Alex Grzankowski

This paper investigates the normative asymmetry between recalcitrant emotions – emotions that conflict with our judgments – and emotional responses to fiction. While the former generate normative tension, the latter do not, but given their structural similarities, this asymmetry is surprising. We argue that the difference arises from the compartmentalization of emotion within imaginative contexts. Drawing on the concept of a cancellation context, we propose that the norm of emotional correctness is suspended when emotions remain compartmentalized. This framework accounts for both the permissibility of fiction-directed emotions (as compared to recalcitrant emotions) and their occasional spillover into real-world normativity.

As AI capabilities grow, the field faces a choice in how to define success. We argue that the primary objective of AI research should be Human Amplification -- aiming for systems that augment human abilities and preserve human agency. While the current focus on intelligent agents drives significant progress, treating replication of human agency as the ultimate metric of success emphasizes delegation over augmentation. We discuss how this presents distinct challenges regarding control and power dynamics. By contrast, anchoring AI development in the service of human agency opens up a much broader design space that blurs the distinction between non-agentic tools and sophisticated delegates. This shift directs research toward cognitive extensions that ideally spark a virtuous cycle of recursive human amplification, where advances in AI further empower us to build safer and more beneficial systems.

Many people feel compelled to interpret, describe, and respond to Large Language Models (LLMs) as if they possess inner mental lives sim- ilar to our own. Responses to this phenomenon have varied. Inflation- ists hold that at least some folk psychological ascriptions to LLMs are warranted. Deflationists argue that all such attributions of mentality to LLMs are misplaced, often cautioning against the risk that anthropomor- phic projection may lead to misplaced trust or potentially even confusion about the moral status of LLMs. We advance this debate by assessing two common deflationary arguments against LLM mentality. What we term the robustness strategy aims to undercut one justification for believing that LLMs are minded entities by showing that putatively cognitive and humanlike behaviours are not robust, failing to generalise appropriately. What we term the etiological strategy undercuts attributions of mental- ity by challenging naive causal explanations of LLM behaviours, offering alternative causal accounts that weaken the case for mental state attribu- tions. While both strategies offer powerful challenges to full-blown infla- tionism, we find that neither strategy provides a knock-down case against ascriptions of mentality to LLMs simpliciter. With this in mind, we ex- plore a modest form of inflationism that permits ascriptions of mentality to LLMs under certain conditions. Specifically, we argue that folk prac- tice provides a defeasible basis for attributing mental states and capacities to LLMs provided those mental states and capacities can be understood in metaphysically undemanding terms (e.g. knowledge, beliefs and de- sires), while greater caution is required when attributing metaphysically demanding mental phenomena such as phenomenal consciousness.

Perception can provide us with a privileged source of evidence about the external world – evidence that makes it rational to believe things about the world. In Reasons First, Mark Schroeder offers a new view on how perception does so. The central motivation behind Schroeder's account is to offer an answer to what evidence perception equips us with according to which it is what he calls world-implicating but non-factive, and thereby to glean some of the key advantages of both externalism and internalism, respectively. He answers this motivation by developing a more specific view that he calls the Apparent Factive Attitude view, which pairs an answer to what evidence is provided by a perceptual experience with an answer to why having that perceptual experience provides you with that evidence. In this paper, we advance two interconnected problems for Schroeder's Apparent Factive Attitude view. A traditional intuitive judgment that often motivates internalists is the idea that internal duplicates must necessarily be equally rational in whatever beliefs they have. Schroeder's arguments rely on a weaker claim – that people who are both internal and historical external duplicates but differ only in the veridicality of a single perceptual experience must be equally rational in whatever beliefs they have. In this way he preserves what he argues to be a more compelling internalist intuition. But our arguments will show that Schroeder's view is committed to denying an even more compelling internalist intuition yet – that internal duplicates must have the same phenomenology.

LLMs are statistical models of language learning through stochastic gradient descent with a next token prediction objective. Prompting a popular view among AI modelers: LLMs are just next token predictors. While LLMs are engineered using next token prediction, and trained based on their success at this task, our view is that a reduction to just next token predictor sells LLMs short. Moreover, there are important explanations of LLM behavior and capabilities that are lost when we engage in this kind of reduction. In order to draw this out, we will make an analogy with a once prominent research program in biology explaining evolution and development from the genes eye view.

The view that moods are dispositions has recently fallen into disrepute. In this paper, we want to revitalise it by providing a new argument for it and by disarming an important objection against it. A shared assumption of our competitors (intentionalists about moods) is that moods are ‘diffuse’. First, we will provide reasons for thinking that existing intentionalist views do not in fact capture this distinctive feature of moods that distinguishes them from emotions. Second, we offer a dispositionalist alternative that we think does better. Finally, we disarm the objection that is responsible for dispositionalism’s bad press.

LLMs are statistical models of language learning through stochastic gradient descent with a next token prediction objective. Prompting a popular view among AI modelers: LLMs are just next token predictors. While LLMs are engineered using next token prediction, and trained based on their success at this task, our view is that a reduction to just next token predictor sells LLMs short. Moreover, there are important explanations of LLM behavior and capabilities that are lost when we engage in this kind of reduction. In order to draw this out, we will make an analogy with a once prominent research program in biology explaining evolution and development from the gene's eye view.

Do emotions represent values? The dominant view in philosophy has it that they do. There is wide disagreement over the details, but this core commitment is common. But there is a new comer on scene: the attitude view. According to it, rather than representing value properties, there is a value-relevant way you represent the targets of emotion. For example, in feeling angry with someone you stand to them in the relation of representing-as-having-wronged-you. Although a recent view, it has quickly generated discussion. But the central considerations in favour of each view are left wanting and it is hard to see how to choose amongst these alternatives. I argue that there is an empirical path to a decision.

In order to uncover the inner workings of our capacities, we look to ‘effects’. Most of us have the capacity to distinguish between spoken ‘ba’ and ‘fa’ sounds. One thought is that this is achieved through aural sensitivities that detect changes in vibration picked up by the eardrum. But the McGurk Effect suggests that there is more to the story. Without changing the incoming vibrations, sound experience can be modulated by showing a video of a mouth making a ‘ba’ sound or a ‘fa’ sound with a consistent sound overlaid. We learn that our overall auditory experiences are at least in part determined by visual cues in addition to what’s first picked up by our eardrums. The McGurk Effect gives us a hint into the inner workings of audition and helps us better understand the capacity to discriminate sounds of a certain sort. In the present paper, the focus is on emotional capacities and a well known effect – recalcitrance. Recalcitrant emotions, such as fearing the dog even though one knows that the dog is harmless or being angry with one’s partner even when one realises it was only in a dream that the partner was nasty, have played the role of effect in much theorising about emotions. But in my view, we’ve stayed a bit too close to home, aiming to fit the effect into a paradigm – the representationalist paradigm – that isn’t fit for purpose. I will use this criticism as a launching off point to introduce a different way of thinking about emotions that is better suited to making sense of recalcitrance. I will argue that emotions are transitions between representational states rather than being representational states themselves. The view is better suited to make sense of recalcitrance and, at the end of the paper, I will offer reasons for thinking that main points that speak in favour of a representationalist approach to emotion can be recaptured or explained away by the transitions view.

The present paper looks at one of the most thorough articles on the intelligence of GPT, research conducted by engineers at Microsoft. Although there is a great deal of value in their work, I will argue that, for familiar philosophical reasons, their methodology, ‘Black-box Interpretability’ is wrongheaded. But there is a better way. There is an exciting and emerging discipline of ‘Inner Interpretability’ (also sometimes called ‘White-box Interpretability’) that aims to uncover the internal activations and weights of models in order to understand what they represent and the algorithms they implement. In my view, a crucial mistake in Black-box Interpretability is the failure to appreciate that how processes are carried out matters when it comes to intelligence and understanding. I can’t pretend to have a full story that provides both necessary and sufficient conditions for being intelligent, but I do think that Inner Interpretability dovetails nicely with plausible philosophical views of what intelligence requires. So the conclusion is modest, but the important point in my view is seeing how to get the research on the right track. Towards the end of the paper, I will show how some of the philosophical concepts can be used to further refine how Inner Interpretability is approached, so the paper helps draw out a profitable, future two-way exchange between Philosophers and Computer Scientists.

Your belief that Obama is a Democrat would not be the belief that it is if it did not represent Obama, nor would the pain in your ankle be the state that it is if, say, it felt like an itch. Accordingly, it is tempting to hold that phenomenal and representational properties are essential to the mental states that have them. But, as several theorists have forcefully argued (including Kripke (1980) and Burge (1979, 1982)) this attractive idea is seemingly in tension with another equally attractive thesis, namely, the token-identity thesis; the thesis according to which every mental state token is identical with some or other token physical state. In this paper, we show that these seemingly incontrovertible essentialist intuitions are in fact compatible with “token physicalism” regarding the mental. Given a suitably plentitudinous ontology of objects, we argue that there are physical things with which our token mental states can be identified. This is preferable to existing views that give up the essentiality claims or simply reject the token-identity thesis.

One of the most influential traditional objections to Adverbialism about perceptual experience is that posed by Frank Jackson’s ‘many property problem’. Perhaps largely because of this objection, few philosophers now defend Adverbialism. We argue, however, that the essence of the many property problem arises for all of the leading metaphysical theories of experience: all leading theories must simply take for granted certain facts about experience, and no theory looks well positioned to explain the facts in a straightforward way. Because of this, the many property problem isn’t on its own a good reason for rejecting Adverbialism; and nor is it a puzzle that will decide amongst the other leading theories.

How we evaluate things and what we desire are closely connected. In typical cases, the things we desire are things that we evaluate as good or desirable. According to evaluation theories of desire, this connection is a very tight one: desires are evaluations of their objects as good or as desirable. There are two main varieties of this view. According to Doxastic Evaluativism, to desire that p is to believe or judge that p is good. According to Perceptual Evaluativism, to desire that p is to perceive p as being good (or for p to seem good). The present paper poses a puzzle for such views. The puzzle should be familiar to theorists interested in the normativity and metaphysics of the emotions, but I am unaware of its application to desire. The aim of the present paper is to present the puzzle as it applies to desire, which should be of independent interest, but I also hope that by shining a light on the puzzle in this domain, we might put ourselves in a better position to offer a solution in all cases. At the end of the paper, I gesture towards a promising way ahead that departs from relying on contradictory contents.

Propositions are the things we believe, intend, desire, and so on, but discussions are often less precise than they could be and an important driver of this deficiency has been a focus on the objects but a neglect of the attitudinal relations we bear to them. In what follows, we will offer some thoughts on what it means for a proposition to be the object of an attitude and we will argue that an important part of the story lies with the attitude relations rather than the propositions. As we will see there are infinitely many relations we might bear to a proposition, but the propositional attitude relations are special amongst them. Accounting for what makes them special will be an important component in the discussion that follows. We will argue that once one appreciates certain facts about propositional attitude relations, various claims that metaphysicians often make regarding propositions themselves begin to look undermotivated. In fact, many views on the metaphysical nature of propositions come to look like plausible candidates for being that to which our propositional attitudes relate us. As will emerge, we will see that the principle role of propositions in the theory of mind is simply to keep track of how our attitudinal states represent things as being. But, we argue, in order to do this work, very few constraints must be placed on the nature of propositions themselves. In particular, contra much of the recent work on the metaphysics of propositions, they need not represent nor must they be structured. In light of these observations, we conclude by sketching our own favored minimalist view of propositions.

How fine-grained are the contents of our beliefs and other cognitive attitudes? Are the contents of our beliefs individuated solely in terms of the objects, properties, and relations that figure in their truth conditions, or rather in terms of our concepts, or modes of presentation of those objects, properties, and relations? So-called Millians famously maintain the former whereas their Fregean rivals hold the latter. Though much ink was spilled on the question of grain, relatively little was ever achieved by way of consensus. We think the lack of consensus itself cries out for explanation. In this paper, we sketch a pluralist resolution (or, better, a dissolution) of the debate that flows from some extremely minimal commitments regarding the metaphysics of propositions and the attitudinal relations we bear to them. In doing so, we focus on the Act-type conception of propositions of Hanks (2015) and Soames (2010, 2016) and our own (2019) favored deflationary account, Minimalism.

In discussions of the emotions, it is commonplace to wheel out examples of people who know that rollercoasters aren’t dangerous but who fear them anyway. Such cases are well known to have been troubling for cognitivists who hold the emotions are judgments or beliefs. But more recently, it has been argued that the very theories that emerged from the failure of cognitivism face trouble as well. One gets the sense that the theory that can accomplish this will win a crucial point over its competitors. In the present paper I offer a new approach to making sense of the normative tension to which recalcitrant emotions give rise. Interestingly, the approach is one that can be adopted by anyone willing to grant that emotions are themselves governed by norms.

According to the classical account, propositions are sui generis, abstract, intrinsically-representational entities and our cognitive attitudes, and the token states within us that realize those attitudes, represent as they do in virtue of their propositional objects. In light of a desire to explain how it could be that propositions represent, much of the recent literature on propositions has pressured various aspects of this account. In place of the classical account, revisionists have aimed to understand propositions in terms of more familiar entities such as facts, types of mental or linguistic acts, and even properties. But we think that the metaphysical story about propositions is much simpler than either the classical theorist or the revisionist would have you believe. In what follows, we argue that a proper understanding of the nature of our cognitive relations to propositions shows that the question of whether propositions themselves represent is, at best, a distraction. We will argue that once this distraction is removed, the possibility of a very pleasing, minimalist story of propositions emerges; a story that appeals only to assumptions that are shared by all theorists in the relevant debate.

In her black and white room, Mary doesn’t know what it is like to see red. Only after undergoing an experience as of something red and hence acquainting herself with red can Mary learn what it is like. But learning what it is like to see red requires more than simply becoming acquainted with it. To be acquainted with something is to know it, but such knowledge, as we argue, is object-knowledge rather than propositional-knowledge. To know what it is like one must know an appropriate propositional answer to the question ‘what is it like?’. Despite this mismatch between object-knowledge and knowing an answer, we believe that acquaintance is crucial to Mary’s epistemic progress. When Mary leaves her black and white room, her new knowledge tempts one to think that she must come to know a candidate answer (a coarse-grained fact) that she didn’t know in her room. Since Mary already knows all the physical facts in her room, any additional facts she might learn appear to threaten physicalism. In reply, many physicalists have been attracted to the phenomenal concept strategy according to which Mary can come to have new knowledge and hence know a new answer to the question ‘what is it like to see red?’ by entertaining a coarse-grained fact under a concept she didn’t possess in her room – Mary learns a new fine-grained fact. We believe both of these accounts of Mary’s epistemic progress are mistaken. As we argue, Mary could know every fact (coarse-grained and fine-grained) that might serve as an answer to the question ‘what is it like to see red?’ and still not know what it is like. The physical world leaves no leftover coarse-grained facts for Mary to learn and because concepts are sharable, easy to possess, and easy to introduce, there are possible situations in which Mary, while in her black and white room, has every concept that might make a fine-grained difference. In short, even when Mary is granted a great deal of factual knowledge and vast conceptual resources, she may still not know an appropriate answer to the question ‘what is it like to see red?’. But in any such situation, Mary lacks acquaintance with red and on this basis we argue that in order to know what it is like, in order for Mary to know an appropriate answer, Mary’s propositional knowledge must be appropriately related to her acquaintance with red.

Cognitivists about the emotions minimally hold that it is a necessary condition for being in an emotional state that one make a certain judgement or have a certain belief. For example, if I am angry with Sam, then I must believe that Sam has wronged me. Perhaps I must also elicit a certainly bodily response or undergo some relevant experience, but crucial to the view is the belief or judgement. In the face of ‘recalcitrant emotions’, this once very popular view has come under heavy criticism that has led many theorists to either abandon the view or to offer more nuanced representational views of the emotions. Against what seems to now be received wisdom, I argue that cognitivists have tools at their disposal that allow them to alleviate the apparent conflicts presented by cases of recalcitrance. But I also believe that cognitivists are still in trouble. Although cognitivists have a range of underexplored resources, their use comes at a high cost. In particular, cognitivists must adopt a widespread and thoroughgoing inaccessibility to our own thoughts and judgements that should strike one as implausible. It is mental opacity rather than mental conflict that is the real problem posed by recalcitrance.

Book synopsis: Our mental lives are entwined with the world. There are worldly things that we have beliefs about and things in the world we desire to have happen. We find some things fearsome and others likable. The puzzle of intentionality — how it is that our minds make contact with the world — is one of the oldest and most vexed issues facing philosophers. Many contemporary philosophers and cognitive scientists have been attracted to the idea that our minds represent the world. This book explores an important assumption about representation, namely, that when we represent things in the world, we represent them as having properties, and in this way our representations have "propositional" structure. The contributors examine what the commitment to propositionalism amounts to; illuminate why one might find the thesis attractive ; and consider ways in which one might depart from propositionalism. The hope is that this will lead towards a more complete understanding of how the mind and world are connected.

Book synopsis: Our mental lives are entwined with the world. There are worldly things that we have beliefs about and things in the world we desire to have happen. We find some things fearsome and others likable. The puzzle of intentionality — how it is that our minds make contact with the world — is one of the oldest and most vexed issues facing philosophers. Many contemporary philosophers and cognitive scientists have been attracted to the idea that our minds represent the world. This book explores an important assumption about representation, namely, that when we represent things in the world, we represent them as having properties, and in this way our representations have "propositional" structure. The contributors examine what the commitment to propositionalism amounts to; illuminate why one might find the thesis attractive ; and consider ways in which one might depart from propositionalism. The hope is that this will lead towards a more complete understanding of how the mind and world are connected.

To understand what non-propositional content is and whether there are any such contents, we first need to know what propositional content is. That issue will be the focus of the first section of this essay. In the second section, with an understanding of propositional content in hand, we will consider representations that fail to have propositional content. In contrast to recent literature, it will be argued that metaphysical considerations concerning what's represented, rather than linguistic considerations, are a more promising way of establishing non-propositional contents. To keep the discussion containable, focus will be on representational mental states, though many of the considerations can be extended to other forms of representation.

Adverbialist theories of thought such as those advanced by Hare and Sellars promise an ontologically sleek understanding of a variety of intentional states, but such theories have been largely abandoned due to the ‘many-property problem’. In an attempt to revitalize this otherwise attractive theory, in a series of papers as well as his recent book, Uriah Kriegel has offered a novel reply to the ‘many-property problem’ and on its basis he argues that ‘adverbialism about intentionality is alive and well’. If true, Kriegel will have shown that the logical landscape has long been unnecessarily constrained. His key idea is that the many-property problem can be overcome by appreciating that mental states stand in the determinable–determinate relation to one another. The present paper shows that this relation can’t save adverbialism because it would require thinkers to think more thoughts than they need be thinking.

Most contemporary philosophical discussions of intentionality start and end with a treatment of the propositional attitudes. In fact, many theorists hold that all attitudes are propositional attitudes. Our folk-psychological ascriptions suggest, however, that there are non-propositional attitudes: I like Sally, my brother fears snakes, everyone loves my grandmother, and Rush Limbaugh hates Obama. I argue that things are as they appear: there are non-propositional attitudes. More specifically, I argue that there are attitudes that relate individuals to non-propositional objects and do so not in virtue of relating them to propositions. I reach this conclusion by not only showing that attempted analyses of apparently non-propositional attitudes in terms of the propositional fail, but that some non-propositional attitudes don’t even supervene on propositional attitudes. If this is correct, then the common discussions of intentionality that address only propositional attitudes are incomplete and those who hold that all intentional states are propositional are mistaken