Abstract

One of the characteristics of what has been called “dogwhistle politics” is the presence of a rhetoric that targets minority groups implicitly. For example, terms like ‘illegals’ and ‘illegal immigrants’, used to target Latin-Americans, have come to permeate the American political discourse as well as everyday conversations. Here I focus on how such expressions, which I call illegality frame code words (IFCW, for short), can be countered by recalcitrant hearers. I begin with the assumption that IFCWs are racial code words, conversational devices that convey implicit racial appeals while allowing for deniability. I then discuss how the existence of an Illegality Frame in the American Immigration debate supports their deniability. Lastly, I discuss how recalcitrant hearers can counter utterances that contain them. In particular, I propose reframing as an adequate strategy. In reframing, agents take control of the goals or the QUD (question under discussion) of the conversation to exclude certain topics from the common ground and include others. The idea is to “change the conversation” to neutralize problematic moves. This maneuver is advantageous not only because it helps to hinder veiled discriminatory practices, but also because it affords control over “the terms of the conversation”.

Keywords:
Code Words; Deniability; (Re)Framing; QUD-shifting

1. Introduction

Framing is defined by Chong & Druckman (2007Chong, D., & Druckman, J. N. (2007). Framing theory. Annual Review of Political Science, 10:103-126. https://doi.org/10.1146/annurev.polisci.10.072805.103054
https://doi.org/10.1146/annurev.polisci.... :104) as the process by which people develop a particular conceptualization of an issue or reorient their thinking about it depending on how the topic is characterized: a powerful communicative tool to set political agendas as well as conceptual grounds for public debates. By way of illustration, consider the comprehensive immigration reform presented by Bush in 2006. As Lakoff & Ferguson (2006Lakoff, G. & Ferguson, S. (2006). The framing of immigration. [accessed 20 November 2022] [accessed 20 November 2022] https://www.huffpost.com/entry/the-framing-of-immigratio_b_21320 .
https://www.huffpost.com/entry/the-frami... ) point out, in framing the “problem” of immigration in terms of citizenship laws and border patrol, Bush excluded the need for public measures around its humanitarian and civil right dimensions. As a result, his reform contributed to circumscribe the public debate about immigration mainly around the notion of illegality, a tendency that has remained centerstage, especially in the last decade, with the Trump Era and the intensification of polarization in the American Congress (Alamillo et al 2020Alamillo, R., H, C., & Madrid Jr, R. (2019). Framing and immigration through the trump era. Sociology Compass, 13(5). https://doi.org/10.1111/soc4.12676
https://doi.org/10.1111/soc4.12676... ).

Experts refer to the conceptual and linguistic frame that associates immigration to illegality as the Illegal or Illegality frame. Sociologists, in particular, have highlighted that the current process of “racialization of illegality”² 2 In other words, that seemingly neutral laws and rhetoric target specific social groups as illegal. See Chavez 2020. targets mainly Latino/as³ 3 Mexicans, Guatemalans, Hondurans, and Salvadorans make up 70% of the undocumented population and 88.6 % of detained and deported immigrants in America today (Menjívar 2021). . In the same line, psychologists have found evidence that race predicts negative attitudes towards immigration policies involving legalization. Perez (2010Perez, E. O. (2010). Explicit evidence on the import of implicit attitudes: The IAT and immigration policy judgments. Political Behavior, 32(4), 517-545. https://doi.org/10.1007/s11109‐010‐9115‐z
https://doi.org/10.1007/s11109‐010‐9115‐... ), using implicit association tests (IAT)⁴ 4 IAT is a computer-based measure involving the rapid sorting of stimuli. It has been successfully used to identify biases related to other out-group categories, such as African-Americans. For more on IAT, see Greenwald et al 1998. , tested subjects about measures such as citizenship for illegal immigrants and found that previous references to Latina/os prompted negative attitudes towards such policies.

This relation between race, anti-immigration sentiments and illegality is reflected in the fact that, in the current context of what Haney-López (2014Haney-López, I. (2014). Dog whistle politics: How coded racial appeals have reinvented racism and wrecked the middle class. Oxford University Press.) calls “dogwhistle politics” - roughly put: politics that is carried out through veiled persuasion -, the rhetoric of some public figures⁵ 5 Sections 2 and 3 include uses of such words made by American authorities, like late justice Scalia, former-president Trump and former-sheriff Joseph Arpaio. has come to include terms that explore the Illegality frame to implicitly target the Latino community. One such term is ‘illegals’, as used, for example, in (1).

In this work, I focus first on how users of terms like ‘illegals’ - call them illegality frame code words (IFCW, for short)⁶ 6 In addition to ‘illegals’, the category includes words like ‘illegal immigrants’, ‘illegal aliens’, ‘anchor baby’ etc. Some of them, like ‘illegals’ and ‘anchor baby’, are also considered pejoratives because of their dehumanizing connotations. - convey racial appeals. I begin with the claim that IFCWs are code words (Khoo 2017Khoo J (2017) Code words and political discourse. Philosophical Topics 45(2):33-64. https://doi.org/10.5840/philtopics201745213
https://doi.org/10.5840/philtopics201745... ): discursive devices that allow for deniability and are used in strategic conversations (Camp 2018Camp, E. (2018). Insinuation, common ground. In Fogal, Daniel, Daniel W. Harris, and Matt Moss, eds. New work on speech acts. Oxford University Press, 40-66.) to make implicit racial appeals while avoiding liability and the risk of confrontation. I then suggest that this deniability is supported and licensed by the existence of the Illegality frame in public debate.

Secondly, I focus on how recalcitrant hearers can counter utterances of IFCWs efficiently in conversational contexts. In this regard, I propose that they have at least two good strategies: calling-out challenges and reframing. The second one, in particular, aims to take control of the goals of the conversation in order to exclude certain topics from the common ground and include others. The goal in such cases is to “change the conversation” mid-discourse to neutralize conversational moves that are potentially problematic. To explain this, I resort to the concepts of topic-elimination (Picazo 2022Picazo, C. (2022). Distorted Debates. Topoi, 1-11.) and QUD-shifting (Roberts 2012Roberts, C. (2012). Information structure: Towards an integrated formal theory of pragmatics. Semantics and pragmatics, 5, 6-1.; Keiser 2022Keiser, J. (2021). The “All Lives Matter” response: QUD-shifting as epistemic injustice. Synthese, 199(3-4), 8465-8483.).

The article is structured as follows: in the next section, I define code words and their deniability relatively to other variants of manipulative speech, like dogwhistles and racial figleaves (in subsection 2.1.). In subsection 2.2., I suggest that one of the sociological roots of their deniability is the process of framing in political discourse. In section 3, I turn to conversational interactions, focusing on reframing in subsection 3.1.

2. Code words in the illegality frame

Following Khoo (2017Khoo J (2017) Code words and political discourse. Philosophical Topics 45(2):33-64. https://doi.org/10.5840/philtopics201745213
https://doi.org/10.5840/philtopics201745... ), I take IFCWs to be racial code words. Racial code words are expressions used to convey implicit racial messages without committing the speaker to explicit racism; ‘welfare’, ‘inner city’ and IFCWs are common examples. Khoo claims that, in using an expression such as ‘illegals’ or ‘illegal immigrant’, a speaker exploits her audience’s stereotypical beliefs to have them draw certain racial inferences.

Consider his example of when late justice Antonin Scalia publicly argued that “illegal immigrants” were a problem for Americans because they “invade their property, strain their social services, and even place their lives in jeopardy” (Khoo 2017Khoo J (2017) Code words and political discourse. Philosophical Topics 45(2):33-64. https://doi.org/10.5840/philtopics201745213
https://doi.org/10.5840/philtopics201745... :49). Any competent speaker that has a stereotypical belief of the form S below, where C stands for illegal immigrants (illegals, illegal aliens etc.) and R for Latin-Americans, is able to infer that what Scalia said about illegal immigrants was said about Latin-Americans.

S: If something is a C, then it is a R.

Now, Khoo is interested in showing that, despite conveying a secondary message, (racial) code words are not multi-dimensional terms - as argued by Stanley (2015Stanley J (2015) How Propaganda Works. Princeton: Princeton University Press.)⁷ 7 See Khoo 2017: 43-46. . Multi-dimensional terms are expressions with two levels of meaning: at-issue and not-at-issue, where the distinction between the two is often put in terms of availability for truth-conditional operators - the at-issue level is available, while the not-at-issue is not -, cancelability and backgroundness.⁸ 8 See Simons et al 2010 and Barbosa 2021 for more details. Take (2) as an example:

In (2), the not-at-issue content, He used to smoke, appears in a parenthetical use - i.e., it is backgrounded - and cannot be cancelled without oddity or contradiction, as shown by (3).

In the case of ‘illegals’ or ‘illegal immigrant’, first, the not-at-issue level of meaning has to be a racial content (e.g., Latin-Americans), and, secondly, it has to be non-cancelable - that is, any attempt to cancel it should cause oddity or be contradictory. Nevertheless, (4) below,

in which ‘most of whom are not Latin-Americans’ is supposed to cancel the alleged not-at-issue content of the IFCW, does not strike us as contradictory, indicating that whatever this racial component is, it is not not-at-issue. This is one of Khoo’s arguments to dismiss the assumption that code words are multi-dimensional and to put forward his own “simple theory of code words”.⁹ 9 To be more specific, Khoo rejects two options, that they are ambiguous (2017:40-42) and that they are multidimensional, favoring instead a simple semantics, according to which they have only one conventional meaning. He also rejects an option based on speaker’s meaning (Grice 1957) - i.e., meaning that depends on the speaker’s intention. He thinks that the racial appeal is conveyed but not communicated by the speaker.

In his proposal, the racial appeal conveyed by an IFCW is an effect¹⁰ 10 Here too Khoo considers two options: whether the effect is produced by an inferential process or an associational one. He is more inclined towards an inferential approach, but he takes this inclination to be provisional since there is still not enough empirical research on code words to allow for a reliable conclusion on the matter. See Khoo 2017:51-52. of the use of the expression. To understand this a little better, take (1) again. In Khoo’s account, (1) activates the pre-existing stereotypical belief (5) below.

The hearer then discloses the final racial message by inferring (6).

Notice that this under-the-radar inferential process depends on how the expression impacts the hearer’s previous cognitive status and not on the speaker intending to communicate something about race. Even in the absence of a communicative intention, the speaker still generates the inference from the stereotypical racial belief, provided, among other things, that the hearer has it.¹¹ 11 Khoo suggests that this inferential process could be understood in terms of perlocutionary effects (Austin 1962), given its independence from intentions. After all, perlocutions do not depend on the speaker’s intention to be produced, firstly, because the speaker may have the intention and ultimately fail to produce the effect. Secondly, because, in performing an illocutionary act, the speaker may bring a perlocutionary effect about unintentionally. However, Khoo does not elaborate on this suggestion.

Khoo is especially convincing in his use of linguistic tests (e.g., the tests in (2)-(4)) to show that this comes as a consequence of the failure of meaning-based approaches (like Stanley’s) to successfully show that code words linguistically encode implicit meanings. So, my goal here will be to further develop Khoo’s account by elaborating on aspects concerning the deniability of IFCWs. In particular, I want to discuss the political and communicative practices that enable it. Because IFCWs share this defining feature with dogwhistles and racial figleaves (Saul 2017Saul J (2017) Racial figleaves, the shifting boundaries of the permissible, and the rise of Donald Trump. Philosophical Topics 45(2):97-116. https://doi.org/10.5840/philtopics201745215
https://doi.org/10.5840/philtopics201745... ), I will first say a few words about if and/or how they differ from these similar phenomena.

2.1. Code words, dogwhistles, figleaves and deniability

In very general lines, dogwhistles are expressions designed to transmit covert messages. It is therefore common to see them characterized as coded speech (Quaranto 2022Quaranto, A. (2022). Dog whistles, covertly coded speech, and the practices that enable them. Synthese, 200(4), 330.), since, like paradigmatic codes, they send one message to the overall audience and another to a particular subset of it. To give an example, the words ‘Google’ and ‘Yahoo’ were recently used as codes for African-Americans and Mexicans (respectively) by racist groups on Twitter (Kantrowitz 2016Kantrowitz, A. (2016) Racist Social Media Users Have a New Code to Avoid Censorship. Buzzfeed. [Online] [Accessed on 18 April 2017] [Online] [Accessed on 18 April 2017] https://www.buzzfeednews.com/article/alexkantrowitz/racist-social-media-users-have-a-newcode-to-avoid-censorshi
https://www.buzzfeednews.com/article/ale... ). These groups added in-group meanings to the names of two well-known corporations and the maneuver allowed them to dogwhistle violent racist messages without having their accounts suspended by the platform’s algorithms.

Dogwhistles are mainly defined by two general features: directionality and deniability (Lo Guercio & Caso, 2022Lo Guercio, N., & Caso, R. (2022). An account of overt intentional dogwhistling. Synthese, 200(3), 1-32.); that is, the hidden messages they send are directed at subsets of the audiences; and their users can deny having issued them. Saul (2018Saul J (2018) Dogwhistles, political manipulation, and philosophy of language. In: Fogal, D., Harris, D. and Moss, M., (eds.) New work on speech acts. Oxford University Press, 360-383. https://doi.org/10.1093/oso/9780198738831.003.0013
https://doi.org/10.1093/oso/978019873883... ) classifies them into different categories. First, the category of overt intentional dogwhistles (OIDs): roughly put, OIDs are coded messages intentionally addressed to specific subsets of the audience that are capable of recognizing them. The Twitter case is a good example. Secondly, the category of covert intentional dogwhistles (CIDs, for short), which are defined as utterances that seem innocuous and unrelated to race but that, in reality, attempt to manipulate the salience of the hearer’s racial attitudes without her awareness.¹² 12 Saul also includes unintentional dogwhistles in her classification. These dogwhistles are re-showings or replayings of intentional dogwhistles that have, nevertheless, the same results. I will not discuss them here.

Saul draws on the study of political racial priming¹³ 13 Racial priming in politics is a rather pernicious strategy that involves, for example, using imagery in political campaigns to prompt racial attitudes without being detected. by Mendelberg (2001Mendelberg T (2001) The Race Card: Campaign Strategy, Implicit Messages and the Norm of Equality. Princeton University Press, New Jersey.) to build her philosophical account of this last category. According to Mendelberg, the post-civil rights American society adheres to what she calls The Norm of Racial Equality, which stablishes that racism is wrong and instances of racist speech or behavior should be socially sanctioned. However, her empirical studies also found high levels of racial resentment towards African-Americans among the population. Racially resentful agents tend to think that: a) African-Americans no longer face discrimination, b) they already have more than they deserve for having poor work ethics; and c) their demands for social justice are not justified. Mendelberg then concluded that the combination of the norm of racial equality and the prevalence of racial resentment explained why politicians that seek to gain support from racist voters resort to implicit rather than explicit racist appeals.

The main example of a CID Saul borrows from her is the Willie Horton add, used in George H. W. Bush’s presidential campaign against Dukakis in the 1980s. The add showed an image of Willie Horton, a black man who was also a furloughed convict. The add was meant to criticize the furlough program during Dukakis time as the governor of Massachusetts. After the add began to be used, Dukakis, who was ahead in the opinion polls, started to plummet. Later on, after the add was publicly called-out as racist, Dukakis regained his previous advantage in the polls.

According to Saul, this case evidences that Americans tend to change attitudes depending on whether racism is implicit or explicit in a given context as a consequence of the norm of racial equality: when the audience is made aware that a racist message has been sent, the CID loses its power. In this sense, one important condition for something to count as a CID is that the intended effect of bringing certain racial attitudes to prominence is achieved but the intention to do so is not recognized. Saul defines them then as covert perlocutionary speech acts. They are perlocutionary, on the one hand, because they consist of effects produced by the utterance, and they are covert, because the dogwhistle effect will fail to succeed if the intention to provoke it is recognized by the hearer. So, CIDs have directionality only in the sense that they are directed at bigoted audiences, but not in the sense that the targeted audience is aware of the message - as in the case of OIDs.

Racial code words do not have directionality. Consider again ‘illegals’: it seems that not just the sub-audience composed by anti-Latino bigots will have the pre-existing belief expressed by (5), but rather that (5) will be at least accessible to other speakers who are linguistically and culturally competent with the American immigration debate. What distinguishes the anti-Latino and the non-anti-Latino audiences more relevantly in this case are the attitudes of their respective members relatively to the racial inference. To bring home the point, compare an utterance of (1) with the Twitter case, in which the target sub-audience differs from the wider audience because its members know something that the members of the wider audience don’t - namely, the words’ in-group meanings. In such a case, there is a difference in linguistic competence between the two audiences. In the case of the IFCW, the most relevant difference is attitudinal.

Nevertheless, racial code words will share the property of (granting) deniability with dogwhistles. Dogwhistlers can deny having issued hidden messages by alleging things like ignorance, lack of intention, unawareness of the message, misquotation, coincidence etc. In the case of paradigmatic OIDs, this will depend on the fact that the meanings attached to the utterances are “under-the-radar”. Overall, these contents are deniable because they are implicitly conveyed, so trivially speakers can more easily avoid being held accountable for them than for what they explicitly said.

Before moving on to discuss the deniability of IFCWs, there is one last kind of conversational device which grants deniability that I would like to mention: racial figleaves. As defined in Saul (2017Saul J (2017) Racial figleaves, the shifting boundaries of the permissible, and the rise of Donald Trump. Philosophical Topics 45(2):97-116. https://doi.org/10.5840/philtopics201745215
https://doi.org/10.5840/philtopics201745... ), a figleaf is “an utterance made in addition to an otherwise overtly racist one, that serves the function of calling into question the racism of the speaker and the utterance” (:98). Basically, the purpose of a figleaf is to try to turn an explicit racist appeal into something that a non-racist would say. The most common type of racial figleaf takes the form of (7):

Saul calls this a (synchronic) Denial figleaf; it contains a denial statement (e.g., ‘I am not a racist’), call it ‘F’, followed by a racist utterance. In such cases, if the speaker is accused of racism, F gives her grounds to say something like: “I have been careful to point out that I am not a racist person …”. The role of F is thus to block inferences concerning the speaker’s intentions, beliefs and attitudes and, in this sense, disentangle her from racism with a preemptive “personalistic” claim.¹⁴ 14 The personalistic ideology (about racism) assumes that racism is an individual - as opposed to systemic - matter. See Hill 2008. Unlike the dogwhistler or the user of an IFCW, the utterer of a figleaf cannot plead ignorance, unawareness, coincidence etc., since saying F evidences that she is aware of the racism conveyed by her words¹⁵ 15 Another difference between figleaves and IFCWs is that although the former serves to divert assumptions regarding the speaker’s racism, racism is explicitly acknowledged, i.e., it is part of the main conversation. In uses of the latter, in contrast, racism is kept off the record. . Deniability in paradigmatic figleaves then stems from the preemptive move made to point out something about oneself.

To recap, racial code words - and IFCWs in particular - are similar to both dogwhistles and figleaves in their property of granting deniability. They introduce race in conversations in ways designed to be deniable, though the ways in which deniability manifests itself are slightly different in each category. The deniability of IFCWs has characteristics that are built on processes like political framing. I turn to this issue next.

2.2. Framing in politics and the deniability of IFWCs

As I mentioned earlier, framing is a powerful communicative tool in politics. Consider the power of slogans - like Trump’s ‘MAGA’ (‘Make America Great Again’) or Obama’s ‘The Change we need’ - to present campaigns. Slogans serve mainly to showcase the core assumptions and commitments of a campaign, but they also help to frame it by setting up and cutting-off the concepts and values that are relevant for the candidacies - e.g., nationalism or social justice. Similarly, when governments and authorities present reforms (bills, etc.), they frame problems and issues with the purpose of circumscribing and reinforcing the preferred concepts, values and vocabulary. Back to Bush’s immigration reform, for instance, while he wasn’t the first to use the Illegality frame, its use made sense considering the types of measures that his government was interested in putting forward as well as the discussions it wanted to leave out. As Lakoff & Ferguson put it: a frame “imposes a structure on the current situation, defines a set of ‘problems’ with that situation, and circumscribes the possibility for solutions” (2006Lakoff, G. & Ferguson, S. (2006). The framing of immigration. [accessed 20 November 2022] [accessed 20 November 2022] https://www.huffpost.com/entry/the-framing-of-immigratio_b_21320 .
https://www.huffpost.com/entry/the-frami... :1). This is a highly goal-oriented and strategic process, made to direct the public debate towards certain choices and preferences and away from others, even though it is often presented as objective.

Everyday uses of the illegality rhetoric too present illegality as an “objective”, “plausible” or “neutral” (meaning, non-racist) justification to antagonize undocumented immigrants, most of whom are Latin-American. Take the case of Joseph Arpaio, the Sheriff of Maricopa County, Arizona, who became notorious in the U. S. for his vocal opposition to “illegal” immigration. In responding to accusations of targeting Latino communities, Arpaio claimed that he was not a racist and his only goal and responsibility as a sheriff was to enforce the law.¹⁶ 16 In an interview given to Cecilia Menjívar, mentioned in Mejívar 2021.

This kind of denial is also common when the use of words like ‘illegal immigrant’ and ‘illegals’ is directly challenged by a recalcitrant hearer. So, if, in a hypothetic example, the utterer of (1) is biased against Latina/os but knows that racism is frowned upon (as a consequence of a rule like the Norm of Racial Equality), she may choose to use ‘illegals’ instead of ‘Latinos’/‘Latin-American’ (or a slur for Latina/os) as a way to avoid the risk of an objection on the grounds of racism and bigotry. Such hypothetical situation illustrates what Camp (2018Camp, E. (2018). Insinuation, common ground. In Fogal, Daniel, Daniel W. Harris, and Matt Moss, eds. New work on speech acts. Oxford University Press, 40-66.) calls strategic conversations: interactions that follow the same rules as fully cooperative conversations, but in which one party crafts her contributions in a way that minimizes the risks of objection and conflict.

Deniability is a typical mark of such conversations. It is enabled in the case of IFCWs by three factors: 1) the fact that the interaction itself is risky; 2) that an alternative “neutral” way of framing the group is available, and 3) that such an alternative is deployable in approximately the same contexts as the more explicitly racial rhetoric. With respect to 1), by ‘risky’ I mean conversations in which there is a palpable risk of violating rules like the Norm of Racial Equality. The risk here is of conflict and loss of cooperativity - which I will discuss in more detail in the next section.

Now, the illegality frame will be crucial to explain 2) and 3). First, because it offers the alternative vocabulary and concepts mentioned in 2), which includes the contested code word. Particularly, the Illegality frame warrants that the racial code word operates as a “code” in Khoo’s sense: as an expression that can be used to talk about race veiledly, while the “main” conversation is presented as being about illegality. I take the relevance of 3) from Quaranto’s (2022Quaranto, A. (2022). Dog whistles, covertly coded speech, and the practices that enable them. Synthese, 200(4), 330.) account of deniability. She argues that coded speech, which she defines partially in terms of deniability,¹⁷ 17 To that extent, then, her account can be applied to IFWCs. However, IFCWs are not really coded speech because of their lack of directionality. In their case, the second message (and linguistic practice) that Quaranto talks about is accessible to the “wider audience”. is generally embodied in a structure of two linguistic practices, one that is shared by all speakers and a second one that sends a different message and is only performable in certain contexts. What makes coded speech deniable, then, is the fact that in some contexts both practices are equally deployable, so the speaker can use the first kind of practice (the one that is shared or accepted by most speakers) to deny the second one. In the above case, the anti-Latino speaker uses the illegality frame to deny that her opposition to undocumented citizens is grounded on race, insisting that it has to do with the law, which is a seemingly more plausible motivation than racism. This is facilitated by the racialization of illegality (mentioned in section 1), which makes the topic of illegality deployable in approximately the same contexts as the topic of Latin-American presence. This correlation is on the basis of the stereotypical racial belief of the form S mentioned in section 2.¹⁸ 18 Khoo opts to discuss this issue in terms of social meaning (2017:56-59), which seems to be a promising line of thought, but one I will not have the proper space to discuss here.

Finally, deniability in strategic conversations - as the hypothetical example above - twists our common expectations about the speaker’s conversational liability: the speaker places herself in the position not to be made responsible for the inferred meaning of ‘illegals’ but rather exclusively for its literal meaning. This kind of move is atypical in paradigmatic conversations in which the speaker is expected to acknowledge what she said and respond for it if confronted.

3. (Re)framing in conversations

Paradigmatic conversations have two (or more) cooperative speakers of a natural language taking successive turns in the roles of the sender and the receiver of information. Speakers are generally considered cooperative when they adhere to the Cooperative Principle (Grice 1991 [1975]Grice, Paul. 1991. Studies in the Way of Words. First Harvard University Press paperback edition. Cambridge, MA: Harvard University Press.) and the maxims that are derived from it. In other words, when their contributions to the interaction follow certain rules regarding truth, relevance, non-redundancy and conspicuousness that make them rational and adequate in the stage of the conversation in which they happen (:26-31).

Also, cooperative speakers in paradigmatic conversations share a common ground of assumptions whose normative function is to regulate what contributions are acceptable or not. Those following Lewis (1979Lewis D (1979) Scorekeeping in a language game. In: Bauerle R, Egli U, von Stechow A (eds) Semantics from Different Points of View. Springer, Berlin, p 172-187, https://doi.org/10.1007/978-3-642-67458-712
https://doi.org/10.1007/978-3-642-67458-... ), for example, resort to the construct of the conversational score to explain how this regulation takes place. Basically, the score determines which moves (e.g., speech acts) participants are permitted to perform depending on what presuppositions are in the context. What Lewis calls the rule of accommodation prescribes that presuppositions must be added to the conversational score to charitably accommodate new moves. So, if a participant asserts ‘Smith stopped smoking’, following the rule of accommodation, the other participants will add the presupposition (P) ‘Smith used to smoke’ to the common ground (Stalnaker 2002Stalnaker, Robert (2002): “Common Ground,” Linguistics and Philosophy 25: 701-21.; 2014Stalnaker, Robert (2014): Context (Oxford: Oxford University Press).).

Now, in the previous section, I focused on one kind of non-paradigmatic conversation, namely, strategic conversations, in which the speaker crafts her contribution in a way that avoids risk. I also mentioned the hypothetical example of a bigot with anti-Latino feelings that uses ‘illegals’ to make a racial appeal while avoiding explicit communication about race: call it “Case B”. What is distinctively non-cooperative about this strategic speaker is the fact that she places herself in the position not to take responsibility for what her utterances contribute to the context. To explain this a bit more, I will borrow a distinction proposed by Camp (2018Camp, E. (2018). Insinuation, common ground. In Fogal, Daniel, Daniel W. Harris, and Matt Moss, eds. New work on speech acts. Oxford University Press, 40-66.) for the case of insinuation.

First of all, Camp follows Stalnaker’s distinction between mutual belief and mutual acceptance in defining common ground. According to both authors, the common ground is based on the mutually manifest acceptance of new contributions, not on participants sharing mutually manifest beliefs - which would be a stronger requirement. After all, participants may actualize the common ground for purposes like conjecture, pretense etc., which do not necessarily involve having belief states. Secondly, Camp in particular assumes that there is a difference between what participants mutually assume, and what they acknowledge as assumed in a context - for example, if confronted. Consider Case B again: the speaker conveys a racial message, namely (6), that is manifest to all participants - provided that they are linguistically and culturally competent with ‘illegals’ and accept (1) -, but she is also in a position not to acknowledge having issued a racial appeal through (6), because of the availability of the practices supported by the Illegality frame. In other words, she can manipulate her own conversational liability.

This kind of strategic maneuver concerning liability changes the regular dynamics of acceptance and rejection, since a recalcitrant hearer that rejects (1) - e.g., on the grounds of racism - may have her challenge neutralized by the speaker’s denial, played as a felicitous counter-move. In fact, we do often see counter-moves of this sort (i.e., that try to neutralize direct challenges). Take the example of the public exchange between journalist Tom Llamas and Donald Trump in a press conference during the 2015 GOP race (Smith 2015Smith, C. (2015) Jed Bush, Donald Trump unapologetic for “anchor baby’ language. [Online] [accessed 20/12/2022] [Online] [accessed 20/12/2022] https://abcnews.go.com/Politics/jeb-bush-donald-trump-unapologetic-anchor-baby-language/story?id=33212138
https://abcnews.go.com/Politics/jeb-bush... ), in which Llamas challenged Trump’s use of the IFWC, ‘anchor baby’. The dialogue went like this:

This was not a particularly successful case of conversational challenge - Trump, a rhetorically skillful speaker, uses Llamas’ intervention to introduce “political correctness”, which is an appealing topic for his supporters -, but it showcases both how direct challenges often go and the types of replies that are available for challenged speakers. In particular, Llamas’ part in (8) is a case of calling-out challenge: the purpose of the recalcitrant hearer is to draw critical attention to an inadequate linguistic practice.

As a matter of fact, drawing on Lewis’ rule of accommodation, a number of authors have discussed the social dangers of not challenging instances of toxic and depreciative speech - like Trump’s in (8). Recall that the most important consequence of the rule of accommodation was that each move registered in the score was regulative of future practices. These authors assume then that when the use of racist language is not challenged and ends up being added to the common ground, it becomes a permitted and legitimate practice.¹⁹ 19 See McGowan 2012 for more about the effects of conversational acceptability in the case of racist speech. In this literature, direct challenges and other forms of direct intervention are presented as important tools to help prevent the legitimization of racist and intolerant conducts²⁰ 20 See Cepollaro et al 2023 for more on counterspeech. . For example, Tirrell (2012Tirrell, Lynn. 2012. Genocidal Language Games. In Maitra, I. and McGowan, M.K., Speech and Harm: Controversies Over Free Speech, Oxford: Oxford University Press, 174-221., 2017Tirrell, L. (2017). Toxic speech: Toward an epidemiology of discursive harm. Philosophical topics, 45(2), 139-162.) analyzed how the use of derogatory language led to genocidal practices in the case of Rwanda in the 90s, pointing out the relevance of countering these uses as they happen. Langton (2018Langton, R. (2018). Blocking as counter-speech. New work on speech acts, 144, 156.) highlighted that blocking is a form of counterspeech and Haslanger (2014Haslanger, S. (2014). The normal, the natural and the good: Generics and ideology. Politica & Società, 3(3), 365-392.) has held that metalinguistic negations²¹ 21 This is a concept she borrows from Horn (1985) that refers to a form of negation that blocks a pragmatically implicated falsehood from entering the common ground. Example: “she is not a lady, she is woman”. are the best way to block the pernicious effects of, for example, striking properties generics.²² 22 Generics that attribute harmful or dangerous properties to a kind, like “Muslims are terrorists”.

For expressions like ‘illegals’/‘illegal immigrant’, Khoo suggests that conversational challenges should be formulated, first, as calling-out challenges that bring to light the implicit mechanisms which enable it to operate as a racial code word; and, secondly, in non-personalistic terms, as in (9) below.

In highlighting that the inferred meanings of racial code words do not depend on the speaker’s communicative intentions, the recalcitrant hearer will block turn-table replies that deny intentionality, like (10).

I take Khoo’s suggestion to be correct. But I wish to focus on the alternative strategy of reframing here: that is, giving responses that introduce new words (which invoke alternative frames) to the context without confrontation. My motivation is the fact that conversational challenges break with the default dynamics of cooperation. In so doing, they may invite resistance, and often for reasons other than genuine conviction, like confusion and/or lack of self-knowledge about one’s implicit attitudes (Saul 2017Saul J (2017) Racial figleaves, the shifting boundaries of the permissible, and the rise of Donald Trump. Philosophical Topics 45(2):97-116. https://doi.org/10.5840/philtopics201745215
https://doi.org/10.5840/philtopics201745... ). This tends to compromise the ultimate goal of the rejection move. In reframing, on the other hand, the recalcitrant speaker indicates her rejection of the utterance without confrontation and, more importantly, with the additional benefit of exploring the process of framing to gain control over the goals of the conversation.

3.1. Reframing and QUD-shifting

Reframing is a fairly common practice in our everyday transactions and disputes, especially when the goal is to make one’s position more appealing to an opposing or resistant party. In the context of politics, reframing appears in discussions about polarization, mainly regarding how consensus can be reached over divisive issues (Lakoff 2010Lakoff, G. (2010). Why it matters how we frame the environment. Environmental communication, 4(1), 70-81.; Feinberg and Willer 2015Feinberg, M., & Willer, R. (2015). From gulf to bridge: When do moral arguments facilitate political influence?. Personality and Social Psychology Bulletin, 41(12), 1665-1681.). One example is the study conducted by Feinberg and Willer, which showed that, in the American dispute between liberals and conservatives, agents who used moral reframing - i.e., the framing of arguments in ways that targeted the other party’s moral values - to convince opposers generated more agreement than those that simply restated their own positions and moral values.

What I call reframing here, however, is not the process of entering into the worldview of an adversary. Rather, it is a more localized, conversational maneuver: a kind of reply that proposes to replace a problematic term with a new one without calling the speaker out.²⁴ 24 For Lakoff (2010), a really successful process of building new frames requires a large-scale cognitive policy, that is, changing not just people’s language, but people’s brains (frames are realized in neural circuits, in Lakoff’s view). This involves not using or repeating existing problematic frames. I focus only on reframing in conversational contexts, so the cognitive - as well as the mediatic and communicative - dimension(s) of (re)framing are not discussed here. For that, see respectively Lakoff 2014 and Ryan & Gamson 2006. For example, (11) below uttered in the context of Case B - i.e., in response to (1):

In such a case, the recalcitrant speaker tacitly indicates the rejection of the term ‘illegals’ by uttering an alternative, with the expectation that it is accommodated to the common ground. This is a relatively less “noisy” rejection than (9), because the recalcitrant hearer explores the mechanics of accommodation to transfer the onus of breaking with the default dynamics of cooperation back to the speaker.²⁶ 26 Of course, this attempt can backfire. The speaker may resist (11) and either challenge the recalcitrant hearer back (and start a meta-discussion) or simply ignore the attempt. Calculating whether reframing is the adequate strategy will be highly contextual and will involve predicting the interlocutor’s behavior. However, what really makes this strategy more interesting than direct challenges in some contexts is the fact that the accommodation of a reframing utterance like (11) shifts the conversational goals or what Roberts (2012Roberts, C. (2012). Information structure: Towards an integrated formal theory of pragmatics. Semantics and pragmatics, 5, 6-1.) and others call question under discussion (QUD) in the context.

Roberts, in the spirit of Lewis’ ideas, takes conversations as game-like cooperative endeavors whose scores are organized around what interlocutors take to be the question under discussion, i.e., what the discourse is “about” or, better still, the question which represents the most immediate aim of the discourse at a time t. Rational cooperative interlocutors know that their goal with every new speech act will be to address the QUD in the context. Relevance²⁷ 27 Relevance is an organizing principle which supports coherence and also processing. It ensures that all participants attempt to answer the question as soon as possible. is pivotal here: a new move will be considered relevant if it contributes to the strategy of inquiry of the game, i.e., the cooperative plan for the resolution of the accepted QUD. Let us take an example. If at a moment t, the immediate QUD of a given conversation is “who used to smoke?”, the intention to answer it becomes part of the common ground and a strategy of inquiry is established, with a stack of sub-questions and sub-goals. Since the QUD denotes a set of alternatives (partial and complete answers) and assertions are the typical payoff moves, if Smith is part of the domain, (2) will count as a good move, and its content will be incorporated to the common ground.

Importantly, the QUD framework purports to capture the ways in which conversations normally evolve, with participants raising QUDs implicitly by means of new utterances. In such cases, the other participants will rely on cues, such as previous speech acts,²⁸ 28 Another cue might be intonational focus (Beaver & Clark 2008). I believe that, in the case of (11), the fact that the hearer responds with ‘Yes’, but changes the wording (while still addressing the previous QUD) might be interpreted as a cue for QUD-shifting. to single out which QUD, from among a set of salient options, better fits the new contribution. For example, an utterance of (2) discloses at least three questions as relevant candidates to the role of the immediate QUD: (a) ‘did Robert use to smoke?’; (b) ‘what did Robert use to do?’ and (c) ‘who used to smoke?’.²⁹ 29 Question (a) is a polar question, which means that it partitions the space of possibilities in two: worlds compatible with Yes answers and worlds compatible with No answers. Questions (b) and (c) are wh-questions, which means that they partition the space of possibilities is multiple cells. Participants will then use the score record (or other cues) to determine which one should be accommodated, thus generating a QUD-shifting.

My suggestion is that reframing utterances give recalcitrant hearers some control over the process of QUD-shifting. Suppose that Case B is a conversation which, at a moment t, is about whether the U.S. is threatened by massive immigration. We have two participants, D, the person with anti-Latino sentiments that does not want to be perceived as a racist, and L, a person with no anti-Latino sentiments.

Assume that, at a time after t, the sub-question at the top of the stack of questions in the context is the polar question ‘are there too many immigrants in the U.S. today?’. The utterance of (1) will be both a felicitous move and a complete answer to it, causing the sub-question to be removed from the stack. In furthering the goals of the game, the QUD may shift. Only now the strategy of inquiry includes the characterization of immigrants as illegals. Since the rule of relevance is a coherence principle, this update authorizes new sub-questions that are about illegality, law enforcement etc.

What the utterance of (11) does is precisely to prevent that the conversation unfolds in this way. On the one hand, this results in a localized process of topic-elimination: the removal of the topic (and the frame it evokes) from the context³⁰ 30 I borrow this terminology from Picazo (2022), but she is more interested in a more enduring process of topic-elimination that can produce communicative harm for excluding topics from public deliberation. Here, I am interested in a more localized discursive phenomenon. . On the other hand, in uttering (11), speaker L proposes a new topic that opens a new strategy of inquiry around sub-questions about, for example, immigration bureaucratic status - provided, of course, that (11) is accommodated. The common goal of the participants can now be about citizenship for immigrants, humanitarian action etc. Any attempt to reverse the conversation back to the subject of illegality and law enforcement will be costly to the other speaker³¹ 31 Reframing is not enough to completely rule out the illegality frame, since speaker D can still try to revert to the previous topic. Nevertheless, it affords speaker L additional control (via the rule of relevance) over the costs, for speaker D, of being less than cooperative. .

Now, it can be objected that (11) does not work as well as (9) as a rejection of (1), because it still makes a racial appeal: most speakers that are linguistically and culturally competent with the immigration debate also have a stereotypical belief of the form S in which C is undocumented citizen and R is Latin-American. This objection, however, is less problematic than it seems. First, because I am not advocating that reframing should replace direct challenges altogether. In some contexts of ethically problematic speech, the correct strategy will still be to defy the interlocutor’s choice of words and worldview directly. In particular, calling-out challenges like (9) will be an effective option to prevent follow-ups that contain personalistic denials. But in other contexts, in which a recalcitrant hearer feels compelled to reject a contribution, but it is more productive (given practical limitations) to promptly reframe what she deems problematic, as in political debates or in social situations that inhibit direct confrontation, reframing is an option that affords additional control.

Secondly, because even if the term ‘undocumented citizen’ is also stereotypically associated to Latin-Americans, this specific use of (11) as a reply to (1) does not aim to explore a racial inference to manipulate the utterer’s conversational liability in order to avoid imputations of racism. On the contrary, (11) purports to block the mechanisms of deniability that are enabled by the illegality frame and explored by the manipulative speaker to convey a racist message.

Another objection would be to question whether exploring the mechanisms of QUD-shifting in the way proposed here is not itself a form of manipulative communication. In particular, recent works, like Keiser (2022Keiser, J. (2021). The “All Lives Matter” response: QUD-shifting as epistemic injustice. Synthese, 199(3-4), 8465-8483.) and Picazo (2022Picazo, C. (2022). Distorted Debates. Topoi, 1-11.), have focused on how QUD-shifting can be used to promote hermeneutic injustice and discursive distortions in the ‘Black Lives matter’ (BLM) versus ‘All lives matter’ (ALM) debate, for example. According to Keiser, activists of the ALM movement promote epistemic injustice because they shift the QUD of the debate mid-discourse from the polar question ‘do black lives matter?’ to the wh-question ‘which lives matter?’. This obscures the intended meaning of the BLM movement and disincentivizes future uses of the BLM slogan, which, on its turn, contributes to the perpetuation of racial injustice. Picazo, on the other hand, argues that what she calls retroactive distortion, i.e., the distortion of the speaker’s meaning by a subsequent speech act made by another speaker, can eliminate important topics from public deliberation. This generates harmful effects, like the erosion of democratic ideals.

Though this has been the dominant tendency in the recent literature and one with great merits, I see no reason to oppose the idea that the same mechanisms of QUD-shifting and topic elimination can be explored to promote communicative justice. They can be used to eliminate topics and practices that potentially erode democratic ideals, such as the practice of making implicit appeals that feed on racism and bigotry for mere political gain. Whether these mechanisms are explored in unwholesome ways depends more on who uses them and for what purposes than on the mechanisms themselves.³² 32 This point is acknowledged by Picazo in her discussion of Langton’s concept of accommodation blocking as counter-speech.

4. Conclusion

I focused here on how the deniability of IFCWs relates to (political) framing and how IFCWs can be countered in conversational contexts via QUD-shifting. As we saw, deniability is what makes IFCWs fit to work as conversational devices that aid dogwhistle politicians, for example, to veiledly discriminate against Latina/os and harness anti-Latino support. The influence of the history of immigration frames here is indirect, but still present. In the best-case scenario, politicians put forward their agendas (in campaign slogans, reforms etc.), making choices of how to refer to certain groups and the result can be coopted by flawed ideologies. As I mentioned in the introduction, in the case of immigration, the process of racialization of illegality from the last decades seems to have permitted this cooptation by anti-Latino bigots. Due attention to how framing affords discursive control helps then, first, to encourage politicians and the civil society to have more critical interest for how agendas are presented. Secondly, it offers some insight as to how a recalcitrant agent can interfere in the process and change its course to help prevent potential communicative harm. Here, I suggested that reframing is one way to achieve that by exploring the rule of relevance to eliminate problematic topics from conversations.

References

Publication Dates

History