Abstract
A satisfactory theory of linguistic communication must explain how it is that, through the interpersonal exchange of auditory, visual, and tactile stimuli, the communicative preconditions for the acquisition of testimonial knowledge regularly come to be satisfied. Without an account of knowledge-yielding communication this success condition for linguistic theorizing is left opaque, and we are left with an incomplete understanding of testimony, and communication more generally, as a source of knowledge. This paper argues that knowledge-yielding communication should be modelled on knowledge itself. It is argued that knowledge-yielding communication occurs iff interlocutors coordinate on truth values in a non-lucky and non-deviant way. This account is able to do significant explanatory work: it sheds light on the nature of referential communication, and it allows us to capture, in an informative way, the sense in which interlocutors must entertain similar propositions in order to communicate successfully.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
The spread of knowledge is not the only important function of language. Language functions, for example, as a means for social bonding and collective action, and also as a means for degradation, subordination, and control. However, the spread of knowledge is clearly an important function of language. So one central task of the theory of language is to explain how linguistic communication enables the spread of knowledge.
Heck (1995) provides a view along these lines, although they merely require knowledge of truth conditions, allowing that interlocutors may entertain propositions containing slightly different senses.
A note on methodology before we begin: The argument relies on judgments about cases. Usually these judgments will be given theoretical support. In particular, I will repeatedly motivate a diagnosis of knowledge or failure of knowledge by considering whether or not a belief is held and true in a non-lucky and non-deviant way. It is an assumption of this paper that knowledge is, roughly, non-deviant non-lucky true belief (where non-deviance is taken to imply justification, and where each clause can be spelled out in different ways, including ways which ultimately refer back to a basic unanalyzable concept of knowledge).
It may seem strange to say that ‘knowledge-yielding communication’ has occurred when knowledge has not been acquired. But ‘knowledge-yielding communication’ is merely a label. If the reader prefers, they may substitute ‘potentially knowledge-yielding communication’ as a label for our target concept, and save the label ‘knowledge-yielding communication’ for the relation as it occurs in communicative exchanges which actually result in knowledge. This labeling convention may, ultimately, be more accurate. However, the convention employed here makes for snappier presentation.
This state of comprehension may be the result of some quasi-inferential process, or it may be a matter of decoding some packaged message. This is a matter with which we need not concern ourselves.
Thus, the type of belief we are concerned with relates closely to what Goldberg (2007), following Audi (1997), calls ‘belief through testimony’. Goldberg and Audi both note that there are cases in which an audience’s belief that p is somehow dependent on a speaker’s testimony, without in any sense being a testimonial belief. For instance, if a speaker says something in English we may, as a result of our understanding the content of their utterance, come to learn that they speak English. Likewise, if somebody states that they are a baritone in a baritone voice we can learn that they are a baritone without, in any sense, relying on their testimony. This prompts Goldberg to focus on knowledge through testimony, which he defines as “knowledge involving reliable belief in what was attested, formed on the basis of its having been attested” (Goldberg 2007, p. 15). In this paper we are concerned with cases where the audience comes to believe that p on the basis of their representation of the speaker as having presented p as true. This rules out beliefs like the English speaking or baritone beliefs as testimonial. However, it builds in less than the notion of knowledge through testimony: knowledge through testimony requires that the belief the audience forms be in a proposition attested to by the speaker. However, it is unclear precisely what it is for a speaker to attest to a particular proposition. Moreover, it is an open question, at this point in the investigation, whether an audience’s testimonial belief must be in a proposition attested to, or rather in a proposition in some sense related to one attested to by a speaker. It may well be that, given the potentially messy and quasi-inferential nature of linguistic communication, true belief through testimony is rare. It may be that most so called ‘testimonial knowledge’ merely approximates knowledge through testimony. This is an issue which I do not wish to prejudge at this point.
Our core question should be distinguished from that of when we communicate well enough to transmit knowledge. A speaker transmits their knowledge or justification when, in some robust sense, their knowledge or justification becomes the audience’s knowledge or justification. Transmission views are highly controversial (see Lackey 1999; MacFarlane 2005; Barnett 2015; Fraser 2016; Peet and Pitcovski 2017; Leonard 2018 for objections). It is a truism that we have a great deal of testimonial knowledge, so much so that any view of linguistic communication that predicts otherwise should be treated with skepticism. This is not true of transmission based knowledge. Hence the focus on a broader question of knowledge-yielding communication. Our question should also be distinguished from that of the general conditions for communicative success. Communicative success is a multifaceted notion. For example, it is sometimes thought of in terms of knowledge transmission (Evans 1982; Heck 1995), the enabling of successful action (Paul 1999; Carston 2002; Bezuidenhout 2002), rational engagement (Heck 2002), or as an intuitive notion (Pagin 2008, 2019). These distinct notions of communicative success, I believe, require distinct treatments. Here I focus on what I take to be one of the most important notions of communicative success: knowledge-yielding communication. Although, as we will see, there are lessons to be drawn about other forms of communicative success (insofar as they also preclude luck).
There are alternative views of sense against which Byrn and Thau’s example has less force. For example, if we deny that the aboutness properties of Tony and Alex’s “Winston” beliefs are explained by reference to the description they attach to the name, and identify senses with explanations of the aboutness properties of token thoughts [in a manner somewhat akin to Evans (1982)], then we can deny that they each think of Winston under the same mode of presentation. The important point is that, as things stand, Loar cases have little dialectical force against the direct reference theorist. In order for these cases to have force it needs to be established that the elimination of luck requires coordination on senses (of some sort). At this point this has not been shown. We will return to this point in Sect. 5.2 when it will be argued that, in order for communicative luck to be eliminated there must be a robust explanation for the relational fact that the interlocutors’ token thoughts share certain aboutness properties. This provides some support for the Fregean thought that communication requires thinking of referents in suitably related ways, although it does not (by itself) require us to build modes of presentation into the contents of the interlocutors’ thoughts.
It might be objected that in HOSPITAL the interlocutors are not justified in taking their usage to be shared. However, it can easily be modified to account for this. We need merely adjust the case such that each interlocutor has a gettiered belief that their meaning is shared.
For example, the radical contextualism of Sperber and Wilson (1986), Bezuidenhout (1997), and Carston (2002). It is also often acknowledged that coordination on finer grained contents is not necessary for successful communication. This claim is held by some fregeans such as Bezuidenhout (1997) and Heck (2002), and several approaches to indexical content (see Weber 2015 for an overview). I will focus on coarse grained contents, as the failure of coarse grained coordination implies failures of fine grained coordination.
The exact meanings assigned by the interlocutors in this case are not essential. In particular we can swap the meanings assigned and gain the same result (as long as we suppose that in the reversed case Branden would, to avoid risk, not predicate tallness of agents who only just count as tall for him). The important fact is that the object of the predication (Michael Jordan) falls safely outside of the margin for error for both interlocutors’ uses of “tall”. As a result, Emily’s belief is rendered safe.
It is not clear in general how introducing indeterminacy makes coordination easier. Indeed, on some ways of modeling indeterminacy (for example in terms of fuzzy meanings) indeterminacy makes precise coordination even harder (MacFarlane 2016).
That interlocutors have such discursive responsibilities is illustrated by the fact that we will generally hold others responsible for failing to correct harmful misunderstandings or false testimony if they are easily able to do so.
One might object as follows: If asked why he believed that all the beer at the party had been consumed Sam would respond “Marie told me so”. However, this is not what Marie intended to tell him. Thus, his belief is based on a falsehood. This objection faces several problems (besides presupposing the impossibility of knowledge from falsehood). Firstly, the question under discussion was that of whether there was any beer left at the party. Marie simply fails to realize this. Insofar was what is said is determined by the question under discussion (Schoubye and Stokke 2016), or similar factors such as reasonable interpretation, we should say that Marie did (accidentally) assert that there isn’t any beer at the party. Thus, we can maintain that Sam’s report is accurate. Moreover, Sam is providing a retrospective rationalization of his belief. The actual process by which he formed his belief took place largely at the subpersonal level: Marie uttered the sentence “there isn’t any beer left”, Sam perceived her utterance, and the sub-personal cogs of interpretation started turning. The output of this process was a representation of the party as lacking beer. Sam’s belief state was automatically updated with this representation (since he possessed no defeaters). It is the actual process of belief formation, not the retrospective rationalization, which matters for knowledge. And this process, embedded in this environment, produced a justified and safe true belief. Thus, Sam gains knowledge.
Similarly, Loar cases have been used to motivate the need for knowledge of truth conditions (Heck 1995), or knowledge of co-reference (Onofri 2018), for successful communication. There are independent reasons to be skeptical of such approaches. For example, Hawthorne and Manley (2012) provide reasons to doubt that knowledge of reference is required. And Peet (2018) argues for the possibility of testimonial knowledge without knowledge of what is said.
Note that if Molly was absent in BEER knowledge-yielding communication would not have occurred. Yet the propositions entertained would be equally similar along all the relevant dimensions to the propositions entertained in BEER.
Indeed, as an anonymous referee points out, since we are not strictly concerned with knowledge through testimony it is unclear what principled reason there could be for maintaining a similarity requirement at all. I am sympathetic to this concern, and my final account makes no mention of content similarity. However, it is predicted by my positive account that, as it happens, knowledge-yielding communication will be vanishingly rare or impossible in cases where interlocutors entertain unrelated contents. This explains the common intuition that successful communication requires content similarity.
For my purposes it does not matter how we individuate propositions. They can be thought of as coarse grained entities representable as sets of possible worlds, or more fine grained structured entities.
I wish to remain neutral on what it takes for a hearer to understand a speaker to be asserting a proposition p. It may involve believing that the speaker intended to communicate p, or perhaps a perception like experience of the speaker as having said p. It does seem clear that neither the audience’s interpretation nor the speaker’s framing of their assertion can be unjustified or irrational if the exchange is to result in knowledge. Providing an account of the justification of interpretation and speech is part of the project of explicating the notion of non-deviant coordination. If the speaker’s framing of their assertion or the hearer’s interpretation is unjustified then, if they nonetheless coordinate, they will have done so in a deviant way. See Sect. 6 for some brief comments on the non-deviance condition. A full account of the rationality of communication requires a more extended treatment than can be provided here.
BASIC LUCK allows that knowledge-yielding communication is prevented in cases where the interlocutors coordinate on truth values, but easily could have failed to do so by entertaining different propositions in relevantly similar situations. It also rules out knowledge-yielding communication in cases where the actual propositions entertained only correspond in truth value as a matter of luck. However, it allows for knowledge-yielding communication in cases where the interlocutors entertain false propositions. In such cases, assuming the coordination on truth values is not lucky, it will be the falsity of the recovered proposition, not some communicative defect, which prevents knowledge acquisition.
It may be that hearers typically entertain several propositions, either simultaneously or in sequence before settling on a final interpretation. If there is a significant duration during which one misinterpretation persists before being corrected we might say that knowledge-yielding communication has failed to occur until the final interpretation has been reached.
Although I will be drawing on proposed analyses of luck I don’t claim that luck is ultimately analyzable. Even if luck is unanalyzable, attempts at analysis are useful in that they illuminate the structure of certain forms of luck.
When assessing the truth of (3) it is important to hold (1) fixed across all the worlds considered. Otherwise we get communicative luck when some event could easily have occurred which would have prevented the communicative exchange from taking place.
The modal account bears similarities to an approach to communicative success with general terms developed by Pagin (2019) according to which such success requires that the intended and recovered propositions be such that they could not easily have differed in truth value. This is not quite the same as MODAL LUCK since it allows for the requisite form of success to occur as a matter of luck. For example, it would allow for lucky communicative success in Loar cases (with general terms). This is due to its formulation in terms of the propositions believed, rather than the representational states of the agents (which could have different propositions as objects in nearby worlds).
The Fregean could avoid this issue by denying that names rigidly designate. However, Jones will still form the following actualized belief alongside his non-actualized belief: the actual person we see on the train every day is a stockbroker. The same puzzle can be generated with respect to this belief, which is plausibly communication based [Blome-Tillmann (2017) observes that we can generalize standard Gettier cases in a similar way to generate problems for simple modal anti-luck conditions on knowledge].
As an anonymous referee points out, not everyone shares the intuition that Vincent’s discovery is lucky. Pritchard (2014), for example, holds that it is merely accidental. This does not undermine the point I am making in this paper however. The important point is that there is some luck-like instability in the event. Such instability is incompatible with knowledge: if it is a mere accident that a belief is true then it does not constitute knowledge. If one were to double down and hold that a belief constitutes knowledge whenever it is non-deviant and non-lucky in the modal sense, allowing for accidentally true beliefs to constitute knowledge, then there is little basis for denying that knowledge-yielding communication occurs in LOAR 2. Anybody who goes this way can embrace the simpler modal version of my account.
Broncano-Berrocal (2018) provides a similar counter example to Pritchard’s modal account of luck and is explicit about the role coincidences play in such cases.
As an anonymous referee points out, it could be maintained that we have, in some sense, explained the relational fact once we have explained the relata. This worry is considered by Lando. She points out that it is that it may be OK to say that, in some sense, we have explained the relational fact by explaining the individual relata and conjoining them. The important point is that mere coincidences are distinguished from non-coincidences by the type of explanation we are able to give. In cases of mere coincidence we are stuck explaining the individual events and conjoining the explanations. In non-coincidental (or, non-merely coincidental) cases we can provide an independent explanation for the relational fact. We can explain the relational fact without simply stating and conjoining the explanations for the individual relata. The latter form of explanation is robust in a way in which the former is not. And it is the lack of such explanation which renders a correspondence merely coincidental [this issue is also discussed by Gamester (2018), who offers a similar solution].
I have presented the anti-luck approach as a helping with Loar cases only insofar as they are a problem for knowledge-yielding communication. However, I believe the solution extends to other forms of communicative success insofar as such successes constitute achievements, and achievements in general preclude luck.
Importantly, safeguarding against lack of coordination is not the same as positively causing coordination. As we will see in Sect. 6, there are cases in which HYBRID LUCK is satisfied despite the propositions entertained by each interlocutor being unrelated. These cases involve a third party who not only safeguards against miscommunication, but positively brings about the coordination.
I don’t claim that modal and explanatory instability exhaust the forms of instability which can render an event lucky. By further investigating the notion of luck we will be able to better understand the ways in which interlocutors must relate in order for knowledge-yielding communication to occur. However, I will not go beyond HYBRID LUCK here, as we already have enough to solve the problems we encountered with the simple view.
It might be thought that COORDINATION DEMON can be captured by appeal to coincidence. After all, the fact that Obama is not the president is explanatorily unrelated to the fact that the great barrier reef is not immune to pollution. However, the fact that Bjorn and Berit’s actual thoughts correspond in truth value is not, in this case, explained by the truth of the particular propositions they entertain. This relational fact can be fully and robustly explained without ever mentioning the contents of their token thoughts. The coordination demon’s presence explains the correspondence in truth value without the need to mention the interlocutors’ token thoughts.
References
Audi, R. (1997). The place of testimony in the fabric of knowledge and justification. American Philosophical Quarterly, 34(4), 405–422.
Barnett, D. (2015). Is memory merely testimony from one’s former self? Philosophical Review, 124(3), 353–392.
Bezuidenhout, A. (1997). The communication of de re thoughts. Noûs., 31(2), 197–225.
Bezuidenhout, A. (2002). Truth-conditional pragmatics. Noûs, 36(16), 105–134.
Blome-Tillmann, M. (2017). Sensitivity actually. Philosophy and Phenomenological Research, 94(3), 606–625.
Broncano-Berrocal, F. (2018). Purifying impure virtue epistemology. Philosophical Studies, 175(2), 385–410.
Byrne, A., & Thau, M. (1996). In defence of the hybrid view. Mind, 105(417), 139–149.
Carston, R. (2002). Thoughts and utterances: The pragmatics of explicit communication. Oxford: Blackwell.
Evans, G. (1982). The varieties of reference. Oxford: Oxford University Press.
Fraser, R. (2016). Risk, doubt, and transmission. Philosophical Studies, 173, 2803–2821.
Gamester, W. (2018). Truth: Explanation, success, and coincidence. Philosophical Studies, 175(5), 1243–1265.
Goldberg, S. (2007). Anti-individualism: Mind and language, knowledge and justification. Cambridge: Cambridge University Press.
Graham, P. (2016). Against actual-world reliablism: Epistemcally correct procedures and reliably true outcomes. In Fernandez (Ed.), Performance epistemology: foundations and applications (pp. 83–106). Oxford: Oxford University Press.
Graham, P. (2017). Normal circumstances reliabilism: Goldman on Reliability and Justified Belief. Philosophical Topics, 45(1), 33–61.
Greco, J. (1999). Agent reliabilism. Philosophical Perspectives, 13, 273–296.
Hawthorne, J., & Manley, D. (2012). The reference book. Oxford: Oxford University Press.
Heck, R. (1995). The sense of communication. Mind, 104(413), 79–106.
Heck, R. (2002). Do demonstratives have senses? Philosopher’s Imprint, 2(2), 1–33.
Lackey, J. (1999). Testimonial knowledge and transmission. The Philosophical Quarterly, 49(197), 471–490.
Lackey, J. (2008). What luck is not. Australasian Journal of Philosophy., 86(2), 255–267.
Lando, T. (2017). Coincidence and common cause. Noûs, 51, 132–151.
Leonard, N. (2018). The transmission view of testimony and the problem of conflicting justification. American Philosophical Quarterly, 55(1), 27–36.
Loar, B. (1976). The semantics of singular terms. Philosophical Studies, 30(6), 353–377.
MacFarlane, J. (2005). Knowledge laundering: Testimony and sensitive invariantism. Analysis, 65(286), 132–138.
MacFarlane, J. (2016). Vagueness as indecision. Proceedings of the Aristotelian Society, 90, 255–283.
Manley, D. (2007). Safety, content, aprioity, self-knowledge. The Journal of Philosophy, 104, 421–441.
Onofri, A. (2018). The publicity of thought. The Philosophical Quarterly, 68(272), 521–541.
Owens, D. (1993). Causes and coincidences. Cambridge: Cambridge University Press.
Pagin, P. (2008). What is communicative success? Canadian Journal of Philosophy, 38, 85–116.
Pagin, P. (2019). When does communication succeed? The case of general terms. In T. Marques & Å. Wikforss (Eds.), Shifting concepts. Oxford University Press.
Paul, M. (1999). Success in referential communication. Philosophical Studies Series Vol. 80. Dordrecht, The Netherlands: Kluwer Academic Publishers.
Peet, A. (2017). Referential intentions and communicative luck. Australasian Journal of Philosophy, 95(2), 379–384.
Peet, A. (2018). Testimonial knowledge without knowledge of what is said. Pacific Philosophical Quarterly, 99(1), 65–81.
Peet, A., & Pitcovski, E. (2017). Lost in transmission: Testimonial justification and practical reason. Analysis, 77(2), 336–344.
Peet, A., & Pitcovski, E. (2018). Normal knowledge: Towards an explanation based theory of knowledge. The Journal of Philosophy, 155(3), 141–157.
Plantinga, A. (1993). Warrant: The current debate. Oxford: Oxford University Press.
Pritchard, D. (2005). Epistemic luck. Oxford: Oxford University Press.
Pritchard, D. (2007). Anti-luck epistemology. Synthese, 158(3), 277–297.
Pritchard, D. (2012). Anti-luck virtue epistemology. The Journal of Philosophy, 109, 247–279.
Pritchard, D. (2014). The modal account of luck. Metaphilosophy, 45(4–5), 594–619.
Schoubye, A., & Stokke, A. (2016). What is said? Noûs, 50(4), 759–793.
Sosa, E. (2007). A virtue epistemology. Oxford: Clarendon Press.
Sperber, D., & Wilson, D. (1986). Relevance: Communication and cognition. Oxford/Cambridge, MA: Blackwell/Harvard University Press.
Weber, C. (2015). Indexical beliefs and communication: Against stalnaker on self-location. Philosophy and Phenomenological Research, 90(3), 640–663.
Williamson, T. (2000). Knowledge and its limits. Oxford: Oxford University Press.
Acknowledgements
This paper has been in the works for a long time, and it has benefited from discussion with, and comments from too many people to name. However, I will do my best to name a few: Mark Bowker, Herman Cappelen, Alex Davies, Anna Drozdzowicz, Rachel Fraser, Lizzie Fricker, Mikkel Gerken, Olav Gjelsvik, Sandy Goldberg, Patrick Greenough, Josh Habgood-Coote, Torfinn Huvenes, Matt McKeever, Andrea Onofri, Peter Pagin, Kim Phillips Pedersen, Eli Pitcovski, Joey Pollock, and an anonymous reviewer for Philosophical Studies. This paper has also benefited greatly from the discussion it received when presented at the University of St Andrews ‘Testimony in Context’ workshop, the University of Hamburg ‘New Trends in Epistemology’ workshop, and the ConceptLab work in progress seminar at the University of Oslo. I thank the audiences who were present at these events, and the organizers.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Peet, A. Knowledge-yielding communication. Philos Stud 176, 3303–3327 (2019). https://doi.org/10.1007/s11098-018-1175-7
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11098-018-1175-7