The Epistemology of Linguistic Understanding Workshop

University of Oslo, November 6th and 7th

The workshop is open to all, but please register to K.P.Pedersen.

Programme for the workshop:


Tuesday November 6th- G.M. 652


09:00-09:30 Welcome

09:30-11:00 Guy Longworth

'Thinking Together'  

We often take ourselves to hear (or see, or feel) people say things. I discuss some of the perceptual and cognitive abilities that are involved in those (and associated) achievement(s). Focusing on the case of thinking out loud, my aim is to explore Ryle’s suggestion that the positions of speaker and audience are in certain significant ways symmetrical.

11:05-12:30 Anna Drożdżowicz

'Experiences of understanding as reasons'

It is often observed that competent language users experience perceptual-like states of utterance understanding (e.g. Fodor, 1987; Pettit, 2002; Siegel, 2006; Dodd, 2014). Recent debates on the epistemology of language understanding point to the potential justificatory role of such experiences in forming beliefs about what speakers say, on the assumption that such states have perceptual-like nature and that their epistemic contribution is roughly similar to that of perceptual states (e.g. Fricker, 2003, Brogaard, 2016). However these accounts say very little about the exact relation between such experiences and corresponding beliefs about what speakers say. In my talk I plan to address this issue by making two related points. First, I will argue that the exact role(s) that such states can play can be captured in terms of (a) doxastic justification where such experiences provide reasons that hearers can treat as reasons for their beliefs and (b) discursive justification that is crucial for many cooperative tasks based on linguistic communication. Second, I will argue that we should abandon a simple perceptual-like model of epistemic reasons that such states can provide. Drawing on observations about the phenomenology and origin of such experiences, I will argue that the nature of epistemic reasons they can provide is different from those we often get from ordinary cases of perception.

12:30-13:30 Lunch

13:30-15:00 Natalia Hickman

'Semantic Normativity Revisited: Semantic Deviance, Reasons-sensitivity and Illusions of Thought'

The paper reconsiders the question whether meaning is normative, picking up on a sensible suggestion of H-J Glock’s, viz. that the motivation for any plausible normativism must begin from the idea of a semantic mistake. §1 Provides a theoretical rationale for this starting point, clarifying a notion of semantic deviance and articulating a normativist interpretation of it. §2 explores semantic deviance in connection with failures of sense or thought, with particular reference to the semantic system of Wittgenstein’s Tractatus, but also in connection with the (less niche, though still controversial) idea of a semantic category-mistake. The tentative conclusion is that illusions of thought and sense provide an argument for semantic normativism, insofar as one ought, unconditionally, to avoid failures of sense and thought. §3 Attempts to make progress in the assessment of normativism from a different angle, while still understanding normativism in terms of semantic deviance. I consider whether a general knowledge norm of practical rationality, instead of any semantic normativity, could explain why a speaker is criticisable for failures to employ linguistic expressions in a way that is guided by the facts, rather than fictions or confusions, concerning their meaning. My first conclusion, again tentative, is that it can’t: such a norm explains why a speaker’s choice of linguistic expression ought to be guided by a known fact, but not why it should be guided by the relevant semantic fact in particular, i.e. by what the expression means. My final speculation is that a speaker’s language use will qualify as rational only if we can explain the special pertinence of a semantic fact to her semantic acts (i.e. her uses of expressions), and that this requires us to postulate a normative connection between fact and act: between meaning and use.

15:00-15:10 Coffee

15:10-16:40 Jennifer Hornsby

'Knowing what I say (or, as it might be, what you yourself say, or she herself says, or .. !)'

I’ll try to update arguments for a suggestion I’ve made before. I suggested that when one thinks of oneself as speaking—i.e. as playing the role of agent in, say, a testimonial exchange—the merits of thinking of some linguistic knowledge as a kind of know-how or competence whose possession is not a matter of knowing facts/propositions can be clear. (The idea is to take the focus off audiences of speech, often treated as needing to “interpret” utterances.)

I shall need to defend what I take to be Gilbert Ryle’s conception of knowledge how against what seems to me Jason Stanley’s misreading of Ryle. (Stanley replied to me when I first made my suggestion. And in his 2011 book, Stanley presents his own “intellectualism” as the only alternative to Ryle.)

I don’t think that there needs to be very much wrong with the formula: “To know the meaning of an utterance is to know the meanings of its constituent words and to know how the words are put together.” But I think that it’s possible to distinguish between knowledge of these two sorts without treating knowledge of either sort as propositional au fond.

16:45-18:10 Andrew Peet

'Communication, Reliability, and Understanding' 

In this talk I present some objections to reliability based approaches to understanding. That is, views committed to the following:   RELIABLE WHAT IS SAID: Understanding involves reliable recovery of what is said (i.e. the speaker's intended meaning).   I then consider how we might develop a view of understanding in the spirit of the reliability based approach, but which avoids the problems I raise.  


Wednesday November 7th- G.M. 452


09:00-09:30 Coffee

09:30-11:00 Sandy Goldberg

'The Epistemology of Testimony: Comprehension vs. Trust vs. Assurance'

One of the central choice points in the epistemology of testimony concerns whether to model the phenomenon in individualistic or interpersonal terms.   Those who favor the latter (like myself) hold that the epistemic goodness or badness of a testimonial belief sometimes reflects the epistemic goodness or badness of the testimony (or the relevant cognitive state of the testifier).  But what is the mechanism by which facts pertaining to one person – the speaker – bear on the epistemic goodness of a belief in another person – the recipient of the testimony?   To date, the two main answers have appealed either to the phenomenon of trust (of a speaker by an audience) or else to the phenomenon of assurance (of an audience by a speaker).  In this paper I suggest that the state of comprehension (of a speech act by an audience) may provide all that we need in the way of such a mechanism.

11:05-12:35 Joey Pollock

'Holistic understanding, disagreement, and verbal disputes'

Traditional accounts of disagreement require that different subjects can often entertain the same propositional contents. For example, a simple case of disagreement is one in which two subjects take opposing attitudes to the same content. For most views of mental content, this ‘shared content’ approach is an obvious choice. However, for the radical holist, this framework is problematic: the holist claims that different subjects cannot, in practice, share thought content.  

I have two aims in this talk. The first is to suggest an account of agreement and disagreement for the holist: this account treats both agreement and disagreement as graded notions. My second aim is to show that the structure of the holist’s account should be attractive to all philosophers who work on disagreement, regardless of their semantic commitments. Disagreement (and agreement) between subjects cannot be adequately characterised without appeal to a holistic account of linguistic and conceptual understanding.

12:35-13:35 Lunch

13:35-15:00 Kim Pedersen

'Knowing What Speakers Say'

Hearing utterances in a familiar language typically provides us with knowledge of what was said by the speaker in making those utterances. For example, hearing your friend utter the sentence ‘hedgehogs have about five thousand spines’ you straight away come to know that she said that hedgehogs have about five thousand spines. How does one come to know what was said in cases like this? Given the phenomenological immediacy with which the knowledge is formed - one seems to just hear the speaker as saying such-and-such - one might think that the knowledge is immediate, not dependent on further knowledge. Many philosophers are committed to this claim. In this talk I will argue that the knowledge is not immediate: knowledge of what is said in ordinary cases depends on further pieces of knowledge - specifically, knowledge of word meanings. I'll show that knowing what words mean plays an epistemic role, not just a causal one.

15:00-15:15 Coffee

15:15-16:45 Anders Nes

'A nonconceptual content of understanding'  

I argue that fluent utterance comprehension, as a personal-level achievement, has a nonconceptual content, concerning semantic aspects of the utterance understood. Specifically, I argue that this so because such comprehension involves a nonconceptual representation of the relation _x means y_, represented to hold between an utterance and whatever it is taken to mean. I call this view 'metasemantic nonconceptualism'. My argument rests on two main claims. First, 'Comprehension as Metasemantic', according to which understanding a certain utterance, U, as having a certain meaning, M, requires not merely grasping M but grasping M as a meaning of U. Second, 'Semantic Concepts Unrequired', according to which understanding an utterance as having a certain meaning does not require possession of semantic concepts, such as the concepts of meaning, reference, or truth. I defend the Comprehension as Metasemantic by reflections on cases where hearers entertain a content that may or may not be the meaning of an utterance that they perceive, but do not understand the utterance as having that meaning, and, I claim, fail to do so because they do not attribute the content to the utterance as its meaning. I defend Semantic Concepts Unrequired by appeal to developmental studies indicating that at least some speakers, notably three-year olds, have such limited metalinguistic awareness that they do not meet plausible possession conditions for semantic concepts. I end by speculating on what form the posited nonconceptual content in comprehension might take, distinguishing two options. On the first, comprehension implicates a purely nonconceptual content of the form, roughly, _this utterance means that_, where the final 'that' is understood as a nonconceptual referential element, functioning perhaps something like a demonstrative, picking out a conceptualised content. On the second, comprehension implicates an essentially only partially nonconceptual content, to the effect that _this utterance means M_, where the subject needs to possess the concepts that specify the meaning M, but not the concept of meaning specifying its represented relation to the utterance. 


Published Oct. 23, 2018 2:37 PM - Last modified Feb. 11, 2020 9:32 AM