concurring, joined by RIPPLE, Circuit Judge:
The majority faithfully applies this court’s precedent to hold that petitioner Rocío Henriquez-Rivas’s proposed social group — comprised of people who testified against gang members in El Salvador — is not a “particular social group” under INA § 208(b)(1). I write separately not because I quarrel with the majority’s application of our precedent, but because I question whether those prior holdings best track the language and intent of the INA, or follow the BIA’s published interpretations of § 208(b)(1) — which, given the patent ambiguity in the term “particular social group,” are entitled to Chevron deference. Were I writing on a clean slate — an honor reserved only for an en banc panel — and to accord proper deference to the BIA’s interpretations, I would hold that Henri-quez-Rivas’s proposed group (persons who have risked their lives by testifying in open court against murderous gangs) has sufficient “social visibility” and “particularity” to qualify as a particular social group for purposes of asylum.
At the age of twelve, while living in her native El Salvador, Henriquez-Rivas witnessed her father’s murder at the hands of the Mara Salvatrucha (“MS”) gang. Hen-riquez-Rivas testified in Salvadoran court against three MS members — two of whom were convicted of her father’s murder. After the trial, suspected MS members visited Henriquez-Rivas’s home and school looking for her. Fearing for her safety, Henriquez-Rivas fled El Salvador. She was detained at the U.S.-Mexico border and charged as removable. She filed an application for asylum, withholding of removal, and CAT relief, contending that she had a well-founded fear of future persecution based on her membership in a particular social group comprised of individuals who had testified against gang members in Salvadoran court.
The Immigration Judge granted Henri-quez-Rivas asylum, holding that her proposed social group was cognizable, and that she had a well-founded fear of persecution should she be removed to El Salva*629dor.1 However, the government appealed, and the BIA reversed. The BIA held that Henriquez-Rivas’s proposed social group was not cognizable under § 208(b)(1) because it lacked “social visibility” and “particularity” under the test set forth by the BIA in Matter of C-A-, 23 I. & N. Dec. 951 (BIA 2006).
The majority here denies Henriquez-Rivas’s petition for review of the BIA’s ruling, relying on our holdings that groups of government informants who provide information or testify against criminal elements do not constitute particular social groups. See, e.g., Soriano v. Holder, 569 F.3d 1162, 1165 (9th Cir.2009) (holding that a class of government informants against Filipino criminal gangs is not a cognizable social group); Velasco-Cervantes v. Holder, 593 F.3d 975, 978 (9th Cir.2010) (finding that material witnesses for the U.S. government’s prosecution of alleged alien smugglers do not constitute a particular social group). I agree that our panel is bound by this analogous precedent, but several factors cause me to question whether this line of cases is indeed faithful to the INA and to the BIA’s interpretations of the statute.
First, it must be acknowledged that the term “particular social group” is ambiguous. The INA does not define the term, and the BIA has defined it only at a high level of generality. In Matter of Acosta, 19 I. & N. Dec. 211 (BIA 1985), the BIA reviewed the IJ’s denial of asylum for a Salvadoran cab driver who claimed that he faced future persecution on account of his membership in a particular social group consisting of Salvadoran cab drivers who were threatened by anti-government guerrillas for refusing to partake in work stoppages. Id. at 232. In affirming the IJ’s ruling that Acosta’s proposed group was not cognizable, the BIA explained:
[W]e interpret the phrase “persecution on account of membership in a particular social group” to mean persecution that is directed toward an individual who is a member of a group of persons all of whom share a common, immutable characteristic. The shared characteristic might be an innate one such as sex, color, or kinship ties, or in some circumstances it might be a shared past experience such as former military leadership or land ownership. The particular kind of group characteristic that will qualify under this construction remains to be determined on a case-by-case basis. However, whatever the common characteristic that defines the group, it must be one that the members of the group either cannot change, or should not be required to change because it is fundamental to their individual identities or consciences.
Id. at 233-34 (emphases added). From this language, our court has identified two types of qualifying social groups:
1) a voluntary association which imparts some common characteristic that is fundamental to the members’ identities, or
2) an innate characteristic which is so fundamental to the identities or consciences of its members they either cannot or should not be required to change it.
Soriano, 569 F.3d at 1165. To determine whether a proposed social group falls within one of these categories, this court looks to the two primary factors identified by *630the BIA: (1) whether a group’s shared characteristic gives members “social visibility,” and (2) whether the group can be defined with sufficient “particularity” to delimit its membership. See Matter of C-A-; see also Soriano, 569 F.3d at 1165.
But, instead of clarifying the “particular social group” analysis, identification of these two factors has only compounded the confusion. Neither this court’s opinions, nor those of the BIA, have clearly defined the social visibility and particularity factors, nor provided reasoned applications of those factors to the facts of each case. Instead, we have practically ignored the first prong, social visibility — a prong which Judge Posner, writing for the Seventh Circuit, has concluded “makes no sense,” see Gatimi v. Holder, 578 F.3d 611, 615 (7th Cir.2009).2 Our opinions note that “social visibility requires that the ‘shared characteristic of the group should generally be recognizable by others in the community.’ ” Id. at 1166 (quoting Santos-Lemus v. Mukasey, 542 F.3d 738, 745-46 (9th Cir.2008)). But we have not specified the relevant community for this analysis (Petitioner’s social circle? Petitioner’s native country as a whole? The United States? The global community?).3 Nor have we specified whether “social visibility” requires that the immutable characteristic particular to the group be readily identifiable to a stranger on the street, or must simply be “recognizable” in some more general sense to the community at-large.
Moreover, although our opinions include greater discussion of the particularity factor, we have not applied that factor consistently, and our reasoning is frequently at odds with the BIA’s published opinions on the matter. Most notably, our cases disagree whether unity of origin, or some other demographic tie, is a sine qua non for a finding of particularity. For example, in Soriano, a Filipino man petitioned for review of the BIA’s decision affirming the IJ’s denial of asylum, withholding of removal, and CAT relief. 569 F.3d at 1163. Soriano contended that he had a well-founded fear of future persecution in the Philippines because he had acted as a police informant against a Filipino criminal gang while living in Los Angeles. Id.
The IJ and BIA denied Soriano relief, and this court denied his petition for review. The panel held that Soriano’s proposed social group of government informants against gangs lacked sufficient particularity to delimit the group’s membership because:
A person who identifies as a “government informant” can be anyone of any demographic description who passes information to government authorities for any purpose. There is no “innate characteristic which is so fundamental to the *631identities or consciences” of government informants that identifies them as a particular social group. The purported group, therefore, “naturally manifest[s] a plethora of different lifestyles, varying interests, diverse cultures, and contrary political leanings.”
Id. (internal citations and quotation marks omitted). See also Velasco-Cervantes, 593 F.3d at 978 (rejecting a proposed group of former material witnesses for the United States government for lack of particularity because “any person of any origin can be involuntarily placed in that role in any type of legal proceeding”) (emphasis added).
But this reasoning conflicts with the BIA’s interpretation of § 208(b)(1), namely that members of a particular social group need not share kinship ties or origin, or have identical interests, lifestyles, or political leanings, to qualify for asylum. In Acosta, the BIA stated that — in addition to sex, color, or kinship — “shared past experience such as former military leadership or land ownership” could be a sufficient immutable characteristic for a particular social group. 19 I. & N. Dec. at 233-34. These paradigmatic groups based solely on shared past experiences (former military leadership or land ownership) could include “any person of any origin” and people with “a plethora of different lifestyles, varying interests, diverse cultures, and contrary political leanings.”4
Furthermore, our own cases are not consistent on the issue of shared kinship or origin as a sine qua non of particularity. For example, in Hernandez-Montiel v. INS, 225 F.3d 1084 (9th Cir.2000), petitioner — a Mexican national — applied for asylum on the basis that he had a well-founded fear of future persecution based on his membership in a particular social group comprised of homosexuals with female sexual identities living in Mexico. Id. at 1087. The IJ and BIA denied petitioner asylum, but this court granted his petition for review. Id. at 1099. This court held, as matter of law, that homosexuals with female sexual identities qualified as a particular social group because they shared a common immutable characteristic “so fundamental to one’s identity that a person should not be required to abandon [it].” Id. at 1093. This despite the fact that members of the proposed group: (1) could come from any origin or bloodline, and (2) did not necessarily share any beliefs or interests other than their sexual identity— the stated reasons for rejecting the proposed social groups in Velasco-Cervantes and Soriano, respectively.
In a word, some of our cases take as a touchstone for the “particularity” requirement, some aspect of identity (birth location, sexual orientation, kinship), which may be immutable, but reject others (such as voluntary inclusion in an informant group) that, once accomplished, are similarly immutable. Given the current confusion in our law, there is no discernible basis for these divergent outcomes — other than, perhaps, a given panel’s sympathy for the characteristics of the group at issue. Somalian women threatened with female genital mutilation are a particular social group, Mohammed v. Gonzales, 400 F.3d 785 (9th Cir.2005), while government witnesses threatened with death for their testimony against violent gang members are not, Soriano, 569 F.3d at 1166. Mexican men with female sexual identities are a particular social group, Hernandez-Montiel, 225 F.3d at 1084, while teenage boys in Honduras threatened with death for *632resisting MS-13 recruitment are not. . Ramos-Lopez v. Holder, 563 F.3d 855, 861 (9th Cir.2009). A petitioner fighting for her right to remain in this country and avoid persecution in her native land deserves a legal system governed not by the vagaries and policy preferences of a given panel, but by well-defined and consistently-applied rules.
Finally, our “particular social group” cases are not just internally inconsistent, but are also at odds with holdings of the Fourth, Sixth, and Seventh circuits in factually similar cases. For example, in Crespin-Valladares v. Holder, 632 F.3d 117 (4th Cir.2011), petitioners were Salvadoran nationals who applied for asylum because they had received death threats after one of their relatives testified against MS-13 gang members in a murder trial in El Salvador. Id. at 120. Petitioners contended they had a well-founded fear of future persecution based on their membership in a particular social group consisting of “family members of those who actively oppose gangs in El Salvador by agreeing to be prosecutorial witnesses.” Id. at 121. The IJ granted petitioners asylum, but the BIA reversed, holding that the proposed social group was not cognizable. Id.
The Fourth Circuit reversed the BIA. Id. The court held that the proposed group was socially visible because the petitioners’ cousin had publicly cooperated with the government and testified in court, and the gang knew that petitioners were related to the witness. Id. at 126. The court also held that the group satisfied the particularity element because the witness’s testimony in court and the kinship ties between petitioners and the witness could be readily verified; thus, membership in the group could be delimited. Id.5
Similarly, in Ramos v. Holder, 589 F.3d 426 (7th Cir.2009), petitioner was a former member of the Mara Salvatrucha gang in El Salvador (the gang against which Hen-riquez-Rivas testified). Id. at 428. Petitioner had quit the gang, become a born-again Christian, and emigrated to the United States. Id. Petitioner applied for asylum and withholding of removal, contending that he had a well-founded fear of future persecution based on his membership in a particular social group consisting of former MS gang members. Id. The IJ and BIA found the proposed group was not cognizable. Id. The Seventh Circuit reversed. Id. Judge Posner held that the group was socially visible to Salvadoran society as a whole, and that former membership in a “specific, well-recognized, indeed notorious gang” (MS) was an immutable — and verifiable — characteristic which was impossible to change. Id. at 429; see also Urbina-Mejia v. Holder, 597 F.3d 360 (6th Cir.2010) (holding that former members of the “18th Street Gang” in Honduras constituted a particular social group for asylum purposes).
In conclusion, our panel is left with little direction on how to apply the two-factor test for particular social groups. Because the relevant factors have been neither ex*633plained nor consistently applied by our court, I agree with the majority that the holdings in Soriano and Velasco-Cervantes are difficult to distinguish and must be followed under our “law of the circuit” jurisprudence. See Miller v. Gammie, 335 F.3d 889, 900 (9th Cir.2003) (en banc) (holding that a three-judge panel’s resolution of a legal issue in a published opinion binds all subsequent three-judge panels— unless the earlier panel is overruled by the Supreme Court or an en banc court).
However, were I writing on a clean slate, interpreting § 208(b)(1) with proper Chevron deference to the BIA’s language in Acosta and Matter of C-A-, I would follow the reasoning of our sister circuits to find that Henriquez-Rivas’s proposed group has sufficient social visibility and particularity to qualify as a particular social group. In addition to the specific threats against Henriquez-Rivas, there is significant evidence that Salvadoran society recognizes the unique vulnerability of people who testify against MS gang members. Most notably, as Henriquez-Rivas noted in her briefing before the IJ and BIA, the Salvadoran legislature enacted a special witness protection law in 2006 to protect people who testify against violent criminal elements, such as MS, in Salvadoran court. See U.S. Dep’t of State, Country Report on Human Rights Practices: El Salvador at 6 (2006); See also MCA, Inc. v. U.S., 685 F.2d 1099, 1104 n. 2 (9th Cir.1982) (noting that this court may take judicial notice of foreign laws under Fed. R.CivJP. 44.1 if the parties give written notice of their intent to raise an issue of foreign law). It is difficult to imagine better evidence that a society recognizes a particular class of individuals as uniquely vulnerable than that a special witness protection law has been tailored to their characteristics. Furthermore, as in Crespin, membership in Henriquez-Rivas’s proposed class can be easily verified — and thus delimited — through court records documenting the petitioner’s testimony against a gang.
Given the number and significance of these cases in our circuit, I believe this issue merits further examination to ensure that our precedent accurately tracks the language of the INA and affords proper deference to the BIA’s proclamations on the matter.
. There can be little doubt that Henriquez-Rivas’s fear of future persecution in El Salvador is well-founded. The IJ so found, and the BIA did not disturb, or even address, that holding. The issue before us is solely whether that persecution in El Salvador is based on one of the five recognized grounds in INA § 208(b)(1): race, religion, nationality, membership in a particular social group, or political opinion.
. Judge Posner noted:
[N]or has the Board attempted, in this or any other case, to explain the reasoning behind the criterion of social visibility. Women who have not yet undergone female genital mutilation in tribes that practice it do not look different from anyone else. A homosexual in a homophobic society will pass as heterosexual. If you are a member of a group that has been targeted for assassination or torture or some other mode of persecution, you will take pains to avoid being socially visible; and to the extent that the members of the target group are successful in remaining invisible, they will not be "seen” by other people in the society "as a segment of the population.”
Gatimi, 578 F.3d at 615.
. Some may think that we held the relevant community for purposes of "social visibility” is the alien’s home country, the country he seeks to avoid. See Santos-Lemus v. Mukasey, 542 F.3d 738, 745 (9th Cir.2008). Two things: (1) we held no such thing; we recited a BIA decision; and (2) that language had to do with the "particularity” prong, not the "social visibility” prong.
. To show the diversity of origin of persons who have been military leaders, compare Napoleon to the Duke of Wellington, or Fidel Castro, from a Spanish landowning sugar plantation family, to his mestizo, lower-class predecessor, Fulgencio Batista.
. It should be noted that the proposed social group in Crespin included only family members of the witnesses and not the witnesses themselves. However, to my mind, if the family members of witnesses are deemed socially visible and particular, the witnesses themselves — a more particular and socially visible and smaller class of people — must, a fortiori, meet those requirements as well. It may be that family members of a government witness or former gang member become subject to persecution through no act of their own, while a witness, informant, or former gang member chooses his path or paths. But once on that road, his persona is immutable, and newly laudable. Should we not encourage the gang members to turn against their murderous former cohorts with at least the same concern for their safety as we show for the former gangster’s family?