Philosophers Among Swedish Research Council Grant Winners

---

Several projects led by philosophers are among the recent recipients of large grants from the Swedish Research Council.

[Sigrid Sandström, untitled (detail)]

They are:

**Natural Properties – a Causation-First Approach **Bram Vaassen (Umeå University), Jennifer McDonald (Columbia University), Tom Wysocki (Georg-August-Universität Göttingen) 4,530,000 SEK (approx. $489,300), 3 years

The world seems to come with natural categorizations, such as electron, proton, moose, and blue things. Other categorizations are clearly less natural, such as taller than the Sibelius monument, having a cousin named Sarah and a dog named Bruce, or being either a proton or a moose. Determining which properties are natural is crucial across philosophy, science, and everyday life. For example, a key debate in all three areas is whether contested concepts like ‘race’ and ‘gender’ correspond to cogent natural properties rather than arbitrary collections of traits – ones similar in kind to having a cousin named Sarah and a dog named Bruce.

This project aims to uncover what makes properties natural by comparing the causal roles of stereotypically natural and stereotypically unnatural properties. Our working hypothesis is that naturalness is determined by causal role: a property like being a proton or a moose is less natural than being a proton because it plays a more limited causal role in the world, making it less suited for good explanations. Such a causation-first account of naturalness is reductive, nominalist-friendly, refrains from ruling properties out by fiat (such as extrinsic ones), delivers natural properties directly relevant to the special sciences, and accounts for the explanatory and predictive value of intuitively natural properties.

**How Words Mean: Lessons from Large Language Models **Jessica Pepp (PI; Uppsala University), Dimitri Coelho Mollo (Umeå University), Pär Sundström (Umeå University), and Bram Vaassen (Umeå University) 4,536,000 SEK (approx. $489,950), 3 years

Given the right prompt, ChatGPT will output, “Ibuprofen can increase the risk of gastric ulcers”. The same output may be produced by a human doctor. It is natural to assume that these two outputs have the same meaning. But do they?

Earlier research has by and large argued in one of two ways about this question. One argument has been that outputs of ChatGPT and similar large language models are meaningless since ChatGPT lacks the mental states that according to established theories are required to ground meanings. Another argument has been that the doctor’s and ChatGPT’s utterances have the same meaning on the ground that established theories can, after all, deliver that result.

This project explores two paths not taken in earlier research. It has two aims. The first aim is to test the hypothesis that ChatGPT’s output is meaningful but has a meaning that differs from the doctor’s utterance. The second aim is to work from the assumption that the two outputs have the same meaning and develop a novel account of what makes this the case. Specifically, we will explore the hypothesis that ChatGPT expresses the same meanings that we express, though not because of any underlying mental states, but because ChatGPT has “quasi-mental states” that ground its meanings.

**Scientific Authorship and Collective Responsibility **Olle Blomberg (PI; University of Gothenburg), Gert Helgesson (Karolinska Institute), and Björn Petersson (Lund University) 4,368,000 SEK (approx. $471,800), 3 years

In cases of scientific misconduct, authorship is the most salient indicator of responsibility. This raises a problem when scientific publications are co-authored: it is often unclear who should be held responsible. The purpose of the project is to determine how the academic community can hold coauthors, individually or collectively, responsible for research misconduct in a fair way, acceptable within that community and to the public. This will be achieved by doing a qualitative study of arguments and reasons appealed to in decisions regarding co-author responsibility in misconduct cases in Sweden (subproject 1), by reviewing positions and arguments in the research ethics literature on co-author responsibility (subproject 2), by analysing the relationship between co-author responsibility and justifications of institutional sanctions (subproject 3), and by building and applying a framework for evaluating models for assigning responsibility of various kinds to co-authors (subproject 4). Some of the models to be evaluated in subproject 4 can be found in the research ethics literature, but we will also develop a genuinely collectivistic model of co-authorship responsibility that this literature has overlooked.

Some authorship guidelines refer to co-authors’ “joint” or “collective” responsibility, but it is unclear how these notions should be applied in practice, potentially leading to arbitrariness. Improved ethical guidelines for authorship and a better understanding of how these are applied in actual cases are needed. The project will address this need and contribute to strengthening ethical guidelines for scientific authorship and facilitate adherence to them. Moreover, by using scientific authorship as a test case for accounts of moral responsibility in collective contexts, the project will contribute to theory development in this area of philosophy.** **

**The Truthmakers for Statements Expressing Negative Causation **Tobias Hansson Wahlberg (Lund University) 4,374,000 SEK (approx. $472,450), 3 years

In daily life and in the sciences, we normally do not hesitate to speak of non-occurrences or absences as causes and effects. Many philosophers think that such statements expressing ‘negative causation’ can be and often are true. However, philosophers have not yet provided plausible and detailed accounts of what in the world makes such statements true. The so-called ‘truthmakers’ cannot involve causal relations, because relations require relata to exist, and absences cannot function as relata. This project will make a thorough investigation of what—if anything—can plausibly be taken to be the truthmakers.

**Social Indeterminacy **Johan Brännmark (Stockholm University), Miguel F. Dos Santos (Stockholm University), and Thomas Brouwer (University of Leeds) 4,539,000 SEK (approx. $490,000), 3 years

Social indeterminacy is prevalent: wherever there are social rules in place there are questions about just what falls under them, and there will also be cases where it is indeterminate if something belongs to certain social category or not. But despite its ubiquity it is still an understudied phenomenon in social ontology. This three-year project seeks to address this lack by inquiring into (i) the different ways in which social indeterminacy can arise, (ii) the ways in which different models of social ontology can account for all its varieties, (iii) the more general nature of social indeterminacy, and (iv) the possible harms and benefits that can come with it, as well as when it can make sense to reduce or increase social indeterminacy.

**The Evolution of Online Virality **Emmanuel J. Genot (PI; Lund University), Patricia Rich (University of Bayreuth) 2,378,000 SEK (approx. $257,000), 2 years

Our project aims to understand why harmful online content (such as “fake news,” conspiracy theories, hate speech, and dangerous challenges) can spread so easily and to understand how we can better act to stop it. To date, no studies have rigorously examined how users are inclined to “walk a mile in someone else’s shoes” and try to think like others when sharing content online. At the same time, efforts to combat harmful content with fact-checking and moderation have often failed. We believe we can demonstrate how these two problems are interconnected and propose better solutions to prevent harmful online content.

Before starting this project, we developed a mathematical analysis of how people share content on social media platforms. We used this analysis to build a computer model that mimics how people share content online while anticipating other users’ reactions. Our simulations showed that a small group of people seeking attention from a majority can trigger a chain reaction of harmful content spreading. With this project, we aim to improve our model to better reflect real-world online interactions.

An improved model will allow us to predict how harmful content can spread online more accurately and realistically. We can then compare our model’s predictions with data from experimental attempts to drive people away from harmful content. Then, we can evaluate whether our model provides a better understanding of observed behavior, particularly compared with strategies proposed so far to nudge online users away from harmful content. Finally, armed with a better understanding of how users think about each other, propose new strategies.

A total of 89 grants were awarded, from 1131 applications. You can learn more about the grants here.