In the United States, police interrogators are often trained to prioritize getting a custodial suspect’s statements (Reference Inbau, Reid, Buckley and JayneInbau et al., 2001). What maximizes a police interrogator’s strategy is to get incriminating statements from suspects. Now, the law requires that statements be voluntary, knowing, and freely provided, hence police interrogators must read suspects the Miranda rights prior to the onset of custodial interrogation and may not overtly or coercively get suspects to waive them. This chapter challenges the notion that custodial suspects freely and voluntarily waive their Miranda rights. In a police interrogation, perception and preference, such as invoking or not invoking the right to counsel, are subject to manipulation. Manipulation is the process by which an interrogator may alter a custodial suspect’s preferences for the benefit of the interrogator.

This chapter provides the theoretical framework for suspects’ invocations for counsel in the corpus. The analysis will focus on the discursive features of suspects’ invocations for counsel during custodial interrogation, as well as the sequences of talk that follow the suspects’ invocations, including the police interrogators’ use of requests for clarification. This part of the analysis will apply game theory, specifically hypergame theory (Reference BennettBennett, 1980), to describe police-suspect exchanges after a custodial suspect invokes counsel. This is a novel approach to analyzing institutional discourse and, in particular, police interrogations.

4.1 Invoking Counsel: A Constitutional Right Handicapped by the Law

There is a presumption in the law that suspects who wish the assistance of counsel must invoke so unequivocally. This creates a distinction between a valid invocation for counsel, that is, one that can stop an interrogation, and an ambiguous one that may not prompt such action from the police. The validity of the invocation is, for the most part, dependent on the police’s assessment of the invocation. Hence, police interrogators may end an interrogation at any point when the suspect invokes their rights, albeit directly or indirectly – with indirectly formulated invocations being the most common and least successful at invoking rights (see Reference MasonMason, 2013, Reference Mason, Mason and Rock2020). The manner in which police approach the invocation of rights stage of an interrogation is often guided by the law and training, and strategically framed, rather than by discursive frameworks.

4.2 Communication As a Type of Game

Linguists have employed game theory to understand how speakers strategically convey meaning and intent in communication. Players have preferences for outcomes that motivate strategic linguistic choices made during the game. Theoretically, in communication as in other games, a player cannot control the outcome of the game just by one’s actions. Instead, a player must take the other player’s linguistic actions into account. This process has traditionally been studied in linguistics under rationalistic game theory (see Reference van Rooijvan Rooij, 2008). Rationalistic game theory examines “when is it rationalizable to communicate at all” and “how can credibility of messages be established” (Reference JägerJäger, 2008, p. 410).

Some common approaches to the credibility question include the application of costly signaling games and cheap talk. Reference RabinRabin (1990), for example, takes a game theoretic approach that is reminiscent of Reference Grice, Cole and MorganGrice’s (1975) cooperative principles that all messages are credible assuming cooperativity. Rabin developed Credible Message Rationalizability (CMR) by combining Reference Crawford and SobelCrawford and Sobel’s (1982) standard equilibrium analysis and Reference Farrell and GibbonsFarrell and Gibbon’s (1989) rich language assumption that players are rational. Players tend to believe statements that are consistent with rationality, allowing for some meaningful communication to be predicted, including those that require implicatureFootnote ³ to derive the players’ intent.

Other approaches predicting communication include Reference BlutnerBlutner’s (2000) Bidirectional Optimality Theory and Reference ParikhParikh’s (1991) Strategic Discourse Model (SDM). Reference Dekker and Van RooyDekker and van Rooij (2000) reformulated Blutner’s theory as the exchange of information between sender and receiver characterized as a strategic signaling game of complete information. The actions of speakers are described as the choice of expressions and the actions of the hearers as the choice of interpretations. These models examine cooperative exchanges, focusing on form-meaning relationships.

Reference ParikhParikh (1991, Reference Parikh2000, Reference Parikh2001) expanded on the analysis of sender and receiver exchanges by adding the sequential features of a communication game of partial information with a unique solution. Reference ParikhParikh (1991) argues that most languages are “situated” which makes language an efficient system of communication: “Once we allow situations a role in the determination of content, it becomes clear that there are certain aspects of utterances that are constant across utterances and others that vary from one situation to another” (p. 475). He notes that part of the task of a theory of communication is to explain how we can get from meaning to content. His model of strategic communication under conditions of ambiguity introduced the concept of “strategic inference” in communication. The central assumption of Parikh’s Strategic Discourse Model (SDM) is that the transfer of information between players involves a strategic interaction between them. When the strategic interaction is common knowledge between the players, that is, when it is a game with a unique solution, the flow will be communicative. Parikh notes “that this game-theoretic structure is situated somewhere, either in the ‘minds’ of players or in their ambient circumstances or both” (p. 478).

Parikh’s SDM, as other models previously discussed, describes how players reach cooperative, meaningful communication, including situations where ambiguity may arise. These approaches, however, do not address noncooperative situations, where players may not have complete information of the signaling strategies of other players (see Reference RabinRabin, 1990; Reference FarrellFarrell, 1993; Reference Asher and LascaridesAsher and Lascarides, 2013). As Reference Pinker, Nowak and LeePinker et al. (2008) note, models of conversation that only assume cooperative agents fail to explain the dynamics of strategic conversations.

Discursive games have also been modeled through game-theoretic constructs such as Bayesian games (Reference Frank and GoodmanFrank and Goodman, 2012; Reference Stevens, Benz, Reuße and KlabundeStevens et al., 2015; Reference BurnettBurnett, 2019; Reference HarsanyiHarsanyi, 1967). In Bayesian games, an agent’s utility is determined not only by actions but also by types of the agents. In these types of games, “the existence of private information leads naturally to attempts by informed parties to communicate (or mislead) and to attempts by uninformed parties to learn and respond” (Reference GibbonsGibbons, 1997, p. 138).

In the analysis of reference, Reference StevensStevens (2016) discusses, through a Bayesian and cognitive lens, whether strategic and nonstrategic thinking play independent roles in explaining pragmatic behavior. Stevens notes that a model that explains interaction must account for why people think strategically and, alternatively, why they may not. Stevens argues that the failure of agents to work out the best strategy for themselves may be rooted in reconciling strategic considerations with situational or cognitive biases. Stevens adds that pragmatic behavior does not always tilt toward strategic optimality and, hence, may only provide partial explanations for complex (linguistic) situations where the choices of strategy may not be obvious.

Rational choice theory may not provide a framework for accounting for an agent’s variances in preferences owing to misperceptions. According to Reference Tversky and KahnemanTversky and Kahneman (1986), the normative analysis in game theory makes four assumptions to develop expected utility theory: cancellation, transitivity, dominance, and invariance. Cancellation eliminates trivial outcomes in which action does not produce different results. Transitivity has a standard mathematical meaning (A≻B, B≻C → A≻C) in which A is preferred (≻) to B, and B is preferred to C, transitivity implies (→) that A is preferred to C. Dominance postulates that if one option is better than another in one state and as good as other options in all other states, it is preferred.

Invariance, which assumes that different representations of the same choice problem should yield the same preference, creates a problem for rational choice theory. Because invariance and dominance are normatively essential, Tversky and Kahneman argue rational choice theory cannot provide a description of choice if an agent shifts their preference unless a fundamental change in perception arises. They find that the normative axioms of rational choice are often violated in situations lacking transparency. To illustrate this point, Tversky and Kahneman show that individuals respond differently to alternative frames of a risky choice, first as the prospect of gaining utility, and second as the prospect of minimizing loss. Even though the alternate statements yield the exact same utility, individuals in their study made different choices depending on how the choice was phrased.

Reference Cason and PlottCason and Plott (2014) assert that the absence of an integrated theory of perception masks the existence of a specific form of misperception: the “failure of game form recognition” (p. 1236). Suffering this type of misperception causes a decision maker to fail “to recognize the proper connections between the acts available for choice and the consequences of choice, which, by necessity, are associated with the method of measuring preferences” (p. 1237). They warn that errors in utility maximization, owing to misperceptions of the game form, lose a separate identity when mistakes are considered generically as an error term, as do errors arising from misinformation, opaque rules, and misperceptions of a situated environment.

Parting from framing theory, the analysis of an unbalanced game model, such as a police interrogation, allows for the examination of differences in each player’s understating of the game and their choices. In these types of games, a player may not be able to recognize the other player’s manipulation of their preference to become that of the other player. Excerpt 4.4 shows a representative excerpt from the corpus that illustrates linguistically how the custodial suspect’s, $CS$’s, preferences are guided by the interrogator’s, $I$’s, discursive actions.

The interrogation continued with the officer vaguely discussing the crime and possible charges. Information is not forthcoming in police interrogations as a matter of training and strategy. The discussion of the attorney resumed to establish that Rodriguez did not want counsel and he waived his rights.

The discursive analysis of Excerpt 4.4 begins with $I$ reading $CS$ the Miranda rights. The suspect affirms understanding of the rights and in turn 11 proceeds to invoke the right to counsel: “Can I speak to an attorney?” The invocation is phrased using a hedged indirect request, which may have linguistic moorings in politeness (Reference Brown and LevinsonBrown & Levison, 1987), face (Reference GoffmanGoffman, 1967), risk aversion, as well as in cognitive frameworks and sociolinguistics (see Reference GibbsGibbs, 1983; Reference HoltgravesHoltgraves, 1994; Reference BrooksBrooks, 2000; Reference Holtgraves and AshleyHoltgraves & Ashley, 2001). It may also be natural in the context of asymmetric bargaining and exchanges. Regardless of the explanation, indirectly invoking the right to counsel shows $CS$’s preference to invoke counsel and do so in a manner that fits $CS$’s understanding of formulating requests in conversation, particularly asymmetric ones (Reference MasonMason, 2013).Footnote ⁴

The use of an indirect request, in light of the Davis ruling, places $CS$’s ability to invoke counsel in the hands of $I$. As noted previously, $I$ is not required by law to inform $CS$ that the invocation does not rise to the level of unequivocalness to end an interrogation. $I$, however, cannot coerce $CS$ into not invoking rights. Statements must be provided voluntarily or have the appearance as such. This creates a situation in which $I$ does not provide information to $CS$, but has to appear to address $CS$’s indirect request.

The next turns of talk (turns 12–15) constitute $CS$’s repeated attempts at invoking counsel and restating his preference, which is met with $I$’s lack of acknowledgment of such attempts. Of note, follow-up moves are commonly used in conversation to clarify, repair, and/or respond to a prior turn of talk. $I$’s “whatever you want” or “you tell me what you want” does not address directly $CS$’s request for counsel or further $CS$’s ability to invoke counsel. $CS$, hence, continues to assert his preference for counsel and repeatedly invokes counsel in turns 11, 13, and 15, albeit successfully only in turn 15.

In turn 16, $I$ changes his follow-up response to: “That’s fine bro we stop because we can’t talk to you anymore, okay, so,” which is the first follow-up to address the outcome of $CS$ invoking counsel (“we can’t talk to you anymore”), but without overtly acknowledging $CS$’s invocation. In turn 17, $I$ interjects with: “You’re going to be charged with murder today.” This statement leads $CS$ to inquire “why?,” which brings $CS$ back to the interrogation table and away from his initial, and sustained, preference to request counsel. The talk continues and eventually $CS$ waives his rights and provides incriminating statements. This is $I$’s preferred outcome based on his training: to get a statement from $CS$.

In Excerpt 4.4, $CS$’s continued attempts to invoke counsel indirectly, coupled with $I$’s strategic use of follow-up moves, serve to misdirect $CS$ from invoking counsel. The ultimate success of the strategy shows $CS$ exhibiting what Reference Gharesifard and CortésGharesifard and Cortés (2010, p. 1077) note is an “imprecise understanding about the objectives and true intentions of other players.” $CS$ attempts to invoke counsel repeatedly, yet $CS$ does not question the follow-ups by $I$ that do not address his attempts at invoking counsel, nor does $CS$ change strategy to use another type of request for counsel (e.g., invoke directly). These options, otherwise available in talk, are not pursued by $CS$ because $CS$ does not recognize that $CS$’s utterances were deemed equivocal and $I$ does not have to accept (by law) these utterances as invocations for counsel, regardless of the reading of the Miranda rights. $CS$ continues to participate as if the conversation abides by commonality of understanding where indirectly formulated requests are understood as unambiguous requests.

On $I$’s third attempt at invoking counsel, $I$ indirectly acknowledges $CS$’s invocation by stating the consequences of the decision. However, $I$ follows up by telling $CS$ “You’re going to be charged with murder today.” This “distracts” $CS$ into engaging with $I$ as a cooperative conversational participant. $CS$ is lured back into conversation because $CS$ believes $I$ will share information about why $CS$ will be charged in the case. The information is not forthcoming, but it did return $CS$ to the invocation stage of the interrogation, and, in this case, $CS$ waived the right to legal counsel and agreed to talk with the interrogator.

Incorporating perceptions into the analysis of discursive strategies prompts the question of whether Bayesian analysis, commonly used in linguistics, or an alternate model should be considered to describe the actions observed in Excerpt 4.4. Bayesian analysis is based on rational utility maximization measured as the sum of the subjectively assessed utility of each outcome weighted by the probability of its occurrence. The surface game, depicted in Figure 4.1 (Reference Mason and MasonMason & Mason, 2021), is a simplified Bayesian model in which $CS$ has 100 percent certainty of obtaining either outcome: choosing to talk to an interrogator or invoking the right to counsel.

Figure 4.1 The surface game

If $CS$ chooses to invoke the right to legal counsel, shown as the arrow pointing to the lower circle on the right, whenever counsel arrives the invocation game becomes the interrogation game, and all future decisions will be generated from this lower node. $I$ has a preference that $CS$ chooses the other path, and $I$ has strategies to shift $CS\prime s$ preference.

The surface game is the game as presented to the $CS$. However, as observed in the corpus (see Excerpt 4.4), something more than the surface game takes place. $I$ has institutional powers that grant $I$ a choice of outcome, the linguistic and procedural response in conversation with $CS$, to shape $CS$’s perceptions. These powers include $I$’s ability to limit $CS$’s discursive options to express preference; to control new information to influence $CS$; and to impose the least preferred outcome on $CS$. This is accomplished through $I$’s initiated follow-up moves. The observed shifts in $CS$’s responses, from invoking counsel to talking to an interrogator without counsel, may not be the product of individual free will, as cited in numerous legal cases, but the product of $CS$’s “failure of game form recognition” that accommodates misperceptions arising from other psychological, discursive, and deceptive influences.

The nuances that institutional power brings to game theory, as summarized by Reference PratesPrates (2014), incorporates into game theory the use of power by one agent to alter the options of another agent:

Prates points out what is essentially applicable to any successful hypergame employed by $I$ and embedded within an invocation game: preferences shift. The shift in preferences can be thought of as new conditions that must be met for $CS$ to invoke a particular action because the environment changes and not the assessed risk.

In Figure 4.1, the switch from the lower node to the higher node (i.e., from invoking the right to counsel to waiving that right) entails a different risk environment for $CS$. As Prates notes: new game, new player, new standards. In an extensive form rendition of the game, subsequent action switches from one pathway to another. The shift creates a new game outside of Bayesian sequential analysis. The hypergame creates the shift observed in Excerpt 4.4.

This chapter proposes an alternative to rational choice theory to investigate why discursive preferences shift in real world applications, such as police interrogations (see also Reference Mason and MasonMason & Mason, 2023). The analysis this approach produces will provide insights into the effect of institutional power in framing “the invocation game” as a type of hypergame:

4.3 Hypergame Dynamics

Hypergame Theory (HgT) expands traditional game theory and noncooperative game theory (see Reference Fudenberg and TiroleFudenberg & Tirole, 1991) by incorporating players’ perceptions of other players’ preferences into their strategic action (see Reference BennettBennett, 1977; Reference Wang, Hipel and FraserWang et al., 1989; Reference Basar and OlsderBaşar & Olsder, 1999). HgT accommodates a player with the power to shape the game while it is being played to exercise that power strategically guided by that player’s perceptions of other player’s preferences. In the invocation game, $I$ controls the discursive exchanges, which confers to $I$ an advantage over $CS$. During the game, $I$ accumulates information on $CS$’s preferences. $I$ uses that information strategically in choosing outcomes to influence $CS$’s perception of the invocation game and to shape the game, while it is being played, to shift $CS$’s preferences to conform with $I$’s. Within the invocation game, each turn of talk produces an ordered pair of ranked preferences for $CS$ and $I$ $\left(P_{CS},P_I\right)$ over $CS$’s choices: (a) waive rights and agree to interrogation, (b) stay silent, (c) engage in talk, and (d) invoke the right to legal counsel. $CS$’s preferences are given initially by nature, are fluid to the situation, and are updated after each turn of $I$’s talk.

For $I$, it is easiest to obtain a statement if no legal counsel is present with the $CS$. An interrogation with a lawyer present is preferable to silence. Talk is preferred to an interrogation with a lawyer present because talk is an opportunity for $I$ to persuade $CS$ to consent to speak without an attorney. It should also be noted that anything $CS$ says can be used against $CS$ in a court of law, making all conversation, by default, in the realm of an interrogation without legal counsel, the most preferred outcome of all.

The legal framework in which police interrogation operates creates two hypergames in which at least one player’s strategy is determined by their perception of the other player’s preferences. The next sections provide the notation and modeling of the hypergames and the application of hypergame modeling to representative excerpts from the corpus.

4.3.1 Hypergame Notation of the Invocation Game

The invocation game adopts Reference Gharesifard and CortesGharesifard and Cortés’s (2014) notation where a game $G=\left(V,S_{outcome},P\right)$, consists of a vector of two players $V=\left\{CS,I\right\}$, an outcome vector $S_{outcome}=S_{CS}$ with cardinality $N=\left|S_{outcome}\right|=2$, where $S_{CS}$ is one of two outcomes arising from $CS$ taking action $X\in \left(x_1\dots x_4\right)$:

Table 4.01

Variable	Description of action
$x_1$	Agree to an interrogation,
$x_2$	Remain silent,
$x_3$	Engage in talk,
$x_4$	Invoke right to counsel

$CS$ will make a choice to invoke or waive, and both $CS$ and $I$ will derive utility from $CS$’s choice generating a preference rank $\left(P\right)$ of outcomes for $CS$ and $I$: $P=\left(P_{CS},P_I\right)$. Preference is measured along a sequence of events.

The surface game is a single move game. The game begins formulaically with the reading of Miranda rights. The choice by $CS$ either to invoke $x_4$ or to waive $x_1$ generates the outcome preference rank $P$. Within the game, two other moves $x_2$ (silence) and $x_3$ (talk) are allowed, but these additional moves occur around $CS$’s choice: $x_1$ or $x_4$. Since both $x_1$ and $x_4$ entail conversation $x_3$, they are subsets of talk: $x_1,x_4\ni x_3$. Given the specific inclusion of $x_1$ and $x_4$ as choices in the game, we distinguish $x_3^{\prime }$ to be talk that excludes invoking and waiving the right to counsel: $x_3^{\prime }=x_3\backslash x_1,x_4$.

I’s preference among outcomes, $P_I=x_1\succ x_3\succ x_4\succ x_2$, is assumed constant by training and will guide $I$’s strategy when engaging with $CS$. $I$ most prefers an interrogation without a lawyer present because the absence of a lawyer makes obtaining a statement from $CS$ easier. $I$ prefers conversation $\left(x_3\right)$ to a decision by $CS$ to invoke the right to a lawyer $\left(x_4\right)$ because it leaves open the possibility of persuading $CS$ to talk without a lawyer, and $I$ least prefers silence as silence does not further $I$’s investigation.

The invocation game situates discourse for $CS$. $CS$ has exactly two powers: the right to remain silent and the right to have a lawyer present for questioning. The power dynamics are evident from what transpires in an interrogation: $CS$ has lost the right to freedom and faces the possibility of imprisonment. $CS$’s preference rank, $P_{CS}={\left(x_1,\dots ,x_4\right)}^T$, is fluid to the situation and is updated after each turn of $I$’s talk $T\in \mathbb{N}\ge 0$. Any $CS$ who prefers $x_2\succ x_3$ will end the game by not engaging in the interrogation, albeit while being detained in jail, or, in some instances, with the release of $CS$ (e.g., at the discretion of $I$).

For those $CS$ willing to talk, $CS$’s preference rankings are given as a state of nature $P_{CS}=x_3\succ x_2$. Therefore, $CS$’s preferences manifest from nature as either $P_{CS}=x_4\succ x_1\succ x_3^{\prime }\succ x_2$ or as $P_{CS}=x_1\succ x_4\succ x_3^{\prime }\succ x_2$. Both $x_1$ and $x_4$ are presumed to be preferred to $x_3\prime$ owing to a preference for relevant conversation over irrelevant conversation, because engaging in talk may lead to providing incriminating statements. Therefore, $x_3^{\prime }$ is held in reserve until $x_3^{\prime }$ is perceived to hold strategic value (e.g., request information or clarification) at which point $CS$ reveals $P_{CS}=x_3^{\prime }\succ x_4\succ x_1\succ x_2$ or $P_{CS}=x_3^{\prime }\succ x_1\succ x_4\succ x_2$. Once $x_3^{\prime }$ loses strategic value, preferences revert to their previous state unless the discourse altered $CS$’s preference ranking.

4.4 Two Hypergames

4.4.1 Type 1 Hypergame: CS Initiates ${\mathit{x}}_{\mathit{3}}^{\prime }$

Type 1 hypergame originates after $CS$ initiates conversation $x_3^{\prime }$. The strategic value of $x_3^{\prime }$ for $CS$ ranges from human need (e.g., a desire for water) to probing for information. The success or failure of $CS$’s initiative depends on the willingness of $I$ to grant such success. Leaving aside human rights violations (e.g., denying requested water), $I$ has the ability to shift the topic of $x_3^{\prime }$. This topic shift is the type 1 hypergame depicted in Figure 4.2 (Reference Mason and MasonMason & Mason, 2021). In this subgame, the shifted conversation continues with the participants denoted as $C{S}^{*}$ and $I^{*}$ until either $I$ reposes the Miranda question, or $CS$ declares $x_1$ or attempts to invoke $x_4$ which returns the conversation to the $CS$ decision node.

Figure 4.2 Type I Hypergame: $CS$ initiates $x_3^{\prime }$

4.4.2 Type 2 Hypergame: I Initiates ${\mathit{x}}_{\mathit{3}}^{\prime }$

In the type 2 hypergame, $CS$ makes an indirectly formulated assertion of $x_4$, and $I$ initiates conversation $x_3^{\prime }$ to divert $CS$ away from $x_4$. The conversation continues until $CS$ asserts $x_1$ or $x_4$, or until $I$ becomes convinced $CS$’s position has changed at which point $I$ reposes the Miranda question. The type 2 hypergame is depicted in Figure 4.3 (Reference Mason and MasonMason & Mason, 2021).

Figure 4.3 Type 2 Hypergame: $I$ initiates $x_3^{\prime }$ to divert $CS$ away from $x_4^{\prime }$

4.5 Analyzing the Invocation Game and Its Hypergames

The two hypergames illustrated in Figures 4.2 and 4.3 show the assertion of power in which $I$ takes footingFootnote ⁵ from $CS$ to influence the outcome of the game by controlling the topic of conversation. In a noncooperative asymmetric exchange, the person who controls the talk may direct the exchanges without regard for mutuality of understanding. In the invocation game, $I$, guided by perceptions of what motivates $CS$, works to influence $CS$ to $I$’s desired preference. $CS$ will be handicapped from achieving discursive goals if $CS$ does not recognize how the talk is situated. $I$ protects $CS$’s misunderstanding instead of eliminating it. An equivocal/ambiguous invocation for counsel allows $I$ to ask follow-up questions aimed at “clarifying” the invocation. $I$ is not required to take any action, such as initiating the process $x_4$, making the follow-up an empty rhetorical clarification. The analysis of Excerpt 4.5, previously cited as Excerpt 4.2, will show how $I$ manipulates $CS$’s preferences through a series of commonly observed extensive moves, with multiple iterations, to match $I$’s own.

The surface game commences at the end of turn 1. After listening to the Miranda rights, $CS$ asserts $x_4$ in turn 2 revealing $P_{CS}=x_4\succ x_1\succ x_3^{\prime }\succ x_2$. In turn 3, $I$ initiates a type 2 hypergame with $x_3^{\prime }$ in which $I$ provides unsolicited information. Viewed as a cooperative move, $I$’s “clarification” implies that $CS$ can choose $x_1$ and assert $x_4$ “at any time” implying $x_4$ is a subset of $x_1:x_4∍x_1$. This “clarification” reduces the perceived cost of $x_1$ because, by implication, $x_4$ is already a part of $x_1$.

$CS$ responds to $I$’s type 2 hypergame in turn 4 as in turn 2 except $CS$ couches this assertion of $x_4$ by first presenting a cooperative ethos. The effect is a short apology made prior to exercising self-interest. This statement should end the game for a second time, but in turn 5 $I$ engages the type 2 hypergame a second time by presenting $x_3^{\prime }$ that appears at first as a summary but changes into what appears to be a clarifying question. Almost word for word, $CS$ repeats the preference expressed in turn 4 and now again in turn 6, and once more the invocation game should end.

In the third iteration of the type 2 hypergame, $I$ shifts strategy to employ an ominous tone. After an initial misstep in turn 7, $I$ returns the conversation directly to $CS$’s decision node by presenting alternative $x_1$ with its procedure of signing a waiver, or, if $CS$ does not choose $x_1$, an outcome with no certainty and an unknown procedure which will be no better than $x_2$. $I$ concludes turn 7 by asking “ok?” Here “ok” is a request for $CS$ to confirm that $CS$ understands $CS$’s choices which do not explicitly include $x_4$.

$I$ has reshaped $CS$’s perception of the game. In turn 6, $CS$ asserts $x_4$ with conviction, just as in turn 4. In turn 8, $CS$ reveals that the perceived value of the alternative to $x_1$ has become worse than it was in turn 6 because the alternative to $x_1$ in turn 6 was $x_4$ but in turn 8 the alternative to $x_1$ is one of uncertainty, as alluded to with the reference to betting, except instead of a payout of $x_4$ there is certainty of going into the “system.” After turn 8, $CS$ signs a waiver revealing the shift to $P_{CS}=x_1\succ x_4\succ x_3\succ x_2$ brought on by $I$’s statement in turn 7.

Excerpt 4.6 provides another example of the hypergames played in a police interrogation.

Excerpt 4.6 is situated at the initial stage of the invocation game initiated by $I$’s reading of the Miranda rights. In turn 1, $I$ seeks affirmation from $CS$ that $CS$ understands these rights and warnings. $CS$ affirms in turn 2. $I$ then imposes a decision on $CS$ in turn 3: $x_1$ or not $x_1$? $CS$ responds by indirectly invoking the right to counsel in turn 4 which reveals $P_{CS}=x_4\succ x_1\succ x_3\succ x_2$. $I$ appears to acknowledge the invocation in turn 5.

In turn 6, $CS$ initiates a type 1 hypergame with $x_3^{\prime }$ which reveals a shift in preference $P_{CS}=x_3\succ x_4\succ x_1\succ x_2$. Strategically, it seems that $CS$ is apologizing for, or justifying, $CS$’s assertion of $x_4$, perhaps responding to the perceived power dynamics between $CS$ and $I$. $I$ acknowledges the expression, then in turn 8 $CS$ executes a strategic probe for information. $I$ defers providing information, and then $CS$ asks a procedural question in turn 10, $I$ expresses a need for clarification, $CS$ repeats the question in turn 12, and then $I$ makes the strategic move in turn 13 to shift the topic back to the question of waiving or invoking the right to counsel, as if $CS$ had not expressed an opinion.

Turn 14 shows that in $x_3^{\prime }$ $CS$ has gone from seeker to provider of information by acknowledging the procedure for the dispreferred choice $x_1$ and that $x_1$ does not negate $x_4$. In turn 15, $I$ does not acknowledge the lingering invocation and instead asserts that $CS$ is in control and must yet decide what to do. In taking this action, $I$ asserts a perception of how power is situated in the talk that is contrary to reality. Had $CS$ been able to assert footing, the invocation of turn 4 would have been acknowledged. Instead, it stands as a powerless utterance.

$CS$ agrees four times in turn 16 to the illusion $I$ gives in turn 15, but then $CS$ expresses confusion. $I$ provides a (mis)guiding reaffirmation that $CS$ is in control. $CS$ assumes control by twice starting an assertion of what $I$ can do in turn 18 to which $I$ helpfully completes the thought asking if $CS$ wants $x_1$. Turn 20 reveals the important change: $P_{CS}=x_1\succ x_4\succ x_3^{\prime }\succ x_2$, and with a request for clarification in turn 20, $CS$ confirms in turn 21.

But, why did $CS$’s preferences change? Key to understanding the change is the fact that during the conversation no new information about circumstances or procedure were provided, despite $CS$’s probe for this information. The only notable topic that arises is one of power: who is in control of the situation. Within the conversation, $CS$’s tone changes from one of certainty about preference, to a small hesitation in the probe for information in turn 8, to a desire for affirmation of understanding of the procedure in turns 10 and 12, and then to the shift in footing from being a requester of information to a provider.

Perhaps the incongruous nature of an exchange in which the one who gives up information is portrayed as being in control is as confusing as it is comforting. After the reassuring illusion of control, $CS$ switches to trust the situation and relinquish all control by waiving the right to legal counsel. It may seem that $CS$ has switched from skepticism of $I$ to trusting $I$, but in probing for information in turn 8 it shows that $CS$ believes $I$ is willing to be helpful, an illusion furthered by the illusion of control. $I$ has reassured $CS$ to see the talk as situated cooperatively with $CS$ in the position of power. $I$ implies that $x_2,x_4\ni x_1$ because if $x_1$ is chosen both $x_4$ and $x_2$ remain as options, and $x_1$ is the way to satisfy the curiosity $CS$ expressed in turn 8. $I$ affirms the change in preference by requesting clarification consistent with a cooperatively situated exchange in turn 21, and $CS$ cooperatively affirms in turn 22.

4.6 Implications of the Invocation Game: Identifying Manipulation of Preferences in Police Interrogation

The application of hypergame theory to empirical linguistic data provides a method for analyzing the strategic aspects of a police interrogation. This game theoretic approach to analyzing linguistic actions describes how $I$ develops linguistic strategies that reshape $CS$’s perceptions to shift $CS$’s preferences to conform with $I$’s preferences. This chapter’s game theoretic approach expands the analysis of discourse by treating the reading and waiving of Miranda rights as a noncooperative game with potentially multiple preference manipulating moves. When applied to linguistics, the hypergame approach gives manipulation a coherent structure. Its distinct structure accommodates binary states consistent with discursive success or failure as well as a continuum of many states consistent with stages in discursive persuasion. It can incorporate motives, incentives, boundaries, etc., to expand the range of analysis to an infinite number of turns, accommodating any number of digressions relevant or not.

Moreover, this chapter considers the invocation game to be a particular example of a more broadly identified mode of a game which recognizes many avenues by which perceptions may be disturbed causing changes in preferences. This game theoretic approach may also be applied to the minimization and maximization strategies of police interrogation, as well as other types of police interviewing techniques in which changes of preferences are observed. Noncoercive police interrogation techniques are not impervious to police interviewers changing suspects’ preferences. The application of hypergame theory to the invocation game presented in this chapter can identify instances of manipulation of preferences at all stages of a police interrogation or interview – broadly defined and applied, including the interview of witnesses and victims of crime.