounds that the Board of
    Immigration Appeals did not consider, because we “may only reach issues decided by the”
    Board. Id. at 429; see also Majority Op. at 5; Arevalo Quintero v. Garland, 
    998 F.3d 612
    ,
    621–22 (4th Cir. 2021). So, even if we think an applicant might fail on one of the other
    numerous aspects of the test, we cannot consider it. Amaya, 986 F.3d at 438 (“There may
    be other legal problems with the proposed [particular social group] . . . . But the only issue
    before us today is whether that [particular social group] lacks particularity.”). Where, as
    here, the Board “only considered the particularity requirement, the only issue before
    us . . . is the narrow question of whether” the proposed particular social group “is
    sufficiently particular.” Id. at 429 (emphasis added).
    Our dissenting colleague fails to recognize this jurisdictional limitation. And that
    failure leads him to frame his opinion using language pointing to other aspects of the
    analysis. Thus, our colleague “impermissibly conflates” the particularity requirement with
    other elements. Id. at 432.
    For example, the dissent raises what he terms “[a] host of common scenarios” that
    he claims demonstrate the difficulty of showing particularity, but these questions conflate
    particularity with persecution, nexus, and membership in the group. Dissenting Op. at 32.
    He asks whether it is enough to have seen a murder on a crowded street, on a live television
    broadcast, or in a variety of other ways. Id. Perhaps, in light of these and similar questions,
    it will be difficult for a specific applicant to show whether or not he is a member of the
    group (element #6). And in many of the dissent’s hypothetical scenarios, the applicant is
    unlikely to be able to prove persecution (element #1) or nexus (element #2). If the applicant
    17
    was “entirely anonymous to the perpetrators,” id., he will surely have a hard time
    demonstrating that any persecution he suffered was on account of his witnessing the
    murder. But again, these other elements are not at issue here. Our only question is
    particularity. 6
    C.
    Turning to the issue of particularity itself, our dissenting colleague raises a number
    of attacks against the particularity of Escobar Gomez’s proposed social group. Because our
    review is limited to the grounds relied upon by the agency, we cannot consider new
    arguments raised by judges of this Court. Portillo Flores, 3 F.4th at 637 (“[O]ur standard
    of review . . . does not authorize us to excuse misapplication of the law or to create
    a post hoc justification for an unexplained conclusion.”). In any event, the dissent’s
    analysis cannot undermine the fundamental fact that Amaya governs this case. 7
    6
    This error is not just a jurisdictional one. Collapsing different aspects of the test
    into particularity renders other elements surplusage, something we held in Amaya is
    inappropriate. Amaya, 986 F.3d at 433. Further, we emphasized in Amaya that
    “[p]articularity is not an evidentiary question—it is a legal question to ensure group
    determinacy.” Id. at 436. By contrast, nexus, for example, is an evidentiary question.
    Cedillos-Cedillos v. Barr, 
    962 F.3d 817
    , 824 (4th Cir. 2020).
    7
    We also may not consider new arguments raised by the Government on appeal.
    See Cordova v. Holder, 
    759 F.3d 332
    , 337 (4th Cir. 2014) (noting that we may only affirm
    on the “grounds upon which the agency acted” (quoting Nken v. Holder, 
    585 F.3d 818
    , 822
    (4th Cir. 2009))). Yet, on appeal, the Government raises arguments not relied on by the
    agency below: that Escobar Gomez’s proposed group is “very vague, given that it does not
    distinguish between Salvadoran citizen-witnesses in El Salvador, or foreigners who
    witnessed a murder while visiting El Salvador,” “does not distinguish between witness[es]
    of solved murders and witnesses of murders under investigation,” “includes witnesses to
    murders committed by criminals who have been dead for decades,” and “includes
    murderers themselves.” Response Br. at 13; see also Oral Arg. at 22:15–40,
    http://www.ca4.uscourts.gov/OAarchive/mp3/20-1654-20210922.mp3              (Government
    (Continued)
    18
    1.
    As noted, we held in Amaya that the group of “former Salvadoran MS-13 members”
    was sufficiently particular despite the Government’s contention that “it is difficult to
    ascertain who exactly is a former gang member.” Amaya, 986 F.3d at 429. We concluded
    that “[a] group can be clearly defined and still have difficult applications, no differently
    than a clear rule can sometimes be difficult to apply.” Id. at 435.
    For that reason, the dissent’s litany of judge-raised questions about who qualifies
    for Escobar Gomez’s proposed particular social group is unpersuasive. He asks, “Is it
    enough to have seen a murder on a crowded street, along with dozens of other bystanders?
    To have overheard a killing but not to have seen it? To have watched one from a nearby
    apartment building, even if entirely anonymous to the perpetrators? To have witnessed one
    while visiting the country? To have seen one on a live television broadcast?” Dissenting
    Op. at 32. When considering particularity, the answer to all but the last question is very
    raising for the first time at oral argument a hypothetical involving a murder occurring on a
    soccer field in front of thousands of fans). Those arguments were not relied on by the
    agency and are not before us. And even if they were, none of those aspects of the proposed
    group make its boundaries vague. They may make the group larger than it could be;
    Escobar Gomez could have proposed, say, the group of “Salvadorans who witnessed, but
    did not participate in, solved murders in El Salvador committed by still-living perpetrators
    in front of a group of ten or fewer identifiable witnesses.” But again, the fact that a group
    can be subdivided does not change the firmness of its boundaries. Amaya, 986 F.3d at 434.
    One could properly characterize a Granny Smith as a piece of fruit or, more narrowly, as
    an apple. That does not make “fruit” a vague category.
    19
    clearly “yes.” 8 In all of those situations, the individual has witnessed a murder in El
    Salvador.
    2.
    Nevertheless, the dissent seeks to distinguish Amaya by asserting, in a conclusory
    fashion, that the proposed social group in this case is “far more open-ended than the group
    of ‘former Salvadoran MS-13 members’ recognized as particular in Amaya.” Id. In my
    view, that claim blinks reality. Gang membership, as well as the even more tenuous notion
    of having left a gang, is inherently socially defined. Witnessing a murder is not; it is based
    on specific experiences. This makes witnessing a murder a more objective, and thus more
    particular, dividing line than former gang membership is.
    Yet we held in Amaya that it was “sufficiently clear who is a former member of a
    group—it is someone who (1) joined the group and (2) is no longer in the group,” making
    the proposed group in that case “stand in contrast to other terms this Court has held to be
    amorphous, like ‘wealth’ and ‘opposition to gangs,’ because the terms here provide
    objective goalposts delineating the boundaries to the group.” Amaya, 986 F.3d at 435
    (quoting Lizama, 
    629 F.3d at 447
    ). The same is true here. A witness to a murder is someone
    who (1) witnessed (2) a murder. Again, those terms are no more ambiguous than the notion
    8
    As for the last question, about viewing a murder on a live television broadcast, this
    again falls under the category of a difficult application, which does not eliminate or blur
    the boundary line. Further, it is hard to imagine a scenario where the television viewer
    would not be anonymous, meaning that his claim would likely fail regardless on grounds
    other than particularity.
    20
    of joining or leaving a criminal organization. In fact, they are less so; they always depend
    on objective facts about what a person has seen or heard.
    By contrast, decisions to join or leave a group can be made in a variety of ways that
    may or may not involve objective criteria that outside observers can point to. The actual
    record evidence in Amaya is instructive—“Amaya d[id] not remember a moment when he
    officially left the gang, and he never told any gang leader that he was leaving.” Id. at 429.
    Escobar Gomez, of course, can point to the precise moment when he became a witness to
    a murder—it is a matter of fact, not social definition. If the group in Amaya satisfied
    particularity, so does the group asserted here.
    3.
    The dissent departs from Amaya in other key ways, as well. Recall that, under
    Amaya, the size of the group and the fact that there may be gradations within the group are
    irrelevant to the question of particularity, as the majority properly recognizes. Majority Op.
    at 6–7. Yet the dissent rests its rejection of the particularity of Escobar Gomez’s proposed
    social group in part on both of these grounds. See Dissenting Op. at 30 (claiming that
    groups whose “defining characteristics” are “highly variable” cannot be particular); id. at
    31 (suggesting that size is relevant to the particularity determination); id. at 33 (disputing
    the particularity of the group of “witnesses to a murder in El Salvador” by calling it “a
    large class”).
    4.
    Instead of relying on the on-point precedent of Amaya, the dissent turns to our
    decisions in Moreno-Osorio and Zelaya. Id. at 34. However, both are distinguishable.
    21
    Moreno-Osorio v. Garland rejected the proposed particular social group of
    “returning migrants from the United States” as insufficiently particular. Moreno-Osorio, 2
    F.4th at 255. It did so on the basis that “returning” and “migrants” were vague terms, such
    that a “returning migrant” “could include a person returning to Honduras after a vacation
    in the United States.” Id. This post-Amaya analysis arguably misses a logical step required
    by Amaya: under that precedent, it does not matter if there are a variety of meanings within
    the group as long as the boundaries of the group are objectively delineated, rather than
    fuzzy and subjective. Amaya, 986 F.3d at 434–35. But, at bottom, Moreno-Osorio rests on
    the undisputed point that subjective terms create insufficiently clear boundaries.
    “Returning migrant” is a subjective term in the same way that, for example, “wealthy
    person” or “opponent of a gang” is: whether someone is a “migrant” depends on the
    observer’s analysis of a variety of factors, such as the person’s intent in leaving and how
    long they’ve been away. By contrast, again, whether someone has witnessed a murder
    depends on objective facts.
    The dissent’s reliance on Zelaya v. Holder fares no better. There, we rejected as
    insufficiently particular the proposed particular social group of “young Honduran males
    who refuse to join MS-13, have notified the authorities of MS-13’s harassment tactics, and
    have an identifiable tormentor within MS-13.” Zelaya, 668 F.3d at 165. The Court held
    that “[r]esisting gang recruitment” was too amorphous to satisfy particularity. Id. at 166.
    But the Court went on to state that “the fact that Zelaya’s conduct in resisting recruitment
    included complaining twice to the police adds little to the particularity equation in the face
    of the common sense proposition that MS-13 would look unfavorably upon anyone who
    22
    complained about its harassment tactics to the police,” and that “the concept that a person
    who is victimized by one gang member more than by other gang members somehow serves
    to particularize all such persons into a targeted social group is just nonsensical.” Id. at 166–
    67. In other words, the applicant’s attempt at narrowing the non-particular group of those
    who refuse to join MS-13 with more specific elements was unsuccessful.
    Yet the fact that the gang would dislike anyone who reported harassment to the
    police does not make the line between the in-group and out-group fuzzy, which Amaya
    held is the only question in the particularity analysis. And the Court made no argument that
    having “an identifiable tormentor” created an undefined boundary, either. So, after Amaya,
    we must use more precise language in clarifying how the boundary lines of the group are
    or are not clear. Regardless, it made no difference in Zelaya because “resisting gang
    recruitment” was already insufficiently particular, and the added elements did nothing to
    change the fuzziness of that boundary. Additionally, Zelaya relied in part on the Board’s
    decision in Matter of S-E-G-, 
    24 I. & N. Dec. 579
     (B.I.A. 2008). Zelaya, 668 F.3d at 165–
    67. As discussed above, see supra note 3, Matter of S-E-G- presents a prime example of
    the error explained in Amaya. So Amaya’s thorough analysis of the particularity
    requirement is the most on-point precedent here. And, for the reasons explained above,
    Amaya points to the inevitable conclusion that Escobar Gomez’s proposed particular social
    group is sufficiently particular.
    23
    III.
    The only question before us in this appeal is whether the group “witnesses to a
    murder in El Salvador” is defined with sufficient particularity such that an observer can
    determine who is in the group and who is not. Whatever intuitions we may hold as to
    whether that group will ultimately satisfy the other prongs of the particular-social-group
    analysis, or whether the applicant will ultimately be able to prove the other elements of his
    asylum and withholding-of-removal claims, can have no bearing on our analysis. As we
    put it in Amaya, “there are many ways one can become a [witness to a murder], and those
    differences may be fatal to an argument that the alleged persecution was on account of
    membership in the [particular social group of witnesses to a murder in El Salvador]. But
    they are irrelevant to the particularity inquiry.” Amaya, 986 F.3d at 435.
    The majority properly limits its analysis to the sole issue before us, and properly
    remands for further consideration by the Board. Accordingly, I concur. I would, however,
    go further: as in Amaya, I would hold that Escobar Gomez has established particularity and
    remand for the Board to consider the Immigration Judge’s other holdings with respect to
    his asylum and withholding-of-removal claims. Id. at 438.
    24
    WILKINSON, Circuit Judge, dissenting:
    I would deny the petition for review. The immigration judge and BIA rejected the
    petitioner’s application for asylum on the grounds that his proposed social group,
    “witnesses to a murder in El Salvador,” failed to meet the particularity requirement of the
    asylum statute. The Board was right to do so and I would affirm its decision on precisely
    that lack-of-particularity ground.
    In fact, the proposed group may well be the least particular social group that this
    court has yet seen. We have rejected many groups as insufficiently particular that were
    nonetheless far more particular than what is before us here. The majority not only
    disregards that precedent but would deny the word “particular” in the asylum statute any
    independent force.
    I.
    Because the majority is carrying the asylum statute beyond the understanding of
    Congress and the Executive Branch, I am compelled to spend some time underscoring the
    extent of its error.
    It is well to quote at the outset the text of the asylum statute. The Attorney General
    may, at his discretion, grant asylum to:
    any person who is outside any country of such person’s nationality . . . and
    who is unable or unwilling to return to, and is unable or unwilling to avail
    himself or herself of the protection of, that country because of persecution or
    a well-founded fear of persecution on account of race, religion, nationality,
    membership in a particular social group, or political opinion.
    25
    
    8 U.S.C. § 1101
    (a)(42); see also 
    id.
     § 1158(b)(1)(A). “The burden of proof is on the
    applicant to establish that the applicant is a refugee” within the meaning of that section. Id.
    § 1158(b)(1)(B).
    Of all the listed grounds in the statute, courts have had the greatest difficulty with
    claims of persecution on account of “membership in a particular social group.” Unlike
    “race, religion, nationality,” or “political opinion,” the term is obviously susceptible of
    many different interpretations. The most expansive approach could encompass “[v]irtually
    any set including more than one person.” Fatin v. INS, 
    12 F.3d 1233
    , 1238 (3d Cir. 1993).
    As asylum claims mounted, however, it became clear that more definite rules were needed
    lest “the social group concept . . . virtually swallow the entire refugee definition.” Matter
    of M-E-V-G-, 
    26 I. & N. Dec. 227
    , 231 (BIA 2014) (quoting In re R-A-, 
    22 I. & N. Dec. 906
    , 919 (BIA 1999, A.G. 2001), remanded for reconsideration in Matter of R-A, 
    24 I. & N. Dec. 629
     (A.G. 2008)). The particularity requirement thus developed to “put ‘outer
    limits’ on the definition of a ‘particular social group.’” 
    Id. at 238
     (quoting Castellano-
    Chacon v. INS, 
    341 F.3d 533
    , 549 (6th Cir. 2003)). This requirement mandates that a
    proposed group be “discrete and have definable boundaries” rather than be “amorphous,
    overbroad, diffuse, or subjective.” Id. at 239. The requirement thus “clarifies the point, at
    least implicit in earlier case law, that not every ‘immutable characteristic’ is sufficiently
    precise to define a particular social group.” Id.
    The Board is owed Chevron deference on its interpretation of ambiguous terms in
    the immigration laws, and we have applied that deference to the Board’s interpretation of
    the term “particular social group.” See Chevron, U.S.A., Inc. v. Nat. Res. Def. Council, Inc.,
    26
    
    467 U.S. 837
    , 842–43 (1984); Lizama v. Holder, 
    629 F.3d 440
    , 446–47 (4th Cir. 2011). 1
    We, along with a majority of the circuits, have concluded that the particularity requirement
    reasonably follows from the statutory term. See Lizama, 
    629 F.3d at 447
    ; Orellana-Monson
    v. Holder, 
    685 F.3d 511
    , 520 (5th Cir. 2012) (collecting cases). After all, “the particularity
    requirement flows quite naturally from the language of the statute, which, of course,
    specifically refers to membership in a ‘particular social group.’” Rivera-Barrientos v.
    Holder, 
    666 F.3d 641
    , 649 (10th Cir. 2012) (quoting 
    8 U.S.C. § 1101
    (a)(42)(A)) (emphasis
    in Rivera-Barrientos). The question before us, then, is whether the proposed group of
    “witnesses to a murder in El Salvador” has been “defined with particularity.”
    I begin the analysis with the plain meaning of the word “particular.” It is quite plain
    that particularity stands in contrast to generality or universality. See Particularity, Oxford
    English Dictionary (online ed. 2021) (“The fact or quality of being particular as opposed
    to being general or universal; the fact of being or relating to one or some, but not all, of a
    class.”). And a “particular” thing means a “distinct, individual, specific” element identified
    from a more universal category. See Particular, Oxford English Dictionary (online ed.
    2021) (“[A] unit or one among a number; taken or considered as an individual, apart from
    the rest; single; distinct, individual, specific.”). Particularity thus denotes a preference for
    1
    By contrast, the Board’s determination in this case, which was rendered by an
    individual member and not by a panel, is not entitled to Chevron deference. See A.R. 5;
    Amaya v. Rosen, 
    986 F.3d 424
    , 430 (4th Cir. 2021). However, for the reasons noted, I
    would find that Escobar’s proposed social group lacks particularity under any standard.
    27
    the sharper not the duller, for the defining not the diffuse. This preference need not be
    absolute, but neither can it be denied.
    The textual context of the phrase “particular social group” is informative, as “a word
    is known by the company it keeps.” Yates v. United States, 
    574 U.S. 528
    , 543 (2015).
    Congress specifically articulated four other grounds of persecution that merit asylum:
    “race,” “religion,” “nationality,” and “political opinion.” None of these are boundless
    concepts; each is sufficiently definite that it often can be easily determined whether any
    given individual belongs to the category or not. This is not to say that edge cases will never
    exist, just that they will be less common. And these words, “immediately surrounding” the
    term “particular social group,” “cabin the contextual meaning of that term.” 
    Id.
     It would be
    awkward to read the phrase “particular social group,” juxtaposed textually with these more
    concrete categories, as admitting entirely indefinite groups. That is not how we generally
    read statutes. See United States v. Williams, 
    553 U.S. 285
    , 294 (2008) (“[A] word is given
    more precise content by the neighboring words with which it is associated.”).
    It is noteworthy that Congress chose the phrase “particular social group” rather than
    just “social group” or “other social group.” The natural inference is that Congress intended
    to cabin the reach of the term and to ensure that proposed groups would not be entirely
    amorphous. After all, concrete terms are particularly important in immigration statutes
    given the “need for uniformity in immigration law,” Cazarez-Gutierrez v. Ashcroft, 
    382 F.3d 905
    , 910 (9th Cir. 2004), and the risk that the courts of appeals will otherwise come
    to inconsistent and hence unfair results. And this explanation corresponds well with the
    Board’s intent in interpreting the particularity requirement, which was expressly designed
    28
    to “put ‘outer limits’ on the definition of a ‘particular social group.’” M-E-V-G-, 26 I. &
    N. Dec. at 238 (quoting Castellano-Chacon, 
    341 F.3d at 549
    ).
    The caselaw interpreting the particularity requirement bears out the above
    principles. We have repeatedly characterized particularity as requiring both that a proposed
    social group “provide an adequate benchmark for determining group membership” and that
    it “embody concrete traits that would readily identify a person as possessing those
    characteristics.” Amaya v. Rosen, 
    986 F.3d 424
    , 429 (4th Cir. 2021) (quoting Lizama, 
    629 F.3d at 447
    ). Thus we have rejected groups that were “too broad and amorphous to have
    well-defined boundaries,” as was the case for the category of “young, Americanized, well-
    off Salvadoran male deportees with criminal histories who oppose gangs.” Lizama, 
    629 F.3d at
    446–47. Along similar lines, we have declined to recognize such amorphous groups
    as young Honduran males who have resisted gang recruitment, see Zelaya v. Holder, 
    668 F.3d 159
    , 165 (4th Cir. 2012), or “returning migrants from the United States” to Honduras,
    see Moreno-Osorio v. Garland, 
    2 F.4th 245
    , 255 (4th Cir. 2021).
    By contrast, we have held that the category of family members of prosecutorial
    witnesses against gangs in El Salvador could qualify as a particular social group, since a
    family unit possesses “particular and well-defined boundaries.” Crespin-Valladares v.
    Holder, 
    632 F.3d 117
    , 125 (4th Cir. 2011) (quoting Matter of S-E-G-, 
    24 I. & N. Dec. 579
    ,
    582 (BIA 2008)). And similarly, we have held that individuals with bipolar disorder who
    exhibit erratic behavior could qualify, given the well-defined and identifiable
    29
    characteristics of bipolar disorder. See Temu v. Holder, 
    740 F.3d 887
    , 891, 895 (4th Cir.
    2014). In short, where the boundaries of a group have been clear, and where a group has
    possessed common and concrete characteristics, we have not hesitated to hold that it
    satisfies particularity. But where the boundaries of a group are impossible to determine, or
    where the defining characteristics of the group are amorphous or highly variable, we have
    rejected the proposed group.
    The other courts of appeals have come to similar results, frequently endorsing the
    Board’s conclusion that particularity requires social groups to have “well-defined
    boundaries.” E.g., De Pena-Paniagua v. Barr, 
    957 F.3d 88
    , 95 (1st Cir. 2020); Rivera-
    Barrientos, 666 F.3d at 648–49; Ucelo-Gomez v. Mukasey, 
    509 F.3d 70
    , 73 (2d Cir. 2007).
    And our sister circuits have also emphasized that the group itself should be precisely and
    sensibly defined—just as we require that a group “embody concrete traits” that “readily
    identify a person as possessing those characteristics.” Lizama, 
    629 F.3d at 447
    . For
    instance, the Fifth Circuit has rejected groups that are so “exceedingly broad” as to
    “encompass[] a diverse cross section of society.” Orellana-Monson, 685 F.3d at 521. The
    First Circuit has noted that “when claims based on a particular social group are proffered,
    ‘they usually are based on discrete classes such as gender.’” De Pena-Paniagua, 957 F.3d
    at 97 (quoting Silva v. Ashcroft, 
    394 F.3d 1
    , 5 (1st Cir. 2005)). And while the size of a
    particular group is never dispositive on its own, see Amaya, 986 F.3d at 433 n.6, the Second
    Circuit has noted that a group is less likely to have been defined with particularity if the
    size of the proposed group would vary dramatically depending on how the group is defined,
    30
    see Ucelo-Gomez, 
    509 F.3d at
    73 n.2 (noting that the category of “affluent Guatemalans”
    could be anything from one to twenty percent of the country’s population).
    The majority fails to do justice to the BIA’s decision. The Board did not reject the
    proposed social group because it was subdivisible. It rejected the group because it was too
    vague, too amorphous, too devoid of “sufficient limiting features,” and because it too
    obviously failed to describe “a discrete class of persons,” as required under our precedent.
    A.R. 4–5 (citing Zelaya, 668 F.3d at 162, 166). The more infinitely open and indeterminate
    a proposed social group is, the less particular it is, as both we and the Board have repeatedly
    recognized. It is not right for this court to chastise an agency for what can only be described
    as an admirable fidelity to circuit precedent. 2
    2
    The Board’s application of the particularity requirement reads as follows:
    The United States Court of Appeals for the Fourth Circuit, in whose jurisdiction this
    case lies, has held that the proposed particular social group of people who report
    gang harassment is not particularly defined because it does not contain sufficient
    limiting features, and thus does not constitute “a discrete class of persons.” See
    Zelaya v. Holder, 
    668 F.3d 159
    , 162, 166 (4th Cir. 2012). The respondent’s group
    of “witnesses to a crime” includes even fewer limiting features than people who
    report gang harassment: it includes both people who report crime, and those who,
    like the respondent, do not (IJ at 8). Because this proposed particular social group
    is not particularly defined, it does not constitute a protected ground under the Act.
    For the same reason, the respondent’s proposed particular social group of “witnesses
    to a murder in El Salvador,” is not, as a matter of law, particularly defined.
    A.R. 4–5.
    31
    The fundamental problem with this group is that it simply is not clear who belongs
    in the group and who does not. While any proposed group will have close cases, here there
    is practically no center. It is not clear whether the group counts only those witnesses who
    testify, or only those who cooperate with police, or if it extends much further. And the
    result is that a vast swath of Salvadorans—and some non-Salvadorans as well—are neither
    clearly inside the group nor clearly outside. A host of common scenarios defies easy
    categorization. Is it enough to have seen a murder on a crowded street, along with dozens
    of other bystanders? To have overheard a killing but not to have seen it? To have watched
    one from a nearby apartment building, even if entirely anonymous to the perpetrators? To
    have witnessed one while visiting the country? To have seen one on a live television
    broadcast? Simply put, the group fails to satisfy the basic criterion of particularity: that it
    be “evident from the group’s description who is in and who is not.” Amaya, 986 F.3d at
    434.
    It is far more open-ended than the group of “former Salvadoran MS-13 members”
    recognized as particular in Amaya. See id. Take the facts of this very case. Escobar testified
    that he witnessed a murder while bicycling with his friends in his hometown. Assume that
    Escobar qualifies as a witness by looking over and viewing the event. But suppose that one
    of Escobar’s friends had kept bicycling forward without turning his head and had
    experienced the murder as earwitness rather than as eyewitness—would he qualify? If
    residents along the street saw or heard the murder from their houses, would they be
    witnesses, even if the gang members never knew that they existed? Most strikingly, are the
    killers themselves witnesses to the murder? There is no easy answer to these questions.
    32
    Indeed, the tragic frequency of homicides in El Salvador ensures that a large class
    of citizens, as well as some visitors, could conceivably qualify as a witness to a murder in
    the country, regardless of whether they reported the crime, whether they feared retaliation,
    or even whether the perpetrators know of their existence. In 2016, for instance, the national
    police chief stated that there were three to four gun battles a day between security forces
    and gang members. See Elise Ditta, ‘3 to 4’ Gun Battles a Day: El Salvador Police Chief,
    InSight Crime (Feb. 15, 2016), https://insightcrime.org/news/brief/three-to-four-gun-
    battles-a-day-el-salvador-police. Murders have taken place in busy marketplaces in broad
    daylight and aboard crowded buses. See, e.g., Alex Papadovassilakis & Steven Dudley,
    The ‘Protection Racket,’ Gangs and Violence in San Salvador, InSight Crime (Oct. 1,
    2020),     https://insightcrime.org/investigations/racket-gangs-violence-in-salvador;   Alex
    Renderos, 16 killed in El Salvador bus attacks, L.A. Times (June 21, 2010),
    https://www.latimes.com/archives/la-xpm-2010-jun-21-la-fg-salvador-bus-20100622-
    story.html. Each of these events is a calamity in itself. But it cannot be that a group that
    may or may not encompass so many different individuals, depending entirely on how
    broadly it is read, can be said to possess “discrete” and “definable boundaries.” Moreno-
    Osorio, 2 F.4th at 254 (quoting Amaya, 986 F.3d at 427).
    In fact, Escobar’s group is far more amorphous and overbroad even than other
    groups whose particularity we have rejected in the past. For instance, in Zelaya we found
    insufficiently particular the category of “young Honduran males who refuse to join [the
    gang] MS-13, have notified the authorities of MS-13’s harassment tactics, and have an
    identifiable tormentor within MS-13.” 668 F.3d at 165. We determined that resisting gang
    33
    membership was too amorphous to provide “an adequate benchmark for determining group
    membership” and that the additional criteria failed to make the group more particular. Id.
    at 166. And most recently, in Moreno-Osorio, we rejected the category of “returning
    migrants from the United States” into Honduras. 2 F.4th at 255. We noted that this was
    “precisely the type of amorphous and overbroad group that cannot satisfy the particularity
    requirement,” since the words “returning” and “migrant” could encompass a variety of
    different individuals depending on how it was defined. Id. Escobar’s proposed social group
    is even more amorphous and overbroad than these.
    I repeat that the simple question here is whether this proposed “particular” social
    group has at least a modestly clear edge. That is to say, an observer should be reasonably
    apprised of who is within the group and who is not. Once again to make a simple point, the
    boundaries of this group are so vague and indeterminate as to go beyond Amaya and indeed
    so far beyond it as to all but remove the term “particular” from the statute. Under the
    majority’s approach, I have trouble envisioning any proposed group that would fail to
    qualify as particular. I have underscored this point above and need not reiterate it here other
    than to note that a more diffuse and various proposed grouping is difficult to imagine. And
    while the statutory criteria may not be hermetically sealed, each retains its distinctive
    meaning, which my good colleagues would deny.
    In sum, there is no cohesion to this asserted group. It could include almost anyone
    or everyone. I would deny the petition.
    34