The Inclusion of a Subjective Component in “Well-Founded Fear” and the Insufficiency of Strict Textualism

By Emma Shoucair

Third-year student at the University of Michigan Law School.

There are two basic approaches to assessing the “well-founded fear” requirement in the Refugee Convention.  On the one hand, the United States, many other common law jurisdictions, and UNHCR require that the applicant have a subjective fear of persecution and that this fear be grounded in objective reality.   On the other hand, New Zealand, civil law jurisdictions, and the Michigan Guidelines require either a softer inquiry into subjective emotional state or no subjective fear inquiry at all.  The subjective fear element in common law jurisdictions is a product of strict textualist interpretation.  The de-emphasis on the subjective fear in civil law jurisdiction stems from the approach in the Vienna Convention on the Law of Treaties, which mandates interpretation of the Refugee Convention’s text in light of its context and object and purpose. Interpreting “well-founded fear” in a strictly textual way, as the United States has done, fails to effectuate the goals of the Refugee Convention and misses a small but salient shift in the meaning of the word fear. This interpretation should be revisited.

I. Background and the Legal Landscape

This note will use as a starting point the United States Supreme Court’s decision in Immigration & Naturalization Service v. Cardoza-Fonseca, in which the Court announced that “well-founded fear” encompasses both subjective (that the applicant experience the emotion of fear) and objective (that this fear be reasonable and grounded in fact) elements. Through an examination of the Court’s reasoning, as well as an examination of the text, context, and object/purpose of “well-founded fear” in the Refugee Convention, this Note argues that the objective prong of the “well-founded fear” test, without a separate, outcome-determinative assessment of subjective fear, is the appropriate legal standard.

a. Split Among Top Courts: Bipartite or Unitary Interpretation of “Fear”

The world’s courts can be divided into roughly two camps. Some courts conduct a bipartite fear inquiry first into whether the applicant subjectively fears persecution upon return to his or her home country and then into whether that subjective fear is objectively reasonable.[1] Other courts take the applicant’s subjective fear into account in the course of determining whether that fear is objectively well founded, but do not conduct a separate dispositive inquiry into the applicant’s subjective emotional state.[2] Many common law jurisdictions, including the United States, fall into the former camp, while the civil law jurisdictions and New Zealand fall into the latter.[3]

i. Bipartite Interpretation

1. The United States

In Immigration and Naturalization Service v. Cardoza-Fonseca, the Supreme Court addressed the case of a Nicaraguan national who had overstayed her visitor’s visa and failed to depart voluntarily.[4] She entered removal proceedings and subsequently filed for withholding of removal and asylum, claiming that she had a “well-founded fear” of persecution upon her return to Nicaragua on account of her political views and her brother’s political activities.[5] Engaging first and foremost with the text, the Supreme Court held, inter alia, that the “well-founded fear” standard laid out in the Refugee Convention (as interpreted through the INA) contained an “obvious focus on the individual’s subjective beliefs” demonstrated through the “ordinary and obvious meaning of the phrase.”[6] In his concurrence, Justice Blackmun also noted that “the very language of the term ‘well-founded fear’ demands a particular type of analysis—an examination of the subjective feelings of an applicant for asylum coupled with an inquiry into the objective nature of the articulated reasons for the fear.”[7] Justice Scalia also concurred on the “plain meaning of ‘well-founded fear.’”[8]

Thus, using a primarily textualist reading of “well-founded fear” rather than a textualist-purposive reading, which takes into consideration the context and purpose of the original text, the Court concluded that “well-founded fear” should be assessed using the two-part test: (1) whether an applicant is subjectively fearful; and (2) whether that fear is objectively reasonable.[9]

The United States is not alone in interpreting “well-founded fear” to include both subjective and objective components. The UNHCR Handbook unambiguously states that “[t]he term ‘well-founded fear’…contains a subjective and objective element.”[10] The high courts of Australia, Canada, Ireland, and the United Kingdom have adopted the bipartite approach to “well-founded fear” consistent with the test laid out in the UNHCR Handbook.[11]

ii. Unitary Interpretation

1. New Zealand

The outlier in the common law world, New Zealand’s Refugee Status Appeals Authority (RSAA), has explicitly rejected the subjective element of the “well-founded fear” inquiry.[12] Its courts have not overturned this holding, noting that “the Authority…is not bound to speculate on matters which in the end are no more than guesswork.”[13]  Difficulties in assessing the subjective emotional state of applicants have led common law courts outside of New Zealand to use credibility to determine subjective fear, contravening the purpose of the Convention and risking rejection of legitimate claims.[14] 

2. Civil Law Jurisdictions

Unlike common law jurisdictions, civil law jurisdictions interpreting the equally authoritative French text of the Refugee Convention have not imposed a subjective fear requirement as an essential part of the craignant avec raison test.[15] These jurisdictions have often incorporated an inquiry into subjective fear into their overall assessment of whether or not an applicant’s fear of persecution is reasonable, but civil law courts have not treated this as a hard and fast requirement.[16]

3. The Michigan Guidelines

The Michigan Guidelines go further than the civil law courts and pushes to abolish the subjective inquiry:

Most critically, the protection of the Refugee Convention is not predicated on the existence of “fear” in the sense of trepidation. It requires instead the demonstration of “fear” understood as a forward-looking expectation of risk. Once fear so conceived is voiced by the act of seeking protection, it falls to the state party assessing refugee status to determine whether that expectation is borne out by the actual circumstances of the case. If it is, then the applicant’s fear (that is, his or her expectation) of being persecuted should be adjudged well-founded.

The United States’ approach stands in opposition to the approach taken by the Guidelines, and, to a slightly lesser extent, that taken by civil law courts and New Zealand.

II. Proper Interpretation Using Vienna Convention Principles

There is a tendency in American courts to prioritize textualism above all other methods of interpretation, sometimes conflating a strict reading of a text with its original purpose.  Textualist analysis can stand as proxy for original intent, as though a careful adherence to text will invariably lead to the originally intended outcome.  However, determining the original purpose of a law is a difficult undertaking better delegated to historians than judges.[17] Not all judges in America are strict textualists, but in our current moment this interpretive method has become the currency of legitimacy. It also ignores reality in a very important way: words regularly undergo changes in meaning independent of the law.  A purely textualist approach risks, over time, a completely faithful adherence to the text that is nonetheless unmoored from what the law was intended to accomplish. Truly originalist interpretations must, in some circumstances, detach from text as the semantics of words changes over time.

The context, object, and purpose of the Refugee Convention counsel in favor of abandoning the subjective fear analysis. Articles 31 and 32 of the Vienna Convention dictate that text must be construed in light of its context, object, and purpose to ensure that treaties respond to changing global circumstances and continue to serve the purposes for which they were enacted.  Addressing purpose protects textual interpretation from departing too far from the original goals of the treaties in a changing world. When the Vienna Convention principles are properly applied here, “well-founded fear” should contain only an objective assessment of the risk of persecution.  This is because the text of the Refugee Convention—both French craignant and English fear—plainly permits an anticipatory meaning.

a. Text

A strong textualist argument can be made that, as long as we understand that the semantics of words can change over time and we interpret the text as it would have been interpreted at the time of drafting, “well-founded fear” clearly allows for an anticipatory interpretation. The United States Supreme Court did not use this diachronic textualist interpretive approach in Cardoza-Fonseca. Because the phrase “well-founded fear” has such an “ordinary and obvious meaning” to the justices, they did not consider whether that meaning differed at the time of drafting from its meaning at the time of interpretation. The justices here imposed their synchronic speaker intuitions on the meaning of fear without going further.  In reality, the semantics of words are fluid over time and can shift in measurable ways over relatively short periods.

Fear is such a word. In modern English, there are two main meanings: first, fear can mean “a form of emotional response,” and second, it can be an “anticipatory appraisal of risk.”[18] To the modern ear, this second meaning sounds more affected and less natural, undoubtedly part of the reason the justices found their interpretation so “obvious.”  I conducted a small corpus study to determine whether the distribution of these two meanings has shifted in American English since 1950. Using the Corpus of Historical American English to examine the frequency of each meaning circa 1950 and the Corpus of Contemporary American English to examine their frequencies in modern English, I used the following protocol: in each corpus I searched for instances of verbal forms of fear.[19]

For each of the first 100 tokens, I tallied whether the token demonstrated an emotional or anticipatory meaning.  In 1950s American English, 45 of the first 100 tokens carried an emotional meaning, while 55 carried the anticipatory meaning. By the time we get to 1985–90 (the time of interpretation), however, the proportion has shifted: 72 of the first 100 tokens are emotional, and only 28 are anticipatory. The difference is persistent and starker still in modern American English (2017): 85 of the first 100 tokens are emotional, and 15 are anticipatory. While both meanings are clearly still present in modern American English, the anticipatory use of fear appears to be less common. If interpreters of legal texts are not sensitive to these trends, it is easy to impose our synchronic understanding of lexical semantics in the place of a truly purposive understanding of meaning.

This evidence, especially when combined with the interpretations of the equally authoritative French craignant of the civil law jurisdictions, speaks in favor of common law jurisdictions reconsidering their subjective fear requirement and interpreting “well-founded fear” as purely objective and anticipatory. The text the drafters of the Convention memorialized may indeed have been intended to reflect an anticipatory use of fear rather than an emotional one.

b. Context, Object and Purpose

Within the context of the Refugee Convention itself, assessing fear objectively fits with the overall approach of the Convention.  As the Michigan Guidelines points out, we terminate refugee status not when the refugee’s subjective fear of harm has ceased, but rather when the objective risk to that refugee has ceased.

Furthermore, given that the object and purpose of the Refugee Convention is to provide surrogate national protection to those at risk of persecution, there is an unseemliness to an inquiry into subjective fear (apart from the difficulties in determining someone’s true subjective state of mind).:[20] It feels as though we are saying an applicant is deserving of assistance only if they are sufficiently afraid.  The drafters of the Refugee Convention, however, could not possibly have meant to condition a person’s worthiness for asylum on that individual’s level of emotional expression.[21] Anyone who is at risk of persecution in their home state and meets the other criteria laid out in the refugee definition is worthy of protection, not just those who emote on command sufficient to satisfy a host state.  The framers of the Refugee Convention intended to lay out a test that delineated those at risk of persecution in their home states from those not at risk, and the subjective fear element only increases the risk that states using it will fail to recognize a valid refugee.[22]

III. Conclusions

In the final analysis, the Vienna Convention principles requiring states to interpret the text of treaties in light of their context, object, and purpose are 1) a far superior way of ensuring the Refugee Convention continues to serve its intended purpose than the strict textualist approach taken by, e.g., the United States’ court system and 2) embodied best by the conclusions reached by the Michigan Guidelines.  The United States Supreme Court should revisit its language requiring the assessment of the applicant’s subjective emotional state when determining whether that applicant meets the “well-founded fear” test.

In Cardoza-Fonseca specifically, one might make the argument that the discussion of the “plain language” of “well-founded fear” in the majority opinion is dicta and not necessary in the narrowest sense to decide the issue before the Court (the difference between the standard of proof required for asylum versus withholding).[23] However, the United States and the other common law jurisdictions can point to heavy domestic stare decisis concerns, so a fruitful first step would be for the UNHCR to formally adopt the approach to “well-founded fear” laid out in the Michigan Guidelines for signatories to the Convention to point to.  Alternatively, the civil law jurisdictions’ adoption of a subjective element in the overall inquiry is less faithful to the object and purpose of the Refugee Convention than the absence of a subjective inquiry altogether, but common law states might feel more comfortable shifting to this middle approach because they may feel that it somewhat constrains their obligations, and thus might make them more amenable to dropping the strict requirement of subjective fear.


[1] See James C. Hathaway & Michelle Foster, The Law of Refugee Status 91–92 (2d ed. 2014).

[2] See id. at 106–08.

[3] Id. at 91–92.

[4] Cardoza-Fonseca, 480 U.S. at 424.

[5] Id. at 424-25.

[6] Cardoza-Fonseca, 480 U.S. at 431.

[7] Id. at 450 (Blackmun, J., concurring).

[8] Id. at 452 (Scalia, J., concurring).

[9] Id. at 425, 440

[10] U.N. High Commissioner for Refugees, Handbook on Procedures and Criteria for Determining Refugee Status Under the 1951 Convention and the 1967 Protocol Relating to the Status of Refugees, ¶ 38, U.N. Doc. HCR/IP/4/Eng/REV.1 (1979).

[11] See also Hathaway & Foster, supra note 4, at 92 nn.6–9.

[12] K v. Refugee Status Appeals Authority [No. 2] [2005] NZAR 441 (HC) at [12] (N.Z.).

[13] Id. at [27].

[14] For a discussion of the pitfalls of assessing subjective fear through the applicant’s credibility, see Hathaway & Foster, supra note 4, at 100–02.

[15] Hathaway & Foster, supra note 4, at 106.

[16] Id.

[17] See, e.g., District of Columbia v. Heller, 554 U.S. 570 (2008) (showing the back and forth on the historical record between Justices Scalia and Stevens, who were unable to reach any consensus on the historical presence or absence of an individual ownership right in the Second Amendment to the United States Constitution).

[18] Hathaway & Foster, supra note 4, at 105–06.

[19] Data is on file with the author. A brief word on methodology: I limited the token count to verbal (rather than nominal) tokens of fear in the interests of consistency. Additionally, the difference between the emotion and anticipatory uses of fear is somewhat easier to see when used as a verb than as a noun. A more rigorous study of the corpus would entail examination of a higher number of tokens (possibly all the tokens in the corpus). I deliberately did not limit the register or context fear appeared in for either search, since the everyday meaning of the word influences the senses in which it is used legally. Admittedly, the line between emotional and anticipatory can be somewhat blurry, so in instances where a token was on the line, I excluded it from the count; this introduced some subjectivity into the token count, which could be ameliorated by providing the tokens to additional testers and averaging out the final tallies. Much more could be said about the methodology of a full and rigorous study on the topic; I attempt here only to illustrate the semantic range and rough distribution.

[20] For discussion of the interpretive pitfalls of subjective fear, see Hathaway & Foster, supra note 4, at 95–104.

[21] James C. Hathaway & William S. Hicks, Is there a Subjective Element in the Refugee Convention’s Requirement of ‘Well-Founded Fear’?, 26 Mich. J. of Int’l L. 505, 507 (2005).

[22] Id. at 507, 514–17.

[23] Immigration & Naturalization Service v. Cardoza-Fonseca, 480 U.S. 421, 443 (1987).


Submit A Comment

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>