What’s the linguistic meaning of delusional utterances? Speech act theory as a tool for understanding delusions

ABSTRACT

Delusions have traditionally been considered the hallmark of mental illness, and their conception, diagnosis and treatment raise many of the fundamental conceptual and practical questions of psychopathology. One of these fundamental questions is whether delusions are understandable. In this paper, we propose to consider the question of understandability of delusions from a philosophy of language perspective. For this purpose, we frame the question of how delusions can be understood as a question about the meaning of delusional utterances. Accordingly, we ask: “what meaning(s) can delusional utterances possibly have?”. We argue that in the current literature, there is a standard approach to the meaning of delusional utterances, namely the descriptive account which assumes that a delusional utterance “p” means that p is the case. Drawing on Speech Act Theory, we argue that solely relying on the descriptive account disregards essential ways of how linguistic meaning is constituted. Further, we show that Speech Act Theory can prove a helpful addition to the theoretical and clinical “toolbox” used for attempting to understand delusional utterances. This, we believe, may address some of the theoretical and clinical shortcomings of using only the currently predominant descriptive account.

KEYWORDS:

• Delusions
• meaning
• speech act theory
• philosophy of language
• philosophy of psychiatry
• philosophy of psychopathology

The paradox disappears only if we make a radical break with the idea that language always functions in one way, always serves the same purpose: to convey thoughts – which may be about houses, pains, good and evil, or whatever.

— Ludwig Wittgenstein, Philosophical Investigations, (Wittgenstein, 2009: §304)

Delusions have traditionally been considered the hallmark of mental illness, and their conception, diagnosis and treatment raise many of the fundamental conceptual and practical questions of psychopathology. Over the decades, these questions have produced a host of literature on delusions from various disciplines ranging from neuroscience to clinical psychiatry, from psychotherapy to philosophy. This resulted in a great variety of often conflicting theoretical accounts as well as clinical approaches. The conceptual riddle and clinical challenge posed by delusions is thus far from being settled.

Regarding the conception of delusions, let us – necessarily very sketchily—, summarize that much recent work analyses delusions in terms of propositional attitudes. Doxastic accounts conceive of delusions as beliefs (e.g., Bortolotti, 2009; Davies & Coltheart, 2000) whereas non-doxastic accounts view them in terms of other propositional attitudes such as acts of imaginations (Currie, 2000) or acceptances (Frankish, 2012). The current clinical conceptions used for diagnostic purposes follow doxastic accounts and define delusions as a “belief that is demonstrably untrue” (World Health Organization, 2018: MB26.0) or a “false belief based on incorrect inference about external reality that is firmly held” (American Psychiatric Association, 2013, p. 819).¹ Phenomenological approaches, on the other hand, object to this propositional understanding of delusions arguing that the propositional attitudes in question are only secondary to more fundamental changes in someone’s existential orientation (e.g., Feyaerts et al., 2021; Fuchs, 2020; Sass et al., 2013). On all accounts, the question of interpersonal and intersubjective understandability arises (Eilan, 2000; Fuchs, 2020; Fulford & Thornton, 2017; Jeppsson, 2021; Van Duppen & Sips, 2018). Any account, it seems, has to negotiate between understanding and utter strangeness as Naomi Eilan (2000) influentially put it.

In this paper, we frame the question of understandability as a question about the meaning of delusional utterances and ask: “what meaning(s) can delusional utterances possibly have?”. Our focus on delusional utterances is motivated by theoretical as well as clinical considerations: Firstly, delusional utterances are often the first interpersonal manifestation of delusions and the clinician or indeed anybody engaging in conversation with a person experiencing delusions has to respond to them. Also, most theoretical analyses of delusions use delusional utterances as examples both for developing as well as for testing their conceptions. In light of their clinical as well as theoretical relevance, it is surprising that delusional utterances have hardly ever been addressed in their own right in the philosophical literature. This is even more true seeing that many answers to the conceptual question are based on preconceived notions of what delusional utterances mean (e.g., Berrios, 1991; cf., Fulford & Thornton, 2017). While there is recent work that deals with the meaning of delusions in general (Deamer & Wilkinson, 2021; Hinzen et al., 2016; Radden, 2010; Ritunnano & Bortolotti, 2021; Ritunnano et al., 2022), the question of the linguistic meaning of delusional utterances has been overlooked so far. Our paper attempts to remedy this gap.

After spelling out assumptions and working definitions (§1), we analyze the way in which meaning is ascribed to delusional utterances on the currently predominant descriptive account (§2). We show a relevant problem for this account (§3) before re-analyzing possible meaning(s) drawing on Speech Act Theory (§4) as proposed by J.L. Austin (§4.1.) and others (§4.2.). In the main part of our paper (§5), then, we show that Speech Act Theory provides us with conceptual resources to complement a descriptive reading of delusional utterances and explain how this can help us make progress on understanding delusional utterances (§§5.1., 5.2.). We discuss objections to this approach (§6) before concluding with tentative remarks on conceptual and clinical implications (§7). Importantly, we don’t wish to argue for any definitive answer of what delusional utterances mean. Instead, we suggest that turning to Speech Act Theory allows us to enrich the “toolbox” which we have at our disposal when attempting to understand delusional utterances both theoretically as well as when interacting with people experiencing delusions.

1. Introductory Remarks

Before getting started on analyzing possible meanings of delusional utterances, some assumptions and working definitions need to be spelled out.

First of all, what are delusional utterances anyways? By “delusional utterance” we understand an utterance produced by a person with delusions which has as its content the content of the delusion. As a result, what counts as a delusional utterance ultimately depends on the contested question of the nature of delusions. Yet, even if we had a clear theoretical conception of delusions, in most real-world situations, an element of doubt would remain when we evaluate an actual utterance as delusional (cf., Wilkinson, 2020). So, for the purpose of our analysis, we do not commit to any specific definition, but just assume that utterances such as “I am being persecuted” and “I am God”—if uttered in a context such that in virtue of them, we can ascribe delusions to the person uttering them (ibid.)—are delusional utterances.

Second, our approach is shaped by two fundamental assumptions:

A1 Delusional utterances have (some sort of) meaning.

A2 Persons suffering from delusions and other people share the same language.²

The present paper aims at a clarification of what (A1) could amount to.

Third, some researchers have cashed out (A1) in terms of subjective meaning. Understanding the meaning of a delusional utterance, then, is taken as understanding the meaningfulness of what’s expressed in the broader context of the utterer’s life and illness (Ritunnano & Bortolotti, 2021; Ritunnano et al., 2022). For instance, someone’s saying “I am God” may be understood as a counter reaction to feelings of powerlessness in response to a distressing situation in this person’s life (e.g., Fulford & Thornton, 2017, p. 2). If this interpretation is correct, it may show us where the delusional utterance “is coming from” and provides an embedding in the biographical context of this person. Such approaches usually focus on the meaning of delusions—as opposed to the meaning of delusional utterances. We, too, believe this to be of crucial importance. However, we think that such an interpretation does not cover all relevant aspects of (A1), since we do not yet seem to be justified in saying that we know what the utterance “I am God” means. Our focus in this paper, then, is on the linguistic meaning of delusional utterances. In summary, we’re hoping to enhance intersubjective understanding by improving our understanding of the linguistic meaning of a person’s delusional utterances.

2. The Descriptive Account

Given the current theoretical and clinical dominance of doxastic conceptions of delusions, we take the doxastic approach as our starting point for analyzing the meaning of delusional utterances.

Now, it is important to notice that the most straight-forward route to a theory which grasps delusions as false beliefs rests on a specific assumption about the meaning of delusional utterances. While we want to stay neutral on the metaphysics of beliefs (cf., Bortolotti, 2009; Quilty Dunn & Mandelbaum, 2018; Schwitzgebel, 2002), it’s plausible on any account that thinking that a person holds a false belief must be grounded on behavioral observations. Among those observations an utterance of the believer with false propositional content will occupy a central position. Yet, thinking that an utterance has false content presupposes knowing what the utterance means. Seen from this perspective, a natural account of the meaning of delusional utterances is thus the descriptive account (DA).

(DA) A delusional utterance “p” means that p is the case.

According to (DA), delusional utterances have their literal content as their meaning. So, delusional utterances such as

(1) “I am God.”

or

(2) “I am being persecuted.”

mean that the speaker is God or is being persecuted respectively. Moreover, what a speaker says is the case – neglecting the possibility of lie and pretense – generally indicates what they believe to be the case (Bortolotti, 2009, p. 12; Moran, 2001). It is assumed that people uttering (1) or (2) believe that what’s expressed by (1) or (2) actually obtains. Thus, there is an intuitive connection between (DA) and the prevailing doxastic account which understands delusions as false beliefs.

Therefore, ascribing false beliefs to sufferers of delusions on the basis of their utterances naturally lends itself to the idea that delusional utterances mean that their literal content obtains. Given the centrality of the doxastic and other propositional conceptions, we think of the descriptive account as a diagnostically integral part of any conceptual and clinical toolbox dealing with delusions. However, this doesn’t mean that we should only have the descriptive account in our toolbox. This is especially so since we identify a problem with (DA)—the “dead-end problem”.

3. Delusion and Disagreement

We believe that (DA) can lead to what we call “the dead-end problem”. Consider Ms. D. (adapted from Küchenhoff, 2019: 57f.):

Ms. D has a malignant tumor. The doctors are worried and want to get her to have the tumor treated. Ms. D develops an unconditional and delusional attitude against this help and says that any treatment is nonsensical, because she is not ill³ at all, but rather the offspring of royal parents with huge power.

Simplified, we can say that in this case we are dealing with the following delusional utterance:

(3) ”I am not ill at all.”

How are we to understand this utterance based on (DA)? An answer can be sketched as follows: When Ms. D says “I am not ill at all” she means, that she is not ill at all. Thus, she expresses her belief about how things are in the world. She describes certain public facts and is therefore making a rather typical descriptive statement. If that’s the case, Ms. D believes something that is false – and this with great stubbornness. The fact that she has a tumor can be clearly proven and, moreover, she herself has no conclusive evidence for her dissenting belief. Importantly, this means that we disagree with her about whether she has in fact a tumor. This has grave consequences for the communicative situation in which the delusional utterance is made. If we assume the falsity of the delusional utterance according to (DA), then the conversation is blocked at least with respect to this utterance. Moreover, unlike political or other forms of potentially “deep” disagreement, the areas of life we disagree about with someone experiencing delusions are often unusual, i.e., they are not common matters of contention. Accordingly, we find ourselves in a situation in which mutual communicative understanding becomes problematic. Understanding delusional utterances solely descriptively thus can lead into a communicative dead-end.

One way to deal with the dead-end problem is to consider the possibility that we may not yet have understood the utterance in question correctly and that we should therefore consider alternative linguistic meanings of the utterance in question besides (DA). Such an approach can be based on reflections by philosopher of language H. P. Grice (1989). Grice shows that utterances which seem to violate basic conversational maxims do not necessarily reflect that these maxims have been intentionally violated or cannot be observed because of a deficit. Instead, it may be the case that these conversational maxims have been exploited to implicate something conversationally with the utterance in question (Grice, 1989: 30f.). For example, with ironic statements we exploit the maxim of quality, which dictates that our utterances should be true. By saying something obviously false, we can indicate in certain situations that we mean exactly the opposite – which also means that the maxim in question, contrary to first appearances, is actually fulfilled. If we were to rely strictly on what is actually said (i.e., something obviously false), we would miss the relevant content of such an utterance (cf., Grice, 1989, p. 34).

Our basic attitude toward delusional utterances is analogous. In the context of a direct, literal reading of delusional utterances (i.e., a descriptive reading), certain conversational maxims can appear to be violated; our conversation is blocked as a result. Therefore, we consider the possibility that these maxims are indeed recognized, but that we have not yet adequately understood the utterance. As a result, it is of central relevance to consider alternative meanings of the utterance.

Importantly, we’re not suggesting that hitting a communicative dead-end always implies that the literal interpretation must be wrong and that some non-literal interpretation must be true. What we suggest, rather, is that break-downs in conversation should motivate us to at least look for alternative meanings. Of course, these alternatives must ultimately be plausible and independently justifiable. Our approach is thus procedural; we may not be able to recognize the meaning of a delusional utterance at first glance, but in an intersubjective process in which more than one possible meaning is considered alternative interpretations can be substantiated.⁴

In the next sections (§§4,5), we will show that Speech Act Theory is a helpful tool for considering other than descriptive meanings of delusional utterances. And we will argue that such an interpretive procedure is not artificial, but indeed part and parcel of our everyday conversational behavior.

4. Speech Act Theory

Speech Act Theory, originally introduced by Austin (1962), can help us see which other kinds of meaning we can attribute to delusional utterances such as (1), (2) or (3). In this section, we briefly give an overview over Austin’s original conception of Speech Act Theory (§4.1.) and some recent developments (§4.2.).

4.1. Austin’s How to Do Things with Words

Speech Act Theory aims essentially at counteracting a merely descriptive understanding of linguistic utterances by showing that language can be used to do quite different things. At the heart of Austin’s (1962) conception of Speech Act Theory lays the notion of the illocutionary act. The illocutionary act denotes the communicative purpose or point pursued by an utterance. It is also defined by a distinction from the so called locutionary act, which describes what is said with an utterance – that is, what is spoken about and what is said about it. The locutionary act focusses on the descriptive part of an utterance, whereas – roughly speaking – the illocutionary act denotes the performative part of an utterance (ibid.: 94ff.).

Let us illustrate this distinction with the following utterance: “My cat has little patience”. The locutionary act of this utterance consists in the fact that I speak about my cat and say something about her, namely that she has little patience. Concerning the illocutionary act of the utterance, several come into question. It is conceivable that I inform my interlocutor with this utterance in a casual conversation about the “personality” of my cat. But I could, e.g., also warn my interlocutor, who is currently petting my cat rather impetuously. Accordingly, different illocutionary acts can be performed with the same locution. Which illocutionary act is performed depends essentially on the context of an utterance.

It’s one of Austin’s central findings that the context relevant to the execution of illocutionary acts can be described in terms of conventional procedures (ibid.: 121 and passim). In other words, whether an illocutionary act is performed can be decided based on whether certain conventions have been followed. This conventionality of illocutionary acts is also reflected in the fact that they can usually be performed in a particularly explicit way. For example, I can explicitly warn my counterpart by saying “I warn you, my cat has little patience – and may soon smack you!”

However, not all effects of utterances are of this conventional kind. For example, I may have scared my interlocutor with the utterance about my cat, but I cannot do this explicitly. It would be rather nonsensical to say: “I scare you of my cat!”.⁵ Based on this observation, Austin (ibid.: 99) makes yet another distinction; he names these non-conventional effects of utterances perlocutionary acts.

4.2. Speech Act Theory since Austin

Since the publication of How to Do Things with Words, Speech Act Theory has undergone many developments and refinements (e.g., Brandom, 1994; Grice, 1989; Lewis, 1979; Stalnaker, 1978). For instance, John Searle (1969) has proposed a revised taxonomy of illocutionary acts. It will give us essential clues as to what kinds of illocutionary acts there are, how they are characterized, and what is required for their accomplishment. Searle established his taxonomy based on clearly specified classification criteria. The three most important criteria are (1.) the illocutionary point, (2.) the direction of fit, and (3.) the psychological state expressed by the speech act. Searle understands the illocutionary point to be the communicative intention pursued with an utterance. The direction of fit indicates the type of correspondence between the words and the facts. If the words are to represent the facts (as in a classical descriptive statement), then the direction of fit runs from the words to the world. If, on the other hand, the facts are to change in the way indicated by the words (e.g., by means of an order or a wish), then the direction of fit runs from the world to the words. As a third criteria Searle refers to the psychological state expressed by the utterance (e.g., belief, intention, or desire).

With this framework, Searle distinguishes between illocutionary acts as given in the following table (cf., Searle, 1979):

Besides this classification, Speech Act Theory has been developed in many other directions. For instance, one thriving literature looks at how speech acts modify the “common ground”, i.e., the set of beliefs, assumptions or pieces of knowledge shared between the participants of a conversation (Clark, 1996; Haslanger, 2012; Langton, 2013; Lewis, 1979; Stalnaker, 2002; Swanson, 2022). Another related strand of Speech Act Theory applies insights from Austin and others to topics of social and political philosophy such as pornography or slurs (Bauer, 2015; Langton, 1993; McDonald, 2021).

5. Delusional Utterances as Speech Acts

Having introduced the conceptual resources of Speech Act Theory we can now reformulate a – potentially stronger – version of the descriptive account:

(DA*) The meaning of a delusional utterance is exhausted in the execution of an assertive.

In this section we will argue against (DA*). Evidently, this leaves something weaker intact, namely that the meaning of delusional utterances includes or even is primarily given by an assertive. Yet often it is assumed that either we accept (DA*) or we reject the idea that delusional utterances have meaning at all, i.e., reject (A1). For instance, psychiatrist and conceptual historian German E. Berrios claims that

Delusions are likely to be empty speech acts, whose informational content refers to neither world nor self. They are not the symbolic expression of anything. Its “content” is but a random fragment of information “trapped” in the very moment the delusion becomes crystallized. (Berrios, 1991, p. 12 cited in Thornton, 2008, emphasis added)

In this quote Berrios neglects – mistakenly, we believe – the possibility that utterances might not be the “symbolic expression of anything” and still have meaning. That meaning overflows what is said, is, of course, the central tenet of Speech Act Theory. We can do different things with utterances, i.e., there are different kinds of illocutionary acts (Austin, 1962: 98ff.). Not all nonassertive speech acts have to be “empty”. In this section, we show that delusional utterances – exemplified through (1), (2) and (3)—can be plausibly understood as instances of different illocutionary acts. For convenience we rely on Searle’s (1969) taxonomy in doing so. Other ways of conceptualizing these speech acts are also possible

5.1. “I am God” and “I’m being persecuted”

If we understand utterances such as (1) or (2) as assertives only, they are assumed to represent certain facts. The direction of fit thus runs from the words to the world. But, as discussed above (§3), with this an analysis alone we might face the dead-end problem. The tools of Speech Act Theory, however, make it possible to conceive of such utterances in terms of different illocutionary acts. We will look at utterances (1) and (2) tentatively in terms of directives and expressives. We are not proposing any conclusive interpretation. Our aim is rather to show how delusional utterances can be interpreted using the conceptual tools of Speech Act Theory. Importantly, interpretations of this kind are commonplace in everyday conversations. Thus, if the interpretations in the following appear convoluted, this is because we are trying to make inferences that we usually undertake in an unconscious and automatic fashion explicit. The focus here is on a conceptual level and thus on the justification of such interpretations from a philosophy of language point of view.

When (1) is uttered in a situation of intense personal distress understanding it as a directive allows us to see it as the expression of a wish to be able to divert the sources of stress and regain power over the situation (cf., Fulford & Thornton, 2017, p. 2). Thus, we no longer understand the utterance as primarily in the service of describing the world; rather, it expresses how the world should be in the eyes of the speaker. Thus understood, the utterance “I am God” expresses a wish by depicting a situation in which this wish is actually fulfilled. Accordingly, we also think of the direction of fit differently, it now runs from the world to the words, while the propositional content remains the same. In short, if we understand (1) as directive, it rather says I shall be God, instead of I am God.

With respect to (2) similar things apply. Besides an assertion, (2) can be understood as a directive expressing a wish – e.g., a wish to be safe or to be helped. The addressee of this directive can be thought of as the person whoever is actually addressed. After all, if you feel persecuted it is reasonable to ask for help where you can get it. With respect to the propositional content some rearranging has to be done: clearly the person does not express a wish that she should be persecuted. Instead, it seems more plausible to assume that she wishes to not be persecuted.⁶

Additionally, we can see (1) and (2) as expressives by which a certain mental state is manifested. There are multiple theories of such expressives (e.g., Blackburn, 2010; Price, 2013; Simpson, 2020). A strong theory claims that they have no direction of fit. Consider “ouch!”—a scream in reaction to pain. Expressive utterances are like such screams in that they are not descriptions – they do not represent facts, nor do they create an image of the world that is desired or to be brought about. Rather, the truth of certain propositions is presupposed – here, that I am in pain—, toward which one has a specific psychological attitude that is expressed – namely that it is painful. A weaker theory, on the other hand, would leave it open for expressives to additionally encompass descriptive aspects, such as describing the utterer’s mental life. Accordingly, “I’m in pain” is different from “ouch!” in that it is both a manifestation of pain and some kind of description of the person’s mental state.⁷

Resultantly, “I am God” can be seen to express feelings of empowerment, joy but also fear or even terror caused by a tremendous responsibility and a loss of humanity. Evidently, then, there are multiple expressives potentially encompassed in (1)’s meaning. Which one it is, whether it is only one alone and whether we have access to these supposed facts will depend on the context of the utterance and more background information. Also (2) can be straight-forwardly seen as expressive by which the speaker’s fear is manifested. Indeed, doing so seems like an obvious reaction if one encountered (2) in an actual clinical context.

By this point, it should have become clear that the interpretations we have outlined are neither novel nor flamboyant. Rather, the framework of Speech Act Theory enables us to pay explicit attention to aspects of meaning (i.e., illocutionary acts) other than descriptive ones (cf., Lycan, 2019). So, our approach conceptualizes interpretations that might also seem intuitively obvious and, by doing so, puts them on safer ground.

Importantly, interpretations in terms of other illocutionary acts, as we have presented, can counteract the dead-end problem. If it becomes apparent that a delusional utterance has directive aspects, for example, we might respond to these aspects: we can meet the person’s wish, adjust to it, or address it. So, the conversation is less blocked than on a merely descriptive view.

In the next section, we want to further elaborate on our interpretations by turning again to the case of Ms. D. So far, we have been concerned to show that interpretations in terms of other illocutionary acts than assertives can be plausible. In the next subsection we want to demonstrate that such non-descriptive interpretations can be substantiated by evidence. Derivations of this kind are not arbitrary but obey patterns of everyday linguistic interpretation.

5.2 “I’m not ill”

It is fair to say that delusional utterances at first glance take the form of mere assertives—(DA*) takes this fully into account. However, as Searle (1979: 30ff.) points out, we perform a variety of so-called indirect speech acts in our everyday communication. In such speech acts “one illocutionary act is performed indirectly by way of performing another” (ibid.: 31). To take an example from Searle: Someone responds to the suggestion to go the movies tonight with the utterance “I have to study for an exam”. If we consider this utterance outside of its context, it appears to us as a mere assertive; the person seems to convey certain information to us. However, it should not be difficult for any competent speaker to see this utterance in the context presented also as a rejection of the suggestions to go to the movies. We can note that the direct speech act is performed by making an utterance in its literal meaning, whereas in the indirect speech act the utterance is taken non-literally (ibid.: 34f.). This non-literal treatment can also change the perception of the propositional content of the utterance – as we have seen above (§5.1) with the example of “I’m being persecuted” as directive.

Analogously, we can now imagine that with delusional utterances speech acts are performed in an indirect way. How do we find out if indirect speech acts are being performed, and if so, which ones? Searle offers us schematic hints to understand such interpretations, which are often employed automatically in everyday conversations. Such interpretations have three essential ingredients: 1) conversational principles, 2) contextual information, and 3) information about the conditions of felicitous performance of illocutionary acts (ibid.: 32).

Let us now look at what such an interpretation might amount to in the case of Ms. D’s utterance “I am not ill” (3). For this purpose, we will do well to look at more information about her case (again adapting from Küchenhoff, 2019: 57f.):

Time passes, the threat of the tumor becomes greater and the attitude of the therapists and doctors more energetic and urgently caring. They are not successful with this. Only when Ms. D can shake off this influence, because the treating therapists give up, she considers participating in further diagnostic and therapeutic measures at all.

How does this background influence our interpretation of (3)? An inference of an indirect speech act could now be guided by the following steps:

1. We first note that Ms. D seems to violate conversational principles (Grice, 1989: 26ff.). In particular, she seems to say something with her utterance for which she lacks adequate reasons: the doctors can convincingly demonstrate that she is ill, whereas Ms. D cannot present evidence as to why she is not ill. If there is no suspicion that she intentionally violates the principles of conversation, it is natural to assume that more is meant by her utterance. As already mentioned above (§3), the apparent falsity of delusional utterances on a descriptive reading can give us a prima facie reason to look for alternative interpretations. Thus, we pursue the idea that with the utterance an additional indirect speech act is performed.

2. If we now ask ourselves what speech act might have been performed indirectly, contextual information comes into play. We know that the cancer diagnosis is emotionally stressful for Ms. D, and that she rejects any further treatment steps. We also learn that her symptoms improve when the doctors let go of her. Furthermore, it is evident from Ms. D’s biography that she has often struggled with intrusive persons (cf., Küchenhoff, 2019: 57f.). Against this background, it seems plausible to assume that Ms. D has a wish to be left alone and not to be pushed. If we assume that this wish could also be relevant in the delusional utterance of Ms. D, we accordingly become aware of the class of directives, since this class is characterized by the expression of a wish (cf., Table 1).

3. Finally, we have to consider whether her actual utterance “I am not ill” points to the fulfillment of certain conditions that indicate a felicitous performance of a directive. Consider the statement “You are standing on my foot”. Saying this, I can perform an indirect directive. I am asking you, or even ordering you, to move away by giving you a good reason to do so. Referring to good reasons is one way of performing an indirect directive: “S can make an indirect directive by either stating that or asking whether there are good or overriding reasons for doing A” (Searle, 1979, p. 45; cf., 38, 43ff.). Looking at the utterance of Ms. D, we see that in fact she gives good reasons for leaving her alone by stating that she is not ill. For there would be no reason for the doctors to treat her if she wasn’t ill.

Table 1. Speech acts according to Searle (1969).

Type of illocutionary act	Description
Assertives (e.g., statements, descriptions, swears)	1. Illocutionary point: commit the speaker to the truth of the expressed proposition
	2. Direction of fit: words to world
	3. Psychological state: belief
Directives (e.g., orders, suggestions, insistences)	1. Illocutionary point: getting the interlocutor to do something
	2. Direction of fit: world to words
	3. Psychological state: wish
Commissives (e.g., promises, guarantees, agreements)	1. Illocutionary point: committing the speaker to a future action
	2. Direction of fit: world to words
	3. Psychological state: intention
Expressives (e.g., apologies, congratulations, condolences)	1. Illocutionary point: expression of a psychological state
	2. Direction of fit: none
	3. Psychological state: diverse
Declarations (e.g., appointments, promotions, nominations)	1. Illocutionary point: bringing about institutional facts
	2. Direction of fit: both
	3. Psychological state: none

These three points exemplify how pieces of evidence can be used to substantiate the interpretation of an utterance like (3) as an indirect directive. Thus, non-descriptive interpretations of delusional utterances are not arbitrary, but can be differentially evaluated with the resources of everyday communication.

Note that we’re not giving an algorithmic rule about how to interpret delusional utterances in actual clinical practice. Given the deeply contextualized and individualized nature of mental health as well as of psychiatric practice, any generalization about the adequate interpretation of delusional utterances wouldn’t be prudent enough. Our aim here is to propose conceptual tools to complement a merely descriptive understanding of delusional utterances. By this we want to underline a clinical attitude which does not understand delusional utterances readily as expressions of false and thus to some extent deficient beliefs – but which strives for an interpretation that can possibly bring to light further communicative relevance of these utterances.

It’s time to summarize. Our considerations of the utterances “I am God” and “I’m being persecuted” (§5.1) have shown the extent to which the framework of Speech Act Theory can help us pay explicit attention to aspects of meaning other than descriptive ones. As a result, we have come across interpretations that have a certain intuitive plausibility and, moreover, counteract the dead-end problem. Our in-depth consideration of indirect speech acts based on the case of Ms. D (§5.2.) has, moreover, exemplified that such interpretations can be substantiated by procedures that underlie our everyday communication but in most cases are not consciously gone through.

6. Objections

Before we conclude we want to address two objections.

6.1. (DA) is Necessitated by Evidence

For starters, you may think that we cannot suspend (DA), as we propose, because it is mandated by empirical findings that point to deficits in perception, judgment, or related capacities. If we must assume that some delusional utterances are indeed based on deficient perception or the like, we must understand these utterances as false.

To this we have two responses. First, empirical evidence on whether and to what extent delusional persons actually do have systematic deficits in the relevant capacities is heterogenous. For instance, it has been pointed out that differences in perception can be explained by shifted priors in a predictive coding framework (Williams & Montagnese, 2020). Importantly, however, a different distribution of priors does not alone constitute a deficit. Indeed, this may allow people with schizophrenia to be even better at certain cognitive tasks, e.g., recognizing patterns (e.g., Teufel et al., 2015). Moreover, the identification of psychological deficits essentially depends on delusional utterances actually being the result from the supposedly impaired faculties. However, whether the delusional utterance appears to us as a deviation in epistemic terms depends on its meaning. In that sense, assuming (DA) will be in many cases a precondition for ascribing a deficit in the first place. Therefore, without denying the significance of empirical research confirming perception- and judgment-deficits, we are skeptical that empirical research alone can vouch for the correctness of any conceptions about the meaning of delusional utterances.

But even if it turned out that delusional utterances are influenced by deficits in the related capacities – and this is our second point – this does not rule out the possibility that these utterances have other aspects of meaning that need to be considered and which are possibly obscured by (DA)—and most certainly obscured by (DA*). We don’t deny the possibility of deficits explaining delusional beliefs and delusional utterances. Our point is simply that sometimes it will be useful to avoid immediately ascribing a deficit and instead focus on other aspects of the linguistic meaning of the delusional utterance. These other aspects could be, for instance, important for therapeutical progress. The conceptual resources provided by Speech Act Theory can help us to do so.

6.2. Suspending (DA) is a Moral Failing

Next, you may point out that to understand delusional utterances in a descriptive sense also means to take deluded individuals seriously and treat them as full agents (Bortolotti, 2009). Suspending the descriptive account, on the other hand, implies to deny a certain authority to the deluded patient – after all we don’t take them to mean what they say. Delusional persons, however, often do seem to mean what they say. To suspend the descriptive conception is presumptuous in this respect. It patronizes delusional persons and implies that they are not clear about how they mean their own utterances.

In response to this important objection, we must once more emphasize that we do not deny that often delusional utterances do indeed have their direct literal meaning as assertives and are to be understood accordingly. Again, our direct target is (DA*). Indeed, also Searle’s analysis of indirect speech acts provides that normally the indirect speech act is performed on top of the direct one. We are more interested in showing that their meaning or communicative role is not necessarily exhausted therein and suggest focusing – where possible – on those aspects of meaning that keep the conversation open.⁸

Next, it’s clearly not always a moral failing to interpret utterances non-literally. If I say “Pass me the salsa” but there is none, giving me the salt does not imply a moral failure of taking me seriously. Often we employ charitable interpretation (cf., Davidson, 1973) diverting from the literal meaning. Also, consider jokes or ironic sentences. If I’m saying “Nice weather” when it’s raining cats and dogs, you ought not to assume an epistemic deficit on my behalf but instead some sense of sarcasm (cf., §3). But you also don’t fail to respect me when you disregard the literal meaning of my utterance (cf. Grice, 1989, p. 34). Indeed, I want you to!

Of course, what separates those examples from delusional utterances is that with misspeakings or jokes we intend to communicate something else than what’s literally said. We can’t make this assumption with delusional utterances.

As a result, it requires great caution and self-reflexivity when a meaning is ascribed to the utterance of an interlocutor that does not correspond to what we’d usually think is meant by that person. And we must be careful that, unlike everyday interpretations in which we infer indirect speech acts or understand utterances non-literally, we are actually making trial attributions of meaning that the person (in this case, the delusional person) might not agree with.⁹

Deamer and Wilkinson (2021) develop an interesting consideration concerning the possibility to understand a delusional utterance differently than the speaker does herself. The authors convincingly elaborate that delusional assertions might arise from metaphorical (and thus non-literal) thinking that is understood literally by the deluded individuals over time. The authors assume that metaphorical thinking enables one to linguistically capture complex emotions and experiences that are undoubtedly present in deluded individuals. Due to certain factors (e.g., a lack of interpersonal communication or deficits in inhibitory control), then, a formerly expressive phrase takes on an assertive form. Evidently, this agrees well with our analysis of delusional utterances as expressives (§5.1).¹⁰

Against this background, we do believe that in certain circumstances it can be permissible to take delusional utterances non-literally, even when we’re uncertain about the communicative intention. We do acknowledge, however, that doing so is a risky endeavor and the costs and benefits of any given interpretation have to be carefully evaluated on a case-by-case base. In the best scenario, a therapeutic session collectively establishes an interpretation of what a certain utterance means which is agreed to by all participants and facilitates understanding. As we’ve tried to show, Speech Act Theory can have something to contribute to such successes.

7. Conclusion

In this paper, we’ve tried to demonstrate that explicitly addressing the linguistic meaning of delusional utterances is a promising step if we want to make progress on the intricate question about the understandability of delusions. Also, we’ve argued that considering delusional utterances as speech acts has the potential to weaken the dead-end problem which arises on a merely descriptive account of meaning as it is suggested by the currently predominant doxastic accounts of delusions. In particular, we’ve shown that there is a lot of initial plausibility for understanding delusional utterances as directives and expressives, and that considering delusional utterances as indirect speech acts does not require any interpretive resources we do not already possess. Therefore, Speech Act Theory can be an important method to enrich our practical and theoretical toolbox in dealing with delusional utterances.

Let us end by noting two important upshots of our argument. First of all, using Speech Act Theory as a tool to understand the linguistic meaning of delusional utterances allows us to avoid assumptions of a deficit. Recall that (DA) suggests that people with delusions entertain false beliefs. Taking utterances such as (1), (2) or (3) literally suggests the assumption of a deficit in perception and/or judgment on behalf of the utterer which explains why she is having false beliefs and thus why she is uttering a false proposition.

However, deficit-oriented accounts have been criticized since the early years of psychiatry, for example by Karl Jaspers (cf., 1963: 97) and recent philosophy of psychiatry echoes Jasper’s sentiment. For instance, Lisa Bortolotti (2009, 2020) has forcefully challenged the idea that irrational beliefs demarcate sufferers of mental health conditions. She argues that people with and people without psychopathological alterations are on a continuum with respect to their irrational beliefs. After all, we all hold positive illusions and confabulate frequently. In that sense, thinking of people with delusions as being qualitatively more irrational, i.e., affected by a special deficit, is an unfounded contribution to a societal stigma.

We believe with Jaspers and Bortolotti, that the assumption of such deficits can be problematic. Besides the stigma, the postulation of an epistemic deficit is also troublesome in that it applies only focally. People suffering from delusions certainly do not perceive everything incorrectly and it can’t be assumed that their reasoning abilities are completely suspended. Rather, it is often possible to converse easily with people suffering from delusions about many everyday topics (Bleuler, 1950; quoted in Bovet & Parnas, 1993; Sass, 1994). Insofar as such deficits are suggested by (DA), suspending (DA) may also counteract any problems caused by those deficit-ascriptions.

Acknowledgement

All authors contributed equally to this work. Many thanks to Rosa Ritunnano and Henrik Walter for feedback and helpful conversations. Also, we are indebted to two anonymous reviewers of this journal for their excellent comments.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

1. More recently, doxastic conceptions usually take into account that there may be delusional utterances that are true, e.g., in cases of erotomania (cf., Bortolotti, 2009; Coltheart, 2007). Partly for this reason, the main body of the DSM-V now includes a definition of delusion that speaks of a “fixed” rather than a “false” belief (American Psychiatric Association, 2013, p. 87). Nevertheless, we want to emphasize that delusional utterances are usually false or bizarre when taken literally.

2. For simplicity we understand two intertranslatable languages as the “same” (Davidson, 1973).

3. The original utterance occurred in German using the word “krank”. Translating “krank” into English is not straight-forward since it means both “feeling ill” and “having a disease”. In the case of Ms. D., it’s fair to assume that she is denying both statements. Understood in a literal sense, she denies that she feels unwell, and she denies having any form of disease. In what follows, we focus on the second sense.

4. We are thankful to an anonymous reviewer for asking us to make this point more explicit.

5. The so-called “hereby” test is relatively helpful in distinguishing illocutionary and perlocutionary acts: While it is possible to say “I hereby warn you about my cat” (illocutionary act), the perlocutionary act of frightening cannot usefully employ an “hereby” (Austin, 1962; Lycan, 2019).

6. In order to fit this utterance neatly into a direction of fit pertaining to directives, we’d had to include a negation into the propositional content. Since we are not in the business of determining once and for all which speech act (2) “really is”, but merely try to showcase a range of possibilities for interpreting (2) with the resources of Speech Act Theory we shall not be troubled by this. Moreover, the discussion of the notion of indirect speech acts (cf., Searle, 1979) below will show how such interpretations can be justified.

7. Wittgenstein is often seen as proposing the strong version of such a theory (McDowell, 1996; Wright, 1998). Scattered remarks of his indicate, however, that he was sensitive to the fact that avowals also can describe certain states of affairs (e.g., Child, 2011; Hubacher Haerle, 2021; Wittgenstein, 2009: part II). Also, in a very insightful paper, Wilkinson (2020) suggests to understand attributions of delusion as expressives. Our proposal differs in that we assume that delusional utterances themselves might in some cases be understood as expressives. We think, however, that our proposal has important synergies with that of Wilkinson.

8. Moreover, it also seems plausible that only a more complete view of an utterance reveals how to understand the descriptive aspect of the meaning in the first place (cf., Heaton, 2013, pp. 112–116).

9. See Ritunnano (2022) for further development of similar thoughts.

10. We can also imagine an analogous treatment of the directive aspects of delusional utterances. It is conceivable, for example, that certain wishes may be formulated only metaphorically (because they might otherwise seem too presumptuous or threatening to the person, or because he or she regards their fulfillment as too unrealistic). As in the considerations of Deamer and Wilkinson (2021), one could then imagine a process that leads the delusional persons to misrecognize their directive utterances as assertive ones.