As the title suggests, Jeremy’s talk argued that there are obstacles specific to sharing desires, as opposed to beliefs (where to “share” an attitude is for two subjects to have attitudes with the same content). A widespread view of desires, the mirror view, suggests the opposite. According to the mirror view, desires are simply “mirror images” of beliefs: like them in most respects, differing only in direction of fit. On this view, we would expect desires to be as easy to share as beliefs. Against the mirror view, Jeremy defended the rival reward theory of desire and argued that, on this view, desires turn out to be much more difficult to share than beliefs.
According to the reward theory, desires are essentially linked to the processing of reward.
“Reward” is here to be understood in its technical sense, as what underlies reinforcement learning. More specifically, Jeremy argued that to desire that p is to represent that p and for that representation to trigger the right activity in the reward system. There are two types of content that can play this role: propositional content and perceptual contents. Desires of the second sort are what Jeremy called “P-desires.” It is P-desires, in particular, that cause trouble for desire-sharing.
What sort of content do P-desires have? According to Jeremy, these contents are characterized by being a) encoded in the proprietary format of the perceptual system (e.g. iconic format) and b) experientially available. For purposes of Jeremy’s argument, the important thing about these contents is that they “outstrip language”: They cannot be encoded into words without loss. This, argues Jeremy, distinguishes them from beliefs, whose contents are always available for discursive expression. This is why P-desires cannot easily be shared: their contents cannot be linguistically communicated to others. Jeremy nevertheless allowed that P-desires can be shared in exceptional cases. He cited two ways. First, by mere coincidence. Second, potentially, via non-linguistic means of communication and influence, such as music.
Jeremy ended his talk by discussing some upshots of his view. 1) The upshots for the philosophy of action, he argued, are few. Though we cannot share P-desires, we can still coordinate our actions based on desiring the same object (e.g., we both desire food from the same restaurant.) 2) The view may have some upshots for how to understand the notion of a preference. A preference is a standing desire, and the view implies that we cannot share standing P-desires or “P-preferences.” 3) Finally, Jeremy mentioned some big-picture implications. In the thought-sharing literature, there is an interest in the question whether people can connect in a metaphysically special way. Jeremy’s view about P-desires suggest a deflationary answer to this question.
Michael discussed the notion of thought sharing from a naturalistic perspective, and in particular, from the basis of the best empirical theories of reference. One of his conclusions is that there are two theoretically motivated notions of sharing: coarse sharing which concerns reference, and fine sharing which concerns modes of reference, where the latter is also made possible by reference borrowing. Importantly, a Fregean principle for the individuation of thoughts plays no role in the discussion, which is focused on what properties cause behaviour.
He started off by considering cognitive maps in nonhuman organisms such as gobies and honey bees. In these cases, ethologists attribute a certain semantic property to an inner state of an organism to explain the state’s place in the causal nexus. What is sharing in the case of these organisms? Two bees, for instance, share a proto-thought when their proto-thoughts have the one semantic property that explains their place in the causal nexus. All proto-thoughts and thoughts can be classified in more than one way depending on the explanatory task at hand: when it comes to the information used by other individuals, the semantic property in question is representation; when it comes to behavior, the *mode of representation* is relevant. With each way of classifying a thought by a semantic property goes a way of sharing.
In humans, representations are far more complex, involving token representations in the mind (as opposed to abstract “propositions”) with conceptual constituent structure (in harmony with the idea of a language of thought). For X and Y to share a thought or concept is for them to have token representations that have *the same semantic property,* according to a theoretically motivated classification. For thought’s and concept’s role in guiding to reality, *reference* and coarse sharing is enough. In case of human behavior, mode of reference and fine sharing is required. Michael provided a number of applications of this basic framework.
A mode of reference is a network of underlying causal chains made up of groundings, either multiple (proper names) or single (demonstratives) that fix the reference, and reference borrowings. Groundings in an object involve the object’s being the focus of the particular experience that caused the concept. Borrowings are largely unconscious causal mental processes in a borrower in response to the use of a name by a competent speaker. Groundings are token acts of perception by individuals but modes of reference are social. Michael discussed various examples involving demonstratives and indexicals to illustrate the role of these theoretical notions.
The remainder of the talk used these ideas to address sameness of thoughts and concepts in the intrasubjective and intersubjective case. In the intrasubjective case, the semantic property that explains the concept’s place in the causal nexus, and in virtue of which the concept causes behaviors, and exists at all is a certain mode of reference. Several examples are again considered with variations of grounding and modes of reference in intrasubjective reasoning. The key general remark, Michael suggests, is that our theory of sameness of concepts, and hence thoughts, explain their causal role in reasoning without a commitment to a Fregean principle for the individuation of thought.
As regards the intersubjective case, the explanatory relevant semantic property is reference, which is associated with transparent descriptions of thoughts and agreement about the world. However, Michael applied this approach to the classical Paderewski and demonstrative examples where reference borrowing also enables fine sharing of concepts via the same social mode of reference. The framework suggests that we need both notions of sharing with neither notion being more fundamental than the other.
The discussion revolved around a set of issues which included the following: the difference between beliefs and desires when it comes to thought sharing and perceptual content; as well different levels at which we may share our desires, beyond “P-desires”; the rationality of desire and the connection with standing desires; agreement in desires vis-à-vis coordination and sharing.
Part of the debate concerned content or experience in concept and thought sharing, and degrees of fine-grained sharing and the role of epistemically undemanding reference borrowing in determining sharing, together with the concept of file shared in e.g. “Paderewski” scenarios or the “George Smith” example. These point to broader questions about the role of rationality and intentions, intentional behavior, and normativity in the elucidation of communicative interactions and understanding, and the scope of reference borrowing, which may include indexicals such as the first-person pronoun. Another issue is how much a content-based theory of sharing may or may not rely on the language of thought hypothesis or aspects of it.
Work produced with the support of a 2025 Leonardo Grant for Scientific Research and Cultural Creation, BBVA Foundation. The Foundation takes no responsibility for the opinions, statements and contents of this project, which are entirely the responsibility of its authors.
We use cookies to deliver personalized content, analyze trends, administer the site, track user movements on the site, and collect demographic information about our user base as a whole. Accept all cookies for the best possible experience on our website or manage your preferences.
Legal notice
