The account of metalinguistic negotiation to be developed in Sect. 3 has two features that are desirable for any adequate account. First, it is a propositional account, and second, it is based on an independently motivated account of expression focus. I explain each of these features in turn.
Propositional accounts
The literature lacks a comparison of the merits of accounts’ affirming and denying that metalinguistic propositions are conveyed during metalinguistic disputes. I want to draw attention to a compelling reason in favour of propositional accounts. Observe how some metalinguistic negotiations involve the tokening of sentences that contain sentential operators (expressions that apply to a sentence to produce another expression such as a sentence; e.g., ‘sometimes’, ‘Yasma thinks that’, etc.):
These disputes are naturally understood to possess the features that Plunkett and Sundell take to characterise metalinguistic negotiations: the relevant expression (‘athlete’, etc.) is used to communicate information about the appropriate usage of that very expression, and the disputes concern how the expression should be used relative to certain contexts. In canonical examples of metalinguistic negotiations like (2a)–(2e), speakers’ views were simple affirmations or denials of the propriety of using an expression in a certain way at their context. Yet in (3a)–(3e), B is naturally understood to hold more complex views about the propriety of using the relevant expression. Moreover, sentential operators that occur in B’s utterances determine exactly which complex view B appears to hold. For example, in (3a) B is naturally understood to hold the view that, if we should use ‘spicy’ with respect to the chili for certain purposes, then we should use ‘spicy’ with respect to lasagne for certain purposes.Footnote 5 In (3e), B is naturally understood to think that sometimes we should use ‘fruit’ with respect to tomatoes for certain purposes, and sometimes we should use ‘vegetables’ with respect to tomatoes for certain purposes.
It is unclear how the complex views that B appears to hold in (3a)–(3e) could be captured in non-propositional terms, such as through B’s altering semantic norms. In contrast, propositional accounts can easily handle these cases: while participants in the canonical examples of metalinguistic negotiations pragmatically convey propositions that affirm or deny the propriety of using the relevant expressions in certain ways, these propositions can interact with the meanings of tokened sentential operators in order to form propositions consisting of more complex claims about appropriate language use.
Against this reasoning, it might be claimed that pragmatically conveyed content cannot interact with tokened sentential operators. It would follow that, if (3a)–(3e) convey complex metalinguistic propositions to which the meanings of sentential operators contribute, then they must semantically express these propositions.Footnote 6 There are several good reasons to resist this position. First, it is intuitively implausible to claim that (say) ‘Sometimes tomatoes are fruit’ means that sometimes we should use the expression ‘fruit’ with respect to tomatoes for certain purposes; it seems to semantically express something about tomatoes, rather than about appropriate language use. Second, the view would entail either that any sentence that can be used in a metalinguistic dispute (e.g., ‘Tomatoes are fruit’) also semantically expresses a proposition about language use, or that adding a sentential operator to a sentence that expresses a non-metalinguistic proposition is sufficient to cause it to express a metalinguistic proposition, where both of these options seem unnatural. Third, Plunkett and Sundell (2019, p. 5) are clear that during metalinguistic disputes, ‘the content about which [the speakers] disagree isn’t the semantic content of their expressions in their context’; hence if metalinguistic propositions are semantically expressed in (3a)–(3e), then we would be unable to classify them as metalinguistic negotiations, despite the fact that they have the characteristics identified by Plunkett and Sundell. (3a)–(3e) thus justify the inference that pragmatically conveyed content can interact with tokened sentential operators, especially if a plausible account of this interaction is given (see Sect. 3.3).
In sum, anyone who concedes the possibility of metalinguistic negotiations like (3a)–(3e) has a compelling reason to accept a propositional account. Accordingly, the account that I go on to develop will be propositional.
Independent motivations and expression focus
It is desirable for an account of metalinguistic negotiation to be based on an independently motivated account of some broader or related phenomenon, if possible. This not only accords with the aim of theoretical parsimony, but also means that the motivation to accept the proposed account is partially independent of considerations related specifically to metalinguistic negotiations. I will suggest that there are connections between metalinguistic negotiation and the phenomenon of expression focus. After giving an overview of focus and explaining its connection to metalinguistic usage, I will sketch the central features of an account of expression focus and show how they extend to metalinguistic negotiation.
The focus of an occurrence of a sentence is an expression that is marked with vocal emphasis in spoken form, and which indicates that certain alternatives to that expression are relevant to our understanding of the sentence.Footnote 7 Ordinarily, focus is used to draw attention to alternatives to the denotation of the focused item. For example, where ‘[ \(]_F\)’ marks the constituent in focus and capitalised morphemes are to be read with vocal emphasis, alternative properties that Grandpa may have are relevant to our understanding of (4a), whereas alternative individuals that may have died are relevant to our understanding of (4b):
Expression focus (see Wedgwood 2005; Krifka 2007; Li 2017) is a special type of focus that speakers utilise to draw attention to alternative linguistic items rather than denotations. For instance, alternative expressions that may be used as part of conveying that Grandpa died are relevant to our understanding of (5a). In (5b) and (5c), the relevant alternative expressions are ones that may be used as part of conveying that individuals eat rutabaga and that some geese are flying:
Footnote 8
There are two main reasons to postulate a connection between expression focus and metalinguistic disputes. Firstly, whenever an expression involves expression focus, it will meet Plunkett and Sundell’s criteria to count as a metalinguistic usage. For example, in (5a) ‘kick the bucket’ is used to communicate information about the appropriate usage of that expression, specifically that the speaker thinks that the expression is inappropriate to use as part of conveying that Grandpa died. Secondly, when a metalinguistic usage of an expression occurs, it is natural for that expression to be focused. For instance, Plunkett and Sundell’s intended construal of (2b) is most naturally attained via vocal emphasis on the occurrences of ‘athlete’.Footnote 9 On the other hand, the following vocal emphasis suggests that the speakers are concerned with establishing some examples of athletes, rather than how the expression ‘athlete’ should be used:
These observations indicate a close connection between expression focus and metalinguistic usage of expressions. Yet it need not follow that an expression is employed in a metalinguistic usage if and only if it receives expression focus. A reasonable position is that certain contextual features might allow assessors to infer that an occurrence of an unfocused expression is employed in a metalinguistic usage, but this is less natural and straightforward than the inference that an occurrence of a focused expression is employed in a metalinguistic usage.
Accordingly, I adapt the account of expression focus developed in Mankowitz 2020, which describes how speakers can convey metalinguistic propositions about the aptness of focused expressions in certain contexts. Informally, the account of expression focus takes these metalinguistic propositions to concern the aptness of using the focused expression as part of conveying a non-metalinguistic, embedded proposition (so-called to distinguish it from the metalinguistic proposition to which it contributes). More formally, the account defines an aptness relation, which holds between an expression e and a proposition p at a circumstance of evaluation i if and only if e is apt for using as part of conveying p at i, relative to contextually relevant standards.Footnote 10 A metalinguistic proposition stating that an expression e is apt for conveying an embedded proposition \(\alpha (\beta )\) is then formulated as: Apt\((e)(\alpha (\beta ))\).
Two questions about embedded propositions remain to be addressed. First, one might wonder why metalinguistic propositions concerning the aptness of expressions should be thought to include embedded propositions in the first place. The answer is that the use of an expression can reasonably be considered appropriate or inappropriate at a circumstance only relative to the conveying of particular information. While Plunkett and Sundell often imply that speakers hold absolute views about the propriety of using an expression with respect to some non-linguistic item, this cannot be an accurate depiction of speakers’ views. For example, Plunkett and Sundell (2013, p. 15) describe (2a) as a case where A ‘accepts the content that we should use ‘spicy’ in such a way that it applies to the chili’. Yet speaker A would presumably consider it inappropriate to use ‘spicy’ in order to convey that the chili is mild, brown, or a tedious topic of conversation. Speaker A’s view that it is appropriate to use ‘spicy’ with respect to the chili is really the view that it is appropriate to use ‘spicy’ as part of conveying that the chili has certain properties (more specifically, a particular flavour). It is therefore natural to accept that the propriety of an expression is always implicitly relativised to the conveying of some information. This idea is captured by taking metalinguistic propositions about appropriate language use to concern an aptness relation between an expression and a proposition.
Second, one might wonder how embedded propositions are formed. In order to capture the full range of metalinguistic propositions that can be conveyed by means of uses of expression focus, the account under discussion takes the context to supply a salient non-linguistic item that forms the embedded proposition when combined with the denotations of unfocused expressions. An item is salient when it is ‘the focus of perceptual or cognitive attention’ with respect to the discourse participants at a context (Mount 2008, p. 154). Generalising, a metalinguistic proposition Apt\((e)(\alpha (\beta ))\) conveyed by a participant in a metalinguistic dispute will be formed by combining a contextually salient non-linguistic item \(\alpha \) with the denotation \(\beta \) of expressions employed in non-metalinguistic usages.Footnote 11 For example, the speakers in (2b) communicate that ‘athlete’ is, or fails to be, apt for using as part of conveying the proposition that Secretariat has a salient property, such as being a successful racehorse. The fact that the account does not require the denotation of an expression e employed in a metalinguistic usage to contribute to the embedded proposition is appealing in light of Plunkett and Sundell’s view that metalinguistic negotiations often occur at contexts for which ‘there is no antecedently settled matter of fact about [e’s] meaning’ (2014, p. 64) and e thus lacks a unique denotation.
In sum, it is desirable to analyse metalinguistic negotiation by means of an independently motivated account of some related phenomenon. The connections between the phenomena of metalinguistic negotiation and expression focus additionally render it desirable to analyse the former by means of an account of the latter. Adapting an existing account of expression focus allows the views communicated by participants in metalinguistic disputes to be analysed as metalinguistic propositions stating that an aptness relation holds between the expression employed in a metalinguistic usage and a context-dependent, embedded proposition.Footnote 12 The embedded proposition is formed by combining the denotations of expressions employed in non-metalinguistic usages with a salient non-linguistic item.