Problemos ISSN 1392-1126 eISSN 2424-6158

2026, vol. 109, pp. 103–118 DOI: https://doi.org/10.15388/Problemos.2026.109.8

Normativity of Meaning Contextualised: A Davidsonian Account

Bartosz Wultański
University of Warsaw, Poland
b.wultanski@uw.edu.pl
https://ror.org/039bjqg32

Abstract. This paper proposes a novel contextualist account of the normativity of meaning, grounded in Davidsonian metasemantics. It presents the view that the normativity of meaning can arise not from global conventions or prescriptive rules, but from the local evidence-based practice under the constitutive ideal of rationality. By analysing how we achieve what Davidson called ‘first’ meaning, the paper demonstrates how context-local standards emerge within the boundaries set by the principle of charity. This framework preserves semantic systematicity and stability, and explains phenomena such as idiosyncrasy and linguistic mistakes without appealing to conventionalism, thereby breaking out of the dilemma between weak ‘correctness’ and overly strong ‘prescriptive’ construals of the normativity of meaning. The result is a robustly normative, yet moderately anti-conventionalist theory of meaning, which does not rule out the existence of linguistic conventions but rather takes their role to be far less significant.
Keywords: normativity of meaning, Davidson, radical interpretation, principle of charity, contextualism.

Kontekstualizuotas reikšmės normatyvumas: deividsoniška perspektyva

Santrauka. Šiame straipsnyje siūloma nauja kontekstualistinė reikšmės normatyvumo samprata, grindžiama Davidsono metasemantika. Joje teigiama, kad reikšmės normatyvumas kyla ne iš visuotinių konvencijų ar preskriptyvių taisyklių, o iš lokalių, įrodymais grįstų praktikų, vadovaujantis konstitutyviu racionalumo idealu. Analizuojant, kaip pasiekiame tai, ką Davidsonas vadina „pirmąja“ reikšme (first meaning), straipsnyje atskleidžiama, kaip „geranoriškumo (charity) principu“ nustatytose ribose iškyla lokalūs konteksto standartai. Laikantis šios teorinės nuostatos, išsaugomas semantinis sistemingumas ir stabilumas, taip pat paaiškinami tokie reiškiniai kaip idiosinkratiškumas ir kalbinės klaidos, nesiremiant konvencionalizmu. Taip išvengiama dilemos tarp silpnosios reikšmės normatyvumo kaip „taisyklingumo“ sampratos ir pernelyg stiprios, „preskriptyviomis“ taisyklėmis grindžiamos jo interpretacijos. Šių pastangų rezultatas – nuosekli normatyvinė, tačiau nuosaikiai antikonvencionalistinė reikšmės teorija, kuri neatmeta kalbinių konvencijų egzistavimo, bet suteikia joms kur kas mažiau reikšmingumo.
Pagrindiniai žodžiai: reikšmės normatyvumas, Davidsonas, radikali interpretacija, geranoriškumo (charity) principas, kontekstualizmas

__________

Received: 02/10/2025. Accepted: 26/09/2026
Copyright © Bartosz Wultański, 2026. Published by Vilnius University Press.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (CC BY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

1. Introduction

1.1. Semantic Normativity and Semantic Rules

Since the slogan meaning is normative has spread in the philosophy of language due to Kripke’s reconstruction of Wittgenstein’s rule-following considerations (Kripke 1982), many contributions have been made to disambiguate this claim and clarify its actual consequences for semantics and metasemantics. Generally, we can take this slogan to mean that semantics consists in some intrinsic normative component that is essential for meaning. In practice, most semantic normativists claim that meaning is entangled with some kind of semantic norm or standard. There are various ways to formulate such a norm (see, e.g., Zadeh 2015) but, arguably, the two most noticeable and most discussed ones are the approaches based on (1) correctness conditions or (2) prescriptive rules of use (see Fennell 2013; Glüer et al. 2024).

Supporters of the correctness view claim that meaningful expressions come with correct and incorrect ways to use them. For example, if ‘cat’ means a cat, then ‘cat’ applies correctly only to a cat, which sets the normative standard for how the expression ‘cat’ should and should not be used, or, as Boghossian (1989: 513) puts it:

Suppose the expression ‘green’ means green. It follows immediately that the expression ‘green’ applies correctly only to these things (the green ones) and not to those (the non-greens).

It is difficult to challenge the claim that meaningful expressions have correct and incorrect uses, and so anti-normativists chose a different strategy: they claim that normativity does not necessarily follow from correctness. If I incorrectly point at a dog and say that it is a cat, it does not mean that I failed to do something I ought to do (to use the expression ‘cat’ only to cats), but that I merely asserted a false statement (see Glüer & Wikforss 2009). So, even if, in some cases, ‘correctness’ can be a normative notion, it does not always have to be, and, in the particular case of semantic correctness, it does not seem to be.

The prescriptive view defends a somewhat similar but stronger thesis, according to which, normativity is not mediated by correctness; instead, semantic norms directly regulate what we ought to say by virtue of what the expression in question means. Glüer, Wikforss, and Ganapini, in their overview article for the Stanford Encyclopedia of Philosophy (2024), are mentioning the following formulation (among others) of a prescriptive norm concerning the meaning of the word ‘green’:

For any speaker S, and any time t: if “green” means green for S at t, then S ought to apply “green” to an object x iff x is green at t1.

While the correctness view has turned out to be too weak to support the normativity of meaning thesis, the prescriptive view is argued to be too strong to be plausible. If we endorse the well-known Kantian formula that “ought implies can”, we end up with unwanted consequences for the prescriptive view, namely, that we ought to apply ‘cat’ to all cats, or ‘green’ to all green objects – which is an obligation that we simply cannot fulfil (see Hattiangadi 2006: 227). Moreover, it seems absurd that any semantic norm would require anything like that. However, if we weaken the condition back to correctness or use other deontic operators, such as ‘may’, the definition may not be strong enough to support normativity. Thus, the most popular formulations of semantic norms face a dilemma: they are either too weak to support the normativity of meaning or too strong and have unwanted consequences.

1.2. The Alternative Approach

However, to defend the normativity of meaning thesis, one does not have to adhere to either of the two aforementioned construals – and one can reconstruct normativity in semantics in a different way2. One of the most important metasemantic frameworks present in the discussion is the Davidsonian approach based on the principle of charity and the concept of radical interpretation. Davidsonian metasemantics has been widely debated by both normativists and anti-normativists, each presenting Davidson as aligned with their own position. Normativists maintain that, despite Davidson’s rejection of conventionalism (Davidson 1984), normativity is preserved through the principle of charity which requires interpreting an interlocutor’s utterances to maximise their truth, coherence, and agreement with our beliefs. Anti-normativists, by contrast, contend that the principle of charity functions as a constitutive ideal enabling communication but does not impose prescriptive norms or standards of correctness (Wikforss 2001: 215). On this view, Davidsonian (meta)semantics leaves no room for genuine linguistic error, and only affords for atypical or unexpected uses of language. As long as we remain rational agents whose utterances are interpretable under the principle of charity, the correctness conditions are unnecessary for grasping meaning.

The anti-normative reading of Davidson’s philosophy of language has gained popularity, but it has also faced significant criticism. Verheggen (2018: 59)3, for example, defends some form of meaning normativism, contending that Davidson’s account incorporates a form of normativity in the guise of a hypothetical imperative: if our aim is successful communication, then we ought to express ourselves in certain ways rather than others, with standards of correctness arising through triangulation4. Fennell (2015: 83) presents a different normative reading of Davidson’s metasemantics, arguing that even if the principle of charity is constitutive, it nevertheless retains an inherently normative dimension.

In this paper, I undertake to develop a normativist account grounded in Davidsonian metasemantics that avoids both the ‘correctness’ and ‘prescriptive’ construals of the normativity of meaning5. Drawing on Verheggen and Fennell, I argue that normativity must be situated within a context in order to attribute meaning to expressions but, unlike them, I trace the normativity of meaning not to constitutive or hypothetical rules of charity themselves, but to an evidence-guided practice under the constitutive ideal of rationality. In the following section, I shall summarise the relevant discussion of the role of normativity in Davidsonian metasemantics. In Section 3, I present a new iteration of semantic normativism grounded in context-local theories of language and revisit the question of the role of convention in language. I conclude with Section 4, addressing some key challenges that my view may face concerning systematicity, stability, regularity, idiosyncrasy, and linguistic mistakes.

2. Davidsonian Metasemantics: The Constitutive Ideal and Normativity

2.1. The Principle of Charity as a Constitutive Ideal

Anti-normativist readings of Davidson’s philosophy of language often begin with his critique (Davidson 1984) of David Lewis’s conventionalism (1969). Lewis (1975: 8) defended meaning normativism by arguing that conventions are normative foundations of language; by conventions, he understood regularities in action (or regularities in both action and belief) that at least two persons are willing to conform to. Davidson does not deny that conventions are an important part of language, or that language is, overall, a social phenomenon; however, he argues that the relation between language and convention goes the other way around – we do not need conventions to have language, but we need language to have conventions.

In the metasemantic story that Davidson presents, there is no place for conventions – at least not as a foundation of linguistic communication. Radical interpretation is a thought experiment in which one has to establish the meanings of the utterances of another person without any prior semantic knowledge about the language they use and without any other assumptions about their lives, customs, thoughts, culture, etc. The lack of presuppositions is important because it is what distinguishes radical interpretation from interpretation from scratch (see Glock 2003: 168). It also excludes the interpreter’s initial knowledge about the interlocutor’s linguistic conventions and any other regular behaviour, conventionalised or not. To make such an interpretation possible, one has to adhere to the principle of charity. This principle helps us to sort evidence and observations and test them in light of any interpretative theory we may come up with. The principle of charity is not “a rigid set of rules but to a collection of ‘rough maxims and methodological generalities’ that the interpreter must supplement by intelligent guesswork in particular cases” (Ibid.: 194). Despite its generality and many different formulations of the principle6, Davidson distinguishes two aspects of it: the principle of correspondence and the principle of coherence. The first one assumes a causal relationship between objects and our beliefs in the external world (or that the conditions that cause speakers to believe that a given sentence is true are in general the same as the truth conditions of this sentence [Ibid.]), and the other aspect concerns interpretation that maximises consistency in the holistic system of one’s thoughts and utterances (see Davidson 2001).

The problem with the principle of charity is that it does not seem to be normative. Even if it is sometimes presented as a rule, such as “we must attribute mental states to people in such a way that our attributions make them as rational as possible” (Wedgwood 2007: 27), for Davidson himself, and for most of his interpreters, it was not a prescriptive rule but, rather, a constitutive ideal. While prescriptive rules set the standard for ‘right’ and ‘wrong’, constitutive rules, broadly speaking, create new situations or institutions. By the constitutive ideal of rationality in the context of Davidson’s philosophy I mean the set of constitutive rules which are necessary for attributing intentional states – such as thoughts, beliefs, or desires – to an agent in the first place. It seems that there is not much choice in following such a rule as it is not the case that this constitutive ideal obliges us to be rational; we (and our interlocutors) either meet the requirements to interpret and to be interpreted – or not. If we were not to meet those requirements, formulating an interpretative theory of language and communicating would be impossible. The principle of charity constitutes states of affairs and does not seem to say anything about what we should – or should not – do. We cannot say that the principle in any way tells us (or others) what to do in order to be charitable but, rather, states what the requirements are for being considered a rational agent (see Glüer & Pagin 1999).

2.2. Challenging the Purely Constitutive Reading

However, as mentioned above, it is possible to reconstruct the principle of charity differently: not as a constitutive ideal but as a hypothetical imperative. If we want to successfully communicate, we should apply the principle of charity as a means to achieve the desired goal. According to Verheggen (2018: 49), the standard for our success in such an endeavour is set by the process of semantic triangulation which fixes meanings between us, our interlocutors, and the external world, all assuming the principle of charity. We simply start to assign the meanings of utterances we are interpreting to different things in the world, and then we test these theories in conversational practice. In this way, ‘correctness’ is not a conventional linguistic standard but a normative construal of the meanings of our utterances in the process of triangulation which Verheggen also calls “semantic attitudinal normativity”. The problem that this approach has to face is whether we really choose to communicate or interpret. It is certainly not the case that we can simply stop understanding what someone means by their utterances and remain ignorant7. When we hear a meaningful utterance, it is hard to dismiss it as random noise. We can, of course, make a choice about whether we want to say something true or false, but what is important here is our decision to be understood and to understand others. Although we can use some vague language on purpose (e.g., if we want to hide something or extricate ourselves from a conversation), it is difficult to avoid rationally accommodating meaningful expressions8. While interpreting the principle of charity as hypothetical conditional gives us some valuable insights, it involves some sort of volition which seems to be at odds with the notable point from a previous section – that to fulfil the requirements set by the principle of charity is to fit the conditions rather than to make a deliberate decision.

There is also another, more straightforward way to challenge Davidson’s anti-normativist reading, and it is to show that we can derive normativity directly from the constitutive ideal of charity. Fennell (2015: 84) argues that even if Davidson’s metasemantics is anti-conventionalist, we can still label him as a moderate normativist as constitutive charity allows for failure and the ability to do otherwise. He points out that the anti-normativist reading of the principle of charity reconstructs the constitutive ideal too strictly. Davidson does not demand perfect rationality in every scenario but rather reasonability in interpretation. We can allow our interpretation to assume that someone holds irrational or largely false beliefs, as long as it is intelligible in a given context which can include circumstances of an individual’s life, the culture they grew up in, their educational background, etc. Grandy (1973: 443) called it “the principle of humanity” which allows for interpretation that is psychologically plausible – we expect someone to hold beliefs that their epistemic circumstances allow for9. This way, the constitutive ideal of charity is weakened (as the constraints of correspondence and coherence can be bent given the epistemic context of a given individual), and therefore it allows for error in an individual’s beliefs. A point that Fennell does not mention is that this weakening also allows for error in interpretation because we cannot know all the relevant circumstances that can affect our interpretative theory. Thus, we can think that someone “should have believed otherwise”, even if it was reasonable for them to hold that specific belief given those specific circumstances. Additionally, we ourselves can make a mistake in interpreting someone’s utterances. Hence, there is a choice to make about how to interpret, but not whether to interpret. Constitutive charity makes interpretation possible, but it does not determine the outcome of the process. We could say, metaphorically, that we should not look for the normativity of meaning outside the constitutive ideal but within it.

3. Normativity of Meaning in Local Contexts.

Building on Verheggen’s theory and Fennell’s counterarguments against anti-normativism, I believe that we can develop a better picture of the normativity of meaning that arises in the process of interpretation under the constitutive construal of charity, where the standard is set not by global, conventionalised meaning norms (for each individual word or sentence) that create a normative structure of language but, rather, metasemantically speaking, by each local interpretative context which determines the standard relative to given circumstances. To make that possible, we need to return to Davidson’s considerations about conventionalism, starting with what he called prior and passing theories.

3.1. Prior and Passing Theories

In A Nice Derangement of Epitaphs, Davidson (1986: 93) coins the term first meaning and reconstructs, on the basis of what he takes to be common belief among philosophers and linguists, its three principles. First meaning is some variation of what others call literal meaning but, because of philosophical baggage and theoretical implications, Davidson did not want to adopt this term into his own theory. In general, we could say that first meaning comes first in the order of interpretation, so, given a ‘normal’ audience and a ‘normal’ occasion, the first meaning of an utterance is something we can find in the dictionary which describes how words are standardly used. The three principles are the following: The first meaning is (a) systematic, which means that there must be systematic relations between meanings and utterances for them to be interpretable. Then, the first meaning is (b) shared in some way by the speaker and the interpreter. And finally, first meanings are (c) prepared, which means that they are governed by previously learned conventions and regularities that are required for successful interpretation. While he admits that, in some way, his theory can preserve (a) and (b), there is no way that (c) can be upheld (given that we understand malapropisms and linguistic inventions), which leads us to the famous passage:

<…> there is no such thing as a language, not if a language is anything like what many philosophers and linguists have supposed. There is therefore no such thing to be learned, mastered, or be born with. We must give up the idea of a clearly defined shared structure which language-users acquire and then apply to cases (Ibid.: 107).

In his counterproposal to the convention-governed view of language, expressed explicitly in condition (c), Davidson presents the notions of prior and passing theories. The prior theory is a starting point and background of interpretation, a set of expectations that the interpreter has towards how the speaker is going to use language. This can be formed by the speaker’s origin, nationality, gender, social class, etc., but also by various external contextual factors such as the place or nature of the meeting. Linguistic conventions can (but do not have to) be part of the prior theory, as we most likely assume that a speaker (if they are speaking our language) is going to use some conventionalised standard of how to speak in a given context. So, even if conventions do not constitute meanings, they often strongly affect one’s prior theory. On the speaker’s side, the prior theory is their expectation of what the interpreter’s prior theory is going to be – how the speaker is most likely going to be understood. The passing theory, on the other hand, is how we actually interpret the speaker’s utterances after testing and adjusting our prior theory accordingly10. For the speaker, the passing theory is how they want to be interpreted – which theory of language they want the interpreter to use11.

Davidson builds an image of language as something that is not some external structure in which language users participate by learning pre-made rules that govern the meanings of each word or sentence, but as something that emerges every time in the process of interpretation, in specific local contexts and circumstances. The so-called first meaning is what comes out in the passing theory, not the prior theory. First meanings are not ready-made in some given structure but have yet to be constructed in interpretation12, which stands directly against condition (c). Camp (2016) describes this position as radically contextualist and compares Davidson to authors such as Noam Chomsky, François Recanati, or Robyn Carston. However, she labels Davidson’s proposal as more plausible, as he is not sceptical about semantics as a whole (unlike Chomsky, who relocates language from the social space to an intrinsic biological system, or Recanati, who basically replaces semantics with ‘truth-conditional pragmatics’); rather, Davisdson is trying to preserve first meaning, despite rejecting conventionalism which he sees as a threat to semantics.

3.2. Contextualised Local Norms

At first glance it may seem that, in rejecting conventionalism, Davidson is also rejecting the normativity of meaning, as rules such as “apply ‘green’ to and only to green objects” are generally formed arbitrarily by the force of convention. He even writes:

For there are no rules for arriving at passing theories, no rules in any strict sense, as opposed to rough maxims and methodological generalities. A passing theory really is like a theory at least in this that it is derived by wit, luck, and wisdom from a private vocabulary and grammar, knowledge of the ways people get their point across, and rules of thumb for figuring out what deviations from the dictionary are most likely (Davidson 1986: 107).

However, the absence of strict rules does not have to mean that there is no normativity involved in the process of interpretation. The passing theory sets the standard that one has to meet to some degree to arrive at first meaning. It does not matter that there are no strict instructions for how exactly to do it every time. The “rough maxims and methodological generalities” can become more specific given the context and circumstances. Take, for example, an unusual utterance like “Actually, fish do not exist”. If we know that the speaker is a conspiracy theorist, we may assume that they mean that fish are some sort of hoax, maybe that they are actually robots controlled by the government to spy on us. However, if the speaker is a marine biologist or a specialist in the sociology of knowledge, we should rather pursue the interpretation that there is something wrong with ‘fish’ as a biological category (which is actually the case, since organisms that we call ‘fish’ do not form a monophyletic group as, for example, mammals do). Different contexts give rise to different standards and guide us toward the most plausible passing theory at any given time. If that were not the case, it would be impossible to adjust and test our prior and passing theories, as there would not be any standard to which we could adjust them. This standard does not have to be clearly and strictly set, as the interpreter’s and the speaker’s passing theories do not have to be identical (or, indeed, they cannot be, because what would it mean for them to be the same?13), but they have to overlap to the extent that we can claim that the communication was successful. Davidson writes:

For the passing theory is the one the interpreter actually uses to interpret an utterance, and it is the theory the speaker intends the interpreter to use. Only if these coincide is understanding complete. (Of course, there are degrees of success in communication; much may be right although something is wrong. This matter of degree is irrelevant to my argument.) The passing theory is where, accident aside, agreement is greatest (Ibid.: 102).

Our prior and passing theories can be better or worse, as can be our interpretations, but what ultimately matters most is whether they are successful – whether they maximise agreement within the circumstances we are in.

If there is no ultimate set of rules and strict standards guiding our interpretation, this means that we also have a choice while weighing evidence that can corroborate or falsify our prior and passing theories. In face-to-face conversations, the situation is quite simple – if we have two possible interpretations of a speaker’s utterance, we can test both of them and choose the one that fits best. A more problematic example is text interpretation where our testing possibilities are limited, and where we have to weigh evidence more carefully, as the risk of misinterpretation is higher. Ultimately, arriving at a passing theory is more or less a matter of choice14; if we agree that we have met the standard to the point of successful communication, there is not much left to do (from a metasemantic perspective). We can, of course, deepen our understanding of the meaning we came up with, but it does not change the semantics unless we decide to exchange the whole interpretation for an alternative one.

To summarise, if we take the interpretative practice under the construal of charity as a practice which 1) is subject to a context-specific standard (even if somewhat vaguely described), 2) can produce better or worse, but still rational, interpretations that can also be mistaken, and 3) leaves both the interpreter and the speaker a decision to make about which passing theory to choose, and how to weigh evidence that affects their interpretative theories, then we should not have any doubts that the meanings which are the result of this practice are inherently normative. I believe that my approach is also coherent with Verheggen’s (2018) view on normative standards that are set in the process of triangulation, as all that I call the evidence corroborating the passing theory – circumstances, context, and specific position of the speaker – is positioned within the triangular relation of the interpreter, the speaker (or, in general, other people who fulfil the constitutive ideal of charity), and the external world, a point on which Verheggen puts great emphasis in her proposal.

4. Addressing Key Challenges

4.1. Systematicity

Systematicity is the first principle of the first meaning that Davidson mentions. Systematic relations between meanings and utterances, and between complex linguistic expressions and their parts, are what enables us to understand potentially infinite meaningful utterances on the basis of their components. This constraint can be met by appealing to linguistic rules and conventions that would create a holistic system of interdependent semantic elements. Davidson, of course, rejects that possibility but still wants to meet the systematicity requirement15. As Camp (2016: 119) points out, local contexts as separate individual cases are not enough for systematicity to occur, as they include only a single utterance made by a single speaker at a single time. While it is true that each passing theory’s explanatory powers are limited to singular expressions, we cannot say that, overall, our interpretation is strictly restricted to a given local context. We can extrapolate from different cases of interpretation and similar social situations, agents, or circumstances, and base our interpretation not only on the actual case but also on our experiences and broader contexts. Systematicity is needed to explain how we are able to produce and interpret a potentially indefinite number of utterances, and Davidson’s framework meets this requirement, even if in a counterfactual manner. A passing theory is particular, but a given interpretative strategy that helps us achieve the passing theory is a question which can work in many different cases. Particular meaning that arises in context-local interpretation is not like a Leibnizian monad – it can be a part of a more compositionally complex sentence. Systematicity is preserved here differently than usual, but it is preserved nevertheless16.

4.2. Stability, Idiosyncrasy, and Regularity

Another possible obstacle for a Davidsonian account of the normativity of meaning can be the lack of regularity and stability that conventions and general, word-specific meaning rules provide. This is one of the main concerns of both Camp (Ibid.) and Fennell (2015) which makes them lean towards a moderate version of contextualism rather than Davidsonian radical contextualism. However, their positions are not exactly the same. Fennell mentions non-standard usages of language that can be difficult to explain without reference to common, ‘global’17 linguistic structure backed by conventions. Any ‘unusual’ utterance presupposes the existence of an external standard that should be irrelevant from the perspective of meanings embedded in local contexts. The problem appears to be that the theory we want to defend has a hard time making sense of idiosyncrasy and linguistic inventions. Supposedly, we have to weaken our theory in order to accommodate these issues within its framework. This is the path Fennell chooses by presenting a weakened version of Davidson’s anti-conventionalism: it is not the case that we do not need conventions or external, shared standards at all, but we do not need any specific ones. To explain the idiosyncratic use of one word or sentence, we need to assume that most other utterances are in line with their conventional meaning, and, due to their stability, we can make an inference about the unstable, exotic meaning in question. There are no highlighted conventions that we need to keep, but still, we cannot abandon them all.

Camp also defends moderate conventionalism, but her insight is slightly different from Fennell’s. In her view, conventions do not have to be strict and fixed, but they are one of our biggest interpretative clues, and, most of the time, they vastly overlap with an interpreter’s prior theories. This makes them a default interpretational stance. When talking about irony, metaphors, malapropisms, dialects, or other non-standard uses of language, we can establish them as non-standard because we know the conventionalised standard to which these linguistic phenomena are exceptions. Conventions play a big role in the normativity of meaning and have significant explanatory force when formulating our prior theories, but Camp insists that they are not strictly a necessary condition for achieving meaning and successful communication. If radical interpretation is possible (and Camp does not say it is not), there must be a way to meaningfully communicate outside of convention. From the metasemantic point of view of Davidson’s theory, conventions are redundant; however, they surely can play a significant role in the interpretation process, especially since, in most scenarios, we are not in a situation of radical interpretation (or even translation from scratch). But, what is more important in the Davidsonian account I am presenting here it is that the normativity of meaning can be preserved without conventions, as normativity is derived from local, contextualised standards that can, but do not have to, be affected by conventions or general, ‘global’ rules, as the overlapping of conventions and prior theories is coincidental. We can say that, in general, conventions make interpretation easier, but they are not required for interpretation to be possible.

There is still a question to answer: what stabilises meanings, if not external linguistic norms that we apply in communication? The answer is the regularity of our linguistic behaviour. It is worth noting that, even in Lewis’ conventionalism, it is not that conventions make our use of language regular – it is the other way around: the regularities in behaviour are what constitutes conventions, which, for Lewis, are simply the most efficient way to achieve successful communication. Conventions could not be the basis for language, as they are made through language. Meaning is determined through a coordination game, and conventions just happen to be the optimal strategy (see Lewis 1969; Keiser 2023). We can say the same in Davidson’s framework – regularity is needed to fulfil the coherence condition of the charity construal. We cannot interpret someone as a coherent subject if their communicative behaviour is not regular but random. Even in the case where someone used a linguistic invention only once, in order to interpret its meaning, we have to assume that this use is regular and stable, and so that its meaning does not fluctuate and change at random but is intended to be certain in possible future cases. Otherwise, it is not interpretable at all, and the speaker does not meet the requirements of constitutive charity.

I believe that we can agree with Fennell and Camp18 that conventions play an explanatory role in cases of idiosyncrasy, without falling into any form of conventionalism in metasemantics and discourse on the normativity of meaning. Even if we need to refer to conventions to explain idiosyncrasy, we do not have to do so to explain the normativity of meaning or justify our metasemantic theory. It is safe to say that idiosyncrasy is possible only if the prior theory is overlapping at least to some extent with conventionalised rules of meaning. In the case of radical interpretation, we cannot even apply the concept of idiosyncrasy, as we do not have any reason to do so without broader knowledge about the speaker, their life and customs, the culture they live in, and finally the conventions they share in their community (if any). Idiosyncrasy is not essential for meaning and its normativity, just as convention is not.

4.3. Linguistic Mistakes

The argument from the impossibility of making a mistake was probably the most prominent argument against normativity in Davidson’s (meta)semantics. I already addressed that issue briefly, but it may need some further elaboration. Bilgrami (1992: 103) claims that Davidson’s anti-conventionalism means that no word has a built-in correctness condition. For example, if someone calls a ketch a ‘yawl’, this need not be a mistake – it might simply mean that, in their idiolect, ‘yawl’ = ketch. Similarly, Bert (from Burge’s ‘arthritis’ example; see Burge 1979) might not misuse ‘arthritis’ but rather mean something like tharthritis (which is a distinct concept). If that were true, it would be impossible to make a linguistic mistake in Davidson’s framework. All such apparent errors would be just unexpected uses of language, and it would make no sense to evaluate them as incorrect.

My answer to this allegation aligns in some way with Fennell’s (2015: 69, 73) view. Depending on the evidence and the prior theory, we might arrive at a different interpretation of this kind of unusual use. We could say that the speaker actually wanted to call a ketch a ‘yawl’, knowing what both of these words mean (for some specific reason), or maybe they wanted to use that name referentially and made a mistake (in the given context), or maybe they do, in fact, call all ketches ‘yawls’. Making a mistake in the presented framework would not be related to some general, ‘global’ standard but to a local, contextualised standard that our interpretation does not align with. We cannot be mistaken at the level of the general maxims of correspondence and coherence that make interpretation possible, but we can make a blunder within this constitutive ideal, e.g., by attributing to the speaker some particular belief that they, in fact, do not have. It would cover the cases where the interpreter agrees with the speaker on the passing theory without enough testing, while assuming that their passing theories overlap when they do not. Perhaps, more importantly, it would explain the exemplary case of a speaker who wanted to match their utterance with the expectations of the interpreter to maximise the chance of being understood, but they were mistaken about what the actual expectations were, or they had a wrong belief about how the word they used regarding the way it was commonly used by the audience.

It is important to note that many normativists (see, e.g., Glock 2015: 845) distinguish between two types of correctness: semantic correctness, which involves understanding the word’s meaning, and factual correctness, which relies on how the world actually is. From this distinction follow two types of mistakes. Linguistic mistakes (as presented above) are a product of using a word without actual knowledge of its meaning, like when we call a ketch a ‘yawl’, because we do not know the definition of ‘yawl’. Factual mistakes, on the other hand, appear when we do, in fact, know the meaning of the word but are mistaken in the matter of facts, e.g., misidentifying a yawl (while knowing what a yawl is) because of the poor visibility conditions. The former category is the focus of this paper, but I do not see a reason why the latter could not also be governed by the proposed framework.

4.4. Many faces of conventionalism

In the presented view, conventions are unnecessary for the metasemantics, which means that they are not required to ground meaning. However, it can be pointed out that normativists such as Glock (2010) or Schroeder (2024) do not endorse the idea that conventions play a significant metasemantic role in grounding meaning, but rather that they constitute meaning, answering the question “What is meaning?” and not “What grounds meaning?”. It is a valid point. Nonetheless, one can argue that it relies on a narrow interpretation of the notion of metasemantics. According to Burgess and Sherman (2014: 3), we can distinguish three aspects of metasemantics: (1) basic metasemantics, which asks, “What are semantic properties grounded in?”; (2) theory of meaning, which asks, “What is it for an expression to mean something?”; and (3) metaphysics of semantic values, which asks, “What is meaning?”. Given this distinction, we can say that both of the aforementioned questions are metasemantic in nature, they just pertain to different features of it. Nevertheless, metasemantic or not, a question about grounding is different from a question about constitution. As the topic of the ontology of meaning goes beyond the scope of this paper, it has to be said that the proposed idea opposes conventionalism understood as a basic metasemantic position and not the metaphysical one.

That being said, even if convention is presented as something not different from meaning itself, not something ‘external’ that governs it but as its intrinsic component, the alternative position still stands on its own, because convention (however it is formulated) is not a mandatory part of prior theory or passing theory. By stating that conventions are ‘external’ or ‘premade’, I do not mean that they are by any means not arbitrary or intersubjectively shared, as arbitrariness is built into the concept of convention. The ‘externality’ of convention indicates here that it creates a general, ‘global’ standard of correctness independent of our will. As speakers, we are in some way committed to this standard, even if our individual ability to shape this convention is very limited, not to say non-existent19. Conversely, context-local standards of communication are partially dependent on the speaker and interpreter agreement conditions, which can, to some extent, be shaped by themselves20.

5. Conclusion

We should not assume that if a theory of meaning cannot fit into the ‘correctness’ or ‘prescriptive’ account of normativity, such a theory has no place for the normativity of meaning at all. However, there are certain tensions between the specific framework presented in this paper and some aspects of standard normativist theory of meaning. In this paper, I have presented an account of meaning that is undoubtedly normative (without deriving its normativity from ‘global’ norms or conventions) while preserving constraints of systematicity, stability, and regularity, and retaining the ability to explain phenomena such as malapropisms (and linguistic mistakes in general), idiosyncrasy, and novel uses of language. This is achieved without committing to the idea that some utterances are generally and ultimately correct or incorrect, or that meaning norms require of us some strict, specific use, independent of the context, as in the ‘prescriptive’ construal of the normativity of meaning. To summarise, in this paper I have argued that (a) there is an alternative way to understand semantic normativity, (b) this alternative way relies on Davidson’s metasemantics, and (c) it can accommodate certain linguistic phenomena and hold against some noticeable objections.

References

Bilgrami, A., 1992. Belief and meaning: the unity and locality of mental content. Oxford: Blackwell.

Boghossian, P. A., 1989. The Rule-Following Considerations. Mind 98(392): 507–549. https://doi.org/10.1093/mind/XCVIII.392.507

Brożek, B., 2013. The Normativity of Meaning. In: The Many Faces of Normativity, eds. J. Stelmach, B. Brożek, M. Hohol. Kraków: Copernicus Center Press, 147–176.

Burge, T., 1979. Individualism and the Mental. Midwest Studies in Philosophy 4(1): 73–121. https://doi.org/10.1111/j.1475-4975.1979.tb00374.x

Burgess, A. & Sherman, B. (eds.), 2014. Introduction: A Plea for the Metaphysics of Meaning. In: Metasemantics. New Essays on the Foundations of Meaning. Oxford University Press, 1–16. https://doi.org/10.1093/acprof:oso/9780199669592.003.0001

Camp, E., 2016. Conventions’ Revenge: Davidson, Derangement, and Dormativity. Inquiry: An Interdisciplinary Journal of Philosophy 59(1): 113–138. http://dx.doi.org/10.1080/0020174X.2015.1115277

Cappelen, H., 2018. Fixing Language: An Essay on Conceptual Engineering. Oxford, GB: Oxford University Press.

Davidson, D., 1991 [1984]. Inquiries Into Truth and Interpretation. New York: Oxford University Press.

Davidson, D., 1974a. On the very idea of a conceptual scheme. In: Inquiries into truth and interpretation. Oxford: Clarendon Press, 183–198.

Davidson, D., 1974b. Replies to Lewis and Quine. In: Inquiries into truth and interpretation. Oxford: Clarendon Press, 281–285.

Davidson, D., 1984. Communication and convention. Synthese 59(1): 3–17. https://www.jstor.org/stable/20115982

Davidson, D., 2001. Three Varietes of Knowledge. In: Subjective, Intersubjective, Objective. Oxford: Clarendon Press, 205–220.

Davidson, D., 2005. A Nice Derangement of Epitaphs. In: Truth, Language, and History. Oxford, Oxford University Press, 89–107.

Donnellan, K., 1966. Reference and Definite Descriptions. The Philosophical Review 75(3): 281–304. https://doi.org/10.2307/2183143

Donnellan, K., 1968. Putting Humpty Dumpty Together Again. The Philosophical Review 77(2): 203–215. https://doi.org/10.2307/2183321

Fennell, J., 2013. The Meaning of ‘Meaning is Normative. Philosophical Investigations 36(1): 56–78. https://doi.org/10.1111/j.1467-9205.2011.01461.x

Fennell, J., 2015. Davidson: Normativist or Anti-normativist? Acta Analytica 30(1): 67–86. https://doi.org/10.1007/s12136-014-0230-z

Ginsborg, H., 2011. Primitive Normativity and Skepticism about Rules. Journal of Philosophy 108(5): 227–254. https://doi.org/10.5840/jphil2011108518

Glock, H. -J., 2003. Quine and Davidson on Language, Thought and Reality. New York: Cambridge University Press. https://doi.org/10.1017/CBO9780511487514

Glock, H. -J., 2010. Does language require conventions. In: Wittgenstein: mind, meaning and metaphilosophy, eds. P. Frascolla, D. Marconi & A. Voltolini. New York: Palgrave-Macmillan, 85–112.

Glock, H. -J., 2015. Meaning and rule following. In: International Encyclopedia of the Social & Behavioral Sciences (Second Edition), eds. H. -J. Glock & J. D. Wright. Amsterdam: Elsevier, 841–849.

Glüer, K., 2001. Dreams and nightmares: conventions, norms and meaning in Davidson’s philosophy of language. In: Interpreting Davidson, eds. P. Kotátko, P. Pagin, & G. Segal. Palo Alto: Stanford CSLI Publications, 53–74.

Glüer, K., Wikforss, Å., & Ganapini, M., 2024. The Normativity of Meaning and Content. In: The Stanford Encyclopedia of Philosophy (Fall 2024 Edition), eds. E. N. Zalta & U. Nodelman.s Available at: https://plato.stanford.edu/archives/fall2024/entries/meaning-normativity/. [Accessed 09 September 2026].

Glüer, K., Wikforss, Å., 2009. Against Content Normativity. Mind 118(469): 31–70. https://doi.org/10.1093/mind/fzn154

Grandy, R., 1973. Reference, Meaning and Belief. The Journal of Philosophy 70: 439–452. https://doi.org/10.2307/2025108

Hattiangadi, A., 2006. Is Meaning Normative? Mind & Language 21(2): 220–240. https://doi.org/10.1111/j.0268-1064.2006.00312

Keiser, J., 2023. Non-Ideal Foundations of Language. New York, NY: Routledge.

Kripke, S., 1982. Wittgenstein on Rules and Private Language. Cambridge MA: Harvard University Press.

Lewis, D. K., 1969. Convention: A Philosophical Study. Cambridge: Harvard University Press.

Lewis, D. K., 1975. Languages and Language. In: Language, Mind, and Knowledge, ed. K. Gunderson. Minnesota Studies in the Philosophy of Science 7. Minneapolis: University of Minnesota Press, 3–35.

Schroeder, S., 2024. Semantic normativity and moral obligation. In: Language, Mind, and Value: Essays on Wittgenstein. Anthem Press, 43–52.

Verheggen, C., 2018. From Triangulation to Semantic Normativity. In: Donald Davidson’s Triangulation Argument: A Philosophical Inquiry, eds. R. H. Myers, C. Verheggen. New York: Routledge, 41–65.

Wedgwood, R., 2007. The Nature of Normativity. Oxford, GB: Oxford University Press.

Wikforss, Å. M., 2001. Semantic normativity. Philosophical Studies 102(2): 203–226. https://doi.org/10.1023/A:1004746319850

Zadeh, M.S.A.H., 2015. Semantic Normativity. Acta Philosophica 24(2): 385–396.