1. The Nature of Meaningfulness

Let’s start from a pre-theoretic fact: some strings of symbols have a meaning in a particular language whereas others do not. For example, ‘Bertrand is British’ has a meaning in English, but not in Esperanto while ‘All mimsy were the borogoves’ doesn’t have a meaning in either. Furthermore, some expressions have multiple meanings or are ambiguous in a language. For example, ‘Bertrand went to the bank’ has multiple meanings in English. This invites the paradigmatic philosophical question:

This is a question about the nature or essence of having meaning or meaningfulness, calling for a reductive analysis of it in terms of something else.

To understand the question better, we need to say just a bit more about what we have in mind by linguistic expressions, meaning, and language.

Let’s call a string of symbols something that has phonetic and/or orthographic properties. Both ‘Bertrand is British’ and ‘*#&’ are strings. Strings are “types” and must be distinguished from their particular uses, events of producing utterances or inscriptions, the latter of which are standardly thought of as the expression’s “tokens”. A linguistic expression is a string of symbols that has syntactic properties and one or more meanings in some language. Expressions are therefore types and must be distinguished from both their uses and tokens. Consequently, the above question is a question about the linguistic meanings of types, what are sometimes called their context-invariant or standing meanings, not about the meaning-related or “semantic” properties of uses of expressions on particular occasions and with certain intentions etc. For example, it’s a question about the meanings of context-sensitive words like ‘I’ and ‘this’ versus the related properties of their uses in the sentences ‘I am a philosopher’ and ‘This is nicer than this’ on particular occasions.

By an expression’s meaning in a language I mean what fully competent speakers of the language have a grasp of in virtue of their semantic competence. Furthermore, it is what makes it possible to use that expression to speak that language and perform linguistic, meaning-related speech acts like saying something, asking questions, or telling someone to do something. For example, the meaning of ‘Bertrand is British’ is what speakers fully competent with it in English have a grasp of and that makes it possible for them to use that expression to speak English and to say that Bertrand is British.

By a language I mean some sort of public or communal entity, a sociolect rather than an idiolect. I find it natural to think of them as historically embedded, ongoing social practices (Dummett Reference Dummett1991: Ch. 4, Kaplan Reference Kaplan, Almog, Perry and Wettstein1989b, Ridge Reference Ridge2021).

Thus, the question about the nature of meaningfulness I’m interested in here is a question about what it is for expressions qua types to have the properties that fully competent speakers have a grasp of in a public language like English.

To prevent possible confusion, let’s briefly distinguish it from two other types of questions in philosophy of language, those of (i) descriptive semantics and those of (ii) foundational or metasemantics (Kaplan Reference Kaplan, Almog, Perry and Wettstein1989b: 573, Stalnaker Reference Stalnaker, Hale and Wright1997: 535). To provide a descriptive semantics for an expression is to describe its meaning by assigning it some sort of a semantic value (e. g. a referent, a truth-condition, intension, a structured proposition, a Kaplanian character etc.). More generally, to provide a descriptive semantics for a language is to provide a theory that describes the meanings of its expressions, usually by assigning some sorts of semantic values to the atomic expressions and explaining how the semantic values of complex expressions depend on those of the atomic ones together with their syntactic structure (Szabo Reference Szabo, Lepore and Sosa2019). In contrast, to provide a foundational or a metasemantics for an expression is to say what makes it the case or metaphysically determines that it has the meaning that it does, what grounds its meaning. More generally, to provide a metasemantics for a language is to say what makes it the case that its expressions have the meanings they do.

Our question about the nature of meaningfulness is not identical to either of these questions. To do descriptive semantics is clearly not to try to answer our question. Rather, when doing descriptive semantics philosophers and linguists usually presuppose something about the how their preferred semantic values are related to meanings. For example, it’s standardly taken for granted that there’s a connection between a declarative sentence’s meaning and its truth and thus, that we get at least some information about a sentence’s meaning by stating its truth-conditions (Heim & Kratzer Reference Heim and Kratzer1997:1, Soames Reference Soames1992: 17). However, what is presupposed might be very general and compatible with lots of different views about the nature of meaningfulness, even if it rules out some.

Of course, one would expect that our question is instead discussed under the rubric of foundational semantics or metasemantics, and sometimes it is (compare Davies Reference Davies, Devitt and Hanley2006: 22). But we need to be clear about the fact that there are two separate questions here. It is one thing to ask about the nature of meaningfulness. It is another thing to ask about the determination or grounding of meaning-facts either in general or about particular expressions. It is easiest to see this if you consider the fact that primitivism about the nature of property F is compatible with a substantive story about F-making, determination, or grounding. For example, take theories of the nature of truth and theories of truth-making. One could be a primitivist about truth yet think that there’s a theory to be had as to what makes different sorts of true propositions true (Asay Reference Asay2013). Conversely, one could adopt a reductive analysis of truth without thereby settling all questions about truth-making. The same applies in the case of theories of the nature of meaningfulness and metasemantics (= theories of meaning-making). One could be a primitivist about meaningfulness yet think that there’s theory to be had as to what makes expressions in general or particular expressions mean what they do. In fact, this is exactly what Paul Horwich seems to think. Horwich is a sort of deflationary primitivist since he thinks that meaningfulness has no nature or essence and thus no analysis of meaningfulness is to be had, but just a story about determination (Horwich Reference Horwich1998: 6). Conversely, one could adopt a reductive analysis of meaningfulness without thereby settling all questions about meaning-making.

The primary purpose of this section has been to introduce the question about the nature of meaningfulness and to make sure it is properly distinguished from the question about determination of meaning facts. Let’s now proceed to Lewis’s view.

2. Lewis on Languages and Language

Lewis starts with a contrast between two ways of looking at language. On the one hand, we can think of languages as abstract objects that pair expressions with their meanings. On the other hand, we can think of language as a form of social, rational human activity. The question is how the two perspectives can be reconciled.

For Lewis, a language (more intuitively, a possible language) £ is a set of strings of symbols paired with meanings. A sentence’s meaning is something that, minimally, combines with the world to get us a truth-value. It can therefore be thought of as an intention, a function from possible worlds to truth-values. Importantly, no answer is given in the case of languages as to what it is for a sentence to have a meaning. The natural view is that ‘meaningfulness’ is here just used in a stipulated, technical sense (Schiffer Reference Schiffer, Hale, Wright and Miller2017: 50).

Lewis’s quasi-technical setup in relating his languages to language consists in the question:

In other words, Lewis’s question, put in more intuitive terms, is what is it for one of the possible languages to be the actual language of a particular community.

His answer is the following:

A convention is, very roughly, a regularity in action which is such that (more or less) everyone behaves in this way and believes that others do as well, it is self-perpetuating, and arbitrary in that some other regularity could’ve done equally well. We’ll look at a more precise definition in the next section.

To be truthful in £1 is to try to never utter a sentence unless it is true-in-£1. In other words, it’s to utter only those sentences one believes to be true-in-£1. So the convention of truthfulness, generalized, can be thought to be a convention of sincerity in a language. Trust is a bit trickier. Lewis says that to be trusting is both to impute truthfulness to others and to tend to respond to other’s utterances by coming to believe that the uttered sentences are true-in-£1 (Lewis Reference Lewis and Gunderson1975: 7). This runs together two different notions: trust proper, which consists in taking others to be sincere, and reliance, which consists in actually coming to form the same beliefs (for discussion see Kölbel 1998). We could therefore think of trust primarily in terms of trust proper. A convention of truthfulness and trust in £1 is thus the regularity in uttering only those sentences one believes to be true-in-£1 and in responding to people’s utterances by coming to believe that they believe that the uttered sentences are true-in-£1.

This is Lewis’s basic story. It has been widely taken to constitute an answer to the nature question, for example, by Bennett, Davis, Loar, and Keiser whose own related proposals are meant to answer that question (Bennett Reference Bennett1976, Davis Reference Davis2003, Reference Davis2005, Loar Reference Loar, Balog and Beardman1976, Reference Loar1981, Keiser Reference Keiser2022, Reference Keiser2023, see also Blackburn Reference Blackburn1984: 118-127, Schiffer Reference Schiffer, Hale, Wright and Miller2017, Stalnaker Reference Stalnaker1984: 32). To see how this goes, keep in mind that as possible languages £ go, meaningfulness is used in a stipulated, technical sense (Schiffer Reference Schiffer, Hale, Wright and Miller2017: 50). The real question about the nature of meaningfulness pertains to the actual language of a community. Lewis’s view can be taken to answer it by saying that for a sentence to be meaningful in an actual language of community P is for there to be a convention in P of using it in certain conditions and a convention of regularly responding to uses in a certain way (compare Kölbel 1998: 303).

In what follows I won’t focus on Lewis’s own view because it is complex and open to interpretation and it is not my aim to get into interpretive disputes. Instead, I’ll present a simplified, Lewis-inspired Conventions view of meaningfulness which embodies its basic idea and argue that it is inferior to a version of the Rules view.

3. Conventions View

Here’s the main thrust of the simplified, Lewis-inspired Conventions view. For an expression to be meaningful in a language of a community is for there to be a conventional regularity in the community of using it in certain conditions. Meanings can then be identified with conventional use-conditions. To develop it further, let’s first clear some ground and abstract away from certain irrelevant features of Lewis’s own package.

One important question about Lewis’s view concerns how to extend the view from sentences to sub-sentential expressions. This has partly to do with the fact that Lewis seems to accept a version of the sentential primacy thesis and a bifurcated, top-down view on which questions about the meaningfulness of sentences are to be answered first and questions about the meaningfulness of sub-sentential expressions are to be answered on that basis.Footnote ² For the sake of manageability, I will focus here on sentences as well, but this implies no commitment to the sentential primacy thesis (for critical discussion see Davis Reference Davis2003: section 8.4, Keiser Reference Keiser2022: 28-30). We will be abstracting away from that dispute.

Another important question concerns how to accommodate context-sensitivity, interrogatives, imperatives etc. (Lewis Reference Lewis and Gunderson1975: 13-14). The view, interpreted in terms of use-conditions, is ideal for accommodating context-sensitivity and mood, but, again, I will leave that to the side.

On the Conventions view we’re considering, for an expression to be meaningful in a community’s language is for there to be a conventional regularity in the community to use it in certain conditions. When it comes to declarative sentences, this is a conventional regularity in truthfulness, in other words, of using sentences in conditions when one has certain beliefs (compare Appiah Reference Appiah1985: 149, Harman Reference Harman1999: 199, Jackson Reference Jackson2006: 57). Using [[ ]]UC for a function that assigns an expression to its use-condition, we can represent this as follows (where s stands for the speaker):

To take another example, consider the interjection ‘Ouch!’ which is standardly taken to semantically express a sudden, sharp pain (Kaplan MS). Here the relevant conventional regularity would be one of sincerity, in other words, of using it in conditions when one has such a pain. We can then represent it’s use-condition as follows:

How do such conventional regularities figure into language use? The idea is that speakers that are fully competent with the expressions in English grasp them. When a speaker uses ‘Bertrand is British!’ or ‘Ouch!’ with its meaning they are somehow related to the conventional regularity and it is because of this that they count as expressing the belief that Bertrand is British and saying that Bertrand is British or as expressing pain. If the hearer has a grip on the conventional regularity, she can infer from the speaker’s use that the former has expressed the relevant belief and said what they have or expressed pain.

This is the view that I’ll take as my target. To fully understand it we need to say a bit more about what it is for there to be a conventional regularity in use. I will here rely on the conceptual framework for distinguishing social practices, conventions, and norms or rules presented by Brennan, Eriksson, Goodin, and Southwood, in their Explaining Norms (Brennan et al. Reference Brennan, Eriksson, Goodin and Southwood2013). Let’s start with the notion of a social practice. Social practices are regularities in action that are explained by the presence of pro-attitudes towards the action that are a matter of common knowledge (Brennan et al. Reference Brennan, Eriksson, Goodin and Southwood2013: 16). Conventions are a species of social practices that are solutions to coordination problems, cases where two (or more) parties want to behave in a mutually advantageous way conditional on others behaving this way where there are several different ways to do that. A standard example of such a problem is choosing a side of the road to drive on. On Lewis’s view, a regularity R in the behavior of members of a community P when they are agents in a recurrent situation S is a convention if and only if:

1. 1. they conform to R (Regularity)

2. 2. they believe that almost everyone conforms to R (Belief)

3. 3. they expect that it will continue to be true that almost everyone conforms to R and this gives them a reason to continue conforming as well (Reason)

4. 4. they have a general preference for general conformity with R rather than slightly-less-than-general conformity (Preference)

5. 5. there is an alternative regularity R* such that it could be the case that 1-4 are true of R* (Arbitrariness)

6. 6. they have common knowledge of 1-5; (Common Knowledge).

On the Conventions view conventional regularities of use are thus understood to be such solutions to coordination problems.

Conventional regularities do not amount to rules. For our purposes the essential difference between conventional regularities and rules pertains to their different sorts of normativity. Conventional regularities are normative only in the sense that they give people reasons to conform and in the sense that conformity to them is socially enforced in that non-conformity will get you sanctioned (Lewis Reference Lewis1969: 97-100). But such convention-generated reasons to conform are always to be balanced against other reasons not to conform. Even if it is generally true that one ought to conform to the convention, the “ought” simply evaporates, if, on a particular occasion, there’s more reason to do the opposite. A rational agent would in such a case simply conclude that one ought not to conform. Furthermore, the reasons to conform are dependent on people’s own preferences, their seeing the value of having the conventional regularity (Gilbert Reference Gilbert1989: 352-355, Guala Reference Guala2013: 3111). If one’s preferences change, the reason goes away.

In contrast, rules are normative not in giving people reasons to conform, but setting down strict requirements, prohibitions, and permissions (we’ll say a lot more later). Once a rule exists, its normativity is isolated from reasons in the sense that the requirement etc. doesn’t evaporate if, on a particular occasion, there’s more reason to do the opposite (Reiland Reference Reiland2024, Schauer Reference Schauer1991: Ch. 3.4). A rational agent wouldn’t in this case conclude that there is no rule setting a requirement, but rather that they should break the rule.Footnote ³ Furthermore, once the rule exists, the requirement etc. it sets is not dependent on people’s preferences, their seeing the value of having the rule. One’s preferences don’t have any impact on the existence of a rule.

This completes my presentation of the simplified Conventions view of meaningfulness. Let’s now proceed to criticism.

4. Speaking a Language and Use with Meaning

Not every use of an expression amounts to speaking a language, using an expression with its meaning, and performing a linguistic act. The basic notion of a use of an expression is something that must be specifiable in non-semantic terms. After all, it’s the fact that uses so conceived have some further properties, like being conventionally regular in certain conditions, that make it the case that the expressions have meanings. So, a mere use of an expression needs to be understood in terms of the tokening of an expression independently of its meaning. In Austin’s terms, mere uses are phatic acts (Austin Reference Austin1962: VIII). Such uses don’t amount to speaking a language or performing a linguistic act such as saying something. Mere uses need to be distinguished from what Kaplan calls uses with meaning which amount to what Austin called rhetic acts, e. g. using a sentence to say something (Austin Reference Austin1962: VIII, Kaplan Reference Kaplan, Almog, Perry and Wettstein1989b: 603). Such uses do amount to speaking a language and performing linguistic acts.

Let’s look at this distinction more closely to articulate a series of constraints on an adequate conception of use with meaning.

Consider first uses by completely incompetent speakers. Take a speaker who doesn’t know how to speak any Estonian and uses one of its sentences without having the slightest clue what it means. Assuming you don’t speak any Estonian, you can try it yourself by uttering or inscribing a sentence of Estonian like ‘Lumi on valge’. Plausibly, even though you’ve used a sentence of Estonian, you haven’t spoken any Estonian or said anything (compare Austin Reference Austin1962: 97). In our terminology, you’ve merely used the sentence, but you haven’t used it with its meaning. However, when a competent speaker utters the same sentence using it with its meaning she will have spoken Estonian and will have used the sentence to say that snow is white.Footnote ⁴

Importantly, the above cases involving uses by incompetent speakers are not the only cases that show that we need to distinguish between mere uses and uses with meaning. Competence doesn’t guarantee that one’s use will be a use with meaning. Even competent speakers can merely use expressions. Consider, for example, a competent speaker who utters some sentence of English with the purpose of practicing pronunciation or testing a microphone or who inscribes one with the purpose of practicing handwriting or improving typing speed. It should be clear that even though in all these cases they’ve used a sentence of English, they haven’t spoken any English nor used it to say anything. Rather, they’ve merely used the sentence.

We can also see the need to distinguish between mere uses and uses with meaning by reflecting on ambiguity. Suppose that languages contain ambiguous expressions: expressions which have more than one meaning. Consider then an ambiguous sentence like ‘Bertrand went to the bank’ which means either the same as ‘Bertrand went to the financial bank’ or the same as ‘Bertrand went to the riverbank’. When a competent English speaker normally uses such a sentence she doesn’t use it to say two things. Rather, she uses the sentence with just one of its meanings and says a single thing (Bach & Harnish Reference Bach and Harnish1979: 20-23).

We’ve now seen three different cases which demonstrate the need to distinguish between mere uses and uses with meaning. Here they are, again:

1. A. Uses by Incompetent Speakers

2. B. Phonological/Orthographic Uses

3. C. Ambiguity

These cases don’t just show that there’s an obvious need to distinguish between mere uses and uses with meaning. They also help us to articulate initial constraints on an intuitive conception of use with meaning.

For starters, Uses by Incompetent Speakers shows that some amount of semantic competence or at least the speaker’s having some idea about the expression’s meaning, even if a mistaken one, is necessary for use with meaning. This gives us our first constraint:

Next, Phonological/Orthographical uses show that semantic competence is not sufficient. Rather, to use an expression with its meaning even a fully competent speaker has to do something or at least have certain properties (the exact reason for this proviso will become clear later). Furthermore, Ambiguity shows that what the speaker must do can’t just be a matter of something general like trying to communicate with someone etc. Rather, the speaker must do something that results in her activating the expression’s meaning in that language, or the speaker needs to have a property which guarantees activation. This gives us our second constraint:

In sum, we can say that to use an expression with its meaning in a language, a speaker has to be minimally competent with the expression in that language and have some property that guarantees activation of the meaning. In the next section I’ll argue that the standard Conventionalist story of use with meaning fails.

5. Conventions and Use with Meaning

Every theory of meaningfulness in a language must yield a plausible picture of what it is to speak a language, that is, of what it is to use an expression with its meaning. How can the Conventions view do that? Let’s start with a simple, but implausible view that nobody has held, that enables us to illuminate a further aspect of use with meaning:

On this view, to use an expression with its meaning is to use it while being in the conventional use-conditions. It is easy to see that this won’t do since it satisfies neither Competence, nor Activation. Even incompetent speakers can use expressions while accidentally being in the conventional use-conditions and that clearly doesn’t result in them using the expressions with their meanings. For example, an incompetent speaker who is learning to pronounce ‘Ouch!’ could utter it, while at the same time being struck by a sudden, sharp pain. This clearly wouldn’t result in their speaking English, using ‘Ouch!’ with its meaning and expressing pain. After all, they have no clue about its meaning, nor do they do anything to activate it.

Furthermore, at least on certain conceptions of what the use-conditions consist in, it seems possible to speak a language and use an expression with its meaning while not being in the relevant conditions. For example, if the use-conditions consist in a mental state such that you can be mistaken about whether you have it, you could use an expression with its meaning while inadvertently not being in the relevant conditions. In other words, given certain conceptions of use-conditions it is possible to use an expression with its meaning while misusing it – using it while the use-conditions don’t obtain. This gives us a further constraint on an account of use with meaning:

However, if to use an expression with its meaning is to conform to the convention then, necessarily, all uses of an expression with its meaning are also correct uses, uses while the use-conditions obtain. And this means that the Conformity view can’t satisfy Misuse.

Those who have adopted something like the Conventions view have instead standardly opted for a related idea, namely that to use an expression is not to conform to the convention, but to try to conform to it:

On this view, to use an expression with its meaning is not to use it in the conventional use-conditions, but to try to use it in such conditions. This satisfies Competence because only minimally semantically competent speakers have some idea about the conventional regularities and can try to conform to them. Furthermore, it might also be taken to satisfy Activation because it requires speakers to do something that puts them in a relation to the use-condition. And it also satisfies Misuse because trying to conform is not inconsistent with inadvertently failing to do so.

Unfortunately, this view still fails. This is because not only does it seem possible to use expressions with their meanings while inadvertently misusing them, it also seems possible to do this while intentionally misusing them – using them while knowing that the use-conditions don’t obtain.

Here the best cases involve intentional misuses for the purposes of deception. If the conventional regularity for declaratives involves belief, then it seems entirely possible to use a sentence with its meaning while knowing that you do not believe the relevant proposition. For example, suppose I know you trust me on matters philosophical and I want to deceive you into thinking that Ludwig was German. I could then utter ‘Ludwig was German’ with its meaning, expressing the belief that Ludwig is German and saying that Ludwig is German, while knowing full well that he was Austrian. To take a different example, if I want to deceive you into thinking that I just had a sudden sharp pain, I can utter ‘Ouch!’ with its meaning, expressing a sudden sharp pain, even if I know full well that I didn’t just have such a pain.

All of this shows that we should acknowledge a fourth constraint on an account of use with meaning:

It should be clear that the Participation view can’t account for this because it understands use with meaning in terms of trying to conform to the conventional regularity which is incompatible with intentional non-conformity. After all, if one is using an expression while knowing that the use-conditions don’t obtain, then one couldn’t be trying to conform to the regularity.

In the next two sections I will introduce the Rules view and show that it can do better. The key is that the notion of a rule enables us to separate the fact of its being in force for a speaker, governing their performances, from the speaker’s trying to act in conformity with it. On the Rules view use with meaning consists in using it while the rule is in force for the speaker which is compatible with intentionally breaking the rule. It will then also become clear why nothing structurally analogous can be found on the Conventions view without inflating it into the Rules view, completing the case against it.

6. Rules View

On the Rules view, for an expression to be meaningful is for it to be governed by a rule of use that tells you in which conditions you’re permitted to use it. Meanings can be identified with permissible use-conditions (Alston Reference Alston2000, Kaplan MS, Reimer Reference Reimer2004, Reiland Reference Reiland2023b).

A rule, in the sense in play here, is a propositional content that attributes some action-type some deontic status like being required, forbidden, or permitted and that is in force. Rules typically come to be in force due to being enacted by an authority or by being personally or collectively accepted (Alston Reference Alston2000: 251, Hart Reference Hart1961, Reiland Reference Reiland2024, Reinach Reference Reinach1913: Ch. 5, Schauer Reference Schauer1991: Ch. 1). Moral and other normative truths, while they can have the same content, are not rules in this sense. Laws are. Like rules of games which give actions their roles in a game, rules of language are constitutive rules which give expressions their meanings. They do this by specifying the conditions in which it is permissible for a speaker to use the expressions. I will use as the normal form for stating such rules (using s for speakers, e for expressions and C for use-conditions):

We can then think of the expression’s meaning as its use-condition C.

The rule for ‘Bertrand is British’, keeping the Lewisian use-conditions, and the rule for ‘Ouch!’ are the following:

How do such rules figure into language use? The idea is that speakers who are fully competent with the expressions in English grasp them. When a speaker uses ‘Bertrand is British!’ or ‘Ouch!’ with its meaning they are somehow related to the rule and it is because of this they count as expressing the belief that Bertrand is British and saying that Bertrand is British or as expressing pain. If the hearer has a grip on the rule, she can infer from the speaker’s use that the former has expressed the relevant belief and said what they have or expressed pain. Given this brief sketch we can now ask how the Rules view can account for use with meaning.

7. Rules and Use with Meaning

It should be clear that on the Rules view we shouldn’t take use with meaning to consist in acting in accordance with the rule nor with trying to follow the rule. These views would be subject to the same difficulties as the Conformity and Participation views. But the Rules view has a structural feature that allows it to go beyond these views. Namely, rules are things that are in force. The simple idea is that to use an expression with its meaning is to use it while the relevant rule is in force for you:

This idea is a straightforward application of a more general view of constitutive rules and how they constitute activities like playing a game. The basic idea is that rules constitute any activity by being in force (Alston Reference Alston2000, Reiland Reference Reiland2020, Williamson Reference Williamson1996). Following Lewis’s view of languages, we can think of a game like chess as a set of pieces together with their roles in the game. Now, on a rule-based view of games, a “Lewisian” game is nothing but a set of propositional contents of the right sort like:

To start playing a game at a time is to take the contents and put them in force for ourselves at the time plus some further conditions like perhaps aiming to win. Thus, a game by itself is just a mere set of propositional contents of the right sort. However, to play a game at a time is to perform the relevant antecedent actions while treating these contents as rules.

We can think of “Lewisian” languages or language-stages similarly as sets of expressions together with their meanings. On the Rules view, a “Lewisian” language is nothing but a set of propositional contents of the right sort. But speaking a language might not be quite as straightforward as taking the contents and putting them in force for ourselves.

In the case of games, it’s natural to think that starting to play at all and which game we play is usually a matter of voluntary decision.Footnote ⁵ Contrast this with Williamson’s view of assertion. He conceives of asserting as a matter of default presumption. As he puts it: “In natural languages, the default use of declarative sentences is to make assertions” (Williamson Reference Williamson1996: 511). The idea is that whenever we say things, the rule of assertion is in force by default. This would have to be because it is somehow generally accepted among speakers of any language. Hence, to assert you don’t have to put it in force at the time of saying. Rather, you merely have to continue accepting it as being in force at the time of saying. You could opt out or discontinue the acceptance by indicating that you’re merely saying things or rather conjecturing or guessing etc. For example, you could do this by using an explicit performative like ‘I conjecture…’ (Williamson Reference Williamson1996: 496). However, to assert, to continue accepting the rule, you don’t have to do anything beyond not opting out.

In the case of speaking a language there is plausibly an element of both default presumption and voluntary decision. On the Rules view what makes mere uses into uses with meaning is the fact that the rules are in force. It seems plausible to think that whether we speak a language at all and which language we speak is a matter of default presumption (compare Bach & Harnish Reference Bach and Harnish1979: 7). For example, when I utter ‘Ludwig is German’ in someone’s presence, there’s a default presumption that I’m not merely practicing pronunciation, but rather using it with its meaning in English. More generally, when a minimally competent speaker uses expressions of a language in the presence of others there seems to be a dual default presumption. First, that one is not merely uttering the expressions, but speaking a language. Second, that one is speaking some particular Lewisian language that is generally spoken in the community that the conversants are part of. In other words, there’s the presumption that one is speaking the language of the relevant community.

On this picture the rules of a particular Lewisian language are in force by default in a community, which makes that the actual language of the community. They’re in force by default because they’re conventionally accepted in the community (Southwood Reference Southwood, Plunkett, Shapiro and Toh2019). Thus, if someone is a minimally competent speaker and a member of the community then to speak the language, to use its expressions with its meanings, she doesn’t have to themselves do anything to put the rules in force at the time of the use. Rather, she merely has to continue accepting the rules as being in force. She could opt out or discontinue the acceptance by making it clear that she’s merely practicing pronunciation or testing a microphone (defeating the default presumption that she’s speaking a language). Similarly, she could opt out by making it clear that she’s engaging in linguistic innovation in trying to use a familiar word with a new meaning or when she’s speaking a different language which shares the same words (defeating the default presumption that she’s speaking the actual language). However, to speak the language or use its expressions with their meanings, she doesn’t have to do anything beyond not opting out.

On the other hand, consider ambiguity. In that case it seems that the default presumption doesn’t do the trick by itself. Rather, the speaker has authority over which of the rules they accept at the time and can voluntarily choose, subject to the usual constraints on intention-formation. For example, in uttering ‘Bertrand went to the bank’ I accept one of the two rules, and this is what makes it the case that I say one thing and not the other (Alston Reference Alston2000: 60, Bach & Harnish Reference Bach and Harnish1979: 20-23).

To sum up, on the Rules view, for a speaker to use an expression with its meaning is for the rule to be in force for them at the time of their use. For example, for a speaker to use ‘Ouch!’ with its meaning in English is for them to use it while (2*), above, is in force for them. This satisfies Competence because only minimally semantically competent speakers count as members of the linguistic community, as those for whom the rules of English are in force by default. Furthermore, it satisfies Activation because it requires speakers to have a property that guarantees activation of meaning, namely, being such that the rule is in force for them. However, it also satisfies both Misuse and Intentional Misuse. After all, being such that the rule is in force for one is not incompatible with either inadvertently or intentionally breaking it. We break rules both inadvertently and intentionally all the time. For example, using ‘Ouch!’ while (2*) is in force for one isn’t incompatible with using it while knowing that one is not in pain.

It should now also become clear why nothing structurally analogous can be found on the Conventions view. Arguably, the only way of achieving this would be to think of conventional regularities as things that could be in force in the same way. But this would essentially make them into rules and simply inflate the Conventions view into the Rules view.