According to standard assumptions in semantics, (a) ordinary users of a language have implicit beliefs about the truth-conditions of sentences in that language, and (b) they often agree on those beliefs. For example, it is assumed that if Anna and John are both competent users of English and the former utters ‘grass is green’ in conversation with the latter, they will both believe that that sentence is true if and only if grass is green. These assumptions play an important role in an intuitively compelling picture of communication, according to which successful communication through literal assertoric utterances is normally effected thanks to our shared beliefs about the truth-conditions of the sentences uttered in the course of the conversation. Against these standard assumptions, this paper argues that the participants in a conversation rarely have the same beliefs about the truth-conditions of the sentences involved in a linguistic interaction. More precisely, it argues for Variance, the thesis that nearly every utterance is such that there is no proposition which more than one language user believes to be that utterance’s truth-conditional content. If Variance is true, we must reject the standard picture of communication. Towards the end of the paper I identify three ways in which ordinary conversations can be communication-like despite the truth of Variance and argue that the most natural amendments to the standard picture fail to explain them.
This is a preview of subscription content, log in to check access.
Buy single article
Instant access to the full article PDF.
Price includes VAT for USA
Subscribe to journal
Immediate online access to all issues from 2019. Subscription will auto renew annually.
This is the net price. Taxes to be calculated in checkout.
These theses follow from the following standard assumptions. First, that knowledge entails belief. Second, that competent language users know the meaning of utterances in the languages they are competent in. Third, that knowing the meaning of an utterance requires knowing its truth-conditions. For a version of this last assumption, see for instance Heim and Kratzer (1998), who start their famous textbook by stating “To know the meaning of a sentence is to know its truth-conditions” (p. 1). Portner (2005) also takes knowledge of truth-conditions to be the starting point for semantics: “The knowledge of meaning involves (at least) knowledge of the conditions under which a sentence is true, and those under which it’s false” (p. 13). Larson and Segal (1995) motivate similar assumptions in their initial discussion of the relation between meaning and truth (pp. 5–7). Though Heim and Kratzer talk about knowledge of a sentence’s (as opposed to an utterance’s) truth-conditions, they do so only for the sake of simplicity, since they are ignoring the influence of context on a sentence’s semantic content. Some authors (e.g. Fodor 1983; Pettit 2002) think that the mental states underlying linguistic competence aren’t beliefs or knowledge, but something else, such as subpersonal mental states or states of a modular input-output system. Versions of the arguments I will develop below apply to those views as well, but I will leave this generalization for future work. Thanks to an anonymous reviewer for helpful discussion.
Without intending to do exegesis here, it would be fair to attribute a version of the standard picture of communication to philosophers in the tradition started by Grice (1989a, b)—a tradition that includes Bach and Harnish (1979), Schiffer (1972), Strawson (1964, 1970) and, to some extent, Stalnaker (1999c). According to philosophers in that tradition, in making an assertoric utterance a speaker means a proposition (or propositions), and the audience understands the utterance only if she recognizes the proposition(s) the speaker meant. If at least one of the propositions the speaker meant (in Gricean terms, the proposition the speaker said) determines the truth-conditions of the uttered sentence, we should expect that if the audience understands the speaker’s utterance, then speaker and audience believe the uttered sentence to have the same truth-conditions. Heck (2002, pp. 6–8), Evans (1982, p.22), and Dummett (2010) have attributed versions of the standard picture of communication to Frege. See also Portner (2005, pp. 21–22) for an endorsement of the picture sketched in this paragraph.
Defenders of the standard picture normally add further requirements for successful communication. For example, they may claim that successful communication between Anna and John normally requires not only that they both know that ‘Carla likes to run’ as Anna used it has the truth-conditional content that Carla likes to run, but also that they both know that the other knows this, that they know that they know it, and so on. For the purposes of the present discussion, we can do without such additions to the standard picture.
Authors like Stalnaker (1999a) sometimes remark that communication will be possible in some cases even if speaker and audience don’t agree over the truth-conditional content of the utterances used to communicate. This suggests that perhaps the standard view should be taken as a view of ideal communication. Unfortunately, this observation alone will not take us very far, since, as we will see in Sect. 5, it is not obvious how to extend the standard view to non-ideal cases. Furthermore, what the truth of Variance suggests is that non-ideal cases of communication are so widespread that offering a view of non-ideal communication is far more pressing than Stalnaker seems to acknowledge. Thanks to an anonymous reviewer for discussion.
Sentences here should be thought of as disambiguated sentences.
See Abreu Zavaleta (2018). See also Dorr and Hawthorne (2014) for a related argument to the effect that, if the propositions expressed by most utterances in ordinary language depend on microphysical facts, they depend very sensitively on those facts. See also Schiffer (1981a) for a related argument to the effect that the content of an utterance can’t include reference to specific modes of presentation, and Buchanan (2010) for a related argument against Gricean accounts of speaker meaning. I intend to discuss the differences between some of these arguments and the one I present here in future work.
Kennedy (2007, Sect. 3) denies that this is true of so-called “absolute” gradable adjectives, which he takes to come with fixed standards. For instance, according to Kennedy, in order for an object to fall in the extension of ‘impure’ it suffices that it has some minimal degree of impurity which remains constant throughout all contexts; in order for an object to fall under the extension of ‘straight’, it must be completely straight; etc. I am skeptical of Kennedy’s claims: if my only purpose is to drink water that won’t poison me, I will be willing to take an utterance of ‘that water is pure’ to be true even if the water in question has one milligram of sodium, but I will be less willing to take a similar utterance to be true in the context of a delicate chemical experiment. Of course, it might be that the range of acceptable standards for absolute gradable adjectives is more constrained than the range of acceptable standards for tallness or expensiveness, but there are reasons to think that the standards for absolute gradable adjectives can change with context nevertheless. I hope to discuss Kennedy’s arguments elsewhere.
See JBiz et al. (2010) for an example of this kind of disagreement about the definition of the verb ‘to run’.
Continuous dimensions are the prime example of a fine-grained dimension, but even dimensions with a finite number of points can satisfy this requirement. For instance, even if the application of F depends on an object’s properties along a dimension with only 10,000 points, the present argument would go through. All the argument requires is that there are enough points along those dimensions for there to be a large number of very similar and equally natural properties (all with a plausible claim to be the one expressed by F), each corresponding to slightly different cutoff points along those dimensions. Predicates whose application depends on an object’s properties along continuous dimensions (e.g. gradable adjectives) are prime examples of this kind of predicate, but they are not the only such examples.
Taking for granted a treatment of vague propositions along roughly supervaluationist lines, here is how the case for Variance would apply if truth-conditional contents are vague. Let’s say that a vague proposition is a set of classical propositions. Necessarily, a vague proposition P is true just in case every proposition in P is true, false just in case every proposition in P is false, and indeterminate otherwise. Then we can see that, in most cases, there will be many vague propositions one could easily have taken to be the truth-conditional content of a given utterance. For instance, suppose that Anna believes that her utterance of (1) has truth-conditional content P, where P is the vague proposition containing exactly the following classical propositions: that the box she pointed at is heavy at least to degree 0.6, that it is heavy at least to degree 0.61, and so on, up to the proposition that the box Anna pointed at is heavy at least to degree 0.7. There are many vague propositions extremely similar to this one which John could easily have taken to be the truth-conditional content of Anna’s utterance. For example, the vague proposition containing exactly the following classical propositions: that the box Anna pointed at is heavy at least to degree 0.61, that it is heavy at least to degree 0.62, and so on up to the proposition that the box Anna pointd at is heavy at least to degree 0.72. Given the huge number of these vague propositions, and that none of them seems more natural than the rest, it would be extremely unlikely for any two language users to believe exactly the same one to be the truth-conditional content of Anna’s utterance. Thanks to an anonymous reviewer for discussion.
Note that appealing to vagueness in the identity conditions of vague truth-conditional contents will not take defenders of the standard picture of communication very far. Strictly speaking, all we need in order to reject the standard picture is that, for most utterances, there is a large enough number of equally natural (or otherwise eligible) yet not definitely identical truth-conditional contents any language user could easily have believed each of those utterances to have. Given the huge number of not-definitely-identical truth-conditional content candidates each of those utterances has, it would be extremely unlikely for any two language users to definitely have the same beliefs about those utterances’ truth-conditional contents. If this is true, then the standard view of communication would predict that definite cases of successful communication are extremely rare.
Facts may be responsible for other facts in a constitutive or a causal sense (perhaps among others). The strategies I will consider here are implemented in terms of constitutive dependence, but it shouldn’t be difficult to see that the same remarks apply to implementations using the causal notion of dependence. As we will see in the discussion below, the reason the present strategy fails is that, given the large numbers of different propositions language users could easily have believed to be an utterance’s truth-conditional content, even small differences in the facts (causally or constitutively) responsible for someone’s beliefs about an utterance’s truth-conditional content will make a difference in those people’s beliefs about that utterance’s truth-conditional content.
Terms that do not express especially natural properties include some so-called “natural-kind terms”. For example, it is unlikely that there is a most natural way of drawing the line between members and non-members of a given species, so natural-kind terms like “dog” or “cat” are not natural in the metaphysical sense required by the present strategy.
The idea that in every conversation there is a set of propositions all the participants presuppose, that those participants presuppose that they presuppose, and so on, can be traced back at least to Schiffer (1972), is clearly present throughout Lewis (1979), and drives much of Stalnaker’s discussion in his (1999c) and subsequent work.
As Stalnaker (2009) puts it,
[A]n assertion is, in effect, a proposal to shrink the context set [the set of possible worlds compatible with every proposition in the common ground] with the content of the assertion. But the context set represents the information that is presumed to be available for the interpretation of the speech act, and if the asserted content is not determined by this information, then the addressee will not be in a position to tell what is being proposed. (p. 407)
See Stalnaker (1999b) for elaboration of this picture in connection with the standard Kaplanian treatment of context-dependence. According to Stalnaker, “Since the relevant contextual parameters must be available, and presupposed to be available, they will be incorporated into the speaker’s presuppositions, and so will be represented by the set of possible situations that constitute the context set.” (1999b, p. 10)
It is not straightforward to develop a version of the present strategy that addresses the argument for Variance in cases of context independence. So even if the present strategy succeeds (which, as I will soon argue, it does not), it would not count against the observations pertaining to context-independent expressions presented in Sect. 2.2.
The worry arises from the thought that, for many (perhaps most) context-dependent sentences, there is a huge number of different linguistic meanings—understood as functions from Kaplanian contexts to possible-worlds propositions—any language user could have associated with each of those sentences. Given the huge number of such different linguistic meanings, it is unlikely that any two language users associate the exact same linguistic meaning with any of a wide variety of sentences.
See Lederman (2019) for further challenges to the notion of common knowledge.
A prominent defender of social externalism is Burge (1979, 1986), but the general idea may be traced back to Putnam (1975). See Williamson (2007) for appeals to social externalism in objections to analyticity. Ludlow (1995, 1997), Pollock (2015), Wikforss (2001), among others, criticize social externalism.
Expressions written in this font are expressions of mentalese.
Note that, if the present strategy is to have any plausibility, it must leave room for the possibility of misunderstandings between members of the same linguistic community. For example, the view must allow that, in certain circumstances, one of the participants in a conversation may believe that an utterance of ‘Anna is next to the bank’ has the truth-conditional content that Anna is next to a certain financial institution, while another believes that the same utterance has the truth-conditional content that Anna is next to a certain river bank. In the present implementation of the strategy, this flexibility is achieved by taking mentalese sentence-types to be counterparts of natural-language sentence-types, and by individuating expression-types of natural language so that the disambiguation of ‘bank’ that denotes financial institutions and the disambiguation of ‘bank’ that denotes river banks count as different expression-types. If expression-types in natural language are individuated in this way, it is in principle possible for two people to associate mentalese types that are counterparts of different disambiguations of ‘bank’, which in turn allows the view to predict that misunderstandings are possible (though, according to this view, very uncommon).
One possible reaction is to assume that sentences of mentalese are themselves context-sensitive (see Gross (2005) for discussion). But this only pushes the problem back. For suppose that Anna tokens Utterance U of ‘that box is heavy’ has the truth-conditional content that S in her belief box, and John tokens Utterance U of ‘the box is heavy’ has the truth-conditional content that S’, where S and S’ are to be replaced by context-sensitive mentalese counterparts of ‘that box is heavy’. Now the issue is whether both sentences should be interpreted with respect to the same contextual parameters in virtue of being counterparts of the same English sentence. If the answer is yes, then misunderstandings will be impossible, since it will be impossible for any two people to disagree over the truth-conditional content of any context-sensitive sentence. If the answer is no, we still face the problem of explaining why, given the huge number of equally eligible contextual parameters, the sentences that replace S and S’ would be evaluated with respect to the exact same one. Thanks to an anonymous reviewer for helpful discussion.
This assumption is controversial in the literature on anaphoric pronouns, but the controversy does not matter for present purposes. See King and Lewis (2017) for an overview of related issues.
I am assuming for simplicity that the participants in a conversation have determinate beliefs about the truth-conditions of ordinary utterances. See Abreu Zavaleta (2018) for discussion of cases in which this assumption doesn’t hold.
Notice that, for all John takes for granted for the purposes of the conversation, Anna’s house may be any one of colors 1–7. Thus, the proposition that Anna’s house is one of colors 2–5 and the proposition that Anna’s house is one of colors 3–6 are not equivalent relative to the common ground in Anna and John’s conversation.
See Abreu Zavaleta (2018) for further development of these views.
Roughly, this corresponds to the idea that a proposition’s subject matter is determined by the states of affairs that play a direct role in making a proposition true or false.
Here is a rough explanation of why (17) and (18) both relevantly entail that H is not red. All states of affairs in which H is one of colors 2–5 (the possible truthmakers for (17)) and all states of affairs in which H is one of colors 3–6 (the possible truthmakers for (18)) are themselves state of affairs in which H is not red (and so, parts of states of affairs in which H is not red). Second, every state of affairs that is either a truthmaker or a falsitymaker for the proposition that H is red concerns exclusively H’s color, and as such, is either a truthmaker or a falsitymaker for (17), and either a truthmaker or a falsitymaker for (18). Thus, both (17) and (18) relevantly entail that H is not red.
Thanks to an anonymous reviewer for helpful discussion of the issues in this subsection.
In relating authors like Barker, Khoo, and Sundell to this strategy, I do not mean to suggest that they would propose it as a solution to the problems arising from Variance. Rather, my aim is to examine a view that one could construct using the resources they employ for different purposes, and assess the viability of such a view.
As authors like Khoo and Sundell point out, there is a sense in which metalinguistic disagreements are genuine disagreements. But what we’re after is not just the notion of a “genuine” disagreement. Rather, we want to capture conflicts between people’s beliefs about the facts they take an utterance to be about; in this case, conflicts between Anna and John’s beliefs over Carla’s degree of tallness. See Abreu Zavaleta (2018) for further discussion.
I want to insist that I am not attributing the present view of indexicals to authors who have defended the existence of metalinguistic negotiation as a natural phenomenon. Authors like Barker (2002), Khoo (2019), Khoo and Knobe (2016), and other proponents of metalinguistic negotiation, are careful to claim that the strategy does not apply tout court, and often restrict their position to discussions of gradable adjectives or moral terms. Yet the problems arising from Variance are much more general than that, and our aim here is to see if the metalinguistic negotiation strategy can solve those problems.
“If communicated contents are restricted to (or essentially tied to) specific contexts of utterance, then it is hard to envision how speakers who find themselves in different contexts can communicate, i.e. under such circumstances communication between contexts is thrown into doubt” (Cappelen and Lepore 2005, p. 153).
The same is true of an objection to holism about meaning often attributed to Fodor and Lepore. Jackman (2017) presents the objection as follows: “Strictly speaking, informative communication would be impossible [if meaning holism is true]. No one would mean the same thing by any of their terms unless they shared all the same beliefs, in which case, communication would be possible, but uninformative, and truly understanding the utterances of others would be impossible unless you already knew everything that they believed.” Yet if Variance is true, the problem of explaining what informative communication requires arises for nearly everyone.
Abreu Zavaleta M (2018) Semantic variance. Ph.D, thesis, New York University
Bach K, Harnish R (1979) Linguistic communication and speech acts. MIT Press, Cambridge, MA
Barker C (2002) The dynamics of vagueness. Linguist Philos 25(1):1–36
Buchanan R (2010) A puzzle about meaning and communication. Noûs 44(2):340–371
Burge T (1979) Individualism and the mental. Midwest Stud Philos 4(1):73–122
Burge T (1986) Intellectual norms and foundations of mind. J Philos 83(December):697–720
Cappelen H, Lepore E (2005) Insensitive semantics: a defense of semantic minimalism and speech act pluralism. Wiley-Blackwell, Oxford
Cresswell MJ (1977) The semantics of degree. In: Partee B (ed) Montague grammar. Academic Press, New York, pp 261–292
Dorr C, Hawthorne J (2014) Semantic plasticity and speech reports. Philos Rev 123(3):281–338
Dummett M (2010) Language and communication. In: Byrne D, Kölbel M (eds) Arguing about language. Routledge, NewYork
Evans G (1982) Varieties of reference. Oxford University Press, Oxford
Field H (1978) Mental representation. Erkenntnis 13(July):9–61
Fine K (2012) Counterfactuals without possible worlds. J Philos 109(3):221–246
Fine K (2016) Angellic content. J Philos Log 45(2):199–226
Fine K (MS) Truthmaker semantics. https://www.academia.edu/10908756/survey_of_truthmaker_semantics. Accessed 25 Aug 2016
Fodor JA (1983) The modularity of mind. MIT Press, Cambridge, MA
Fodor JA (1987) Psychosemantics: the problem of meaning in the philosophy of mind. MIT Press, Cambridge, MA
Grice HP (1989a) Meaning. In: Studies in the way of words. Harvard University Press, Cambridge, MA
Grice HP (1989b) Utterer’s meaning and intention. In: Studies in the way of words. Harvard University Press, Cambridge, MA
Gross S (2005) Context-sensitive truth-theoretic accounts of semantic competence. Mind Lang 20(1):68–102
Hawthorne J, Magidor O (2009) Assertion, context, and epistemic accessibility. Mind 118(470):377–397
Hawthorne J, Magidor O (2010) Assertion and epistemic opacity. Mind 119(476):1087–1105
Heck R (2002) Do demonstratives have senses? Philos Impr 2(2):1–33
Heim I (2000) Degree operators and scope. In: Jackson B, Matthews T (eds) Semantics and linguistic theory, vol 10. Cornell University, Ithaca, NY, pp 40–64
Heim I, Kratzer A (1998) Semantics in generative grammar. Blackwell, London
Jackman H (2017) Meaning holism. In: EN Zalta (ed), The stanford encyclopedia of philosophy (Spring 2017 ed.). Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/archives/spr2017/entries/meaning-holism/
JBiz, et al (2010) How fast do you have to be going to count it as running? http://www.letsrun.com/forum/flat_read.php?thread=3576961. Accessed 27 July 2017
Kaplan D (1989) Demonstratives. In: Almog J, Perry J, Wettstein H (eds) Themes from kaplan. Oxford University Press, Oxford, pp 481–563
Kennedy C (2007) Vagueness and grammar: the semantics of relative and absolute gradable adjectives. Linguist Philos 30(1):1–45
Kennedy C, McNally L (2005) Scale structure, degree modification, and the semantics of gradable predicates. Language 81:345–381
Khoo J (2019) Quasi indexicals. Philos Phenomenol Res
Khoo J, Knobe J (2016) Moral disagreement and moral semantics. Noûs 109–143
King JC, Lewis KS (2017) Anaphora. In: EN Zalta (ed), The stanford encyclopedia of philosophy (Summer 2017 ed.). Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/archives/sum2017/entries/anaphora/
Landau B, Smith L, Jones S (1988) The importance of shape in early lexical learning. Cogn Dev 3(3):299–321
Landau B, Smith L, Jones S (1998) Object shape, object function, and object name. J Mem Lang 38(1):1–27
Larson R, Segal G (1995) Knowledge of meaning: an introduction to semantic theory. MIT Press, Cambridge
Lederman H (2019) Uncommon knowledge. Mind 127:1069
Lewis D (1975) Languages and language. In: Gunderson K (ed) Minnesota studies in the philosophy of science. University of Minnesota Press, Minnesota, pp 3–35
Lewis D (1979) Scorekeeping in a language game. J Philos Log 8(1):339–359
Lewis D (1983) New work for a theory of universals. Aust J Philos 61(December):343–377
Ludlow P (1995) Externalism, self-knowledge, and the prevalence of slow-switching. Analysis 55(1):45–49
Ludlow P (1997) On the relevance of slow switching. Analysis 57(4):285–86
Markman EM (1990) Constraints children place on word meanings. Cogn Sci 14(1):57–77
Markman EM, Hutchinson JE (1984) Children’s sensitivity to constraints on word meaning: taxonomic versus thematic relations. Cogn Psychol 16(1):1–27
Pettit D (2002) Why knowledge is unnecessary for understanding language. Mind 111(443):519–550
Pollock J (2015) Social externalism and the problem of communication. Philos Stud 172(12):3229–3251
Portner P (2005) What is meaning?: Fundamentals of formal semantics. Blackwell, Oxford
Putnam H (1975) The meaning of ’meaning’. Minn Stud Philos Sci 7:131–193
Schiffer S (1972) Meaning. Clarendon Press, Oxford
Schiffer S (1981a) Indexicals and the theory of reference. Synthese 49(1):43–100
Schiffer S (1981b) Truth and the theory of content. In: Parret H (ed) Meaning and understanding. Berlin, New York, pp 204–222
Sider T (2011) Writing the book of the world. Oxford University Press, Oxford
Stalnaker RC (1999a) Assertion. In: Context and content: essays on intentionality in speech and thought. Oxford University Press, Oxford, pp 78–95
Stalnaker RC (1999b) Context and content: essays on intentionality in speech and thought. Oxford University Press, Oxford
Stalnaker RC (1999c) Pragmatic presuppositions. In: Context and content: essays on intentionality in speech and thought. Oxford University Press, Oxford, pp 47–62
Stalnaker RC (2009) On hawthorne and magidor on assertion, context, and epistemic accessibility. Mind 118(470):399–409
Strawson PF (1964) Intention and convention in speech acts. Philos Rev 73(4):439–460
Strawson PF (1970) Meaning and truth: an inaugural lecture delivered before the university of oxford on 5 november 1969. Clarendon Press, Oxford
Sundell T (2011) Disagreements about taste. Philos Stud 155(2):267–288
Wikforss AM (2001) Social externalism and conceptual errors. Philos Q 51(203):217–31
Williamson T (2007) The philosophy of philosophy. Blackwell, London
Yablo S (2014) Aboutness. Princeton University Press, Princeton
Thanks to Ben Holguín, Ian Grubb, Rose Flinn, Tienmu Ma, Chris Scambler, and two anonymous reviewers for very helpful comments. For helpful discussion, thanks to audiences at Institut d’Études Avancées de Paris, New York University, The New School for Social Research, and Università degli Studi di Torino. Thanks especially to Cian Dorr, James Pryor, Stephen Schiffer, and Erica Shumener for their insightful feedback on serveral drafts of this paper.
Conflict of interest
The author (Martín Abreu Zavaleta) declares that he has no conflict of interest.
This article does not contain any studies with human participants or animals performed by the author.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Abreu Zavaleta, M. Communication and Variance. Topoi (2019). https://doi.org/10.1007/s11245-019-09648-3
- Truth-conditional content
- Metalinguistic negotiation