2.1 The Narrow ‘Component’ View

In its narrow conception, pragmatics is viewed as a component of our ability for language, which is to be approached on par with the other core levels of linguistic analysis; that is, phonetics, phonology, morphology, syntax, and – more crucially, of course – semantics. In this picture, the focus of enquiry for pragmatics falls on the ways in which contexts of utterance can be seen to perform certain functions that complement the output of other performance systems, whose knowledge basis underlies our linguistic competence; most notably semantic interpretation, but, on occasion, morphosyntactic parsing and phonological processing too.

By and large, this view is closely linked to the mainstream way of looking at meaning in language as a function of the truth-conditional content of sentences,Footnote ³ allowing for contextual considerations to enter the picture only when necessary. From this angle, pragmatics falls under the scope of linguistic enquiry only insofar as contextual information can be shown to directly affect the semantic interpretation of sentences, like, for example, in the case of reference assignment (especially to indexicals) or disambiguation; hence its resulting characterisation as narrow. Indeed, even though he did advise some caution when it comes to blindly adopting this viewpoint, Levinson (Reference Levinson1983: 32) concluded his aforementioned overview of the field’s scope characterising as ‘most promising’ those ‘definitions that equate pragmatics with “meaning minus semantics” or with a theory of language understanding that takes context into account, in order to complement the contribution that semantics makes to meaning’. Following suit, work that adopts the component orientation has tended to concentrate almost exclusively on a handful of phenomena (including, naturally, speech acts) which cannot be accounted for with resort to semantic theory alone.

This essentially ‘pick-and-choose’ outlook was even more evident during the early days, right before the recognition of pragmatics as a discipline in its own right. In fact, the haphazard way in which scholars would use the label ‘pragmatic’ back then was what famously prompted Bar-Hillel (Reference Bar-Hillel1971: 405) to refer to pragmatics as a ‘wastebasket’ for linguistic phenomena which appear to not fit ‘syntactico-semantic theory’ and, at the same time, underline the need to ‘bring some order into the contents of this wastebasket’ by ‘clarify[ing] somewhat better the explicandum […] before embarking on the explication’. Understandably, such an undertaking would be neither straightforward nor easy, especially since ‘an “after the fact” definition, when the set of pragmatic topics had already been relatively well established, is basically a mission impossible’ (Ariel, Reference Ariel2010:16). As it happens though, a way of bringing this much-needed conceptual order was already brewing, owing primarily to the collective work of Grice (Reference Grice1989).

Around the same time that Bar-Hillel coined the wastebasket metaphor, Grice was already developing his theory of implicature, with the express aim of accounting for the divergences in meaning that were often assumed to exist between formal devices in logic and their natural language counterparts (see, in particular, Grice, Reference Grice, Cole and Morgan1975). Needless to say, implicature instantly became a staple of pragmatic theory, and for good reason: not only did it help safeguard the purity of the logical approach, by allowing for a (non-truth-conditional) description of aspects of linguistic meaning that were clearly unfit for formal semantic accounts, as in the case of conventional implicatures; it also provided a principled way of explaining how contexts of utterance prototypically allow language users to communicate inferences that go beyond the standing linguistic meaning of sentences, and are thus correspondingly pragmatic in nature, as in the case of conversational implicatures.

In effect, along with the notion of a speech act, to which we will return shortly, this latter contribution of Grice also helped extend the scope of pragmatics in its narrow conception beyond its most immediate ‘add-on to semantics’ role. By including in the picture further inferences, especially implicatures of the particularised conversational kind, which may require linguistic meaning in order to be worked out but are still quintessentially context-driven, it drew even more attention to the study of communicated meaning. More importantly though, when it comes to the delineation of the scope of pragmatics, in developing his approach, Grice offered an ingenious solution to the aforementioned wastebasket problem. Complementing his earlier account of nonnatural meaning (Grice, Reference Paul1957; for a brief exposition, see Section 5.3), the philosopher’s theory of implicature pretty much solidified the field’s emphasis on the intentions of the speaker. Adopting the Gricean rationale, research on the various distinct phenomena that, by that time, would have been taken to fall under pragmatics could now be seen to converge towards a common goal: the identification of the meaning that speakers intend to communicate to hearers when producing utterances, or as this ended up being commonly referred to, the identification of speaker meaning.

Overall, for scholars adopting the narrow view of the discipline, the notion of speaker meaning allowed for a clarification, as well as restriction, of the field’s scope. Obviously, the interpretation of any utterance can give rise to a whole host of meaning-related inferences whose origin can be attributed to the utterance’s context of production (see, e.g., Elder, Reference Elder2024), but only some of those inferences are directly linked to (or derived from) the linguistic content of an utterance’s underlying sentence – that is, the linguistic structure that the speaker chose to token on this specific occasion. Delimiting the focus of attention to just those contextual inferences that have been communicated properly – that is, by virtue of the speaker overtly inviting the hearer to draw them – would be deemed sufficient for recovering those aspects of meaning that are needed to complement the picture captured by semantic theory. As a result, with Grice’s insight, the study of pragmatics in its narrow conception could finally be seen to have both a clear-cut object of investigation under the rubric of speaker meaning and a tightly knit domain of application, that of intentional verbal communication.

2.2 The Broad ‘Perspective’ View

Αs we just saw, the earliest adoptions of a pragmatic outlook for the study of language involved the operationalisation of contextual information in order to explain certain phenomena that would not fit existing semantic and syntactic accounts. Going back to the very coinage of the term ‘pragmatics’, however, it is easy to see that the discipline was not originally conceived of along such lines. Upon introducing the term in his seminal semiotic account, Morris (Reference Morris, Neumeth, Carnap and Morris1938: 6–7) defined pragmatics as the ‘study of the relation of signs to interpreters’, distinguishing it from semantics and syntactics, which respectively study ‘the relations of signs to the objects to which the signs are applicable’ and ‘the formal relation of signs to one another’. From this angle, pragmatics was expected to deal with ‘the biotic aspects of semiosis, that is, with all the psychological, biological, and sociological phenomena which occur in the functioning of signs’ (Morris, Reference Morris, Neumeth, Carnap and Morris1938: 30); clearly a much broader role than the one that would eventually be granted to it by the narrow pragmatics camp.

Against this backdrop, much like the narrow conception of pragmatics emerged out of a need to consolidate the scope of existing linguistic research, the broad one developed as a reaction to the way linguistic pragmatics was evolving at the time. On the occasion of launching the Journal of Pragmatics and in what is generally considered as the first explicit defence of the broad take on the discipline, Haberland and Mey (Reference Haberland and Mey1977: 5) took issue with the overwhelming preoccupation of their contemporaries with ‘formal pragmatics (mostly in the Montague tradition)’, which they suggested may well be thought of as ‘an extremely sophisticated way of doing semantics’ given its complete disregard of socio-cultural considerations. According to them, since ‘language, as the means of societal communication, is transmitted in a social setting’, the human capacity for language ‘cannot be studied and/or measured in the abstract’ and without reference ‘to the societal conditions that language learners and users live by’ (Haberland & Mey, Reference Haberland and Mey1977: 9–10; emphasis in original). Then, drawing inspiration from Morris’ original vision, and more specifically from his suggestion that pragmatic rules are to ‘state the conditions in the interpreters under which the sign vehicle [in our case, some linguistic token] is a sign [i.e. meaningful]’ (Morris, Reference Morris, Neumeth, Carnap and Morris1938: 35), they attributed to pragmatics the role of ‘inquiring into the necessary conditions that govern the use of language’ (Haberland & Mey, Reference Haberland and Mey1977: 6).

From then on, pragmatics would additionally be conceptualised of as a perspective for studying language – or, as Verschueren (Reference Verschueren, Verschueren and Östman2022: 14) has correspondingly put it, a ‘perspective on whatever phonologists, morphologists, syntacticians, semanticists, psycholinguists, sociolinguists, etc. deal with’, rather than a separate component that complements other levels of linguistic analysis. Perhaps more pertinently in view of the present discussion, embracing pragmatics ‘as a general functional perspective on (any aspect of) language’ essentially meant adopting ‘an approach to language which takes into account the full complexity of its cognitive, social, and cultural (i.e., “meaningful”) functioning in the lives of human beings’ (Verschueren, Reference Verschueren, Verschueren and Östman2022: 17). Given this, apart from casting a much wider net in relation to the positioning of pragmatics vis-à-vis linguistic theory, the perspective view is also deemed broad because it significantly extends the discipline’s remit, allowing for the inclusion of multiple phenomena over and above those that the component pragmatics tradition has typically concentrated on. For instance, being generally content with delimiting their research to the domain of intentional verbal communication, the narrow-view camp has tended to disregard aspects of meaning that are not tied to speaker intentions, with ‘phenomena like politeness, face-saving and turn taking’ often being excluded from the remit of pragmatic analysis on the grounds that their inclusion would conflate distinct aims as to what the discipline is about (Blakemore, Reference Blakemore1992: 47–48). By reinstating socio-cultural considerations at the centre of its agenda, the broad take on pragmatics has allowed for a ‘boundless diversification’ (Verschueren, Reference Verschueren, Verschueren and Östman2022: 9) in the field. As a result, a significant amount of research that does not prioritise formal analysis nor concentrates specifically on the study of speaker meaning has managed to additionally flourish therein.Footnote ⁴

In this picture, when it comes to the role of speaker intentions for the study of pragmatics, the perspective view casts doubt not only on their adequacy as a defining factor for the field’s object of enquiry, but also on their very traceability. As Mey and Talbot (Reference Mey and Talbot1988: 747) note, the main problem with restricting communication to its overtly intentional dimension ‘hinges on the extent to which people “know what they’re doing” and the extent to which what they are doing is predetermined and even unknowable’. Again, socio-cultural considerations appear to play a central role here. Regardless of how much communicative behaviour is anchored by an individual’s intentions and goals, for perspective pragmatics, it is still underlain by ‘historically constituted ways of saying and doing’ that a social agent ‘cannot be aware of’ (Mey & Talbot, Reference Mey and Talbot1988: 747). At the same time, drawing on conversation-analytic research that focuses on the dynamics of interaction, the broad pragmatics camp typically takes meaning to be co-constructed on the fly by both interlocutors through tacit negotiations that take place as a verbal exchange unravels, which in turn opens the possibility that the intentions on which verbal communication relies constitute more of ‘a post facto participant resource that emerges through interaction’ than ‘an a priori mental state of speakers’ (Haugh, Reference Haugh2008: 104).

All in all, the perspective take on pragmatics has attempted to shift ‘the pragmatic question par excellence’ from ‘What does this utterance mean?’ to ‘Why has this utterance been produced?’ (Haberland & Mey, Reference Haberland and Mey1977: 8). In this picture, the pragmaticist is called to concentrate on ‘the continuous making of communicative choices, both in speaking and in interpreting’ (Verschueren, Reference Verschueren, Verschueren and Östman2022: 17–18); whereby these choices are by default variable, negotiated in interaction, and can be swiftly adapted to any interlocutor’s communicative needs. Therefore, instead of a clear-cut object of investigation, like the traditional study of the way in which ‘speakers implicate [and] hearers infer’ (Horn, Reference Horn, Horn and Ward2004: 6), broad pragmatics proposes an alternative way of looking at language use as a whole. From this angle, as Verschueren (Reference Verschueren, Verschueren and Östman2022: 19) elucidates, explanations of a pragmatic nature are to be provided within the purview of the following four guiding principles: (a) communicative choices pertain to all levels of linguistic structure; (b) the contribution of contextual factors is ever-changing and comprises socio-cultural considerations; (c) certain choices can be primed on the basis of their salience in particular situations; and (d) even communication principles and strategies can dynamically change over time.

2.3 In Lieu of a State of the Art

Even though the preceding outlines will anchor the discussion of speech act theory in what follows, it is imperative to stress that they are only indicative of a much more nuanced picture. For one, work on pragmatics that seemingly adopts either one of these two positionings abounds with variations of significant theoretical importance, such as the distinction between Grice’s speaker meaning and relevance theory’s intented import (Sperber & Wilson, Reference Sperber and Wilson2015) or that between genre and activity type (Linell, Reference Linell1998: Chapter 12), to name but two. Then, chances are that, confronted with a choice, scholars who self-identify as pragmaticists would only accept being categorised as working solely within one of these two traditions with some pretty hefty caveats. For instance, the vast majority of research that adopts the perspective view of pragmatics still relies heavily on theoretical concepts and analytical distinctions that were originally drawn with a view to accounting for speaker meaning in encapsulation from socio-cultural or interactional considerations. Similarly, approaches that have traditionally been couched in a narrow-view background do not always shy away from taking into account such considerations either, as can be seen, for instance, in the incorporation of a full-fledged conversation-analytic approach to the treatment of speech acts in Levinson (Reference Levinson1983, Reference Levinson and Huang2017; see also Section 7.2) or in recent work on relevance theory that seeks to draw the focus away from speaker intentions (Assimakopoulos, Reference Assimakopoulos, Ifantidou, de Saussure and Wharton2021a, Reference Assimakopoulos2022). Therefore, as Culpeper and Haugh (Reference Culpeper and Haugh2014: 7) rightly point out, ‘one should not over-emphasise differences between the [… two] views’.

Still, given their obvious points of divergence, looking at how these views approach the scope of the field can be useful in assessing its evolution. To wit, narrow pragmatics takes linguistic representations as its point of departure with a view to accounting for those meaning-related phenomena that cannot be fully approached in abstraction from their context of utterance. In contrast, broad pragmatics adopts a top-down approach, in the specific sense that it starts from society and the interlocutors themselves and concentrates on how the conditions of the former and the patterns of interaction of the latter shape the context that mediates meaning-making on all occasions of language use. By the same token, each tradition relies on a rather different conception of context, too. In component pragmatics, we typically talk about utterance contexts, concentrating mainly on those aspects of contextual information which reveal the meaning that the speaker intended to communicate on some occasion of use or another.Footnote ⁵ Conversely, for perspective pragmatics, the focus is placed on the broader discourse context,Footnote ⁶ which focuses instead on all stages of a communicative exchange and includes the vast range of socio-cultural norms and attitudes that underlie them, affecting both interlocutors’ behaviour. Finally, differences can also be located even in the methodological choices of scholars subscribing to the two camps, as is evident in the insistence of researchers adopting the broad view of pragmatics to work on authentic sequences of verbal interaction, when corresponding work on narrow pragmatics has traditionally relied on the analysis of single utterances in invented question-answer pairs.

In conclusion, appreciating the caution that Culpeper and Haugh have advised, perhaps the safest way to approach the distinction at hand is by following Haberland and Mey (Reference Haberland and Mey2002: 1674) in viewing component and perspective pragmatics as two ‘extremes’, whereby ‘narrow but precise description’ is respectively juxtaposed with ‘broad but fuzzy inclusion’. Clearly, being extremes, they both have their weaknesses when taken at face value, with narrow pragmatics appearing too restrictive to many and broad pragmatics being correspondingly deemed ‘too inclusive to be of much use’ (Davis, Reference Davis and Davis1991: 3) to others. Still, as we will eventually see, they are both extremely useful when it comes to the theoretical description of speech acts.

4.1 From Performatives …

The first part of Austin’s lecture series motivates the concept of the descriptive fallacy by drawing attention to the fact that there are several commonly used utterances which remain meaningful despite having a primary function that quite clearly deviates from the mere description of states of affairs. In contrast to constatives, which is what Austin (Reference Austin1975: 7) called descriptive statements that are ‘normally thought of as just saying something’, what the philosopher had in mind here is a specific group of utterances that can be quite straightforwardly seen to be ‘doing something’ (Austin, Reference Austin1975: 12) instead. For example, consider the following:

1. 6. I promise that I will come to your party.

2. 7. I apologise for having kept you waiting.

3. 8. I order you to wash the dishes.

4. 9. I sentence you to 5 years in prison.

Even intuitively, it should be pretty easy to see that the primary aim of the utterances in (6)-(9) is quite different from providing a simple description of a state of affairs; rather, they explicitly seek to perform an action – be it making a promise, as in (6), an apology, as in (7), a command, as in (8), or giving someone a prison sentence, as in (9). In view of this, Austin fittingly called utterances of this sort performatives and dedicated roughly the first half of his lecture series to exploring how they can be identified in contradistinction to constatives.

When it comes to the mechanics of performative utterances, Austin showed that an utterance does not really need to embed a special verb, like ‘sentence’, ‘promise’, ‘apologise’, or ‘order’, in order to constitute a performative. Utterances of this sort can also be produced without such performative verbs, as in (10), which can pretty straightforwardly substitute (6) when uttered in the same context.

1. 10. I will come to your party.

Along these lines, apart from explicit performative utterances like the ones in (6)–(9), Austin recognised the existence of implicit ones too, which carry out pretty much the same types of actions. At the same time, he also acknowledged that the mere presence of what appears to be a performative verb in an utterance does not really guarantee the utterance’s performative character, since verbs of this class can also be used non-performatively. Consider, for instance, a context in which I am producing the utterance in (11), while recounting a hypothetical scene from the movie Titanic, in which Jack momentarily leaves Rose to go save a little boy from drowning.

1. 11. Jack promised Rose that he will find her.

In uttering (11), I am clearly not engaging in an act of promise myself; I am rather reporting on such an act having being performed by Jack, by producing a statement that is ultimately describing a state of affairs which can either be true, if Jack indeed made that promise, or false otherwise. Similarly, the same class of verbs can also be used non-performatively when describing habitual behaviour, as in the case of (12).

1. 12. I promise only when I know I can keep my word.

On top of ‘the verbs which seem, on grounds of vocabulary, to be specially performative verbs’ (Reference Austin1975, p. 61), Austin also considered a number of additional ways in which the performative function of the relevant utterance class can be identified, but again found them all wanting. For one, as shown by the example utterances in (13a) and (13b), which could be used interchangeably to perform the same action, person and voice are immaterial in this context.

1. 1. a. I am letting you go.

   2. b. You are fired.

Then, while typically a good criterion for establishing an utterance’s function, sentence mood was deemed inappropriate too, since there is no one-to-one relation between the form of a sentence and the performative function of its utterance. For example, someone could well advise me to stop smoking by using the explicit indicative sentence in (14a), but they could also do the same by using the conditional in (14b).

1. 1. a. I advise you to stop smoking.

   2. b. If I were you, I would stop smoking.

The same applies to tense. As we can see in the case of (15b), which can perform the same action as (15a) even though it uses the past instead of the present tense, as performatives typically tend to do, tense does not always provide a reliable indication of a performative function,

1. 1. a. I find you guilty of breaking that vase.

   2. b. You broke that vase.

Despite the apparent difficulty in identifying a linguistic or grammatical criterion that can reveal an utterance’s performative character at face value, Austin’s exploration did not leave him entirely empty-handed either. It still provided him with a blueprint for prototypical explicit performatives, enabling him to consider the possibility that any ‘performative should be reducible, or expandible, or analysable into a form, or reproducible in a form, with a verb in the first person singular present indicative active’ (Austin, Reference Austin1975: 61–62), as per our examples (6)–(9). Even though this explicit performative formula does not really guarantee the presence of a performative utterance, as we saw in (12), it can be particularly useful for reframing an implicit performative in explicit terms; simply consider the reformulation of our example in (10) into that in (6). At that, an extra step that can help confirm whether we are in the presence of a pure performative utterance would be to insert the adverb ‘hereby’ between the verb and the subject in this formula: if the insertion is permissible, as in (16), the verb is used performatively; if not, as in (17), the utterance can be classified as descriptive.

1. 16. I hereby promise that I will come to your party.

2. 17. *I hereby promise only when I know I can keep my word.

4.1.1 Conditions for Felicitous Performance

Naturally, Austin’s investigation of the constative/performative distinction did not just concentrate on the form of the relevant utterances. A substantial part of his account also dealt with the relative contribution of performatives, as opposed to constatives, to the speech situations in which they arise. As we saw in (11), while the propositional content of an utterance that reports on a performative can be evaluated as true or false, asking whether the content of an act of this sort itself can be assigned a truth value is very much beside the point. Consider, for example, the utterance in (18), as directed by Jack to Rose in the same context hypothesised in (11).

1. 18. I will find you.

Quite evidently, the extent to which the proposition expressed in (18) can be evaluated for its truth is pretty much irrelevant. Rather, the only aspect of its meaning that could conceivably be judged as true or false in this context is whether the expression of this proposition constitutes an act of promise on the part of Jack.

With truth conditions becoming immaterial in this setting, approaching the relevant contribution of performatives is bound to need a different handle. As Austin (Reference Austin1975: 14) noted in this vein, ‘besides the uttering of the words of the so-called performative, a good many other things have as a general rule to be right and to go right if we are to be said to have happily brought off our action’. Indeed, since performatives are themselves actions, the right question to ask, instead of whether they are true or false, is whether the action they perform has been successfully executed. To this end, Austin (Reference Austin1975: 14–15), suggested a number of conditions that a performative needs to satisfy in order to fulfil its function:

1. Α. (1) There must exist an accepted conventional procedure having a certain conventional effect, that procedure to include the uttering of certain words by certain persons in certain circumstances, and further, (2) the particular persons and circumstances in a given case must be appropriate for the invocation of the particular procedure invoked.

2. Β. (1) The procedure must be executed by all participants both correctly and (2) completely.

3. Γ. (1) Where, as often, the procedure is designed for use by persons having certain thoughts or feelings, or for the inauguration of certain consequential conduct on the part of any participant, then a person participating in and so invoking the procedure must in fact have those thoughts or feelings, and the participants must intend so to conduct themselves, and further (2) must actually so conduct themselves subsequently.

Austin’s categorisation of these conditions into three pairs was meant to highlight the different ways in which performative utterances can fail to complete their associated action, which would ultimately render them ‘unhappy’, as the philosopher himself put it (Austin, Reference Austin1975: 15): the conditions under A and B are meant to capture the successful performance of an action, while those under Γ control for cases where the deployment of the performative utterance may be successful but is at the same time defective, inappropriate or misleading. In this picture, when an infelicity is due to non-adherence to a condition under A and B, we have a misfire of a purported performative, whereas when this infelicity results from a violation of Γ, we are in the presence of an abuse of a performative which has otherwise been successfully launched.

In order to see how these conditions work in practice, let’s consider some scenarios in which they would be taken to not be observed. In relation to condition A.1, suppose that you want to insult someone, but instead of just using any one of the many conventional ways to do so (directly or by resorting to sarcasm, a joke, a slur, etc.), you merely utter (19).

1. 19. I insult you.

Despite your original intention to perform the relevant action, chances are that the action will not be completed, since the verb ‘insult’ is conventionally descriptive, rather than performative. Turning to condition A.2, imagine a situation in which two siblings fight and one tells the other: ‘I sentence you to 5 years in prison.’, as in (9) further up. Here, the speaker may have indeed used a conventional (almost ritualistic) procedure for performing the action at hand, but given the circumstances, and more specifically their lack of the relevant authority to send anyone, let alone their interlocutor, to prison, the action itself is void, or, as Austin put it, misapplied. All in all, on both these occasions of a misfire, the infelicity can be taken to result from what Austin called a misinvocation of procedure.

Along similar lines, infelicities due to the two conditions in B are the result of a misexecution of the performative, since they arise when there are flaws (B.1) or hitches (B.2) in performing the relevant action. Flaws relate to an otherwise appropriate application of a conventional procedure for the act itself, which, however, includes a mistake; as when a priest utters (20) during a wedding ceremony, but the names of the couple getting married are actually ‘John’ and ‘Mary’.

1. 20. Bill and Kate, I now pronounce you husband and wife.

Hitches, on the other hand, have to do with issues that may impede the complete performance of the purported action. For example, I may utter (21) genuinely intending to offer a bet, but unless you somehow acknowledge that this is the case, the bet is not really on.

1. 21. I bet you my lunch money that we will win tonight’s game.

Finally, when it comes to violations of the conditions under Γ, that is, abuses of the performative, these apply to situations where the verbal invocation of the action may have been successful, but its performance is still infelicitous due to a lack of the appropriate psychological states on the part of the speaker or simply because the parties involved do not really follow through with it. Such abuses can be seen to take place on the grounds of insincerities (Γ.1) when, for example, someone promises to do something without having any intention to keep their promise. Similarly, in our familiar example of offering a bet, a performative can be taken to have been abused if, despite both interlocutors having previously explicitly agreed on it, one of them still does not keep their end of the agreement (Γ.2).

All in all, by introducing the constative/performative distinction, Austin did not just challenge the dominant view of his time, which concentrated almost exclusively on the descriptive dimension of the meaning of sentences; he essentially showed that the primary function of a great deal of utterances that are ubiquitously produced out of our linguistic repertoire is to do, rather than describe things. At that, he offered a principled way of accounting for the way in which utterances of this kind remain meaningful, by identifying the conditions that ensure their successful performance. Through this doctrine of infelicities (summarised in Figure 1), Austin further substantiated what it means for an utterance to execute an action, providing a detailed account of the ways in which the actions that the relevant utterances perform may fail to be realised. Still, as we will now turn to see, Austin’s contribution runs much deeper than the identification of just a special class of utterances whose content necessitates an alternative approach to capturing their essence.

Figure 1 The doctrine of infelicities

(after Austin, Reference Austin1975: 18)

4.2 … to speech acts

Halfway through his argumentation, Austin’s own inquiry into the opposition between constatives and performatives led him to altogether question the relevant distinction. What prompted him to seek, as he himself put it, ‘a fresh start on the problem’ was the failure, as per the foregoing discussion, ‘to find a grammatical criterion for performatives’ (Austin, Reference Austin1975: 91), alongside the realisation that even descriptive statements can be straightforwardly reformulated into explicit performatives by means of the explicit performative formula, without this actually compromising their being truth-evaluable. Consider, for example, the relevant reformulation of our prototypical constative utterance in (22a) into an explicit performative in (22b):

1. 1. a. Snow is white.

   2. b. I (hereby) state/assert/suggest … that snow is white.

By the same token, Austin additionally recognised that, by effectively performing an act of asserting or stating,Footnote ⁸ constatives too have conditions for felicitous performance, in the sense that they can also misfire or be abused. For instance, my production of (23) at this very moment would easily be seen as a misinvocation of the procedure for asserting, since I quite obviously have no way of knowing, let alone asserting, that the exact number I mention is correct.

1. 23. There are 10,323 babies being born around the world right now.

Similarly, assertions can also be misexecuted, as is often the case with slips of the tongue, such as in (24), where ‘penguins’ is quite clearly used by mistake in the place of ‘pigeons’.

1. 24. Mary is feeding the penguins in Hyde Park.

Finally, insofar as a ‘statement is liable to the insincerity form of infelicity’ (Austin, Reference Austin1975: 136), assertions are equally liable to abuses of type Γ, as in the case of (25), when uttered by someone who has not actually engaged in the relevant action:

1. 25. I processed the refund to your account two days ago.

Showing, then, that even constatives can be seen to operate on par with performatives, Austin reached the conclusion that all utterances should effectively be considered as action and are thus amenable to the treatment that had up to that point been reserved for the discussion of performatives alone; hence the reappraisal of all utterances as speech acts from this point of his argumentation onwards.

4.3 How to Read Austin

As should have become evident by now, the complexity of Austin’s original argumentation lies primarily in his decision to inaugurate his discussion with the description of a special class of utterances whose treatment is not amenable to a truth-conditional analysis, but then eventually extend the relevant rationale to the treatment of all utterances over and above this – initially introduced as special – class. Indeed, this is something that was criticised by several of Austin’s contemporaries, while even nowadays, Austin is typically thought to have simply changed his mind about pursuing the constative/performative distinction while he was developing his argumentation. As a result, most overviews of speech act theory in the pragmatics literature, tend to only refer to it in passing, concentrating on the philosopher’s eventual description of speech acts instead, which is again only briefly presented before his views are mixed with later elaborations on the topic.

Against this backdrop, the main reason why this Element has dedicated substantially more space to Austin’s original argumentation than other related overviews of speech act theory is that it will eventually return to this argumentation with a view to challenging the mainstream way of theorising about speech acts nowadays. Granted, as we will see in Section 6, Austin’s exposition may have had some exegetical weaknesses, such as his lack of clarity or precision in the use of theoretically-charged terms at times, but lack of coherence should not be considered one of them. As Sbisà (Reference Sbisà2007: 462–463) elucidates in a paper which also inspired the heading of the present subsection, there is actually ‘no trace of a change of mind in the various sets of [Austin’s] notes forming the manuscript’; rather, How to Do Things with Words should be read as ‘a complex argument in support of the claim that all speech should be considered as action and, more specifically, that speech can be described as the performing of actions of the same kind as those performed by means of performative utterances’.

5.1 Speech Act Theory and Component Pragmatics

As already mentioned in Section 2.1, before being recognised as a discipline in its own right, pragmatics would be casually included in formal analyses on an as-needed basis. For component pragmatics, what speech act theory did was pretty much eliminate this optionality. By accepting that all utterances are effectively speech acts, an extra layer of illocutionary force would always have to be added over and above that of propositional content, when it comes to their description. So, even though one might have in principle been able to argue that meaning can be studied in its entirety through the lens of formal theory up to that point, the same could not be said for illocutionary force, given its obvious context-dependence. To put it simply, since all utterances have an illocutionary force – some even explicitly so – and the analysis of this force necessitates input from pragmatics, then pragmatics needs to be recognised as a compulsory part of linguistic analysis.

In this picture, the aspect of speech act theory that has featured most prominently in the disciplinary conceptualisation of pragmatics is the distinction between an utterance’s content and its force. In fact, this distinction is so fundamental that it even underlies a most popular way of altogether distinguishing between semantics and pragmatics, whereby ‘semantics studies the contents of communicative acts [… and] pragmatics studies their force’ (Green, Reference Green and Zalta2021: section 2.1). The relevant literature typically attributes this distinction to Austin and the way in which he correspondingly distinguished between the locutionary and the illocutionary act, as per the discussion in Section 4.2.2. However, as we will now see, its most common understanding in the field is actually tied to Searle’s reappraisal of speech act theory rather than Austin’s own original thinking (for a critical discussion, see Section 6).

Right from the start of a paper dedicated to Austin’s locution/illocution distinction, Searle (Reference Searle1968: 405) stated: ‘In attempting to explore Austin’s notion of an illocutionary act I have found his corresponding notion of a locutionary act very unhelpful and have been forced to adopt a quite different distinction between illocutionary acts and propositional acts’. In his explanation for this move, Searle (Reference Searle1968: 407) posited that the locution/illocution distinction fails to be ‘completely general, in the sense of marking off two mutually exclusive classes of acts’ on the basis of the following rationale (inevitably simplified and adapted to fit the discussion up to this point):

• – The locutionary act consists in producing an utterance with intelligible meaning.

• – The presence of a performative verb in an explicit performative utterance cannot but be part of this meaning, which Austin attributes to locution.

• – The function of the performative verb in such cases is to explicitly indicate the speech act’s illocutionary force, which is something that belongs to the level of the illocution.

• – It therefore follows that, at least in explicit performative utterances, the (purported) illocutionary act forms part of the locutionary act, when it should be independent from it.

Searle’s solution to this apparent incongruity was to replace Austin’s notion of the locution, as the more general act of saying something, with that of the propositional act, that is, the act of expressing the proposition encoded by the sentence that an utterance tokens in its production. Isolating the proposition – that is, just that part of the meaning of the uttered sentence that is truth-evaluable – allowed Searle to posit that any additional element that may be included in the utterance’s form is simply an indicator of the utterance’s force, and thus a determinant of the illocution. From this angle then, the problem that he had identified with locutionary acts largely evaporates, since ‘the propositional act is not represented […] by the entire sentence, but only by those portions of the sentence which do not include the indicators of illocutionary force’, and ‘so construed no propositional act is by itself an illocutionary act’ (Searle, Reference Searle1968: 421). In suggesting this, Searle also provided a formula which seeks to capture the underlying form of the illocutionary act performed by explicit performative utterances, as follows:

The appeal of Searle’s proposal for the narrow view of pragmatics, and thus its associated role as a complement of semantic theory, should be fairly easy to appreciate. Separating the embedded (truth-evaluable) proposition from the performative matrix verb offers a principled way of accounting for the meaning of sentences that do not – at face value at least – fit the traditional truth-conditional approach. More specifically, in the case of an explicit performative, like ‘I promise that I will come to your party.’, as in our familiar example in (6), we can still maintain that the uttered sentence does indeed carry some truth-evaluable content, in the sense of the proposition expressed by ‘that I will come to your party’, without having to worry about the interference of the ‘I promise’ part, which simply indicates the illocutionary force of a promise that this proposition is communicated with.

Within the same spirit, a concept that is central to Searle’s reappraisal of speech act theory is that of an illocutionary force indicating device (henceforth IFID), which, if present in an utterance, can help determine the F in the aforementioned formula. In this respect, following Austin’s original exploration of a grammatical criterion for performatives (see Section 4.1), Searle suggested that, apart from performative verbs (like ‘promise’ in our example), IFIDs also comprise ‘at least word order, stress, intonation contour, punctuation, [and] the mood of the verb’ (Reference Searle1969: 30). By attributing this role to IFIDs, Searle’s proposal became particularly popular for tackling a problem that semantic circles had traditionally characterised as central for any ‘comprehensive theory of meaning for a natural language’; that is, the problem of ‘all the sentences that seem not to have truth values at all: the imperatives, optatives, interrogatives, and a host more’ (Davidson, Reference Davidson1967: 321). Effectively, Searle’s formula allowed for an analysis of sentences in, say, the interrogative or the imperative, as in (34) and (35), by postulating that they express the same proposition as their corresponding declarative sentence, here offered in (33), with the distinct verb moods, word orders and intonation contours indicating each utterance’s illocutionary force – that is, an assertion in (33),Footnote ¹² a question in (34), and an order in (35).

1. 33. You will go to the school party on Friday.

2. 34. Will you go to the school party on Friday?

3. 35. Go to the school party on Friday!

In this vein, even though it would eventually be argued that the postulation of a force marker for mood ‘runs into difficulties’ (Portner, Reference Portner, Watanabe and Robert2004: 240), Searle’s proposal was indeed instrumental for establishing that ‘the semantics of the (major) morpho-syntactic sentence types – imperative, declarative and interrogative – determines the major illocutionary force types’; an assumption that still remains ‘at the core of traditional speech act theory […] and has [had] enormous impact on semantic-pragmatic theorising’ (Ruytenbeek et al., Reference Ruytenbeek, Ostashchenko and Kissine2017: 46).

5.2 Formal Updates to Speech Act Theory

Considering again the primacy that component pragmatics grants to semantics when it comes to the study of meaning, what probably cemented the popularity of Searle’s elaborations on Austin’s original theory was his continuous engagement with ‘bringing formal order’ to the description of speech acts (Culpeper & Haugh, Reference Culpeper and Haugh2014: 6; see also Section 6.1). Continuing – in essence – our discussion of Searle’s contribution to the currently received view of speech act theory in pragmatics, we find his attempts to describe illocutionary acts with more analytical precision. In this regard, there are two more proposals of his which proved so successful in the field that they always accompany Austin’s initial account in overviews of speech act theory– with one of them even demonstrably sidelining Austin’s original position on the matter.

5.2.1 Felicity Conditions: From Regulative to Constitutive Rules

In the first instance, we encounter Searle’s reappraisal of what Austin called conditions for the successful performance of a speech act (see Section 4.1.1). Searle’s main proposal in this regard was that, on top of simply regulating the successful deployment of an illocutionary act, the relevant conditions can also be taken to effectively ‘constitute’ the act, in the sense of rendering its ‘existence […] logically dependent’ on them (Searle, Reference Searle1969: 34). By reassessing felicity conditions as constitutive rules of this sort, Searle returned to Austin’s original list and reworked it into a set which aims at offering ‘a model for explicating illocutionary acts in general’ (Searle, Reference Searle1969: 56) and thus comprises:

• – the propositional content condition, which constrains the form of the utterance that embodies the illocutionary act under question;

• – the preparatory conditions, which specify the preconditions for the illocutionary act’s deployment;

• – the sincerity condition, which stipulates the intention that the speaker needs to have in order to execute the illocutionary act; and

• – the essential condition, which determines the illocutionary act’s purpose.

Consider, for example, the illocutionary force of promising. Within this picture, an illocution counts as a promise if and only if it constitutes an undertaking on behalf of the speaker to do something (essential condition) that (a) the hearer would welcome and that (b) the speaker would not just normally do either way (preparatory conditions). Further, the act would need to refer to an undertaking that will take place in the future (propositional content condition) and which the speaker genuinely intends to follow through with (sincerity condition).

Apart from providing an alternative way of approaching illocutions, Searle’s model was also designed with a view to allowing one to distinguish between them, by specifying the relevant conditions that underlie their differences. For instance, as we can see in Figure 2, this model enables us to determine that the difference between a request and an order is due to a specification of the social positioning of the speaker and the hearer, as stipulated by the extra preparatory condition in the case of the latter, which is included in its essential condition too.

Figure 2 Constitutive rules for requests and orders

(after Searle, Reference Searle1969: 66)

Figure 2Long description

The table consists of three columns: types of rule, request, and order. The rows for each type of rule read the following data: Row 1. Propositional content. Request: Future act A of H, Order: Future act A of H. Row 2. Preparatory. Request: 1. H is able to do A. S believes that H is able to do A. 2. It is not obvious to either S or H that H will do A in the normal course of events out of their own accord; Order: 1. H is able to do A. S believes that H is able to do A. 2. It is not obvious to either S or H that H will do A in the normal course of events out of their own accord. 3. S must be in a position of authority over H. Row 3. Sincerity. Request: S wants H to do A; Order: Request: S wants H to do A. Row 4. Essential. Request: Counts as an attempt to get the H do A; Order: Counts as an attempt to get H to do A in virtue of the authority of S over H. A key below the table reads H equals hearer, S equals speaker.

5.3 Illocutions as Acts of Speaker Meaning

While not directly related to it, Searle’s taxonomy above can also help introduce yet another post-Austinian elaboration that has grown to take centre stage in the contemporary view of speech act theory in pragmatics. As already noted in Section 4.2.2, Austin conceived of illocutions as conventional acts. While discussing his notion of uptake there, we saw that, for an utterance of, say, ‘Close the door!’ to constitute an order, there must be a convention linking the use of these words in this specific way with the issuance of an order. Obviously, the most straightforward way of appreciating this perspective is by considering prototypical examples of what Searle called declarations, as in (40), or, similarly, an utterance of ‘I sentence you to 5 years in prison’, as in (9) quite further up.

Deferring the discussion of Austin’s specific version of conventionalism to Section 6.2, it would be safe to say that his overall adherence to the conventionalist doctrine was met with a fair amount of contention by his peers from quite early on. Most famously, Strawson (Reference Strawson1964) criticised Austin for focusing too much on ceremonial (almost institutionalised) speech acts, like the declarations just mentioned, and failing to see that his approach does not quite apply to the more basic illocutions that we casually perform in our everyday exchanges, and for which no relevant conventional procedure can be identified. To use an example of his own, consider the production of the utterance in (41) with a view to issuing a warning to a skater.

1. 41. The ice over there is very thin.

As Strawson (Reference Strawson1964: 444) pointed out, in cases like (41), it is highly unlikely that ‘there is any statable convention at all [, …] such that the speaker’s act can be said to be an act done as conforming to that convention’. To address this issue, Strawson proposed a move away from Austin’s conventionalism, which could of course be preserved when dealing with ritualised cases, and offered a reappraisal of Austin’s account of the illocutionary act in intentionalist terms, using Grice’s theory of meaning as his springboard. Obviously, at that point in time, Grice had not yet developed his theory of implicature, so Strawson’s approach remained purely at the conceptual level of approaching illocutionary force as belonging to the remit of nonnatural meaning, or – as Grice preferred to refer to it – meaning_NN; that is, what we nowadays commonly refer to as speaker meaning.

As already noted in Section 2.1, Grice’s ‘Meaning’ (Reference Paul1957) was the point of departure for a number of developments that would lead to the eventual recognition of speaker meaning as the remit of pragmatics for the narrow conception of the field. In this account, Grice (Reference Paul1957: 385) defined meaning_NN as follows:

Even though this specific definition has been subject to numerous refinements through proposals that attempted to make its original exposition even more watertight,Footnote ¹⁵ I cannot dwell on it too much here due to space restrictions. In a nutshell though, through his postulation of meaning_NN, Grice is widely recognised to have worked out a more precise way of accounting for the intuition that ‘communication is a matter of intentionally affecting another person’s psychological states’ (Bach, Reference Bach, Wilson and Keil1999: 359). More specifically, he suggested that the way in which we get our interlocutors to understand what we intend to communicate to them is simply by getting them to recognise that we intend to communicate something to them in the first place.

Within this frame, Strawson (Reference Strawson1964) argued that, notwithstanding the ceremonial cases that Austin had focused on, what he had referred to as ‘the securing of uptake’ can only be achieved if the audience recognises the speaker’s relevant communicative intention. Therefore, since, as Austin insisted, this securing of uptake is a requirement for the successful performance of the illocutionary act, Strawson posited that the most reliable way of approaching the act itself would be by viewing it as intentional, and thus as part of meaning_NN.Footnote ¹⁶ In the years to follow, Strawson’s suggestion gained significant traction, and ‘speaker intention began to take the foreground in leading analyses of the illocutionary act’ (Sbisà, Reference Sbisà, Verschueren and Östman2022: 1312). Especially in the aftermath of the publication of ‘Logic and Conversation’ (Grice, Reference Grice, Cole and Morgan1975), and Searle’s simultaneous introduction of the category of ‘indirect speech acts’, to which we will return shortly, it became increasingly commonplace for illocutionary acts like the one performed in the case of (41) to be approached as particularised conversational implicatures. From this angle, one would assume that the skater inferentially works out that the speaker uttered what they did with the intention of issuing a warning (to not skate at some particular spot), simply by following the cooperative principle and considering the utterance context.

The most influential, and perhaps also most systematic attempt to fully explicate this line of reasoning can be found in Bach and Harnish (Reference Bach and Harnish1979). Like Strawson before them, Bach and Harnish recognised the existence of conventional illocutionary acts, but also suggested that the vast majority of illocutions are communicative in nature and thus need to be approached from an intention-based, inferentialist perspective. To this effect, they correspondingly posited that the success of these speech acts depends on the – by then – familiar Gricean process of pragmatic inference which is triggered by the hearer’s recognition of the speaker’s communicative intention. On this basis, they postulated a speech act schema, which constitutes a working out of the inference pattern that underlies the successful performance of speech acts. According to this schema, the hearer (a) recognises that the speaker is uttering an expression, (b) accesses its operative meaning,Footnote ¹⁷ (c) understands what the speaker is saying (by engaging in reference assignment, disambiguation, etc.), and (d) infers the speaker’s illocutionary intention. Obviously, in terms of the present discussion, the crux of this approach is this latter conception of an illocution as intentional action, which rests on salient contextual information (i.e., what Bach and Harnish called mutual contextual beliefs) and is regulated by the following uniformly applicable communicative presumption (Bach & Harnish, Reference Bach and Harnish1979: 7; emphasis my own):

5.4 Further Engagement with Speech Act Theory

Following the recognition of the category of indirect speech acts and the admittedly keen interest it generated, attention to speech act theory gradually started to fizzle out. As Levinson (Reference Levinson and Huang2017: 201–202) notes in this vein, the 1980s witnessed an abundance of studies

As Kissine (Reference Kissine, Allan and Jaszczolt2012: 169) correspondingly asserts, in recent years, ‘the main contemporary pragmatic theories of utterance interpretation devote little space, if any at all, to the way utterances are interpreted as speech acts, that is to the way they are assigned an illocutionary force’. When talking about utterance interpretation here, Kissine alludes of course to the so-called cognitive turn in pragmatics, which started dominating the field in the 1990s by shifting the focus from the mainly philosophical discussion of speaker meaning to the investigation of the inferential processing that underlies its recovery during verbal communication. Still, within this setting too, most of the work that has specifically touched on speech act theory over the past forty years or so has, for the most part, remained consonant with the received view outlined in this section.

A notable exception to this can be found in the case of relevance theory, the proponents of which famously argued that ‘the vast range of data that speech-act theorists have been concerned with is of no special interest to pragmatics’ (Sperber & Wilson, Reference Sperber and Wilson1995: 243). In a nutshell, the rationale behind this position is that, since speech acts are in their essence mainly social or institutional, illocutionary force assignment is not a necessary part of utterance comprehension; that is, hearers do not always need to recognise – let alone (meta)represent – the force with which an utterance was produced in order to work out its intended meaning. Rather, the extent to which this assignment is needed depends on whether the interpretation of an utterance requires such an attribution to achieve relevance, in the sense of yielding enough worthwhile information to offset the cognitive effort that the hearer spends in the process.Footnote ²⁰ In this picture, there are only three generic speech acts whose mental representation is deemed essential for the comprehension process, namely those of ‘saying’, ‘asking’, and ‘telling’. Recall, for instance, our examples in (33)–(35) – repeated in the following as (48)–(50) for convenience.

1. 48. You will go to the school party on Friday.

2. 49. Will you go to the school party on Friday?

3. 50. Go to the school party on Friday!

For relevance theorists, the mood indicators of an utterance’s underlying sentence belong to a wider category of linguistically encoded semantic constraints, dubbed procedures (after Blakemore Reference Blakemore1987; see also Wilson & Sperber, Reference Wilson and Sperber1993), which are specifically designed to guide the inferential comprehension process to some particular interpretive outcome. On this occasion, the relevant procedure would be to lead the hearer to embed the speaker’s explicitly expressed proposition in a higher-order speech act description. Assuming that our examples mentioned previously are addressed by, say, Mary to George, the descriptions for (49) and (50) could be approximated as in (51) and (52), respectively.

1. 51. Mary is asking George whether he will go to the school party on Friday.

2. 52. Mary is telling George to go to the school party on Friday.

From a relevance-theoretic angle then, (51) would be approached as a mental representation that describes a desirable thought, in the sense of marking Mary’s expectation for a relevant answer, while (52) as one that describes a desirable state of affairs.Footnote ²¹

Notwithstanding the extent to which illocutionary force attribution may be compulsory during utterance interpretation, it would appear at first sight that relevance theory’s approach to sentence mood may have something in common with Searle’s position presented in Section 5.1, since, in both cases, the proposition expressed by the declarative sentence in (48), when this sentence is transformed into an interrogative or imperative form in (49) and (50), continues to be embedded in the corresponding speech act representation. This conclusion, however, would be incorrect. Contrary to the Searlean picture, where, as we have seen, illocutionary forces are determined by the semantics of sentence types, in relevance theory they are always ‘recovered by a mixture of decoding and inference based on a variety of linguistic and non-linguistic clues’ (Wilson & Sperber, Reference Wilson, Sperber, Horn and Ward2004: 623; emphasis my own). As one may have guessed by now, reference is made here to the broader literalism/contextualism debate (see, e.g., Recanati, Reference Recanati2004b) that has dominated (narrow) pragmatics over the past couple of decades. For relevance theory, which falls under the contextualist camp, it is not only the illocutionary force carried by an utterance of an interrogative or imperative sentence, but even the very basic proposition explicitly expressed by the declarative form that requires contextual input.Footnote ²² For Searle, on the other hand, illocutionary forces can, in principle at least, be traced at the semantic level too, since, as per his principle of expressibility, ‘whenever one wishes to make an utterance with force F, it is always possible to utter a sentence the meaning of which expresses exactly force F’ (Searle, Reference Searle1968: 418).Footnote ²³

Despite the importance of the literalism/contextualism debate for modern pragmatic theory, neither the relevance-theoretic approach outlined earlier nor the overall debate itself is featured in recent pragmatically oriented overviews of speech act theory. All in all, it would seem that, at least as far as pragmatics is concerned, research on speech act theory has indeed reached a standstill. Given the centrality of the topic for the discipline though, this is rather astonishing, especially considering that ‘contemporary literature is rife with confusions’ in relation to ‘illocutionary force attribution’ (Kissine, Reference Kissine, Allan and Jaszczolt2012: 170) as well as that ‘the most fundamental issues’ surrounding speech acts have also not been ‘resolved at all’ (Levinson, Reference Levinson and Huang2017: 200). As we will now turn to see though, perhaps things are not as grim as they sound at this point.

6.1 On Meaning

Going back to Section 2, we can recall that during its establishment as a discipline in its own right, pragmatics was predominantly conceptualised, on par with the component view, as an ‘add-on’ to traditional semantic theory – the latter being itself an offshoot of the same ‘ideal language philosophy’ tradition that was responsible for the advent of logical positivism around Austin’s time. By the 1960s logical positivism had of course collapsed; yet, the scholarly fascination with formal analysis lived on. On the premise of the famous Montagovian argument that ‘there is […] no important theoretical difference between natural languages and the artificial languages of logicians’ (Montague, Reference Montague and Thomason)1974: 222), the 1970s witnessed the inception of formal semantics, which admittedly dominates linguistic thought to this very day. Ever since, truth conditions, compositionality and the general discussion of propositional content have remained at the centre of attention, as the main, if not exclusive foci of semantic investigation. This time though, those aspects of meaning that could not fit into formal theory were no longer perceived as deficiencies of natural language that render sentences meaningless – perhaps a result of the earlier demise of logical positivism. Instead, they were to be systematically investigated within the purview of (component) pragmatics.

From this angle, it was clear that Austin had made a valid point about the challenges that performatives pose for formal accounts of linguistic meaning, which he solidified further by showing that even constatives behave in the same way; the key challenge being that the meaning of utterances formulated as explicit performatives cannot really be assigned a truth value. Given this, and considering also the discussion in Section 5.1, it should be safe to assume that a main reason why Searle’s account ‘proved attractive to linguists’ (Levinson, Reference Levinson and Huang2017: 201) was that, by replacing Austin’s locutionary acts with propositional acts, it offered a principled way of tackling this very issue, while also safeguarding the purity of the formal approach through the reintroduction of propositional contents in the equation.

Going back to Austin’s original work, however, it is clear that the very notion of a speech act was developed with a view to altogether criticising, rather than endorsing, the traditional understanding of meaning in terms of propositions and truth conditions. For instance, while recapping his overall discussion in his final lecture, Austin (Reference Austin1975: 148; emphasis my own) stated that ‘the need for the general theory [of speech acts] arises simply because the traditional “statement” [i.e., the object of enquiry for traditional truth-conditional accounts of meaning] is an abstraction, an ideal, and so is its traditional truth or falsity’. Then, right after that, while listing some ‘morals’ of his account, he additionally stipulated that

On top of this, Austin had previously even questioned the extent to which truth or falsity themselves are objective measures. More specifically, during his already mentioned ‘objective assessment of the accomplished utterance’, Austin (Reference Austin1975: 143) reached the conclusion that ‘in real life, as opposed to the situations envisaged in logical theory, one cannot always answer in a simple manner whether [a constative …] is true or false’. From this angle, as Sadock (Reference Sadock, Horn and Ward2004: 57) aptly puts it, truth and falsity are in a way best seen as a convenient shorthand for assessing a specific group of illocutions: ‘with illocutionary acts of assertion, statement, and the like, we happen to call correspondence with the facts truth and a lack of it falsity, whereas in the case of other kinds of illocutions, we do not use those particular words’.

Of course, the issue here is not whether Searle’s proposal to substitute locutionary acts with propositional ones was somehow internally incoherent; after all, it provided an elegant alternative to Austin’s original notion of the locution, which apparently resonated better with scholars in the field. It is rather that ‘the introduction of propositions into Austin’s account of speech acts is not only not needed, but also inconsistent with the general picture of speech that Austin intend[ed] to provide’ (Sbisà, Reference Sbisà2024: 258). To make matters worse, Searle’s proposal has somehow managed to trump Austin’s original take to such an extent that, in the collective conscience of pragmaticists today, it is taken to be essentially co-extensive with what Austin himself had in mind when postulating speech acts in the first place. In other words, as Oishi (Reference Oishi, Allan, Capone and Kecskes2016: 337) points out, it is nowadays ‘taken for granted that Austin’s seminal idea of speech acts is developed and completed by Searle, in spite of the fact that their ideas are different at crucial points’. As a result, overviews of speech-act theory, including those found in the pragmatics literature, use the term ‘locutionary act’ to describe what is in essence Searle’s propositional act. On top of this, they also substitute – pretty much without fail – Austin’s explicit reference to ‘meaning’ (in juxtaposition to ‘force’) with reference to ‘content’. Consider, for example, the following formulation by Harris et al. (Reference Harris, Fogal, Moss, Fogal, Harris and Moss2018: 22; emphasis my own) in their very recent overview of the contemporary speech-act-theoretic landscape:

To be fair of course, Austin was not particularly forthcoming about his actual views on what meaning amounts to. The admittedly few scattered comments that he made in relation to his use of the term in How to Do Things with Words were rather cryptic, equating meaning with the traditional conceptions of sense and reference, which he took ‘on the strength of current views’ (Austin, Reference Austin1975: 149). However, he did explicitly state that the ‘notion of correspondence with facts’ which pertains to locutions needs to be ‘over-simplified because essentially it brings in the illocutionary aspect’ (Austin, Reference Austin1975: 146) and additionally recognised that ‘the theory of “meaning” as equivalent to “sense and reference” will certainly require some weeding-out and reformulating in terms of the distinction between locutionary and illocutionary acts’ (Austin, Reference Austin1975: 149).Footnote ²⁴ What is more, upon consideration of Austin’s collective works, one can readily see that, apart from being ‘critical of the identification of meaning with truth-conditions’, he also ‘rejected the objectification of “meanings”’ (Sbisà, Reference Sbisà2012: 5). All in all, as Sbisà (Reference Sbisà2012: 12) points out, for Austin, ‘meaning is […] something dynamic, something people do, and it is misleading to deal with it as with an entity or object’ (Sbisà, Reference Sbisà2012: 12).Footnote ²⁵ Ultimately though, this is precisely what Searle’s proposal to substitute the locutionary with the propositional act brought back into the picture, by essentially suggesting that ‘there must be something – the proposition: mental object, abstract object or whatever it may be – corresponding to the sentence uttered so as to constitute its meaning’ (Sbisà, Reference Sbisà2012: 13).

Perhaps more crucially even, the Searlean tradition did not just stop there. It objectified illocutionary force too, treating it as a ‘component of meaning’ (Searle & Vanderveken, Reference Searle and Vanderveken1985: 7) and eventually incorporating it in logical analysis even, by suggesting that ‘the ideal language of a universal grammar [à la Montague] must contain logical constants and operators capable of generating names for all possible illocutionary forces of utterances’ (Searle & Vanderveken, Reference Searle and Vanderveken1985: 7–8).Footnote ²⁶ In a way, as Burkhardt (Reference Burkhardt and Burkhardt1990: 125) pointed out around the time that pragmatic interest in speech-act theory had already started to dwindle, it seems that, as time went by, ‘speech act theorists, in one way or other, ha[d] put the cart before the horse and […], slowly, the insight began to grow that linguistic action is much more a (lexical) semantic than a pragmatic problem’.Footnote ²⁷ No wonder then that, by the 1990s, the relationship of pragmatics with speech act theory had already started to weaken, with pragmatic interest shifting to the study of speech acts as ‘performance phenomena’ that exhibit variation across ‘different groups of people’.

6.2 On Intention

To my mind, a parallel argument can also be made in relation to the role that the (equally entrenched) Gricean conception of speaker meaning has played in pragmatically oriented appraisals of speech act theory. Again, of course, the issue is not whether intention-based accounts of speech acts are somehow ill-formed in their inherent argumentation, but rather that they specifically cater to completing the received view’s formally oriented description of speech acts, departing again in substantial ways from Austin’s original rationale. In relation to this latter point, Sbisà’s (Reference Sbisà2009: 35) observation in the following is certainly telling:

As we have already seen, intention-based approaches to illocutionary force are largely attributed to Strawson, with Bach and Harnish playing a key role in their solidification. However, Searle also interfered with their contemporary conceptualisation, albeit more indirectly. Recall, from Section 5.3, Grice’s definition of meaning_NN, reproduced next for convenience:

Upon formulating this definition, Grice (1957: 385) further noted that ‘to ask what A meant is to ask for a specification of the intended effect (though, of course, it may not always be possible to get a straight answer involving a “that” clause, for example, “a belief that … ”)’. Considering Grice’s exemplification of this intended effect in terms of a belief and seeing how that belief would ultimately be the hearer’s, in speech-act-theoretic terms, it would follow that, for Grice, what we are dealing with is essentially a perlocutionary effect. Recognising this, Searle (Reference Searle1969: 47) famously argued that the said effect should be conceived of as illocutionary instead, since it ‘consists simply in the hearer understanding the utterance of the speaker’ rather than ‘a belief or response’.

Now, it is widely accepted that Searle’s reappraisal of speech acts in terms of constitutive rules (see Section 5.2.1) was intended to align, even if weakly so, his theoretical exposition with Austin’s conventionalism; the relevant hypothesis being that ‘the semantic structure of a language may be regarded as a conventional realization of a series of sets of underlying constitutive rules, and […] speech acts are acts characteristically performed by uttering expressions in accordance with these sets of constitutive rules’ (Searle, Reference Searle1969: 37; emphasis my own). Still, by semanticisingFootnote ²⁸ illocutionary force in this way while additionally suggesting that the illocutionary effect consists in the hearer ‘simply understanding’ the speaker’s meaning, Searle (like Strawson) can be seen to have directly contributed to ‘an impoverishment of the heuristic potential of the notion of illocutionary act in the analysis of verbal interaction’ (Sbisà, Reference Sbisà2009: 37).

In order to fully appreciate the relevant argument, we need to briefly circle back to the original Austinian notion of illocutionary effect. As already mentioned in Section 4.2.2, for Austin, illocutionary effects comprise the securing of uptake and the recognition of the typical effects that a speech act customarily brings about. Since the securing of uptake amounts to the securing of ‘the understanding of the meaning and of the force of the locution’ in this picture, if one is prepared to overlook the (significant of course) divergence in the contemporary received view with respect to the Austinian locution, some sort of equivalence could be established. What we are still missing, however, are ‘the typical effects a speech act customarily brings about’; that is, the two extra categories of illocutionary effects that Austin (Reference Austin1975: 117) identified: (a) the ‘certain ways’ in which ‘the illocutionary act “takes effect”’ and (b) the invitation ‘by convention of a response or sequel’ by some illocutions. Considering these in turn, we can readily witness an interesting picture emerging.

Unfortunately, as Sbisà (Reference Sbisà2009: 44) points out, the passage through which Austin introduced the ‘taking effect’ of the illocutionary act is ‘rather mysterious’ and the example he used to this effect is ‘not a very helpful one’, since it refers to a highly ceremonial act, which could easily be categorised as ‘conventional’ by an intentionalist:

Still, Sbisà offers an astute argument about how Austin’s insight on this occasion can (and probably should) be extended to cover all kinds of illocutions. Based on a juxtaposition with natural causation, which would be the ‘normal’ way of bringing about consequences (that is, perlocutionary effects in this context), Sbisà (Reference Sbisà2009: 45) suggests that illocutionary effects arise due to ‘changes in states of affairs belonging to the same level of reality as norms’, which in turn belong to ‘the realm of social conventions’. From this angle, understanding the speaker’s utterance as having a particular illocutionary force means taking it as an invocation of a socially determined procedure with a recognisable effect. So, in Austin’s example mentioned just now, and insofar as there are no infelicities in the performance of the illocutionary act at hand, we take the ship to have a new name because we understand the speaker’s utterance as belonging to the recognisable procedure of naming a ship. By the same token, alluding back to Strawson’s example in (41), in the absence of infelicities, we take it that the speaker ‘should not be attributed responsibility for whatever may happen to us in connection with the warning’s content’, because ‘we understand that the speaker’s utterance has the force of a warning’ (Sbisà, Reference Sbisà2009: 49). From this angle, if we are to posit, as Austin would, that these effects are conventional rather than intentional, their uptake would be secured through (tacit or explicit) intersubjective agreement, which also safeguards that they can be annuled in the presence of infelicities in the speech act’s deployment (on this point, see Sbisà, Reference Sbisà2007). Viewed in this way, the securing of uptake can be decoupled from speaker intentions, as an illocutionary act’s ‘taking effect’ would be the result of both the speaker and the hearer agreeing on the speech act’s illocutionary force.

To a large extent, this approach is additionally supported by the third kind of illocutionary effect posited by Austin, that is, the ‘invitation of a response or sequel’. While this specific effect only applies to some speech acts, like questions (which invite answers) or offers (which invite an acceptance or refusal), it can still be taken to indicate that, at least in these cases, ‘to perform an illocutionary act means to initiate a reproduction of an interaction patternFootnote ²⁹ that involves the speaker’s utterance and the hearer’s cooperative response to it’ (Witek, Reference Witek2015: 43). Obviously, the illocutionary effect in this case consists in the invitation of the sequel, with the sequel itself – i.e., the answer, the acceptance or refusal – being a perlocutionary effect; yet, a consideration of the latter can in principle also be deemed essential for the attribution of an illocutionary force to the speech act. As Sbisà (Reference Sbisà, Searle, Parrett and Verschueren1992: 101–102) discusses, that is because,

In order to see how speech act theory could accommodate this angle, the following argument by van Rees (Reference van Rees1992: 40; emphasis in original) can be particularly illuminating:

Despite having been made in the context of the criticism that speech-act theory cannot adequately account for ordinary conversation, launched against it by conversation analysts this time, van Rees’ argument can be seen to further substantiate Sbisà’s emphasis on the hearer’s response. However, the relevant nuance of Austin’s original account that is being referred to on both these occasions has been pushed to the background in the contemporary received view. That is because right from the beginning of his engagement with speech act theory, Searle (Reference Searle1969: 23; emphasis my own) has consistently referred to ‘complete speech acts as asserting, questioning, commanding, etc.’, using ‘the name “illocutionary acts”’; a practice that has now been adopted by most speech act theorists, who use the labels ‘speech act’ and ‘illocutionary act’ interchangeably.

Going back to his original argumentation though, it is pretty clear that Austin would never endorse an equivalence of the total speech act with the illocution alone, since in his model the ‘introduction of illocution between locution and perlocution prevents any conflation between meaning on the one hand, and efficacy in arousing feelings, dispositions, and reactions on the other’ (Sbisà, Reference Sbisà2024: 57).Footnote ³¹ In this regard, Austin’s explicit aim in the second half of his lecture series may have been to ‘fasten on the illocutionary act’, but his reason for doing so, as we also saw in Section 4.2.2, was because he wanted to ‘contrast it with the locutionary and perlocutionary ones’, rather than give it exclusive priority. As Oishi (Reference Oishi, Allan, Capone and Kecskes2016: 337) further explains, for Austin,

Indeed, from early on in How to Do Things with Words and while talking about constatives even (that is, utterances that were either way considered meaningful because of their truth conditions), Austin (Reference Austin1975: 52; emphasis my own) asserted that

This is something that he insisted on with respect to all speech acts, as he notably reiterated this position in his already mentioned final ‘morals’ too, where he emphasised that ‘the total speech act in the total speech situation is the only actual phenomenon which […] we are engaged in elucidating’ (Austin, Reference Austin1975: 148; emphasis in original).

6.3 Speech Act Theory and the Scope of Pragmatics

Considering that the motivation behind its original exposition was a need to show that the analysis of linguistic utterances goes far beyond the formal descriptive dimension, it is truly remarkable how, during its evolution, speech act theory ended up getting highly formalised too. As discussed in this section, this can be straightforwardly attributed to the domination of component pragmatics as the main conceptualisation of the field over a number of decades. After all, when one is to treat pragmatics as an add-on to semantic theory, what better way to incorporate speech acts within its remit than by grounding them on a semantic basis (through the postulation of the propositional act and accompanying IFIDs) and then attributing any extra meaning that they may carry over and above this basis (as in cases of indirection) to the tried and tested Gricean paradigm.

As a result, however, the pragmatic essence of speech acts has been severely impoverished, to the extent that current pragmatic theorising no longer engages with speech act theory, at least as much as its centrality for the discipline should be prompting it to. Within this frame, the recognition of illocutionary force has been reduced to a mainly categorical exercise, with forces themselves having been stripped of much of their role as functions of actions that bring about conventional effects of a varied (social, interpersonal, conversational …) nature. At the same time, the parallel focus on speaker intentions has rendered perlocution irrelevant for pragmatic enquiry, since by ‘stop[ping] so to speak at the recognition by the hearer of the illocutionary intentions of the speaker’ the question of ‘whether the hearer believes an assertion, executes an order, complies with a request, etc is not a subject of pragmatic rules’ (van Dijk, Reference van Dijk1977: 199). Clearly, all this has played a pivotal role in the attested diminishing interest in speech act theory, whereby ‘after the initial enthusiasm of the 1970s, so many scholars doing research on verbal interaction turned to other approaches’ (Sbisà, Reference Sbisà2009: 37).

Against this backdrop, if we are to identify the main reason why the theoretical discussion of speech acts within pragmatics has taken a back seat in recent years, we may well point again to Burkhardt’s (Reference Burkhardt and Burkhardt1990: 125) metaphorical reference to ‘the decline’ of the paradigm:

7.1 Operationalising Context

In order to understand the role that context can be taken to play in the delineation of speech acts from the broad pragmatic angle, we need to first appreciate how, in this conception, it goes beyond the description of a mere frame within which utterances are situated. For perspective pragmatics, contexts of verbal interaction additionally comprise sets of affordances that enable verbal exchanges to unravel in specific ways through the coordinated actions of the interlocutors. Echoing to some extent Levinson’s independent idea of a direct mapping of illocutionary force onto sentences in context, Mey (Reference Mey2001: 228; emphasis my own) correspondingly notes that, for broad pragmatics, speech actsFootnote ³²

Anyone familiar with the relevant literature should recognise the literature should recognise this description as directly linked to Mey’s notion of the pragmeme as a ‘general situational prototype, capable of being executed in the situation’ (Mey, Reference Mey2001: 221). Since for Mey (Reference Mey2001: 228) pragmemes ‘correspond pretty closely to what Levinson, in an earlier article, ha[d] called “activity types”’, however, for the purposes of the present exposition, I will treat these two terms as co-extensive. To elucidate the relevant notion, let’s look at the exchange in (54) between a customer C and a shopkeeper S in the rather familiar context of a short visit to the grocery shop:

1. 54. S (to last customer): Bub-bye.

   C: Some apples please. Just help myself is that alright?

   S: Yes they’re all fine.

   C: Yes they look good.

   S: // // There, that’s eighteen, orright?

   C: uhuh. You’ve just got the one kind of lettuce?

   S: Yes. Cos.

   → C: That’s a nice one.

   S: Yes. They are getting proper now aren’t they. Thirty six please.

   // // thank you very much.

   C: Thanks. Goodbye.

   (Levinson, Reference Levinson1979: 372)

Now, consider C’s utterance of ‘That’s a nice one’, which was presumably produced while pointing to some specific lettuce in this setting. As Levinson (Reference Levinson1979: 372; emphasis in original) points out, on this particular occasion, this single utterance simultaneously ‘counted as selecting a lettuce, requesting that it be wrapped, and undertaking to pay for it’, since ‘there were no further negotiations about the lettuce’ later on, and all these three actions were eventually fulfilled.Footnote ³³ This example effectively shows that, being first and foremost social (rather than simply linguistic) actions, speech acts also rely heavily on the discourse context in which they are situated. What activity types, or indeed pragmemes, add to the picture then is a set of affordances or ‘strict constraints on contributions to any particular activity’ which give rise to ‘strong expectations about the functions that any utterances at a certain point in the proceedings can be fulfilling’ (Levinson, Reference Levinson1979: 377). For Levinson (Reference Levinson1979: 371; emphasis in original), these affordances constitute ‘inferential schemata […, which] are tied to […] the structural properties of the activity in question’, but regardless of the way in which we may approach their specific logical or cognitive instantiation, the upshot is that they ultimately mediate verbal behaviour.

From the point of view of the present discussion, such an appraisal would suggest that the relevant contextual information is not just responsible for working out non-literal illocutionary forces, as per the received view on speech act theory, or for determining whether and how illocutionary force assignment is part of the process that leads to the reconstruction of speaker meaning, as a contextualist alternative, like the relevance-theoretic one presented in Section 5.4 might suggest. Discourse context is rather part and parcel of the way in which interlocutors coordinate their verbal exchanges, for it ultimately provides them with blueprints of the typical procedures for – and effects of – particular speech acts; a position that I take to be fully consonant with Austin’s conventionalism, as presented in Section 6.2.

7.2 Speech Acts in Interaction

Remaining on the discussion in Section 6.2, recall also how the interactional angle should also feature centrally in our theory of speech acts. Obviously, as already seen in Section 2.2, broad pragmatics is well-equipped to handle this, since it either way views meaning-making as an essentially collaborative process, whereby all interlocutors ‘co-construct meanings on the fly through tacit negotiations that take place as their interaction unravels’. In this respect, looking at instances of mundane conversation suggests that this may well apply to the identification of a speech act’s illocutionary force too.

Consider, for instance, the exchange in (55) which also took place in our familiar setting of a casual visit to the grocery shop:

1. 55. C: Do you have pecan Danish today ?

   → S: Yes we do. Would you like one of those ?

   C: Yes please

   S: Okay ((turns to get))

   (Merritt, Reference Merritt1976: 324; as cited in Levinson, Reference Levinson1983: 359)

Note S’s first turn in the transcript above. It starts with an utterance that is a direct answer to C’s question in the previous turn and then continues with the issuance of an indirect act of offering. In the course of the overall interaction, however, it is fairly clear that S’s performance of the speech act of offering is actually prompted by the question in C’s original turn. Exemplifying the role of the discourse context as a blueprint for typical exchanges in specific settings, the adoption of the interactional perspective allows us to explain how C’s seemingly information-seeking question is additionally interpreted by S as carrying the force of a request on this occasion.

As Levinson (Reference Levinson1983: 356–364) famously showed in his reanalysis of indirect requests, focusing on the position of a particular utterance in the course of an exchange can offer an alternative understanding of the way in which illocutionary forces are assigned to utterances during verbal interaction. Obviously, in order to appreciate this alternative, it is essential to first get briefly acquainted with some key terms in the conversation-analytic tradition (for a complete overview, see Levinson, Reference Levinson1983: Chapter 6). The main ingredients that we will need for this are the notions of ‘preference organisation’, ‘adjacency pair’, and ‘pre-sequence’, as well as the analytical term ‘position’. Starting with the latter, given that mundane conversation is typically a matter of continuous turn-taking, the term position is used in this setting to refer to the place of a turn, which typically plays some specific function in relation to other turns within an organised conversational sequence. For example, in the case of a question-answer adjacency pair, which we actually encountered twice in (55), the presence of a question in position 1 is typically expected to be followed by an answer in position 2. Preference organisation now has to do more specifically with the way in which such conversational sequences are customarily organised in the interest of promoting social solidarity. In the case of making a request, for instance, the preferred (that is, the unmarked) next turn would be one of granting the request, with a rejection being dispreferred (i.e., marked) in the relevant structural sense. Finally, in this setting, pre-sequences are turns that initiate a particular conversational move with a view to prefiguring a preferred next turn.

Within this frame, Levinson’s reanalysis attributed to indirect requests the role of a pre-sequence, as follows. From a conversation-analytic viewpoint, a prototypical request sequence exhibits a four-position structure, as in the case of (56), which again takes place in our familiar context.

1. 56.

   Position 1 – C: Hi. Do you have uh size C flashlight batteries ?
   	((PRE-REQUEST))
   Position 2 – S: Yes sir	((GO AHEAD))
   Position 3 – C: I’ll have four please	((REQUEST))
   Position 4 – S: ((turns to get))	((RESPONSE))
   (Merritt, Reference Merritt1976: 324; as cited in Levinson, Reference Levinson1983: 357)

Since, in terms of social solidarity, the dispreferred response in the case of a request is that of a refusal, the typical inclusion of a pre-request in the sequence is meant to check for the most likely grounds for such a refusal, with the presence of such grounds effectively allowing the initiator of the request to abort it altogether, as we can observe in (57):

1. 57. → C: Do you have Marlboros ?

   → S: Uh, no. We ran out

   C: Okay. Thanks anyway

   S: Sorry

   (Merritt, Reference Merritt1976: 325; as cited in Levinson, Reference Levinson1983: 358)

Still, as we already saw in (55), on top of offering grounds for refusal, pre-requests also allow the recipient to change the entire sequence from one of requesting to one of offering. That is because, for conversation analysts, in the presence of ‘preferences between alternative sequence types, and therefore between alternative first parts of adjacency pairs […], a pre-sequence can elicit from its recipient the preferred first part’, which suggests that, since ‘offers [a]re structurally preferred to requests as a way of getting transfers accomplished[, …] a pre-request can get an offer next, obviating the need for actually doing the request’ (Schegloff, Reference Schegloff and Psathas1979: 49). This much can even happen directly, as exemplified in (58):

1. 58. C: Have you got Embassy Gold please ?

   S: Yes dear ((provides))

   (Sinclair, Reference Sinclair1976: 60; as cited in Levinson, Reference Levinson1983: 361)

Following this rationale, Levinson concluded that, when it comes to the structural organisation of requests in verbal interaction – on condition, of course, that these are not aborted, as in (57) – there is a quite marked preference ranking: The most preferred sequence is one that directly goes from a position-1 pre-request to a position-4 response, as in (58), followed by a sequence where the pre-request prompts an offer on the part of the interlocutor, as in (55), with the otherwise prototypical sequence where the pre-request is followed by an invitation to make the request overtly, as in (56), being least preferred in this vein.

Regardless of the actual validity of this analysis of indirect requests as pre-requests (see, e.g., Fox, Reference Fox2015), what it quite clearly problematises is the traditional assumption that a speech act is tied exclusively to its producer’s choice of expression (for direct speech acts) or intention (for indirect ones). In this regard, the adoption of a conversation-analytic perspective can offer a fresh look that conforms with the view of speech acts as context-driven actions, in the sense discussed in Section 7.1. That is because it offers a principled way of identifying whether uptake has been secured through its so-called next-turn proof procedure, whereby ‘the display of [the parties’ understandings of prior turns’ talk] in the talk of subsequent turns affords both a resource for the analysis of prior turns and a proof procedure for professional analyses of prior turns – resources intrinsic to the data themselves’ (Sacks et al., Reference Sacks, Schegloff and Jefferson1974: 729).

At the same time, the relevance of the conversation-analytic method for the broad pragmatic perspective should be self-evident too. After all, as Bilmes (Reference Bilmes1988: 161–162; emphasis my own) underlines, conversation analysis provides a model

Circling back to the discussion in Section 6.2 then, it seems that this approach can be very usefully implemented in the accommodation of the interactional angle within speech act theory in a way that additionally aligns well with Austin’s purported conventionalism. This is something that Levinson (Reference Levinson and Huang2017: 204) also underlines, when he notes that, since ‘for success, the action depends on the uptake’ the extent to which ‘almost all speech acts are joint actions’ is ‘a fundamental aspect of speech acts neglected in Searlian analysis’.

7.3 Speech Acts and the Conditions of Society

Having further motivated the conceptualisation of speech acts as ‘sequentially positioned and contextually sensitive’ (Deppermann & Haugh, Reference Deppermann, Haugh, Deppermann and Haugh2022: 3), we can now turn to the way in which the conditions of society could also feature in a revamped theoretical account of speech acts that embraces the broad view of pragmatics. The main line of argumentation in this vein should be intuitively straightforward: since the execution of speech acts depends on the blueprints of typical procedures and effects that the discourse context engenders, and these blueprints are either way socially determined, speech acts are by default conditioned by society too.

For example, there is by now overwhelming evidence that ‘different cultures find expression in different systems of speech acts, and that different speech acts become entrenched, and, to some extent, codified in different languages’ (Wierzbicka, Reference Wierzbicka1985: 146). Clearly, this is why risks can arise in performing even the simplest speech act in the context of intercultural communication. Mey (Reference Mey, Allan, Capone and Kecskes2016: 122–123) offers a pertinent example:

Technically speaking, what we are dealing with in this case is a request that is not completed due to a hitch in its deployment (see Section 4.1.1). Clearly, what causes the hitch is the difference in the conventional procedure of performing a request in the speaker’s own culture vis-à-vis the Japanese one, which hinders and thus blocks the securing of uptake and, by extension, the intended perlocutionary effect from being realised. The limitation of the received view of speech act theory in this vein should also be evident: focusing solely on the intentions of the speaker may indeed allow us to recognise the speech act as a request, but its apparent misexecution remains unaccounted for, unless we allow for the relevant sociocultural considerations to enter the picture.

To be completely fair, the social dimension is of course not entirely alien to the received view of speech act theory. For instance, as we saw in Section 5.2.1, Searle distinguished between requests and orders on the grounds of the social positioning of the speaker and the hearer. What is more, he has also acknowledged that a ‘sentence only determines a set of truth conditions (or other sorts of conditions of satisfaction) against a background of assumptions and practices’ (Searle, Reference Searle, Searle, Kiefer and Bierwisch1980: 231), many of which are clearly social. Even so, this acknowledgment came with the caveat that we cannot really explicate these assumptions and practices ‘without generating an infinite regress’ (Searle, Reference Searle1983: 148). Therefore, as Pratt (Reference Pratt1986: 60) has noted, ‘while often acknowledging the[ir] theory’s dependence on undeveloped assumptions about social interaction, [speech act philosophers] argue that it is impossible to develop these assumptions in any satisfactory way’.

Pratt’s illuminating discussion in that very same paper reveals yet another shortcoming of the received view of speech acts, which this time applies to the component pragmatics tradition as a whole.

Indeed, one-to-one exchanges may be generally taken as prototypical within the domain of verbal interaction, but in reality they represent a rather idealised scenario. As Pratt (Reference Pratt1986: 61) rightly points out, several verbal exchanges involve ‘multiple participants with multiple intentions toward one another’, while at the same time ‘people have myriad encounters each day with utterances directed at a mass addressee’ too – an angle that has undoubtedly become particularly relevant with the advent of traditional and more recently social media as well. Making a similar observation, albeit in the context of discussing argumentation this time, Lewiński (Reference Lewiński2021b: 435) remarks that ‘this seemingly innocuous fact has the potential to turn speech act theory on its head’, since ‘the whole communicative transaction – the speaker’s complex intention […] and the hearers’ uptake – cannot be described anymore as a dyadic affair’. Obviously, this is something that the narrow take on pragmatics, with its focus on one-to-one exchanges, cannot straightforwardly address. Broad pragmatics, on the other hand, offers a perspective that naturally lends itself to analysing verbal interaction of a more complex nature, since, by starting from the discourse context, it considers ‘pragmatic acting [of any sort …] as adapting oneself, linguistically and otherwise, to one’s world […] and within the affordances it puts at [one’s] disposal’ (Mey, Reference Mey2001: 215). In this picture then, a pragmatic analysis of speech acts in polylogues would begin from a consideration of the discourse context in which they take place (e.g. a party game, a political debate, social media, etc.) and its specific affordances.

Finally, on top of the characteristics of the social setting in which the (dialogical or polylogical) interaction takes place, there is a further sense in which linguistic action is fundamentally influenced by the conditions of society. As Pratt (Reference Pratt1986: 68) has again aptly put it, the received view of speech act theory has largely overlooked the fact that ‘people always speak from and in a socially constituted position, a position that is, moreover, constantly shifting, and defined in a speech situation by the intersection of many different forces’. As one might be able to deduce from this quotation, reference is made here to the more general study of discursive practice as mediated by ideology.

Following Langton’s (Reference Langton1993) analysis of pornography as an illocutionary act of subordination and silencing towards women and Altman’s (Reference Altman1993) appraisal of hate speech as an illocutionary act of treating someone as a moral subordinate, there is by now a rich literature applying speech act theory to the investigation of social issues underpinned, among other things, by the use of language in specific ways. Notably though, a common characteristic of all these approaches is that they base their argumentation mainly on Austin’s original exposition, and specifically on the distinction between illocution and perlocution, rather than on the further elaborations that formed the received view presented in Section 5 (which, as we saw in Section 6.2, also pushed perlocution to the side). In fact, in one of the more recent papers following this trend, Saul (Reference Saul, Fogal, Harris and Moss2018: 361) recognises that the mainstream conception of speech act theory still has limitations when it comes to discussing ideologically-charged language use, as in the case of political manipulation through dogwhistling:Footnote ³⁴

Indeed, as the prolific research within the domain of Critical Discourse Studies has famously revealed, discourse is both ‘socially constitutive’, and ‘socially conditioned’ (Fairclough & Wodak, Reference Fairclough, Wodak and van Dijk1997: 258), in the sense that ‘it does not only influence, but also reflects […] social practices’ (Assimakopoulos, Reference Assimakopoulos2021b: 397). From our present perspective, this realisation alone should grant society a theoretical role that is much deeper than simply that of distinguishing between particular speech acts like requests and orders. Again, as Mey (Reference Mey2001: 320; emphasis in original) has correspondingly argued, the perspective take on pragmatics can accommodate this angle too: