David Plunkett (Dartmouth): Reflections on Some Varieties of Metalinguistic Negotiation

In co-authored and solo-authored work over roughly the past decade, Tim Sundell and I have developed the idea of “metalinguistic negotiation”. On our view, metalinguistic negotiations are a type of dispute in which speakers appear to use (rather than explicitly mention) a term in conflicting ways to put forward views about how that very term should be used. Sundell and I have argued that a number of disputes of interest to philosophers (as well as some disputes amongst philosophers themselves) might well be metalinguistic negotiations. In this paper, I reflect on reflect on a number of interesting distinctions between different kinds of metalinguistic negotiation that we haven’t focused on. I think that understanding these distinctions can help illuminate a number of varieties of metalinguistic negotiation that are interesting as linguistic phenomena. My primary aim in this paper is to get those phenomena on the table. In so doing, I hope that philosophers from different subareas of philosophy might attend to these phenomena to further explore how these phenomena matter for a range of philosophical debates (beyond some preliminary remarks I make on this front here). Moreover, I have a secondary aim as well: I aim to show how better appreciating these varieties of metalinguistic negotiation can help us better understand the capacious nature of Sundell’s and my general account of “metalinguistic negotiation”, and what it is (and is not) committed to. 

Philippe de Brabanter & Bruno Leclercq (Liège, Bruxelles): From semantic deference to semantic externalism to a classification of disagreements 

We distinguish three broad kinds of semantic externalism, which we assume are grounded in three types of semantic deference. Deference to usage (dispositional compliance with the norms of a common language practice) grounds ‘usage-dependent externalism’. Deference to the way current experts define words grounds ‘conventional externalism’. For ‘indexical externalism’, there is ultimate deference to the true nature of a phenomenon, yet ‘proximal’ deference to the people most likely to capture that nature. These distinctions prove fruitful when it comes to classifying disagreements between speakers. First, as the debates of the last 50 years have abundantly shown, there may be (conscious or not) metalinguistic disagreement as to what kind of semantic deference should be at work for the word. Second, even amongst those who agree on the kind of deference for a particular word, there may still be other kinds of disagreement. Within usage-dependent externalism, the extent of the relevant linguistic community remains open for discussion. Within conventional externalism, different speakers may still disagree on who are the experts responsible for fixing meaning. Within indexical externalism, there may be disagreement about who is most likely to capture the true nature of a phenomenon referred to, especially if that phenomenon is not (currently) studied by some specific scientific discipline. The classification just outlined furthermore suggests that the tasks conceptual engineers engage in and the objectives they pursue depend on the type of externalism that a term falls under. 

Poppy Mankowitz (Bristol): Metalinguistic mechanisms 

Those who discuss metalinguistic negotiation generally do not explain what mechanism underlies the phenomenon. This makes it difficult to account for certain complex metalinguistic disputes, and to tell what theories of meaning are compatible with the potential for metalinguistic negotiation. I claim that the most promising account takes the mechanism to consist of speakers’ implicating propositions about which expressions are apt for using as part of conveying contextually relevant information. This account is compatible with both externalist and internalist theories of meaning. 

Pedro Abreu (NOVA): The borders and objects of metalinguistic disagreement

In metalinguistic disagreements, the disagreement between disputants resides not at the level of the literally expressed content, but concerns the particular usage or meaning modulation of some expression deployed in some linguistic exchange. The reinterpretation of disputes as metalinguistic disagreements constitutes a promising strategy for making sense of plausibly many debates, but it still faces important challenges. In this presentation, I consider two intimately related unsolved issues: (1) How to distinguish metalinguistic disagreements from regular, non-metalinguistic disagreements? (2) How to delineate the objects of metalinguistic disagreements?

Any answer to the first question is crucially dependent on our answer to the second; which cases happen to instantiate the relevant kind of meta-level divergence, and which don’t, is directly determined by how we delineate what disputants can disagree about when they disagree metalinguistically. I argue that our best answer to the second question is that the objects of metalinguistic disagreements are fixed and individuated through rationalizing interpretation of a selected fragment of speakers’ linguistic behaviour. I also show how this line of response can help us explain our difficulties in providing a principled answer to the first question, and accept as appropriate the lack of an objective boundary between metalinguistic and non-metalinguistic disagreements. 

Antonin Thuns (Bruxelles): Externalistic functions rather than externalistic contents: The case for a (moderately) anti-realist and (radically) naturalistic account of meaning and metalinguistic disagreement

This paper addresses the question of the compatibility, from a naturalistic perspective, of Sundell’s (2012) Meaning is Use Minus Pragmatics Strategy with a particular form of semantic externalism: Millikan’s (1984, 2005, 2017) teleosemantic theory of semantic content determination. Sundell advocates a usage-based theory of word meaning in which substantial disagreement is possible even if speakers use the same words with different idiolectal meanings. That is, speakers may speak different “languages”, yet there might be independent reasons (metaphysical, ethical) to prefer one “language” over the other. Sundell’s approach thus offers an interesting decoupling between semantic content determination (the basis for the expression of truths or falsehoods) and the broader question of correctness. The proposal is to combine Sundell’s insights with a modified version of Millikan’s teleosemantic theory of meaning, in which “proper functions” are not limited to acts of referring and teleosemantic content need not play the role classical semantic content. On this synchretic account, speakers in disagreement may be using the same terms with different “meanings” while still reproducing the same non-referential proper functions of these terms. Moreover, the idea that a semantic theory should ascribe realistic-intentional contents at all is dropped. In the naturalistic and eliminativistic framework I propose following Rosenberg (1999, 2015), only the functions (“referential” or other) of expressions can be individuated externalistically, from a third-person and purely theoretical standpoint. Semantic contents proper can only be internal, as abstractions of cognitive properties.

Matthew Shields (Wake Forest/UCD): Making conceptual disputes explicit

The recent literatures on metalinguistic negotiation and conceptual engineering have prompted philosophers to revisit the question of whether speakers count as thinking and talking about the same topic, even when their understanding of that topic radically differs. Herman Cappelen has denied that a determinate type of content is necessary to justify our samesaying judgments, while Tim Sundell has pushed back against this view. In this paper, I argue that the debate between Cappelen and Sundell cannot be resolved in the form they frame it. Both philosophers focus on cases where the differences in understanding among speakers are implicit. By ‘implicit’, I mean that the speakers under consideration are utilizing the key term or concept to judge whether it applies to some particular, but are not making the key term or concept itself a matter of direct, explicit discussion. After analyzing the Sundell-Cappelen debate in the paper’s first section, I show in the second section that the focus on implicit cases is pervasive in philosophical discussions, citing examples from debates between Frank Jackson and Tim Williamson, between Hilary Putnam and Noam Chomsky, and recent views defended by Laura and François Schroeter.

In the paper’s third and final section, I argue through a series of examples of explicit conceptual disputes that the assumption that speakers will generally conclude that they are thinking and talking past one another if they are exposed to the different (including radically different) content they associate with a key term or concept turns out to be misguided. These cases of explicit conceptual disputes provide, I claim, a stronger argument for Cappelen’s “no content” view of samesaying. They also show that a more complex philosophical story is needed to make sense of our practices of samesaying.

Delia Belleri (Lisbon): Conceptual engineering, representational skepticism, and defeaters 

Some concepts, at some point in time, turn out as being defective in some given respect. In certain circumstances, revising them seems advisable. Moreover, this could potentially happen with any concept in our representational repertoire. Should we then maintain a sceptical attitude towards all concepts, to be on the safe side? This is what Herman Cappelen calls “representational skepticism” (Cappelen 2018: 5-6). Drawing inspiration from the epistemology literature on perceptual and testimonial justification, I argue that, in a conservative spirit (Chisholm 1989), we are generally warranted in trusting our concepts, provided the absence of certain defeaters, which might lead one to question the assumed epistemic or non-epistemic good quality of a specific concept or set thereof. As an illustration, I focus on two types of defeaters: (i) defeaters that reveal disagreements with certain epistemically reliable communities, such as oppressed groups, or scientific experts, on the epistemic or non-epistemic good quality of a concept. (ii) Defeaters that reveal “peer disagreements” on the epistemic or non-epistemic good quality of a concept, with subjects that (whether or not they are experts) are in one’s same epistemic position. I argue that, when such defeaters are in place, one’s justification to accept and deploy the concept(s) under examination ought to be critically reconsidered. As a result, one need not be a representational skeptic. In general, one can lead one’s cognitive and linguistic life without questioning the concepts one is using, unless one is presented with information that counts as a defeater.

References

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

Chisholm, R. 1989. Theory of Knowledge. Englewood Cliffs, NJ: Prentice-Hall.

Sarah Sawyer (Sussex): External constraints on metalinguistic disagreement in the context of conceptual engineering

In this paper I distinguish three different kinds of phenomena: the phenomena of interest to philosophy as traditionally conceived, such as truth, knowledge, and causation; the phenomena of interest to the natural sciences, such as atomic number, force, and mass; and the phenomena of interest to the social sciences, such as class, poverty, and misogyny. Understanding the distinctions between these kinds of phenomena reveals that the normative question of what the meanings of our terms ought to be is really the question of what the grounds of our kinds ought to be, and that the scope of conceptual engineering is neither as broad as advocates have claimed nor as narrow as sceptics believe. As a result, the Implementation Challenge for conceptual engineering is not, as is typically thought, the semantic challenge of how to change the meanings of our terms but is instead the practical challenge of how to change the grounds of our social kinds and hence is not as intractable as has been supposed.

Mark Pinder (Open): Semantic amelioration as a cause for optimism about conceptual engineering 

Conceptual engineering is in vogue, and Haslanger’s ameliorative project is centre stage. It is thus an embarrassment that, given semantic externalism, Haslanger’s project seems not to be a cause for optimism about conceptual engineering at all. There are two standard interpretations of Haslanger’s project. On the first, Haslanger is seeking to change how people in general use and understand terms such as ‘man’, ‘woman’ and ‘race’. But, given semantic externalism, the project faces an implementation problem and is doomed to fail. On the second interpretation, Haslanger adopts a version of semantic externalism to reveal the semantic meanings of our gender and race terms. But then the project seems not to involve conceptual engineering at all. In this talk, I begin by sketching the two standard interpretations of Haslanger’s project, briefly raising the concerns articulated above. I then articulate and defend an alternative interpretation of Haslanger’s project. On this alternative—I call it the terminological interpretation—Haslanger’s project is a successful conceptual engineering project. Moreover, it is a successful conceptual engineering project independently of whether metasemantic externalism (or internalism) is true. 

Cyrill Mamin (Jena): Realism and Social Kinds in Conceptual Amelioration

How can debates linked to socio-political conceptual amelioration be substantial, despite the different concepts in play? The present account addresses this question by considering the relations between socio-political concepts and their referents, social kinds. In debates involving kind-representing concepts, they are evaluated according to their representational accuracy. For example, any given concept RAPE can be evaluated according to how accurately it represents the social reality of rape. In these debates, the underlying substantial question is, “Is the social kind like this?” In debates involving kind-generating concepts, different social kinds are generated, according to the different concepts in play. The concepts are then evaluated according to their functional adequacy. The corresponding substantial question is, “Is this (or that) the social kind we want?” As a result, the relevant debates have an underlying substantial question in both kind-representing and kind-generating cases. Substantiality can be guaranteed by shifting the focus from word meanings/concepts to social kinds while remaining neutral regarding the externalism-internalism divide. 

Alexander Kocurek (Cornell): Verbal disagreement and semantic plans 

I develop an expressivist account of verbal disagreements as practical disagreements over how to interpret language. On the proposed account, verbal disagreements often involve divergent plans concerning language use rather than conflicting beliefs about what a word means or a mere appearance of disagreement due to verbal confusion. To articulate this more precisely, I introduce the notion of a semantic hyperplan, i.e., a maximally specific plan for interpreting language in any situation (cf. Gibbard 2003). Roughly, a disagreement is wholly verbal if the disagreement would dissolve were either speaker to adopt an interpretation of the disputed claim left open by the other speaker’s semantic plans (cf. Vermeulen 2018). This account enjoys several advantages: it analyzes verbal disagreement solely in terms of the mental states of speakers in a way that’s compatible with semantic externalism; it accounts for cases where speakers are undecided on how exactly to interpret an expression and where a speaker’s choice of interpretation depends on factual matters they may be uncertain of; it can be formally implemented in a neo-Stalnakerian possible worlds framework of communication; and it unifies and clarifies ideas behind other accounts in the literature, including those appealing to topics (Balcerak Jackson 2014; Jenkins 2014), explanatory relations (Chalmers 2011), or charitable interpretations (Hirsch 2009). 

Joanna Odrowaz-Sypniewska (Warsaw): Negotiating boundaries

In my talk I’m going to argue – against Cappelen (2018) – that semantic externalism is compatible with metalinguistic negotiation as witnessed by natural kind terms and vague predicates. Even Putnamian externalist semantics has to admit – due e.g. to the qua problem and open-texture – that we have some saying about the extensions of natural kind terms. And when we focus on vague terms it is simply a fact that people’s disagreements have an effect on their extensions. The only theory of vagueness that makes metalinguistic negotiations over vague boundaries impossible is externalist epistemicism of Williamson’s style. However, there is no reason why externalist semantic theories of vagueness couldn’t allow metalinguistic negotiation. 

It is a separate issue how often we engage in negotiations that are primarily metalinguistic. Plunkett and Sundell regard exchanges like “Philip is tall”/ “Philip is not tall” as metalinguistic and add that such negotiations “can be a way of arguing about object-level issues” (2021: 148). I’d like to suggest a different solution. Namely, we should grant people their folk-linguistic intuitions and agree that they are debating object-level issues directly. We should reverse Plunkett and Sundell’s picture and claim that such debates are primarily object-language and only secondarily metalinguistic. This view allows such debates to be canonical and preserves the intuitive assumption that the contents of the interlocutors’ utterances are contradictory. In order to explain why such disagreements can be faultless I argue that borderline utterances should have a weaker illocutionary force than a full-blooded assertion.

Erich Rast (NOVA): Discussing Word Meaning: The Question of Internalism vs. Externalism

In this talk, I go through some of the standard arguments for and against semantic externalism and argue that (i) explicit metalinguistic disputes illustrate that speakers often discuss different definitions of words, and (ii) that an indexical externalist metasemantics has conceptual problems with explaining productive word composition processes without violating some of Quine’s theses about ontological relativity. In contrast, internalist lexical meaning fares better with metalinguistic disputes and word composition. However, it can be shown that it is not truth-conditionally relevant for the truth conditions of an utterance as a whole unless it becomes “at issue.” It instead tracks world-level beliefs and similar attitudes speakers commonly hold within a speaker community at a time. From the perspective of sentence-level truth-conditional semantics, it may thus not be considered meaning at all. I argue that this type of meaning is nevertheless indispensable for doing lexical semantics and explaining metalinguistic disputes. 

Mark Richard (Harvard): Conceptual engineering: Be careful what you wish for

Herman Cappelen argues that ‘conceptual engineering is about the world…the result of conceptual engineering … [is] an object-level change: we’re changing what gender, freedom, salad, marriage, etc. are.’ Hilary Putnam once somewhat notoriously argued that ‘ “Objects” do not exist independently of conceptual schemes’; it’s clear that he would have extended this claim to kinds, both natural and social.

Cappelen tells us that his view is ‘metaphysically lightweight’ and not a sort of idealism; Putnam told us that his pragmatic realism wasn’t a kind of relativism and was consistent with the idea that there are objective truths. In this talk I look at relations between Putnam’s pragmatic realism and the idea that in engineering our concepts we are changing the world. My primary interests are whether these views do involve a (problematic) sort of idealism, and whether, if we take the idea of conceptual engineering to heart, we should endorse one (or both) of them.