Gricean insinuation and the fake one-way mirror effect

Abstract

The insinuating speaker, when successful, achieves two goals: they introduce a contentious content into a conversation while simultaneously marking it as not being officially stated. This positioning allows the speaker to plausibly deny any intention behind the implied message when challenged. I argue that reconciling the communicative nature of insinuation and its off-record status within the Gricean framework of overt communication appears to present a significant conceptual puzzle. In this paper, I address this challenge by introducing the notion of Gricean insinuation, which I define as communication that is both partially overt and partially covert. I start by exploring the nature and complexity of the insinuation puzzle mentioned earlier and reviewing two existing Gricean solutions to it. Then, I proceed to analyse three concepts relevant to the insinuation phenomenon: indirectness, cancellability, and plausible deniability. Following this, I introduce the concept of Gricean insinuation. In particular, I use the common ground framework to elucidate what I refer to as the ‘Fake One-Way Mirror Effect’, and apply the resulting model to explain some examples of insinuation as acts of off-record communication. In conclusion, I assess the model’s effectiveness and propose directions for future research.

Elizabeth Camp characterizes insinuation as ‘the communication of beliefs, requests, and other attitudes ‘off-record’, so that the speaker’s main communicative point remains unstated’ (Camp, 2018, p. 42). This definition encapsulates the dual nature of insinuation, emphasizing both its communicative nature and its off-record status, which, I believe, are in tension with each other. According to the widely accepted Gricean perspective on communication (Grice, 1989; Strawson, 1964; Schiffer, 1972; Bach & Harnish, 1979; Sperber & Wilson, 1995; Wilson & Sperber, 2012; cf. Witek, 2022a), our communicative interactions consist in expressing and recognizing overt intentions, i.e., intentions intended to be recognized or even intended to be recognized as being intended to be recognized (Camp, 2018; Oswald, 2022). In contrast, when a speaker conveys a certain meaning off-record, their goal is to create a discursive situation in which ‘it is not possible to attribute only one clear communicative intention to [their] act’ (Brown & Levinson, 1987, p. 211). In other words, the meaning in question remains unspoken, and the intention behind its conveyance is essentially unavowable (Strawson, 1964), or even covert (Attardo, 1999; Bach & Harnish, 1979). Consequently, the speaker can plausibly deny intending to imply what they actually communicate, while their interlocutor can choose to intentionally overlook the unstated function and meaning of their act. In summary, reconciling the communicative nature of insinuation and its off-record status within the Gricean framework of overt communication appears to present a significant conceptual puzzle.

In this paper, my goal is to explore a potential solution to the previously outlined puzzle of insinuation through the development of the concept I call Gricean insinuation. Generally, I believe that theorising about insinuation requires a form of conceptual engineering, which can be understood as designing, implementing, and evaluating new categories (Chalmers, 2020). The colloquial notion of insinuation, as used in everyday contexts, lacks precision and fails to capture a uniform class of discursive phenomena comprehensively. Therefore, to adequately address these phenomena, we need a suite of distinct yet closely related theoretical categories, each tailored to address and foreground a different aspect of the communicative practice of insinuation. My aim in this paper is to contribute to this effort by introducing the concept of Gricean insinuation.

Briefly, I define Gricean insinuation as a communicative practice that consists of performing acts of indirect communication, which are both partially overt and partially covert. In this respect, the account I propose differs from models that depict insinuation as a form of overt communication (Camp, 2018; Fraser, 2001; Oswald, 2022) and from theories presenting insinuation as a covert act (Attardo, 1999; Bach & Harnish, 1979; Strawson, 1964). Moreover, I reject the idea, maintained by some scholars including Elizabeth Camp (2018) and David Bell (1997), that the specific form of indirectness germane to insinuating is to be accounted for in terms of conversational implicatures. Specifically, I distinguish between the cancellability of implicatures and the deniability of insinuations, arguing that they should and can be explained independently of each other, and that cancellation and denial are two distinct conversational moves. I contend that the point behind acts of Gricean insinuation is to maintain what I describe as ‘psychologically plausible deniability’, which I explain by reference to what I term the ‘Fake One-Way Mirror Effect’. Through this, I aim to provide a precise framework for discussing at least some forms of insinuation.

In summary, the key assumption of this paper is that the discursive phenomena we describe as insinuations do not form a homogenous class. Instead, they exploit different conversational mechanisms and are designed to achieve distinct ends. The key takeaway from this discussion is that some of these phenomena can be adequately described as acts of Gricean insinuation and accounted for in terms of the Fake One-Way Mirror Effect. However, I contend that a comprehensive account of the diverse and intricate phenomena of insinuation requires more than one conceptual category. In particular, I believe a concept is necessary that would enable us to address the social dimension of insinuating, presenting it as a procedure-governed discursive practice. A detailed construction of this category, which can be called Austinian insinuation, goes beyond the scope of the present paper, and will be addressed in future work (Witek, in preparation).

The paper is organized as follows. Section 1 discusses the central ideas of the Gricean approach to meaning and communication, delves into the intricacies of the puzzle of insinuation, and offers a brief reconstruction of two possible Gricean solutions to it. Section 2 provides a short overview of the central concepts used to discuss the phenomenon of insinuation: indirectness, cancellability, and plausible deniability; specifically, it distinguishes between cancelling and denying as two different conversational moves and introduces the concept of psychologically plausible deniability, which plays a key role in defining Gricean insinuation. Section 3 develops the notion of Gricean insinuation as a partially overt act, uses the common ground model to elaborate on the idea of the Fake One-Way Mirror Effect, and discusses several examples of insinuation using the resulting framework. Section 4 discusses the strengths and limitations of the proposed model and suggests avenues for future research.

Before I proceed with this agenda, I would like to make one additional point. Following Camp (2018), I attempt to account for the communicative mechanisms behind insinuation within the Stalnakerian common ground framework (Stalnaker, 2002, 2014). Specifically, I believe that acts of Gricean insinuation can be defined by reference to how they update the context of conversation. Unlike Camp, however, I do not explain this particular contextual update in terms of the difference between what is mutually believed and what is mutually accepted (Camp, 2018, pp. 56–58). Instead, I characterize it by examining how the insinuating speaker disrupts the iterative structure of the common belief among the participants in conversation. For this reason, in SubSect. 3.3, I employ a formal representation of this iterative structure to precisely identify the locus of the Fake One-Way Mirror Effect. Nevertheless, this representation is not indispensable to my central argument, which primarily focuses on discussing examples of insinuation. Therefore, readers may choose to skip SubSect. 3.3, as it is included solely for the sake of precision.

1 The puzzle of insinuation and existing solutions to it

Before discussing the puzzle of insinuation in SubSect. 1.2 and exploring two possible Gricean solutions in SubSect. 1.3, it is beneficial to briefly revisit the core principles of the Gricean approach to conversational meaning-making in SubSect. 1.1.

1.1 The Gricean perspective on meaning and communication

The main idea behind the Gricean approach is that our communicative transactions consist of performing and interpreting acts of speaker-meaning. In his often cited paper titled ‘Meaning’, H.P. Grice (1989, p. 219) posits that to speaker mean something is to utter a sentence with the intention to get the hearer to entertain or even adopt a certain attitude by means of getting them to recognize this intention; in short, the intention which determines the meaning of the speaker’s act is both reflexive and overt in that it refers to itself and is intended to be recognized, respectively.¹ Grice also suggests that the above-mentioned token-reflexive definition of speaker meaning is equivalent to the following iterative analysis:

(SM₁) In uttering U, S means something if and only if S intends:

• (i₁) To produce by uttering U a certain response R on the part of H,

• (i₂) To get H to recognize (i₁),

• (i₃) That the fulfilment of (i₂) function as H’s reason for his response R.

In short, (SM₁) attributes to speaker S, who attempts to communicate something, a sequence of three distinct intentions, with each one designed to refer to the one immediately preceding it.

However, some scholars argue that (SM₁) fails to provide sufficient conditions for S’s meaning something in uttering U.² For instance, Strawson (1964, pp. 446, 447) considers a scenario in which S, aware that H is observing them but also knowing that H does not realize S is aware of this observation, arranges convincing-looking evidence that p. Strawson contends that this scenario serves as a counterexample to Grice’s original three-part analysis: it meets conditions (i₁), (i₂), and (i₃), yet it cannot be easily considered a situation where S attempts to tell something, where ‘telling’ is a technical term employed by Grice to describe acts of communication (cf. Strawson, 1964, p. 447; Grice, 1989, p. 218). He concludes that a further condition should be appended to (SM₁), stipulating that S intends (i₄) to get H to recognize (i₂). According to Strawson’s extended version of the iterative analysis of speaker meaning, then, intention (i₁) is not only intended to be recognized but also intended to be recognized as being intended to be recognized.

Strawson presents his counterexample in general terms, whereas Grice (1989, pp. 94, 95; cf. Camp, 2018, p. 54) examines its more detailed version known as ‘Giveaway Smile’, which, as I will argue in Sect. 3, serves as an example of Gricean insinuation. Imagine a scenario where an employee named Ann plays bridge against her boss, Peter. Ann wants her boss to win and for him to know that she wants him to win. Yet, Ann does not want to be too obvious about her plan. Thus, she adopts a communicative strategy: whenever she has a good hand, she simulates a spontaneous smile of pleasure. This smile is intentionally produced to make Peter believe that Ann has a good hand. Further, she intends for this intention to be recognized by Peter. Nonetheless, Ann wants to conceal her intention (i₂) for Peter to recognize that her smile is intended to influence his belief. In other words, while she wants Peter to recognize that her simulated smile is produced with intention (i₁) to make him believe that she has a good hand, she doesn’t want him to realize that this intention is overt and, consequently, that her smile is an attempt to tell something. In short, Ann intends for Peter not to recognize her intention (i₂).

Stephen Schiffer (1972, pp. 17, 18) considers an analogous scenario, known as ‘River Rat’ (cf. Green, 2007, pp. 63, 64). Imagine Paul inspecting a house he is considering buying. His friend, Alice, concerned about the home’s condition, wants to convince him that it is infested with rats. Knowing Paul is inside and aware she’s being watched, Alice places a river rat in a conspicuous spot. She hopes Paul will see the rat and conclude: ‘Though staged, Alice wouldn’t place the rat unless she truly believed the house has a rat problem. Given her reliability and honesty, she must be warning me that the house is infested’. According to Schiffer, Alice has a complex set of intentions in placing the rat: she intends (i₁) Paul to believe the house is rat-infested and intends (i₃) to achieve this by making him recognize her intention. This aligns with requirements posed by (SM₁). However, while it can be informally stated that Alice is signalling to Paul that the house is rat-infested, it’s not accurate to say that she actually tells that the house has a rat problem based on the rat placement alone.³

Viewed from the perspective of Strawson’s four-part analysis—which adds to (SM₁) the further condition that S intends (i₄) for H to recognize (i₂)—Ann in ‘Giveaway Smile’ and Alice in ‘River Rat’ cannot be said to be attempting to communicate that Ann has a good hand and that the house is rat-infested, respectively. Interestingly, Schiffer (1972, pp. 18, 19) analyses another scenario, ‘Moon Over Miami’, arguing that it serves as a counterexample to Strawson’s four-part analysis and as such necessitates the introduction of a fifth condition, stipulating that S intends (i₅) for H to recognize (i₄). Imagine a situation in which speaker S, who has a hideous singing voice, intends (i₁) to make A leave the room by singing ‘Moon Over Miami’. Additionally, S intends (i₂) for A to recognize that S is singing ‘Moon over Miami’ with intention (i₁). S also intends (i₄) that A recognizes intention (i₂) as a demonstration of S’s disdain for A’s presence in the room. S plans for A to believe that S’s goal is to drive A away with the repulsive singing. However, S expects and intends (i₃) that A will actually leave the room upon recognizing S’s intention (i₁).⁴ According to Schiffer (1972, p. 18), while S’s behaviour aligns with Strawson’s four-condition definition of speaker-meaning, it cannot be considered as a case of S’s telling A to leave the room.

Intriguingly, Strawson anticipates this development, suggesting that ‘[i]t is possible that further argument could be produced to show that adding [condition (i₄)] is not sufficient to constitute the case as one of attempted communication’. (Strawson, 1964, p. 447) He also points out that the iterative analysis is logically embarrassing in that it ‘seems open to a regressive series of intentions that intentions should be recognized’. (Ibid, p. 454) In summary, the comprehensive iterative analysis of speaker meaning, designed to be resistant to any possible counterexample and to elucidate Grice’s token-reflexive definition, takes the form of an infinite sequence of conditions, which can be articulated through a suitable recursive formula, denoted as (i_n+1), within (SM₂).

(SM₂) In uttering U, S means something if and only if S intends:

• (i₁) To produce by uttering U a certain response R on the part of H,

• (i₂) To get H to recognize (i₁),

• (i₃) That the fulfilment of (i₂) function as H’s reason for his response R,

• (i₄) To get H to recognize (i₂),

• (i_n+1) To get H to recognize (i_n), for n > 3.

Dan Sperber and Deirdre Wilson (1995, p. 31) and Mitchell S. Green (2007, p. 66) claim that the concept of speaker meaning, as characterized by (SM₂), is psychologically implausible. Specifically, they contend that the human mind, with its limited capacity, is incapable of accommodating an infinite hierarchy of conditions. Responding to this criticism of the iterative analysis, it is worth paying attention to the following two points. First, the token-reflexive definition of speaker meaning is also subject to criticism for psychological implausibility (for a discussion, see Siebel, 2003 and Witek, 2009). Second, as Steven Pinker et al., (2008, p. 837) note, the human mind can be said to know an infinite number of things, provided ‘the knowledge is not enumerated as an infinite list (…) but is implicit in a finite recursive formula’. In a similar vein, David Lewis and Bart Geurts argue that the iterative structure of common ground—and the same may be said for the iterative structure of meaning-constituting intentions—‘is a chain of implications, not of steps in anyone’s actual reasoning’ (Lewis, 2002, p. 53; cf. Geurts, 2019, p. 16).

1.2 The Phenomenon and the Puzzle of Insinuation

Consider, following Steven Pinker (2007, pp. 444–446; cf. Pinker et al., 2008, p. 834; Terkourafi, 2011, p. 2863; Camp, 2018, p. 43), a situation in which a driver is pulled over for speeding and receives a ticket. While talking to the officer, he utters the following words:

• (1) a. I’m in a bit of a hurry.

  b. Is there any way we can settle this right now?

Next, following Camp (2018, p. 43), consider a scenario where an estate agent is discussing with a couple interested in purchasing a house in a suburban area. Their ethnic identity differs from that of the majority of the area’s residents. The estate agent suggests:

• (2) Perhaps you would feel more comfortable locating in a more … transitional neighbourhood, like Ashwood?

Finally, picture a scenario where Sally and Harry are seated on a bench by a busy street, taking in the buzz of the city. Suddenly, they notice John alighting from a bus across the street. Sally says:

• (3) This is the third time this week that I have seen John getting off the bus at the stop near Maria’s place.

It is reasonable to assume that in posing question (1b), the speeding driver indirectly offers a bribe to the officer, while simultaneously maintaining deniability: if challenged, he can coherently claim he had no intention of making such an offer. The estate agent’s statement of (2), presented as a positive suggestion, may be interpreted as insinuating that the buyers are not desirable residents for the area under discussion. Finally, Sally’s remark (3) may be interpreted as a veiled and deniable suggestion that John and Maria are involved in an affair.

In summary, the driver, the estate agent, and Sally manage to accomplish two things: they introduce potentially problematic content into the conversation while simultaneously marking it as unofficially stated; this off-record status of their communicative acts can be defined in terms of their ability to deny having intended to perform them and their interlocutors’ option to overlook or ignore what they implicitly convey (Bell, 1997, pp. 41–43). The central challenge posed by insinuation, then, lies in reconciling the following two intuitively appealing ideas: (idea₁) the communicative nature of insinuating utterances and (idea₂) the off-record status of the communicative acts these utterances constitute.

Meeting this challenge appears to be particularly difficult for those who adopt the Gricean perspective on communication (Grice, 1989; Strawson, 1964; Schiffer, 1972; Bach & Harnish, 1979; Sperber & Wilson, 1995; Wilson & Sperber, 2012; cf. Witek, 2022a), as outlined in SubSect. 1.1. The core principle of the Gricean approach is that the meaning of the speaker’s speech act is determined by the intention with which it is performed, this intention being intended to be recognized by the hearer, or even, as argued by Strawson (1964), intended to be recognized as being intended to be recognized (cf. Bach & Harnish, 1979; Camp, 2018; Oswald, 2022). Thus, successful communication involves getting the hearer to recognize the speaker’s intention that determines the meaning of their utterance. As Strawson points out, the speaker who attempts to communicate something ‘has a motive, inseparable from the nature of [their] act, for making that intention clear’ (Strawson, 1964, p. 450); in short, overt communicative intentions have what Strawson describes as ‘essential avowability’ (Strawson, 1964, p. 454).

Strawson uses the idea of essentially avowable communicative intentions to explain the phenomenon called the ‘general suitability [of communicative illocutionary acts] for being made explicit with the help of (…) explicitly performative formula[s]’. (Strawson, 1964, pp. 449, 450) Another linguistic phenomenon that can be accounted for by reference to the essential avowability of communicative intentions is the reinforceability of conversational implicatures, i.e., the fact that is ‘possible to make them explicit without being guilty of redundancy’ (Sadock, 1978, p. 294), or in other words, that they ‘can be conjoined with an overt statement of their content without a sense of anomalous redundancy’. (Levinson, 1983, p. 120) For instance, if you suggest we go to the cinema together and I reply, ‘I have to study for an exam’ (Searle, 1979, p. 33), I conversationally implicate that I will not be able to accompany you. If I add, ‘I will not be able to go with you,’ my utterance, even though perceived as redundant, is not considered anomalously redundant. It does not produce an anomaly because it has a purpose, i.e., it serves to forestall misunderstanding of what I intend to imply.⁵

In short, the Gricean speaker has a motive to make their intention clear and to help the hearer to recognize it. The insinuator, by contrast, is unwilling to reinforce the implicit content of their utterance or, in other words, is unwilling to admit that they have intended to imply it. Quite the contrary: when challenged, they are prepared to coherently deny having any such intention.

Using an idiom borrowed from Penelope Brown and Stephen C. Levinson (1987), I would suggest that the insinuating speaker deliberately ‘goes off record’ in communicating the implicit content of their utterance—that is, they adopt the strategy of off-record indirectness. This creates a discursive situation in which ‘there is more than one unambiguously attributable intention so that [they] cannot be held to have committed [themselves] to one particular intent’ (Brown & Levinson, 1987, p. 69). According to Brown and Levinson (1987, pp. 213, 214), one way to achieve this goal is by inviting conversational implicatures through violations of the Gricean maxims of Relevance, Quantity, or Quality⁶; another way is by violating the maxim of Manner through deliberate vagueness or ambiguity.⁷ To go on record in performing a speech act, in contrast, is to put the conversational participants in a position in which it is ‘clear to [them] what communicative intention led the actor to [perform this act] (i.e., there is just one unambiguously attributable intention with which witnesses would concur)’ (Brown & Levinson, 1987, pp. 68, 69). According to Brown and Levinson (1987, p. 70), a natural way to achieve this is by employing conventional means that constrain the hearer’s inference designed to identify the speaker’s intention; specifically, it is to perform conventionalized indirect speech acts. In my view, it is possible to extend the concept of on-record indirect speech to ‘strong’ conversational implicatures, that is, implicatures required to maintain the assumption that the speaker’s utterance is relevant (Sperber & Wilson, 1995, p. 199; cf. Oswald, 2022, p. 165) or meets the criteria of discourse coherence (Asher & Lascarides, 2001).

All in all, it is worthwhile to distinguish between the speaker’s intention to go on-record or off-record in performing a speech act and the specific means or strategies they employ to achieve this goal. Generally, the speaker goes off-record in performing a speech act only if they aim to create a situation in which there is more than one intention that can be unambiguously attributed to them. Specifically, the insinuator goes off-record in conveying potentially troublesome content only if they intend to produce a situation in which they cannot be held to have committed themselves to one particular intent. To achieve this goal, they can flagrantly violate conversational maxims, be deliberately ambiguous, encourage the hearer to derive a host of ‘weak’ implicatures in the sense outlined by Sperber and Wilson (1995, pp. 198, 199), or employ other techniques of indirect speech.

In summary, acts of Gricean communication meet the criteria set by (SM₂) or at least by Strawson’s four-part analysis of speaker meaning. That is to say, the Gricean speaker aims to produce a situation in which the hearer can easily see through their communicative plan. In contrast, the insinuating speaker deliberately produces and maintains a situation in which the hearer cannot unambiguously attribute one intention to them, thereby maintaining plausible deniability.

The puzzle of insinuation, then, is that it is challenging to reconcile (idea₁) and (idea₂), that is, the communicative nature of insinuating utterances and the off-record status of the communicative acts these utterances constitute.

1.3 Two Gricean solutions to the puzzle of insinuation

There are at least two possible solutions to the puzzle of insinuation aligning with the spirit of the Gricean analysis: solution (s₁) suggested by the model of insinuation as a covert act (Attardo, 1999; Bach & Harnish, 1979; Strawson, 1964), and solution (s₂) indicated by the theory of insinuation as an overt act (Camp, 2018; Fraser, 2001; Oswald, 2022). Let us discuss them sequentially, examining whether they enable harmonizing (idea₁) and (idea₂) within the Gricean framework.

Let us first consider solution (s₁), which is inherent in the model of insinuation as a covert act, different versions of which are proposed and advocated by Strawson (1964), Bach and Harnish (1979), and Attardo (1999). According to Strawson, the ‘whole point of insinuating is that the audience is to suspect, but not more than suspect, the intention, for example, to induce or disclose a certain belief’ (Strawson, 1964, p. 454; italics mine—MW). In making this claim, he suggests that the insinuator’s intention to get the hearer to entertain a certain attitude is intended not to be recognized: merely suspecting that one’s interlocutor has a certain intention means that one does not recognize it. In sum, the ‘intention one has in insinuating is essentially nonavowable’, and hence covert. Likewise, Bach and Harnish (1979, p. 101) classify insinuation⁸ as a covert collateral act. They introduce the concept of collateral acts to denote a broad and heterogeneous class of linguistic or ‘conversational acts [that] can be performed in conjunction with or in lieu of illocutionary acts’ (Bach & Harnish, 1979, p. 97), including kidding, joking, mimicking, reciting a poem, circumlocution, and other uses of language that cannot be adequately explained through their Speech Act Schema.⁹ Some of them are overt, others covert, and yet others neither overt nor covert. Specifically, Bach and Harnish (1979, p. 101) define covert collateral acts, including innuendo (insinuation), deliberate ambiguity, and the so-called ‘sneaky presuppositions’,¹⁰ as acts ‘performed with intentions that are intended not to be recognized’. In a similar vein, Attardo argues that insinuating utterances are best understood as covert speech acts, that is, ‘acts which have as one of their felicity conditions that the hearer not be aware of the speaker’s intention to achieve the speech act in question’ (Attardo, 1999, p. 202).

The term ‘speech act’ can be read either in its narrow or broad sense. That is to say, it can be used either to refer exclusively to illocutionary acts or to denote a broad class of verbal or discursive acts characterized by reference to their effects. In my view, when Attardo discusses covert speech acts, he use the term ‘speech act’ in its broad sense. However, to avoid confusion, in the remaining part of this section, I follow Bach and Harnish and use the term ‘covert acts’ rather than ‘covert speech acts’.¹¹ I also take it that the successful performance of a covert act requires that its underlying intention remains unrecognized.¹² A key aspect of a covert intention represented in (C), then, is a counter-intention (i₂’), which prevents intention (i₁) from triggering the endless cascade of conditions recursively defined in (SM₂).

• (C) In uttering U, S covertly intends to produce R if and only if S intends:

• (i₁) To produce by uttering U a certain response R on the part of H,

• (i₂’) For H not to recognize (i₁).

If insinuation is a form of covert speech, then the insinuating speaker intends (i₁) to elicit a certain response R from their interlocutor, such as inducing a particular belief or attitude, while simultaneously intending (i₂’) for the interlocutor to remain unaware of intention (i₁). The purpose for hiding this intention stems from the speaker’s desire to avoid responsibility for contributing the insinuated attitude into the conversation and, consequently, to maintain plausible deniability. A fundamental assumption of solution (s₁) is that we are only accountable for actions undertaken intentionally. Therefore, when successful, the insinuating speaker achieves two goals: they elicit the intended response R from their interlocutor and simultaneously escapes responsibility for it; in short, they secure the fulfilment of intentions (i₁) and (i₂’), respectively.

In summary, solution (s₁), which is suggested by the model of insinuation as a covert act, offers an intuitively compelling explanation for (idea₂) while simultaneously rejecting (idea₁). Specifically, it categorizes acts of insinuating as non-communicative, with this classification being contingent upon interpreting ‘communicative’ according to the Gricean standards. Bach and Harnish (1979, p. 101) contend that insinuation is a covert collateral act, meaning that it is devoid of communicative character. Similarly, Strawson (1964, p. 454) argues that insinuating is not a communicative illocutionary act because its underlying intention is not overt. In summary, rather than reconciling (idea₁) and (idea₂), solution (s₁) removes the tension between them by dismissing (idea₁). Consequently, it fails to accommodate the intuitively appealing notion that acts of insinuating are communicative in some sense.

Moreover, solution (s₁) fails to accommodate a perceptible distinction between clear cases of insinuating, such as (1, 2, and 3), and certain clearly covert acts, like the ones Mary Kate McGowan (2023) examines as examples of covert audience-indirectness. Consider, following McGowan (2023, p. 411), a scenario where, within hearing distance of their daughter Nora, T says to his partner U:

• (4) I sure hope the kids settle down tonight. Santa only comes if the kids are asleep.

The underlying intention of T’s utterance of (4) is to encourage Nora to go to bed early. This intention is covert in the sense outlined in (C), i.e., its fulfilment relies on its not being recognized by Nora, who is the intended addressee of T’s utterance, or, as McGowan (2023, p. 412) puts it, ‘a covertly intended overhearer’. In my view, however, this scenario should not be regarded as an instance of insinuation. Hence, even if we agree that being a covert collateral act is a necessary condition for insinuation, it should not be regarded as a sufficient condition.

In summary, solution (s₁), by categorizing insinuation as a covert act, fails to accommodate the intuitively appealing notion that acts of insinuating are communicative. Furthermore, it does not establish sufficient conditions for insinuation. These two remarks are not intended as definitive arguments against the models proposed by Bach and Harnish, Strawson, and Attardo, nor are they aimed at undermining the well-motivated concept of covert collateral acts. Nevertheless, they motivate the search for an alternative solution to the puzzle of insinuation—a different model that would enable us to harmonize (idea₁) and (idea₂).

Let us now turn our attention to solution (s₂), which is inherent in the theory of insinuation as an overt speech act. Its proponents maintain that insinuated contents are speaker-meant (Camp, 2018; Fraser, 2001; Oswald, 2022) and conveyed through the mechanisms of conversational implicatures (Camp, 2018) or pragmatic implications (Fraser, 2001). In short, they portray insinuating as a communicative practice understood along the Gricean lines. Consequently, solution (s₂) offers a straightforward explanation of (idea₁). However, it encounters a significant challenge in accounting for (idea₂) within the Gricean model of overt communicative intentions. As Camp puts it, the ‘phenomenon of implicature with deniability lies at [the] core [of insinuation]’. (Camp, 2018, p. 46) Insinuated contents, then, are speaker-meant at the level of what is conversationally implicated. The question then arises: how can these forms of indirect communication be overt and yet remain off-record?¹³

Camp suggests that both (a) the off-record status of the insinuated meaning of an utterance and (b) its deniability are represented in the content of the speaker’s communicative intention. She claims that ‘(…) an insinuating speaker typically intends H to recognize their intention [(a)] that M(Q) be off-record, and [(b)] that they are prepared to deny having meant M(Q) if challenged’. (Camp, 2018, p. 55; ‘M(Q)’ stands for the insinuated attitude.) In my view, this solution possesses considerable explanatory potential and can illuminate at least some forms of insinuation. However, it is theoretically demanding as it requires the acceptance of a further assumption that appears to extend beyond the Gricean perspective outlined in SubSect. 1.1. Specifically, it assumes that both the insinuating speaker and their addressee possess the notions of off-record communication and deniability; without these independently defined concepts, they would not be able to form, express, and recognize intentions which involve the aforementioned aspects (a) and (b).

Let us briefly discuss another attempt to elaborate on (idea₂) within the Gricean model of overt communication. In his paper titled ‘An account of innuendo’, Bruce Fraser states:

Of course, because of unwanted import the speaker may not want to take responsibility for the insinuation. Rejection of this responsibility is within the speaker’s ability. (…) It is an issue separate from that of the communicative status of an insinuation if, for political or other reasons, the speaker does [not] wish to be explicit and direct with the content of the insinuation and takes the implied route. (Fraser, 2001, pp. 330, 331; cf. Oswald, 2022, p. 162)

In short, Fraser does not assume that the off-record status of insinuation is conveyed through the content of the speaker’s intention. Instead, he takes is for granted that rejecting the responsibility for insinuating a certain unwanted content ‘is within the speaker’s ability’. In my view, this claim requires further elaboration and justification. My hypothesis is that Fraser envisions two ideas: first, that insinuated contents are implicated, or in his words, ‘implied’ (Fraser, 2001, pp. 326, 327); and second, that implicatures are cancellable. However, as I will argue in SubSect. 2.2 below, the cancellability of implicatures and the deniability of insinuation are two different phenomena, each requiring an independent explanation. Fraser also refers to extra-linguistic norms and rules, such as those of law, morality, politics, etiquette, and politeness. In my view, however, this reference fails to address the central question of what it is for a certain aspect of a communicative act to have the off-record status. Instead, it addresses the question of what motivates the speaker to communicate certain contents off-record.

It is important to stress that my intention in this subsection is not to argue against the models proposed by Camp and Fraser. Instead, my goal is to examine whether their models suggest a satisfactory solution to the puzzle of insinuation, specifically whether they allow us to maintain and harmonize (idea₁) and (idea₂) within the Gricean framework as outlined in SubSect. 1.1.

In Sect. 3, I develop a model of insinuation as a partially overt act and argue that it allows us to reconcile (idea₁) and (idea₂) within a broadly Gricean perspective on communication. However, before I proceed with this objective, I would like to elaborate on a few concepts central to the discussion of the communicative practice of insinuating.

2 Indirectness and plausible deniability

It is widely acknowledged that insinuation is a form of indirect communication or, more specifically, that the insinuated meaning of an utterance goes beyond its officially communicated content. Typical examples of indirectness in discourse are conversational implicatures (Grice, 1989) and indirect speech acts (Asher & Lascarides, 2001; Searle, 1979). In SubSect. 2.1, however, I argue that insinuation does not necessarily have to be subsumed under one of these two categories. Specifically, in SubSect. 2.2, I elaborate on the notion of the cancellability of implicatures and contrast it with the idea of the plausible deniability of insinuation. Next, in SubSect. 2.3, I distinguish between three types of plausible deniability: contextual, psychological, and epistemic. The key message of this section is that even though insinuation, construed as a form of indirect speech, resembles in certain respects the phenomenon of conversational implicature, it should not be identified with it.

2.1 Indirectness in discourse

The structure of an act of indirect speech involves at least four aspects. Firstly, (a) speaker S produces utterance U of sentence L, the conventional function of which is to present proposition P with illocutionary force F. Secondly, in doing this, (b) S locutes F(P) and, thirdly, provided further conditions are met, (c) illocutes F(P), i.e., performs a direct illocutionary act of F-ing that P. Fourthly, in performing this act, (d) S conveys a distinct proposition Q in mode M or, in other words, indirectly gets across attitude M(Q).

Following Camp (2018, pp. 44, 45), from whom I adopt the description above, I do not address the fact that the linguistic meaning of L typically underdetermines force F and propositional content P of S’s direct illocutionary act.¹⁴ It is also worth emphasizing that I deliberately use the somewhat awkward phrase ‘an act of indirect speech’ rather than ‘an indirect speech act’. The latter is usually used to refer to ‘cases in which one illocutionary act is performed indirectly by way of performing another’ (Searle, 1979, p. 31), whereas the former encompasses any kind of indirect speech. This terminological convention is inspired by Mitchell S. Green’s distinction between speech acts and acts of speech. According to Green (2021, p. 5), ‘an act of speech is any act of uttering meaningful words, [while] ‘speech act’ is a term of art’, meaning that it specifically denotes illocutionary acts, understood as ‘acts that can (though need not) be performed by saying that one is doing so’. I would add that an act of speech, that merits its name, involves not only uttering meaningful words but also affecting the context of its performance in a certain way.¹⁵ Moreover, I believe the distinction between speech acts and acts of speech allows for a broader interpretation of discursive indirectness, irrespective of whether it can be described in illocutionary terms.¹⁶ In essence, then, conveying attitude M(Q) is S’s act of indirect speech which may, but does not necessarily, function as an indirect illocutionary act of M-ing that Q.

Many acts of indirect speech can be adequately explained in terms of conversational implicatures, i.e., those aspects of utterance meaning that are recognized by maxim-guided pragmatic inferences (Grice, 1989, pp. 26, 27). By way of illustration, consider the following exchange, which was briefly discussed in Sect. 1:

• (5) A: a. Let’s go to the movies tonight.

  B: b. I have to study for an exam.

In uttering (5b), B makes a direct statement about themselves and indirectly rejects A’s proposal. According to Searle (1979, p. 33), from whom I adopt this example, the indirect illocutionary force of B’s utterance is communicated through the Gricean mechanism of conversational implicature: the assumption that, in uttering (5b), B rejects A’s proposal is required to maintain the default assumption that B is a cooperative speaker observing conversational maxims in general, and the maxim of relevance in particular. However, there are numerous cases of indirection in speech that do not involve conversational implicatures construed as indirect and cooperative acts of speaker-meaning. For instance, McGowan (2023) discusses two forms of indirection—indirection with respect to other agent re-uses and audience-indirection—that cannot be described in terms of saying one thing and meaning another, or, more specifically, in terms of directly meaning F(P) and conversationally implicating M(Q).¹⁷ In my view, at least some forms of insinuation require a similar treatment or, more generally, their indirect nature cannot be adequately accounted for in terms of conversationally implicated forces and meanings.

In summary, to conceptualize insinuation as a form of indirect speech, we do not need to assert that each act of insinuating possesses an indirect illocutionary force. Nor do we have to adopt the general rule that insinuated contents are conversationally implicated. It is sufficient to make the following assumption:

• (INDIR) For every act of insinuating, there exists a specific illocutionary or, at the very least, locutionary act that serves as its vehicle.

However, it is instructive to emphasize that (INDIR) does not categorically exclude the possibility that at least some types of insinuation employ mechanisms similar to implicature. Furthermore, (INDIR) leaves open the question of whether at least some acts of insinuation should be explained by reference to indirectly conveyed illocutionary forces. I will leave the discussion of these possibilities for another occasion. My aim in this paper is to use the broadly Gricean framework to construct the concept of Gricean insinuation. Before delving into the intricacies of this construction, let me first explore the notion of plausible deniability, which plays a pivotal role in elucidating the rationale behind insinuation. I believe this notion should be clearly distinguished from what Grice calls the explicit cancellability of conversational implicatures.

2.2 Cancellability and deniability

Grice considers cancellability as a possible test for the presence of conversational implicatures; he states:

A putative conversational implicature that p is explicitly cancellable if, to the forms of words the utterance of which putatively implicates that p, it is admissible to add but not p, or I do not mean to imply that p (…). (Grice, 1989, p. 44)

Strictly speaking, cancellability is a diagnostic tool that theoreticians can use to determine whether a specific implication of an utterance is conversationally implicated. If it is feasible to combine, within a single utterance, the sentence whose utterance implicates that p and an appropriate cancellation phrase, without resulting in contradiction, then the implication is deemed to be conversationally implicated by the utterance in question, rather than being entailed by its linguistically determined content.¹⁸ For illustration, consider scenario (5) again. To test the proposition ‘B won’t go with A to the movies tonight’ for its status as a conversational implicature, a theoretician might envisage a modified version of (5) in which B, after uttering (5b), adds ‘but I will go with you to the movies tonight’. If the addition of such a cancellation clause does not lead to contradiction, then the proposition in question is conversationally implicated.

However, the ‘admissible’ explicit cancellation of an implicature can be seen not just as a test available to theoreticians, but also as a conversational move that actual speakers can employ in real conversational settings (Blome-Tillmann, 2008; Geurts, 2010; Weiner, 2006; Zakkou, 2018), thereby influencing the incremental process of meaning recovery (Jaszczolt, 2009). To illustrate, let’s examine the following two scenarios that Geurts (2010) identifies as involving explicit cancellations:

• (6) A: a. I’m out of petrol.

  B: b. There is a garage round the corner,

  c. but it’s closed.

• (7) a. X is meeting a woman this evening.

  b.—his sister, in fact.

Turns (6a) and (6b) constitute Grice’s famous ‘Garage’ scenario (Grice, 1989, p. 32), in which B conversationally implicates that the garage round the corner is open and has petrol to sell. In adding (6c), however, B ‘cancels’ this implicature. Normally, to utter (7a) is to suggest that the woman whom X is meeting this evening is ‘someone other than X’s wife, mother, sister, or perhaps even close platonic friend’. (Grice, 1989, p. 37) Still, this implicature can be ‘cancelled’ by adding (7b). As Geurts (2010, pp. 20, 21) aptly notes, however, the use of the verb ‘cancel’ to describe the dynamics of dialogues (6) and (7) is misleading (cf. Zakkou, 2018, pp. 8, 9). The intent behind turns (6c) and (7b), namely, is to prevent the addressee from deriving the potential implicatures of (6b) and (7a), respectively. In short, we normally use cancelation clauses to prevent misunderstanding rather than to cancel or retract what we mean.

In short, it is useful to distinguish between cancellation seen as a diagnostic procedure used by theoreticians to determine if a specific implication of an utterance is its conversational implicature, and cancellation understood as a conversational move made by actual speakers in real conversational settings, aimed at preventing misunderstandings. To say that conversational implicatures are cancellable implies that they pass the cancellability test. Conversely, to say that a putative or potential implicature of a specific utterance is cancelled is to describe how a specific conversational act performed in uttering a cancellation phrase affects the incremental process of meaning recovery. Cancellability construed as a diagnostic property is a non-gradable or binary category (Mazzarella et al., 2018, p. 16): the addition of a cancellation phrase to the examined utterance either results in a contradiction or yields a coherent discursive contribution. On the other hand, cancellation understood as a conversational move can vary in success. As demonstrated by Jaszczolt (2009), the ease with which a potential implicature can be conversationally cancelled is inversely proportional to its degree of entrenchment.

Let us now proceed to discuss the deniability of insinuation. Like conversational cancellations, acts of denying are typically performed in uttering phrases of the form ‘but not p’ and ‘I do not mean to imply that p’. Despite this similarity, however, cancellations and denials should be viewed as two distinct types of speech acts that require different conversational occasions for their execution and have different intended effects. As illustrated by scenarios (6) and (7), conversational cancellation involves juxtaposing, within a single utterance, a sentence whose utterance purports to implicate that p and an appropriate cancellation phrase. The purpose of this conversational manoeuvre is to contribute to the incremental construction of discourse meaning by preventing the addressee from drawing certain pragmatic inferences. In contrast, the appropriateness of a denial requires that the speaker has been challenged by their interlocutor. Consider again the scenario in which Harry and Sally notice John getting off a bus across the street, and Sally states (3), repeated below as turn (8a).

(8)	Sally:	a.	This is the third time this week that I have seen John getting off the bus at the stop near Maria’s place.
	Harry:	b.	Do you mean that John and Maria are having an affair?
	Sally:	c.	I don’t intend to imply anything of that sort!
		d.	You said it, not me.

Turn (8b) constitutes Harry’s challenge, which prompts Sally to utter (8c), that is, to deny having intended to imply that John and Maria are involved in an affair. The purpose of her denial is to disavow any responsibility for contributing this problematic proposition to the ongoing conversation. Furthermore, she seizes the opportunity and, in uttering (8d), attempts to shift the responsibility onto Harry.

In summary, it is valuable to distinguish between three different acts that can be executed using cancellation phrases: diagnostic cancellation, conversational cancellation, and conversational denial. Diagnostic cancellation is a test available to theoreticians who want to determine if a specific implication of an utterance merits description as a conversational implicature. Conversational cancellation and conversational denial, in turn, are two discourse moves that can be performed by actual speakers in real dialogical settings. They impose different constraints on the discourse context in which they can be made and produce different effects. Keeping these distinctions in mind, let us briefly review a few claims made by scholars working on cancellability and deniability.

Lee and Pinker (2010, p. 791), Mazzarella et al., (2018, p. 16), and Mazzarella (2023, p. 220) maintain that deniability requires cancellability. In my view, this statement means that one can successfully deny having intended to imply that p only if the implication that p is diagnostically cancellable. Blome-Tillmann, in turn, considers the following variant of the Principle of Explicit Cancellability construed as a test for conversational implicatures:

If an utterance of P conversationally implicates q in C, then utterances of ‘P, but not Q’ or ‘P, but I don’t mean to imply that Q’ are admissible in C and they cancel the speaker’s commitment to q. (Blome-Tillmann, 2008, p. 157)

In my view, however, this test is too demanding because it requires that a successful cancellation nullifies the speaker’s commitment to the content of what their diagnosed utterance putatively or genuinely implies. I would argue, instead, that the principle under discussion provides a test that enables us to identify implicatures that are not well-entrenched in Jaszczolt’s (2009) sense. Moreover, the cancellation of the speaker’s commitment to the implied content of their utterance is what Steven Oswald (2022, p. 164), following Álvaro Domínguez-Armas and Andrés Soria-Ruiz (2021), considers as a type of denial. He distinguishes, namely, between cancellation construed as the speaker’s denial of the implied content of their utterance and disavowal understood as their denial of having intended to convey it. Cancellation in Oswald’s sense, then, is to be distinguished from conversational cancellation as characterized above.

2.3 Three types of plausible deniability

Let us recall that the point of insinuation is to contribute a certain problematic content to the conversation while maintaining plausible deniability. In other words, the insinuating speaker wants to place themselves in a situation where they, when challenged, can plausibly deny having intended to imply the insinuated content of their utterance. It is worth noting, however, that there are at least three distinct types or, more accurately, aspects of plausibility that pertain to denials: contextual, psychological, and epistemic.

Consider, first, the notion of contextually plausible deniability. Camp (2018, p. 49) argues that speaker S, who denies having intended to imply M(Q) in uttering U in context C, pretends to be in a marginally altered context C’, emerging from a potentially minimal reconstruction of actual context C (Mazzarella, 2023), wherein their utterance might be interpreted as implying an alternative attitude M(Q)’. Contextually plausible deniability, then, can be described in terms of the contextual availability of such an alternative interpretation of the speaker’s utterance, which would fit the interacting agents’ expectations of conversational correctness. As Mazzarella aptly notes:

A full-fledged denial typically comprises not only an explicit withdrawal of the target content (“I didn’t mean to suggest that p”) but also the presentation of an alternative interpretation of the utterance (“I only meant that q”). (Mazzarella, 2023, p. 222)

For instance, Sally’s denial, made in turn (8c), is contextually plausible only if there is a contextually available set of presuppositions that can be used to infer an alternative meaning M(Q)’, which Sally can be taken to indirectly communicate in her utterance of (8a), and which satisfies the currently held expectations of cooperativity (Camp, 2018) or relevance (Oswald, 2022, p. 165).

Psychologically or cognitively plausible denials, in turn, are perlocutionarily successful or at least are not complete perlocutionary failures. For instance, Sally’s denial, made in uttering (8c), is perlocutionary successful if Harry takes it seriously and comes to believe that, appearances to the contrary, Sally has not intended to imply that John and Maria are involved in an affair. Psychologically plausible denials, then, are perceived by the hearers as credible, sincere, and convincing. Psychologically implausible denials, on the other hand, are perceived as bald-faced lies (Camp, 2018, p. 47; cf. Dinges & Zakkou, 2023, p. 392), and ‘routinely trigger an ascription of bad faith to the speaker’ (Oswald, 2022, p. 167). As Mazzarella et al., (2018, p. 16) and Mazzarella (2023, p. 221) argue, plausibility of this sort is not a binary property, but comes in degrees. In this regard, the psychological plausibility of denials echoes the effectiveness of conversational cancellations as discussed by Jaszczolt (2009). Furthermore, it is closely associated with contextual plausibility: the extent to which M(Q)’ is available as an alternative interpretation of the speaker’s utterance is positively correlated with the extent of the hearer’s acceptance of their denial as a sincere conversational move.

In their insightful paper titled ‘On Deniability,’ Alexander Dinges and Julia Zakkou (2023) argue that deniability is an epistemic phenomenon. However, I believe it is more appropriate to consider the epistemic plausibility of denials as merely one aspect, though not the sole one, of a more complex discourse property. From my perspective, this facet of plausibility can be characterized by reference to the epistemic situations of both the speaker and the hearer, or more precisely, in terms of the allocation of the burden of proof between them. Normally, there is a general presumption in favour of the denying speaker, who possesses the ultimate authority over the intentions underlying their utterances. The speaker’s denial is felicitous if it succeeds in shifting the burden of proof from the denying speaker to the potentially resistant hearer who would like to question and rebut its content.¹⁹ It is additionally epistemically plausible if it positions the hearer in a situation where, given the prevailing epistemic standards, they lack the resources to prove the speaker’s insincerity (Camp, 2018, p. 46) or, in other words, are unable to fulfil the commitment imposed on them by the denying speaker. Even though the speaker’s denial is psychologically highly implausible and the hearer firmly believes that the speaker has intended to imply M(Q), the hearer may not be able to prove their stance, and as a result, they may not be in a position to know that the denial is insincere (Dinges & Zakkou, 2023, p. 289). In short, psychologically implausible denials may be epistemically plausible.

A key idea behind this paper is that the colloquial concept of insinuation encompasses a range of discourse phenomena which, upon closer examination, do not form a homogenous class. One method to distinguish among them is to consider the different strategies of indirect communication they utilize and the various aspects of plausible deniability that insinuating speakers aim to achieve and maintain. In the remainder of this paper, I will explore one possible form of the phenomena under discussion: Gricean insinuation, characterized by what I term ‘partially overt’ indirect communication and psychologically plausible deniability.

3 The concept of Gricean insinuation

My purpose in this section is to introduce and motivate the concept of Gricean insinuation, construed as a partially overt act. Roughly, a successful Gricean insinuator conveys certain risky or troublesome content while maintaining psychologically plausible deniability. I argue that to achieve this goal, they feign indifference as to whether their intention to convey this content is recognized, thereby producing what I term the ‘Fake One-Way Mirror Effect’.

In SubSect. 3.1, I introduce the concept of Gricean insinuation and explain what it means for this type of act to be partially overt. In SubSect. 3.2, I use the common ground framework (Stalnaker, 2002, 2014) to discuss examples of the Fake One-Way Mirror Effect. SubSect. 3.3 offers a more formal presentation of this discussion; it is instructive to stress, however, that it is included for the sake of precision only and can be omitted without missing any essential information. SubSect. 3.4 advocates for the idea that the insinuator’s feigned disinterest in whether their intention is recognized serves as a conversational counterpart to the fake one-way mirror. It also shows how the proposed framework can be applied to analyse instances of Gricean insinuation in triadic communicative interactions.

3.1 Gricean insinuation as a partially overt act

As mentioned in the previous section, the proposed concept of Gricean insinuation aims to identify those instances of insinuating that can be characterised as partially overt acts of indirect communication intended to maintain psychologically plausible deniability. I label it ‘Gricean’ because it helps us resolve the puzzle of insinuation outlined in Sect. 1 by elaborating on the communicative nature of acts of insinuating and their off-record status within a broadly Gricean perspective. A central idea of this solution is that at least some acts of insinuating are partially overt, meaning also that they are also partially covert. In essence, an act is partially overt if the intention behind it is intended to be recognized, but is not intended to be recognized as being intended to be recognized. Intriguingly, this idea is briefly mentioned by Kent Bach and Robert M. Harnish’s (1979) in their model of covert collateral acts. They assert, namely, that covert acts ‘succeed (…) only if their intent is not recognized, or at least not recognized as intended to be recognized’. (Bach & Harnish, 1979, p. 101; italics are mine—MW).

An instance of Gricean insinuation, then, is a partially overt act, the structure of which can be described as follows.

(PO) In uttering U, S performs a partially overt act if and only if S intends:

• (i₁) To produce by uttering U a certain response R on the part of H,

• (i₂) To get H to recognize (i₁),

• (i₃) That the fulfilment of (i₂) function as H’s reason for his response R,

• (i₄’) To get H not to recognize (i₂).

Formula (PO) results from modifying (SM₂) as discussed in Sect. 2. Recall that the latter employs the recursive condition (i_n+1) to represent the infinite sequence of intentions underlying the performance of an overt act of speaker meaning. According to (i_n+1), each subsequent intention necessitates the recognition of its immediate predecessor. In contrast, a pivotal element of (PO) is condition (i₄’), which sets limits on the iterative structure of (SM₂). Similar to condition (i₂’) in (C)—presented below again for convenience—condition (i₄’) in (PO) acts as a counter-intention that halts the endless cascade of intentions.

(C) In uttering U, S covertly intends to produce R if and only if S intends:

• (i₁) to produce by uttering U a certain response R on the part of H,

• (i₂’) for H not to recognize (i₁).

Intention (i₁) in (C) is fully covert. In contrast, intention (i₁) in (PO) is not fully covert, since condition (i₂) requires that it is intended to be recognized. However, it is not fully overt either, because condition (i₄’) necessitates that it is not intended to be recognized as being intended to be recognized. In short, intention (i₁) in (PO) is simultaneously partially overt and partially covert. Importantly, indirect communication through partially overt acts so defined cannot be adequately described in terms of conversational implicatures since conversationally implicated contents are overtly meant in the sense specified by (SM₂).

It is noteworthy that (PO) elucidates the common structure of possible counterexamples to Grice’s original three-condition analyses, such as ‘River Rat’ (Schiffer, 1972, pp. 17, 18; cf. Green, 2007, pp. 63, 64) and ‘Giveaway Smile’ (Grice, 1989, pp. 94, 95; cf. Camp, 2018, p. 55), discussed in SubSect. 1.1. Below, I will describe and discuss the latter again, because I believe it serves as an example of Gricean insinuation.

Ann plays bridge against her boss, Peter. She wants him to win. Moreover, Ann wants Peter to know that she wants him to win. Yet, she does not want to be too blunt about her plan. Thus, she adopts a communicative strategy: whenever she has a good hand, she simulates a spontaneous smile of pleasure. This smile is intentionally produced to induce in Peter the belief that Ann has a good hand. Moreover, this intention is intended to be recognized. Nonetheless, Ann wants to conceal her intention to get Peter to recognize that she intends for him to believe that she has a good hand. In other words, she wants Peter to recognize that her simulated smile is produced with intention (i₁) to make him believe that she has a good hand, but she doesn’t want him to believe that this intention is overt and, consistently, that her smile is an attempt to tell something.²⁰ In fact, Ann has the opposite or counter intention of the (i₄’) type: she intends for Peter not to recognize her intention (i₂).

Unlike fully covert acts that satisfy requirements (i₁) and (i₂’) as outlined in (C), Ann’s smile is partially overt. In this regard, it can be considered ‘communicative’ in a limited sense of the word. Ann intends not only to influence Peter’s beliefs with her feigned smile of pleasure but also intends him to recognize that the smile is intended to produce this effect. However, she simultaneously seeks to maintain plausible deniability. That is, she aims to place herself in a position in which she can credibly deny having communicated or told that she has a good hand. More precisely, her strategy is to arrange matters so she can deny having any intention of making Peter realize she intends him to believe that she has a good hand. A fundamental assumption behind this strategy is that we are normally held accountable for our intentional actions. If Ann’s plan succeeds she can reject and disavow any conversational responsibility for communicating this belief, i.e., she is in a position to reliably claim that she has not communicated that she has a good hand.

The instance of insinuation described above is successful if Ann makes Peter believe that she has a good hand, while preserving psychologically plausible deniability. To accomplish this, she intends to place Peter in a position in which he does not believe, or at least doubts,²¹ that her feigned smile—which, as a matter of fact, he perceives as an attempt to influence his beliefs—is actually an instance of Gricean telling, i.e., an act of overt communication. Generally, a speaker who insinuates, as Ann does in the Giveaway Smile scenario, and aims to maintain psychologically plausible deniability, can accomplish this by creating and utilizing what I term the ‘Fake One-Way Mirror Effect’.

3.2 Fake one-way mirror effect, contextual update, and mutual understanding

The above presented notions of Gricean insinuation and the Fake One-Way Mirror Effect can be elaborated within the common ground framework developed by Robert Stalnaker (2002, 2014). In his paper titled ‘Common Ground’, he defines common ground in terms of acceptance.²² He posits:

It is common ground that ϕ in a group if all members accept (for the purposes of the conversation) that ϕ, and all believe that all accept that ϕ, and all believe that all believe that all accept that ϕ, etc. (Stalnaker, 2002, p. 716)

In the same paper, he introduces the concept of common or mutual belief, which he defines as follows:

a proposition ϕ is common belief of a group of believers if and only if all in the group believe that ϕ, all believe that all believe it, all believe that all believe that believe it, etc. (Stalnaker, 2002, p. 704)

Stalnaker regards common belief as a model for common ground, noting that the latter emulates the iterative architecture of the former.²³ For the sake of simplicity, then, I will use the concept of common belief in the remaining part of this section. The key idea behind the proposed model is that the off-record status of an act of Gricean insinuation can be understood by examining the constraints it imposes not only on the iterative structure of the speaker’s intention, as outlined in (PO), but also on the structure of the context affected by their utterance, regardless of whether the latter is defined as common ground or common belief. In this regard, the model of Gricean insinuation I propose differs from Camp’s theory of communicative insinuation, which relies on the Stalnakerian difference between believing and accepting. Camp (2018, p. 56) argues that acts of insinuating change mutual belief without altering common ground.²⁴ In contrast, I contend that the difference between acts of Gricean insinuation and instances of fully-fledged communication is best understood by reference to the unique structure of the contextual update produced by the former. This uniqueness can be further explained by reference to the Fake One-Way Mirror Effect.

Common belief is a dynamic information structure that evolves in response to what Stalnaker refers to as ‘manifest events’. He explains:

A manifest event is something that happens in the environment of the relevant parties that is obviously evident to all. A goat walks into the room, or all of the lights suddenly go out. In such a case, it immediately becomes common knowledge [or common belief] that the event has happened (…). (Stalnaker, 2014, p. 47)

To illustrate, let me consider again the scenario where Sally and Harry are seated on a bench and notice John alighting from a bus across the street. Let’s assume that it is common belief among them that each knows John well and, consequently, can easily and immediately recognize him in any encounter. John’s getting off the bus, then, becomes a manifest event that updates their common belief. This update can be represented as a recursively defined endless cascade of pairs of beliefs: Sally believes that John is getting of the bus and Harry believes that John is getting of the bus, Harry believes that Sally believes that John is getting of the bus and Sally believes that Harry believes that John is getting of the bus, and so on.

Importantly, John’s getting off the bus updates the common belief among Sally and Harry only if they are appropriately epistemically related to each other: Sally sees that Harry sees John, Harry sees that Sally sees John, Sally sees that Harry sees that Sally sees John, and Harry sees that Sally sees that Harry sees John, and so on. Consider, however, two alternative scenarios—let me call them ‘Separate Rooms’ and ‘Fake One-Way Mirror’—which fail to meet this requirement.

In the first scenario, Sally and Harry are observing the same street from different windows in separate rooms. Each sees John getting off the bus, but they cannot see each other. In this case, John’s getting off the bus can no longer serve as a self-manifest event. Instead of updating the common belief among Sally and Harry, it leads to what can be termed, following James J. Lee and Steven Pinker (2010, p. 796), ‘mere shared individual knowledge’—both Sally and Harry believe that John is getting off the bus, but neither realizes that they share this belief.

In the ‘Fake One-Way Mirror’ scenario, Sally and Harry are seated on a bench and notice John getting off a bus. This time, however, they are separated by a framed glass pane, which Harry mistakenly believes to be a one-way mirror. Consequently, he incorrectly assumes that Sally cannot see him and that she only sees her reflection. In reality, Sally can see Harry and knows that he is unaware of the fact that she sees him as he watches her observing the event. Like in the ‘Separate Rooms’ scenario, John’s getting off the bus cannot be referred to as a ‘manifest event’ in the sense outlined by Stalnaker. That is to say, it fails to produce the endless cascade of pairs of beliefs described in one of the previous paragraphs. Still, the update it engenders goes beyond mere shared individual knowledge. Sally believes that Harry believes that John is getting of the bus and Harry believes that Sally believes that John is getting of the bus. Moreover, Sally believes that Harry believes that she believes that John is getting of the bus. However, due to the peculiar epistemic situation that Harry finds himself in, it is not the case that he believes that Sally believes that he believes that John is getting of the bus. In fact, Harry is convinced that Sally cannot see him, and consequently he believes that she does not believe that he believes that John is getting of the bus. The Stalnakerian iterative mechanism makes its first step, that is, it leads us beyond mere shared individual knowledge, and then jams. This is what I term the ‘Fake One-Way Mirror Effect’.

Importantly, the occurrence of this effect places Sally in an epistemically advantageous position. Due to the deceptive situation in which he is in, Harry doesn’t believe that Sally believes that he believes that John is getting of the bus. In reality, however, Sally is fully aware of this situation and, if required, she would take advantage of it.

As Stalnaker (2014, p. 58; cf. Witek, 2019, pp. 177, 178) observes, speech acts are a subtype of manifest events, and their performance updates the common belief among sufficiently competent and appropriately situated conversational participants. To illustrate, let’s consider again the original Sally and Harry scenario where Sally utters (3), provided below for convenience.

• (3) This is the third time this week that I have seen John getting off the bus at the stop near Maria’s place.

By doing so, she directly and literally states that P, where ‘P’ means that ‘Sally has seen John alighting from the bus at the stop near Maria’s place for the third time that week’. Provided Sally and Harry share the same conversational standards, the performance of this statement is a manifest event, which leads to an update in the common belief between them. This update has the Stalnakerian iterative structure: Sally believes that she states that P and Harry believes that Sally states that P, Harry believes that Sally believes that she states that P and Sally believes that Harry believes that she states that P, and so on. Let’s also assume that the statement in question is an act of successful overt communication: Sally utters (3) with an appropriate iterative intention, the structure of which can be represented by (SM₂), and this intention is subsequently recognized by Harry. Importantly, Sally and Harry’s mutual understanding of the iterative intention behind her statement that P can be represented as a recursively defined endless cascade of pairs of beliefs, where the left element of n-th pair displays Sally’s awareness of her intention (i_n), and its right side represents Harry’s recognition of this intention.

Let’s assume that in uttering (3) Sally not only directly and literally states that P, but also intends to insinuate that Q, where ‘Q’ means ‘John and Maria are involved in an affair’. In other words, she utters (3) with an appropriate partially overt intention whose structure is outlined in (PO). If her intention is successful, Sally contributes the proposition that Q to the ongoing conversation while maintaining psychologically plausible deniability. To achieve this, she evokes and utilizes the Fake One-Way Mirror Effect. Specifically, she restricts the iterative structure of the contextual update produced by her act. She believes that she conveys that Q and Harry believes that she conveys that Q. Sally also believes that Harry believes that she conveys that Q and Harry believes that Sally believes that she conveys that Q. Moreover, Sally believes that Harry believes that she believes that she conveys that Q. However, it is not the case that Harry believes that Sally believes that he believes that she conveys that Q. Simultaneously, Sally disrupts the iterative structure of the mutual understanding of her act. She believes that she intends for Harry to believe that Q and Harry believes that Sally intends for him to believe that Q. In other words, Sally is aware of her intention (i₁); Harry recognizes this intention, thereby fulfilling her intention (i₂). Moreover, Sally believes that she has intention (i₂), that is, that she intends for Harry to recognize her intention (i₁). Nevertheless, it is not the case that Harry believes that Sally has intention (i₂); in other words, Sally’s intention (i₄’) is fulfilled.

Specifically, the fulfilment of Sally’s intention (i₄’) means that Harry remains uncertain about, and thereby does not believe, that Sally’s act of implying that Q is intended as an act of telling.²⁵ This uncertainty allows Sally to credibly deny having intended for Harry to recognize her intention to get across the belief that Q. In short, she is in a position to reliably claim that she has not told that Q. To clarify, Harry is uncertain if Sally believes that he believes that she is conveying that Q. He is also unsure whether her intention for him to believe that Q is overt in the sense of being intended to be recognized. Consequently, Sally finds herself able to credibly assert that she did not intend to tell Q.

The concept to be explored further is what serves, in the discussed scenario, as a ‘conversational counterpart’ to the fake one-way mirror. In my view, this question can be adequately addressed and answered by referring to Sally’s feigned disinterest in whether her intention (i₂) is fulfilled. However, before I delve into the details of this hypothesis in SubSect. 3.4, I would like to provide a slightly more technical discussion of the Fake One-Way Mirror Effect. Specifically, in SubSect. 3.3, I will use formal representations of the iterative structure of contextual updates and mutual understandings to precisely identify the locus of the Fake One-Way Mirror Effect. However, these representations are not indispensable to my argument, which primarily focuses on discussing individual scenarios presented in this subsection. Therefore, readers may skip SubSect. 3.3, which is included for the sake of precision, and proceed directly to SubSect. 3.4 without missing any essential information.

3.3 Fake one-way mirror effect: a formal representation

Consider again the scenario where Sally and Harry, who are sitting on a bench on a busy street, notice John alighting from a bus. Provided John’s getting off the bus is a manifest event in the sense outlined by Stalnaker (2014, p. 47), it automatically updates the common belief between Sally and Harry. Following Geurts (2019), this update can be represented by Diagram (9), where ‘E’ means ‘John is getting off the bus’, formula ‘B_Xa’ denotes ‘X believes that a’, ‘S’ refers to Sally, and ‘H’ to Harry.

In the ‘Fake One-Way Mirror’ scenario as presented in SubSect. 3.2, the occurrence of E, which, in this context, can no longer be referred to as a ‘manifest event’, leads to the following change in the beliefs of Sally and Harry:

The cascade of pairs of beliefs is interrupted at step (10c), and the occurrence of E fails to update the common belief between Sally and Harry. This interruption is what I refer to as the ‘Fake One-Way Mirror Effect’.

Consider again the situation in which Sally, after noticing John, utters (3) to state that P, where ‘P’ means ‘Sally has seen John alighting from the bus at the stop near Maria’s place for the third time that week’. The performance of this statement is manifest event E, which leads to an update in the common belief between Sally and Harry. The iterative structure of this update can be depicted by diagram (9). Let’s also assume that the statement in question is an act of successful overt communication, meaning that Sally's iterative intention is recognized by Harry, as depicted in the right column in (11) below, where ‘I_Sa’ stands for ‘S intends to make it the case that a’; specifically, ‘I_SB_HP’ means ‘S intends to make H believe that P’.

The left side of (11a) depicts Sally’s awareness of her intention (i₁), while the right side represents the situation that fulfils her intention (i₂), i.e., Harry’s recognition of Sally’s intention (i₁). Similarly, the left side of (11b) displays Sally’s awareness of her intention (i₂), and its right side describes the fulfilment of her intention (i₄). Generally, each subsequent line n in the infinite sequence depicted in diagram (11) corresponds to a pair of beliefs, one held by Sally and the other by Harry, concerning aspect (i_n) of Sally’s complex communicative intention. In essence, (11) depicts a situation in which Sally and Harry have achieved mutual understanding of the communicative intention underlying her statement that P.

Recall that in uttering (3) Sally not only directly and literally states that P, but also insinuates that Q, where ‘Q’ means ‘John and Maria are involved in an affair’. To assume this, however, is to take her act of indirect speech to produce and utilize the Fake One-Way Mirror Effect. The structure of this effect can be depicted by diagram (10); this time let’s assume that ‘E’ stands for ‘S conveys that Q’.

Examine the right side of (10c), which interrupts the potentially endless cascade of belief pairs. Given the situation of conversational uncertainty he finds himself in, Harry is not sure whether, and therefore does not believe that, Sally holds the belief that he believes that she conveys that Q. In short, he grasps Sally’s insinuated meaning, but is unsure of whether she recognizes or pays attention to this understanding, and as a result, is unsure if it is mutually acknowledged. In short, Sally’s act of indirect speech is off-record, which means that she maintains plausible deniability.

The off-record status of Sally’s act of implying that Q can be further explained by reference to diagram (12). Recall that ‘I_Sa’ stands for ‘S intends to make it the case that a’. Specifically, ‘I_SB_HQ’ means ‘S intends to make H believe that Q’.

The left side of (12a) displays Sally’s awareness of her intention (i₁) to get Harry to believe that Q, while the right side depicts the fulfilment of her intention (i₂), i.e., Harry’s recognition of her intention (i₁). Similarly, the left side of (12b) displays Sally’s awareness of her intention (i₂). However, Harry fails to recognize this intention, and as a result, Sally’s act of indirect speech does not to constitute a successful act of fully-fledged Gricean communication. Indeed, this is what she aims to accomplish. Specifically, the right side of (12b) describes the fulfilment of her counter intention of the (i₄’) type: Harry remains uncertain about, and thereby does not believe, that her act of implying that Q is intended as an act of telling.²⁶ As I argued in the previous subsection, this uncertainty allows Sally to credibly deny having intended for Harry to recognize her intention to convey the belief that Q.

3.4 Conversational fake one-way mirrors

In my view, the key factor in the mechanism producing the conversational variant of the fake one-way mirror is the insinuating speaker’s feigned disinterest in the fulfilment of their intention (i₂). In the particular case under discussion, Sally feigns indifference to whether Harry recognizes her intention (i₁) to get him to believe that Q. More specifically, Sally acts as if she was uninterested in any signs and cues indicating the fulfilment of her intention (i₂). If Sally’s pretence²⁷ is successful, Harry will erroneously believe that she is uninterested in his recognition of intention (i₁), or at least, he will be unsure of whether Sally is attentive to the way he reads her utterance of (3). Consequently, it is not the case that Harry believes that Sally believes that he believes that she conveys that Q. To frame it in terms of intention recognition, Harry is uncertain whether Sally harbours intention (i₂). In summary, her feigned indifference acts as the conversational equivalent to the physical fake one-way mirror. Consequently, it restricts the iterative structures of contextual updates and mutual understandings, as outlined in SubSects. 3.2 and 3.3.

A key idea underlying the proposed analysis of the conversational Fake One-Way Mirror Effect is that, at least as far as non-conventionalized indirect speech is concerned,²⁸ manifesting one’s intention typically involves demonstrating an interest in its fulfilment. By feigning indifference towards whether Harry recognizes her intention (i₁), Sally implies that she lacks intention (i₂) or, at the very least, introduces doubt about her possession of this intention. This ambiguity serves as a conversational variant of the Fake One-Way Mirror Effect. Therefore, if Harry is unsure about Sally’s interest in his recognition of intention (i₁), he also harbours doubts about her intention (i₂) to make him recognize intention (i₁). He recognizes that she intends for him to entertain the belief that Q. In other words, he recognizes the primary objective behind Sally’s utterance, and in this sense, he can be said to understand her insinuation. However, he remains uncertain about whether Sally intends (i₂) for him to realize that she holds intention (i₁). As a result, Sally can plausibly deny having intention (i₂), thereby credibly stating that she hasn’t told that Q. In conclusion, she effectively conveys the proposition that Q while maintaining its off-record status. In other words, Sally successfully leads Harry to understand that, by uttering (3) and stating P, she intends for him to realize her intention to make him believe that Q. However, she simultaneously prevents him from recognizing that this intention is overt.

So far I have analysed the Fake One-Way Mirror Effect as it manifests in the communicative interaction between Sally and Harry. However, it can also occur in triadic interactions, where, in addition to the speaker and the hearer, a third party is involved which becomes the actual addressee of the speaker’s act of insinuation. To illustrate, let us examine the following two scenarios discussed by McGowan (2023) as examples of indirection with respect to intended audience. In the first one, R and her boyfriend S are waiting in line to enter a nightclub. The boy standing behind them is smoking. R is aware that S has no cigarettes. Nevertheless, she speaks to him, saying:

• (13) Oh, how I wish I had a smoke. You got any? (McGowan, 2023, p. 411)

In fact, the intended overhearer of R’ utterance is the smoker behind her. The second scenario involves three co-workers: X, Y, and Z. Z is known for being frugal, though some colleagues label him as stingy. There was a recent promotion within the office, and flowers were sent to the individual who got promoted. X and Y know that they both contributed to the purchase, but Z did not. Concerned that Z might have incorrectly assumed the company covered the cost of the flowers, and with the intention that Z overhears the conversation, X says to Y:

• (14) You contributed for the flowers right? You might not realize this but the company did not pay for them. (McGowan, 2023, p. 412)

On the face of it, the two scenarios in question seem to constitute examples of covert speech as defined by Bach and Harnish. However, as McGowan notes, speaker R is indifferent as to whether the smoker realizes that she utters (13) intending to prompt him to offer her a cigarette. In this respect, then, her act of audience-indirection does not qualify as a covert act in the sense specified by (C). Similarly, I contend that the face-protecting plan behind speaker X’s utterance of (14) does not require, for its success, that speaker Z fails to recognize X’s intention to encourage him to contribute to the purchase. Therefore, unlike speaker T’s utterance of (4) covertly addressed to Nora, R’s utterance of (13) and X’s utterance of (14), as examples of audience-indirection, do not necessarily need to fulfil the criteria set by (C). Does this mean that they are instances of Gricean insinuation as defined by (PO)? I do not believe so. Instead, they occupy a grey area between fully covert and partially overt acts.

However, consider slightly modified versions of the two scenarios under discussion. Assume, namely, that R, S, and the smoker are not waiting in line but standing in a triangle, and as a result, can see each other. Specifically, each of them is able to perceive, process, and respond to non-verbal cues produced by the other participants. In particular, they have access to the cues that normally signal the speaker’s interest in whether their intentions are recognized.²⁹R, not knowing the smoker well and being unsure if she can address him directly, addresses her utterance of (13) to S while feigning indifference as to whether the smoking boy understands that he is, in fact, its actual recipient. It’s important to emphasize that, contrary to the original scenario presented by McGowan, in the revised version, R’s lack of concern about whether the smoker realizes her intention is actually directed at him is not genuine but feigned. This time, R’s utterance qualifies as an instance of Gricean insinuation. It fulfils the conditions set by (PO): R intends (i₁) to get the smoker to offer her a cigarette and, in situating her utterance in this specific context (see Mey, 2001; cf. Capone, 2005; Jaszczolt, 2009) she intends (i₂) for him to recognize her intention (i₁). At the same time, R pretends to be uninterested in whether the smoker recognizes her intention (i₁). In short, R uses her feigned indifference to produce the conversational Fake One-Way Mirror Effect.³⁰ Similarly, let’s assume that X utters (14) during a trialogue between her, Y and Z. I believe that her utterance can be qualified as an instance of Gricean insinuation, meeting the requirements set by (PO). To maintain psychologically plausible deniability, X feigns disinterest in whether Z recognizes her intention (i₁), thereby producing the conversational Fake One-Way Mirror Effect.

4 Conclusions

In this paper, I have developed the notion of Gricean insinuation and used it to describe and explain several examples of insinuation, the common feature of which is that they are intended to convey certain potentially problematic propositions while preserving psychologically plausible deniability. In my view, Sally uttering (3) to suggest that John and Maria are involved in an affair, speaker R uttering (13) to prompt the smoking boy to offer her a cigarette, and speaker X uttering (14) to encourage Z to contribute to the purchase of flowers, perform partially overt acts of indirect communication. In short, the three utterances under discussion qualify as instances of Gricean insinuation. Being partially overt, these acts are communicative in a broader sense of the term. At the same time, they are partially covert, meaning that their insinuated meanings have the off-record status. Sally, speaker R, and speaker X, when challenged, are in a position to credibly deny having intended to imply what they indeed succeed in conveying. Concurrently, Harry, speaker S, and speaker T have an option to ignore what has been insinuated by their interlocutors. Deniability and ignorability are two sides of the same coin (Bell, 1997, pp. 41–43). Overall, to accomplish these two objectives—namely, to communicate something while maintaining psychologically plausible deniability—Sally, speaker R, and speaker X produce what I call the Fake One-Way Mirror Effect.

It is worth stressing that in describing instances of Gricean insinuation as acts of partially overt communication, I do not intend to suggest that insinuation is a subspecies of speaker-meaning, nor do I imply that it is a type of illocutionary act. Rather, the take-home message of this paper is that explaining the communicative nature of insinuation requires us to revise and extend the traditional Gricean model of meaning-making.³¹

In summary, the framework outlined in Sect. 3 offers a solution to the puzzle of insinuation introduced in Sect. 1, focusing specifically on instances that aim to preserve psychologically plausible deniability, as described in Sect. 2. From my perspective, another scenario that fits this treatment involves the estate agent addressing his utterance of (2) to a couple of potential buyers. In my view, it is at least arguable that by framing her comment as a thoughtful and positive suggestion (Camp, 2018, p. 43), the agent aims for the couple to fail to recognize not so much the real motive behind it, but its overt nature.

The proposed model, however, has certain limitations. First, note that the indirect act of offering a bribe by the speeding driver in utterance (1) cannot be classified as an instance of Gricean insinuation. When challenged, the driver can deny the intention to offer a bribe indirectly, but such a denial is likely to be viewed as a blatant lie. Thus, the deniability he achieves is not psychologically plausible. Nonetheless, the officer and other relevant social agents lack the means to conclusively prove the driver’s denial is insincere and made in bad faith. In essence, his deniability, while not successful in perlocutionary terms, remains epistemically plausible. To address this and similar phenomena, I believe a distinct model needs to be developed, one that utilizes the Austinian framework to depict insinuating as a procedure-governed communicative practice (Witek, in preparation). Secondly, at present, I have little to say about the particular cues that normally signal the speaker’s interest in whether their intentions are recognized, let alone how the insinuating speaker refrains from producing them in order to engender the Fake One-Way Mirror Effect. In my view, elaborating on this topic requires an appropriately designed empirical study. The completion of these tasks, however, is reserved for a future occasion.